diff --git a/internal/cli/sandbox.go b/internal/cli/sandbox.go index bb49fa8e1..892638362 100644 --- a/internal/cli/sandbox.go +++ b/internal/cli/sandbox.go @@ -22,7 +22,7 @@ const permissionProfileScopeNote = "permissionProfile is derived from this proce func runSandbox(args []string, stdout io.Writer, stderr io.Writer, deps appDeps) int { if len(args) == 0 { - return writeExecUsageError(stderr, "sandbox subcommand required. Use `zero sandbox policy` or `zero sandbox grants list`.") + return writeExecUsageError(stderr, "sandbox subcommand required. Use `zero sandbox policy`, `zero sandbox exec`, or `zero sandbox grants list`.") } switch args[0] { case "-h", "--help", "help": @@ -36,6 +36,8 @@ func runSandbox(args []string, stdout io.Writer, stderr io.Writer, deps appDeps) return runSandboxSetup(args[1:], stdout, stderr, deps) case "check": return runSandboxCheck(args[1:], stdout, stderr, deps) + case "exec": + return runSandboxExec(args[1:], stdout, stderr, deps) case "grants": return runSandboxGrants(args[1:], stdout, stderr, deps) default: @@ -174,10 +176,17 @@ func runSandboxSetup(args []string, stdout io.Writer, stderr io.Writer, deps app if !setupHelper.Available() { return writeAppError(stderr, "Windows sandbox setup helper is not available", exitProvider) } + // Resolved here, in the shell the user typed `zero sandbox setup` into, and + // carried in the args. The helper may be launched elevated, and an elevated + // process does not inherit this shell's environment. Stated explicitly rather + // than left nil (which resolves the same way) because this is the call site + // the opt-in is about. + principalOptIn := zeroSandbox.WindowsSandboxPrincipalOptIn(nil) setupArgs, err := zeroSandbox.BuildWindowsSandboxSetupArgs(zeroSandbox.WindowsSandboxSetupArgsOptions{ CommandCWD: workspaceRoot, WorkspaceRoots: []string{workspaceRoot}, PermissionProfile: profile, + PrincipalOptIn: &principalOptIn, }) if err != nil { return writeAppError(stderr, err.Error(), exitCrash) @@ -646,6 +655,7 @@ Commands: policy Inspect active sandbox policy and platform backend setup Run native platform sandbox setup check Evaluate the sandbox decision for a hypothetical tool action + exec Run one command through the real sandbox grants Manage persistent sandbox grants `) diff --git a/internal/cli/sandbox_exec.go b/internal/cli/sandbox_exec.go new file mode 100644 index 000000000..5119be556 --- /dev/null +++ b/internal/cli/sandbox_exec.go @@ -0,0 +1,358 @@ +package cli + +import ( + "context" + "errors" + "fmt" + "io" + "os" + "os/exec" + "strings" + "time" + + "github.com/Gitlawb/zero/internal/config" + "github.com/Gitlawb/zero/internal/execution" + zeroSandbox "github.com/Gitlawb/zero/internal/sandbox" +) + +// runSandboxExec runs ONE command through the real sandbox and exits with its +// status. +// +// This exists because until now the sandbox could only be exercised through a +// full agent turn with a model in the loop. `zero sandbox policy` reports what +// the posture would be and `zero sandbox check` evaluates a hypothetical +// decision, but nothing actually ran a command and let you look at what +// happened on disk afterwards. The practical result is that enforcement is +// covered almost entirely by tests asserting the shape of an ACL plan, and +// almost not at all by tests asserting a write was refused. +// +// A plan can be perfectly correct and never reach the filesystem. That is not +// hypothetical here: the .git rename guard was emitted correctly by the planner +// and silently skipped by the applier, and four tests covering the plan all +// passed while the ACE was absent from disk. Something that runs the real +// binary and then stats the file is the only thing that catches that class. +// +// Deliberately NOT a debug curiosity: it takes the same path a shell tool +// takes, the engine's decision first and then the engine's command plan, so +// what it proves is what users get. It prints the decision, then the resolved +// backend and enforcement level, to stderr before running, so a harness can +// assert the sandbox was actually engaged rather than quietly downgraded. +func runSandboxExec(args []string, stdout io.Writer, stderr io.Writer, deps appDeps) int { + command, err := parseSandboxExecArgs(args) + if err != nil { + if errors.Is(err, errSandboxExecHelp) { + if writeErr := writeSandboxExecHelp(stdout); writeErr != nil { + return exitCrash + } + return exitSuccess + } + return writeExecUsageError(stderr, err.Error()) + } + + workspaceRoot, err := resolveWorkspaceRoot("", deps) + if err != nil { + return writeExecUsageError(stderr, err.Error()) + } + resolved, err := deps.resolveConfig(workspaceRoot, config.Overrides{}) + if err != nil { + return writeAppError(stderr, err.Error(), exitProvider) + } + policy := applyConfiguredSandboxPolicy(zeroSandbox.DefaultPolicy(), resolved.Sandbox) + + scope, err := zeroSandbox.NewScope(workspaceRoot, resolved.Sandbox.AdditionalWriteRoots) + if err != nil { + return writeAppError(stderr, fmt.Sprintf("resolve sandbox write roots: %v", err), exitCrash) + } + + // Through the ENGINE, not a SandboxManager built here. + // + // The engine is what a real tool call goes through, and it does more than + // hand the manager a request: it resolves the permission profile, calls + // prepareSandboxRuntime, and folds the runtime state into that profile + // before planning. Building a manager directly skipped all of it, so the + // command ran without the runtime write root, and cache or temp writes could + // pass or fail differently from the sandboxed command this exists to imitate. + // A harness that exercises the wrong path is worse than no harness, because + // its result still reads as evidence. + engine := zeroSandbox.NewEngine(zeroSandbox.EngineOptions{ + WorkspaceRoot: workspaceRoot, + Policy: policy, + Scope: scope, + Backend: deps.selectSandboxBackend(zeroSandbox.BackendOptions{}), + // The same scrub list a real tool call gets. Without it the engine keeps an + // empty set, and only the hardcoded names plus the provider catalog's own + // AuthEnvVars are removed from the child's environment: a key named by + // `apiKeyEnv` in the user's config would be scrubbed for every sandboxed + // tool call and handed to this one. Reproducing the production environment + // is the entire point of the command, and a credential is the last part of + // it that may differ. + SensitiveEnvKeys: providerSensitiveEnvKeys(resolved), + }) + + // The CLI's shared shutdown context, so Ctrl+C and a directed SIGTERM both + // arrive here rather than only at whatever the terminal happens to signal. + runCtx, stopSignals := signalContext() + defer stopSignals() + + // THE DECISION COMES BEFORE THE PLAN, as it does for every real tool call. + // + // A shell tool is evaluated by the engine and only then planned + // (tools.Registry.RunWithOptions). This went straight to BuildCommandPlan, so + // it took "the same path a shell tool takes" only from the second step on, and + // every refusal that lives in Evaluate did not exist here. The one that + // matters most is the nested-repository guard: a workspace governed by an + // ancestor repository carries NO git carveouts, on the strength of that guard + // refusing to create a repository there, and `zero sandbox exec -- git init` + // created one under the plain workspace grant with a writable config and + // hooks. A harness that runs what production would have refused does not + // prove what users get. Reported by @jatmn. + // + // PermissionGranted is true because the person at the terminal typed the + // command, which is the approval a prompt asks a session user for. It lifts + // exactly what an approval lifts in a session and nothing else: a denial is + // still a denial, and an approved network command still runs inside a sandbox + // that has the network off. + decision := sandboxExecDecision(runCtx, engine, workspaceRoot, command) + fmt.Fprintf(stderr, "sandbox: decision=%s reason=%s\n", decision.Action, decision.Reason) + if decision.Action == zeroSandbox.ActionDeny { + return writeAppError(stderr, decision.ErrorString(), exitCrash) + } + + plan, err := engine.BuildCommandPlan(zeroSandbox.CommandSpec{ + Name: command[0], + Args: command[1:], + Dir: workspaceRoot, + Env: os.Environ(), + }) + if err != nil { + return writeAppError(stderr, fmt.Sprintf("build sandbox command plan: %v", err), exitCrash) + } + // The plan owns resources beyond its construction: on Linux it allocates a + // policy-report file and registers the removal here, so without this every + // invocation leaves a /tmp/zero-sandbox-report-* behind. + defer plan.Cleanup() + + // Printed before the command runs and on stderr, so it survives a command + // that writes to stdout and stays greppable by a test harness. A downgrade + // is reported loudly for the same reason: a smoke test that passes because + // the sandbox quietly stood down is worse than no smoke test. + fmt.Fprintf(stderr, "sandbox: backend=%s enforcement=%s wrapped=%t workspace=%s\n", + plan.Backend.Name, plan.EnforcementLevel, plan.Wrapped, plan.WorkspaceRoot) + if strings.TrimSpace(plan.DowngradeReason) != "" { + fmt.Fprintf(stderr, "sandbox: DOWNGRADED: %s\n", plan.DowngradeReason) + } + + return runSandboxPlannedCommand(runCtx, plan, stdout, stderr) +} + +// sandboxExecToolName is the tool a session would have run this command with. +// Grants and the engine's shell rules are keyed by it. +const sandboxExecToolName = "bash" + +// sandboxExecDecision asks the engine the question a session asks before it +// plans a shell command, for the argv this invocation was given. +func sandboxExecDecision(ctx context.Context, engine *zeroSandbox.Engine, workspaceRoot string, argv []string) zeroSandbox.Decision { + return engine.Evaluate(ctx, zeroSandbox.Request{ + WorkspaceRoot: workspaceRoot, + ToolName: sandboxExecToolName, + SideEffect: zeroSandbox.SideEffectShell, + PermissionGranted: true, + PermissionMode: zeroSandbox.PermissionModeAsk, + Args: map[string]any{"command": sandboxExecCommandText(argv)}, + Reason: "zero sandbox exec", + }) +} + +// sandboxExecCommandText renders argv as the command text the engine analyses. +// +// The engine classifies a shell tool by parsing its command string, so argv has +// to become one without changing what it says. A word made only of characters +// the shell does not interpret is left bare, which keeps `git init` reading as a +// session would have written it. Anything else is single-quoted, so an argument +// such as "git init" stays ONE word and is not mistaken for a command, and a +// payload handed to `sh -c` reaches the analyser as the payload it is. +func sandboxExecCommandText(argv []string) string { + words := make([]string, 0, len(argv)) + for _, word := range argv { + words = append(words, sandboxExecQuoteWord(word)) + } + return strings.Join(words, " ") +} + +func sandboxExecQuoteWord(word string) string { + if word == "" { + return "''" + } + bare := true + for _, r := range word { + switch { + case r >= 'a' && r <= 'z', r >= 'A' && r <= 'Z', r >= '0' && r <= '9': + case strings.ContainsRune("_@%+=:,./-", r): + default: + bare = false + } + } + if bare { + return word + } + // A single quote cannot appear inside single quotes, so it closes them, is + // written escaped, and opens them again. + return "'" + strings.ReplaceAll(word, "'", `'\''`) + "'" +} + +const ( + // sandboxExecShutdownGrace is how long a cancelled child has to exit on its + // own after the graceful request, before it is killed outright. + sandboxExecShutdownGrace = 5 * time.Second + // sandboxExecShutdownPoll is how often that wait re-checks, matching the + // background package's termination policy. + sandboxExecShutdownPoll = 50 * time.Millisecond +) + +func runSandboxPlannedCommand(ctx context.Context, plan zeroSandbox.CommandPlan, stdout io.Writer, stderr io.Writer) int { + // CANCELLING THE WRAPPER HAS TO REACH THE COMMAND. + // + // This started the backend wrapper with a bare exec.Command().Run(): no + // context, no forwarding, no shutdown path. A terminal masks it, because + // terminals signal the whole foreground process group, but a supervisor or + // task runner that sends SIGTERM to the zero sandbox exec PID killed only + // Zero. The wrapper and everything under it kept running, doing filesystem + // and network work after the caller considered the task cancelled, and the + // deferred plan cleanup never ran because the process was gone. + // + // DELIBERATELY NOT execution.ConfigureProcessGroup. Putting the child in its + // own process group closes this hole and opens a worse one: it severs every + // kernel-delivered group signal, so a supervisor escalating to a group kill, + // and a terminal delivering Ctrl+C to its foreground group, would stop + // reaching the child. That trades a path that works today for one that does + // not. Keeping the child in Zero's group leaves group delivery unchanged, + // while the directed-signal case, which reached nothing at all before, now + // reaches the child. + process := exec.CommandContext(ctx, plan.Name, plan.Args...) + // TWO PHASES, AND THE FIRST ONE HAS TO EXIST. + // + // The comments here used to describe a graceful signal followed by escalation + // while the code did neither: Cancel called KillProcessTree, which is SIGKILL + // on Unix and taskkill /T /F on Windows, and WaitDelay only starts counting + // AFTER Cancel returns, so it could never postpone a kill that had already + // happened. A directed SIGTERM reached the child as 137 rather than 143, with + // no chance to flush output, drop a lock file, or run a shutdown handler. + // + // TerminateProcessTree is the two-phase policy the background package already + // uses: SIGTERM, poll for the grace, then SIGKILL what is still alive. On + // Windows it is the single-phase kill, which is correct rather than lazy, + // because there is no signal to send a child that owns no console. + process.Cancel = func() error { + return execution.TerminateProcessTree(process.Process.Pid, sandboxExecShutdownGrace, sandboxExecShutdownPoll) + } + // A backstop only. Cancel above already waits out the grace and escalates, so + // this bounds the case where the tree is gone but Wait is still blocked on an + // inherited pipe a grandchild holds open. + process.WaitDelay = sandboxExecShutdownGrace + sandboxExecShutdownPoll + process.Dir = plan.Dir + if process.Dir == "" { + process.Dir = plan.WorkspaceRoot + } + // SPECIFIED-EMPTY IS NOT UNSPECIFIED. + // + // exec.Cmd treats a nil Env as "inherit this process's entire environment", + // which is a different statement from "run with no variables". The plan owns + // its environment: directCommandEnv and scrubSensitiveEnv return a slice they + // built, and that slice is non-nil with length zero when every entry was + // sensitive. Testing length collapsed those two states and turned the strictest + // possible answer into the loosest one. + // + // Not reachable through `zero sandbox exec` today, because the child + // environment is os.Environ() and an environment holding only sensitive keys + // has no %AppData%, so config resolution fails before the planner runs. Fixed + // anyway: the guard is one assignment, and the next caller that hands the plan + // a deliberately narrow environment would inherit everything instead, silently. + if plan.Env != nil { + process.Env = plan.Env + } + process.Stdin = os.Stdin + process.Stdout = stdout + process.Stderr = stderr + + if err := process.Run(); err != nil { + var exitErr *exec.ExitError + if errors.As(err, &exitErr) { + // The command's own status, not ours. A harness asserting "the write + // was refused" needs the refusal's exit code, not a wrapper's. + // A SIGNALED CHILD HAS NO EXIT CODE TO REPORT. ExitCode() answers -1 + // there, and os.Exit truncates that to 255, so a child that took SIGTERM + // became indistinguishable from one that chose to exit 255. Fold the + // signal into the conventional 128+n a shell would report, which needs + // the ProcessState rather than the integer. + if status, signaled := signaledExitStatus(exitErr.ProcessState); signaled { + return status + } + return exitErr.ExitCode() + } + fmt.Fprintf(stderr, "sandbox exec: %v\n", err) + return exitCrash + } + return exitSuccess +} + +var errSandboxExecHelp = errors.New("help requested") + +// parseSandboxExecArgs takes everything after `--` as the command, so the +// command's own flags are never mistaken for ours. +func parseSandboxExecArgs(args []string) ([]string, error) { + if len(args) == 0 { + return nil, errors.New("usage: zero sandbox exec -- [args...]") + } + // Only the FIRST token is ours to interpret. Everything from the second + // onwards belongs to the child, help flags included. + // + // Written as a straight-line decision on args[0] rather than a loop, because + // a loop here is a lie. An earlier version scanned for `--` "up front" and + // said so in its comment, but every branch of its body returned, so it only + // ever examined index 0 and the promised scan did not exist. The rule below + // is what that code actually implemented, and it is the rule we want: it is + // the wrapper's own prefix that can ask for the wrapper's help, and the + // prefix is at most one token long. + switch args[0] { + case "--": + // Everything after the separator is the command, verbatim, including + // `--help`. That is the documented contract: `zero sandbox exec -- cmd + // --help` must run cmd's help, not ours. + command := args[1:] + if len(command) == 0 { + return nil, errors.New("usage: zero sandbox exec -- [args...]") + } + return command, nil + case "-h", "--help", "help": + return nil, errSandboxExecHelp + } + // The separator-less form, tolerated for interactive use: the first token is + // the command, so nothing after it is ours to read. Without this, + // `zero sandbox exec mycmd --help` printed OUR help and never ran mycmd, + // which is the same contract break as reading past `--`. The help text still + // shows the separator, because anything with a leading dash needs it. + return args, nil +} + +func writeSandboxExecHelp(w io.Writer) error { + _, err := fmt.Fprint(w, `Usage: + zero sandbox exec -- [args...] + +Runs one command through the real sandbox and exits with its status. + +Everything after the -- separator is the command, so its own flags are not +parsed as Zero's. The command is evaluated the way a session evaluates a shell +tool call before anything is planned: the decision is written to stderr, and a +command a session refuses is refused here too, with the same reason. Running it +from a terminal counts as the approval a session would prompt for, and lifts +only what that approval lifts. The resolved backend and enforcement level are +then written to stderr before the command runs, and a downgrade is reported +there explicitly. + +Examples: + zero sandbox exec -- cmd /c echo hello + zero sandbox exec -- powershell -Command "Set-Content out.txt x" + +`) + return err +} diff --git a/internal/cli/sandbox_exec_cancel_test.go b/internal/cli/sandbox_exec_cancel_test.go new file mode 100644 index 000000000..96916b882 --- /dev/null +++ b/internal/cli/sandbox_exec_cancel_test.go @@ -0,0 +1,78 @@ +package cli + +import ( + "context" + "io" + "os" + "runtime" + "testing" + "time" + + zeroSandbox "github.com/Gitlawb/zero/internal/sandbox" +) + +// longLivedPlan returns a plan whose child outlives the test unless something +// terminates it. +func longLivedPlan(t *testing.T) zeroSandbox.CommandPlan { + t.Helper() + plan := zeroSandbox.CommandPlan{Dir: t.TempDir()} + if runtime.GOOS == "windows" { + plan.Name = "cmd.exe" + plan.Args = []string{"/c", "ping -n 120 127.0.0.1 >NUL"} + return plan + } + plan.Name = "/bin/sh" + plan.Args = []string{"-c", "sleep 120"} + return plan +} + +// CANCELLING THE WRAPPER HAS TO REACH THE COMMAND. +// +// The sandboxed command was started with a bare exec.Command().Run(): no +// context, no forwarding, no shutdown path. A terminal masks that, because it +// signals the whole foreground process group, but a supervisor or task runner +// that sends SIGTERM to the wrapper's PID killed only Zero. The command kept +// running, doing filesystem and network work after the caller considered the +// task cancelled, and the deferred plan cleanup never ran. +// +// Driven with a real long-lived child and a real cancellation, asserting that +// the call actually returns rather than that a field is set. +func TestCancellingTheWrapperTerminatesTheSandboxedCommand(t *testing.T) { + ctx, cancel := context.WithCancel(context.Background()) + + done := make(chan int, 1) + go func() { + done <- runSandboxPlannedCommand(ctx, longLivedPlan(t), io.Discard, io.Discard) + }() + + // Let the child actually start, or cancelling proves nothing. + time.Sleep(300 * time.Millisecond) + select { + case <-done: + t.Fatal("SETUP INVALID: the child exited on its own, so cancellation was not exercised") + default: + } + + cancel() + select { + case <-done: + case <-time.After(sandboxExecShutdownGrace + 10*time.Second): + t.Fatal("cancelling the wrapper did not terminate the sandboxed command; it would keep running after the caller gave up") + } +} + +// And an uncancelled command still runs to completion and reports its own +// status, or the fix above would be "kill everything immediately". +func TestAnUncancelledSandboxedCommandStillReportsItsStatus(t *testing.T) { + plan := zeroSandbox.CommandPlan{Dir: t.TempDir()} + if runtime.GOOS == "windows" { + plan.Name = "cmd.exe" + plan.Args = []string{"/c", "exit 3"} + } else { + plan.Name = "/bin/sh" + plan.Args = []string{"-c", "exit 3"} + } + if code := runSandboxPlannedCommand(context.Background(), plan, io.Discard, os.Stderr); code != 3 { + t.Fatalf("exit code = %d, want the child's own 3", code) + } +} diff --git a/internal/cli/sandbox_exec_decision_test.go b/internal/cli/sandbox_exec_decision_test.go new file mode 100644 index 000000000..8214d58b2 --- /dev/null +++ b/internal/cli/sandbox_exec_decision_test.go @@ -0,0 +1,162 @@ +package cli + +import ( + "bytes" + "context" + "os" + "path/filepath" + "strings" + "testing" + + "github.com/Gitlawb/zero/internal/config" + "github.com/Gitlawb/zero/internal/sandbox" +) + +// governedNestedWorkspace is a workspace inside somebody else's repository: an +// ancestor carries .git and the workspace itself does not. Setup plans no git +// carveouts for it, which is why a repository must not be created there. +func governedNestedWorkspace(t *testing.T) (ancestor, workspace string) { + t.Helper() + ancestor = t.TempDir() + if err := os.MkdirAll(filepath.Join(ancestor, ".git"), 0o700); err != nil { + t.Fatal(err) + } + workspace = filepath.Join(ancestor, "packages", "app") + if err := os.MkdirAll(filepath.Join(workspace, "sub dir"), 0o700); err != nil { + t.Fatal(err) + } + return ancestor, workspace +} + +func sandboxExecTestDeps(workspace string) appDeps { + return appDeps{ + getwd: func() (string, error) { return workspace, nil }, + resolveConfig: func(string, config.Overrides) (config.ResolvedConfig, error) { + return config.ResolvedConfig{}, nil + }, + // No native runner. A command that got past the decision would run + // unwrapped, which is exactly what makes a missing refusal visible on + // disk instead of only in a return value. + selectSandboxBackend: func(sandbox.BackendOptions) sandbox.Backend { + return sandbox.Backend{Name: sandbox.BackendUnavailable, Fallback: true, Message: "no native sandbox in this test"} + }, + } +} + +// `zero sandbox exec` REFUSES WHAT A SESSION WOULD REFUSE. +// +// The command is documented as taking the path a shell tool takes. A shell tool +// is evaluated by the engine and then planned; this went straight to the plan, +// so nothing that lives in Evaluate applied. In a workspace governed by an +// ancestor repository that is the nested-repository guard, and that guard is the +// only thing standing in for the git carveouts setup deliberately does not plan +// there. `zero sandbox exec -- git init` created the repository, writable config +// and hooks included. +// +// Driven through the real CLI entry point, with a backend that would run the +// command unwrapped if it were let through, and checked on disk: the refusal has +// to come with the reason an operator can act on, and no repository may exist +// afterwards. Reported by @jatmn. +func TestSandboxExecRefusesToCreateARepositoryInAGovernedWorkspace(t *testing.T) { + for _, testCase := range []struct { + name string + argv []string + made []string + }{ + {name: "git init", argv: []string{"git", "init"}, made: []string{".git"}}, + {name: "git init behind -C and a quoted directory", argv: []string{"git", "-C", "sub dir", "init"}, made: []string{"sub dir/.git"}}, + {name: "git init behind a shell launcher", argv: []string{"sh", "-c", "git init"}, made: []string{".git"}}, + {name: "git init behind env", argv: []string{"env", "GIT_TRACE=0", "git", "init"}, made: []string{".git"}}, + {name: "git clone", argv: []string{"git", "clone", "https://example.invalid/repo.git", "vendored"}, made: []string{"vendored"}}, + } { + t.Run(testCase.name, func(t *testing.T) { + _, workspace := governedNestedWorkspace(t) + var stdout, stderr bytes.Buffer + args := append([]string{"sandbox", "exec", "--"}, testCase.argv...) + exitCode := runWithDeps(args, &stdout, &stderr, sandboxExecTestDeps(workspace)) + + for _, made := range testCase.made { + if _, err := os.Lstat(filepath.Join(workspace, filepath.FromSlash(made))); err == nil { + t.Errorf("%s exists after the command: the repository was created under the plain workspace grant", made) + } + } + if exitCode == exitSuccess { + t.Errorf("exit = %d for a command a session refuses; stderr:\n%s", exitCode, stderr.String()) + } + if !strings.Contains(stderr.String(), "decision=deny") { + t.Errorf("stderr does not report the decision:\n%s", stderr.String()) + } + if !strings.Contains(stderr.String(), "sits inside an existing git repository") { + t.Errorf("the refusal does not carry the nested-repository reason and its remedy:\n%s", stderr.String()) + } + if strings.Contains(stderr.String(), "backend=") { + t.Errorf("a command plan was built for a refused command:\n%s", stderr.String()) + } + }) + } +} + +// The decision is the engine's, asked the way a session asks it. These are the +// controls: what must NOT be refused, so the gate above is not simply refusing +// git, or refusing this workspace. +func TestSandboxExecDecisionLeavesOrdinaryCommandsAlone(t *testing.T) { + _, governed := governedNestedWorkspace(t) + standalone := t.TempDir() + + for _, testCase := range []struct { + name string + workspace string + argv []string + }{ + {name: "an ordinary git command in the governed workspace", workspace: governed, argv: []string{"git", "status"}}, + {name: "the words git init as one argument", workspace: governed, argv: []string{"echo", "git init"}}, + {name: "git init in a workspace nobody governs", workspace: standalone, argv: []string{"git", "init"}}, + } { + t.Run(testCase.name, func(t *testing.T) { + engine := sandbox.NewEngine(sandbox.EngineOptions{ + WorkspaceRoot: testCase.workspace, + Policy: sandbox.DefaultPolicy(), + Backend: sandbox.Backend{Name: sandbox.BackendUnavailable, Fallback: true}, + }) + decision := sandboxExecDecision(context.Background(), engine, testCase.workspace, testCase.argv) + if decision.Action == sandbox.ActionDeny { + t.Fatalf("%q was refused: %s", strings.Join(testCase.argv, " "), decision.ErrorString()) + } + }) + } +} + +// ARGV BECOMES COMMAND TEXT WITHOUT CHANGING WHAT IT SAYS. The engine classifies +// a shell command by parsing its text, so the rendering decides what the +// classifier sees. One argument has to stay one word, and a payload handed to a +// shell launcher has to arrive as that payload. +func TestSandboxExecCommandTextKeepsArgvIntact(t *testing.T) { + for _, testCase := range []struct { + name string + argv []string + want string + gitInit bool + }{ + {name: "bare words stay bare", argv: []string{"git", "init"}, want: "git init", gitInit: true}, + {name: "a spaced argument is one word", argv: []string{"echo", "git init"}, want: "echo 'git init'"}, + {name: "a launcher payload is analysed as a payload", argv: []string{"sh", "-c", "git init"}, want: "sh -c 'git init'", gitInit: true}, + {name: "a quote inside an argument", argv: []string{"printf", "it's"}, want: `printf 'it'\''s'`}, + {name: "substitution syntax is inert", argv: []string{"echo", "$(git init)"}, want: "echo '$(git init)'"}, + {name: "an empty argument survives", argv: []string{"printf", ""}, want: "printf ''"}, + {name: "a directory with a space", argv: []string{"git", "-C", "sub dir", "init"}, want: "git -C 'sub dir' init", gitInit: true}, + } { + t.Run(testCase.name, func(t *testing.T) { + text := sandboxExecCommandText(testCase.argv) + if text != testCase.want { + t.Fatalf("command text = %q, want %q", text, testCase.want) + } + analysis := sandbox.AnalyzeCommand(text) + if analysis.TooComplex { + t.Fatalf("%q does not parse, so the classifier would treat a plain argv as obfuscated", text) + } + if analysis.GitInit != testCase.gitInit { + t.Errorf("%q: GitInit = %v, want %v", text, analysis.GitInit, testCase.gitInit) + } + }) + } +} diff --git a/internal/cli/sandbox_exec_env_test.go b/internal/cli/sandbox_exec_env_test.go new file mode 100644 index 000000000..17624671c --- /dev/null +++ b/internal/cli/sandbox_exec_env_test.go @@ -0,0 +1,72 @@ +package cli + +import ( + "context" + "os" + "runtime" + "strings" + "testing" + + zeroSandbox "github.com/Gitlawb/zero/internal/sandbox" +) + +// envPrinterPlan returns a plan whose child prints its own environment, so the +// assertion is about what the child received rather than about a field value. +func envPrinterPlan(t *testing.T, env []string) zeroSandbox.CommandPlan { + t.Helper() + plan := zeroSandbox.CommandPlan{Dir: t.TempDir(), Env: env} + if runtime.GOOS == "windows" { + plan.Name = "cmd.exe" + plan.Args = []string{"/c", "set"} + return plan + } + plan.Name = "/bin/sh" + plan.Args = []string{"-c", "env"} + return plan +} + +// A PLAN THAT SPECIFIES NO VARIABLES MUST NOT INHERIT EVERY VARIABLE. +// +// exec.Cmd reads a nil Env as "inherit this process's entire environment", which +// is a different statement from "run with no variables". The plan owns its +// environment: directCommandEnv and scrubSensitiveEnv return a slice they built, +// and that slice is non-nil with length zero when every entry was sensitive. +// Testing its length collapsed the two states and turned the strictest possible +// answer into the loosest one. +// +// Not reachable through `zero sandbox exec` today, because the child environment +// is os.Environ() and an environment holding only sensitive keys has no +// %AppData%, so config resolution fails before the planner runs. Pinned anyway, +// because the next caller that hands the plan a deliberately narrow environment +// would silently inherit everything instead. +func TestAnEmptyPlannedEnvironmentDoesNotInherit(t *testing.T) { + const marker = "ZZ_SANDBOX_ENV_MARKER" + t.Setenv(marker, "leaked-value") + + var out strings.Builder + // Specified, and deliberately empty. + code := runSandboxPlannedCommand(context.Background(), envPrinterPlan(t, []string{}), &out, os.Stderr) + if code != 0 { + t.Fatalf("child exited %d: %s", code, out.String()) + } + if strings.Contains(out.String(), marker) { + t.Fatalf("a plan specifying no environment leaked the caller's %s to the child:\n%s", marker, out.String()) + } +} + +// And a nil environment still inherits, which is the meaning the planner relies +// on when it did not build one. Without this the fix above could be "never pass +// the environment", which would break every ordinary command. +func TestANilPlannedEnvironmentStillInherits(t *testing.T) { + const marker = "ZZ_SANDBOX_ENV_MARKER" + t.Setenv(marker, "inherited-value") + + var out strings.Builder + code := runSandboxPlannedCommand(context.Background(), envPrinterPlan(t, nil), &out, os.Stderr) + if code != 0 { + t.Fatalf("child exited %d: %s", code, out.String()) + } + if !strings.Contains(out.String(), marker) { + t.Fatalf("a nil planned environment stopped inheriting, which changes the meaning the planner relies on:\n%s", out.String()) + } +} diff --git a/internal/cli/sandbox_exec_grace_unix_test.go b/internal/cli/sandbox_exec_grace_unix_test.go new file mode 100644 index 000000000..39f8159b3 --- /dev/null +++ b/internal/cli/sandbox_exec_grace_unix_test.go @@ -0,0 +1,187 @@ +//go:build !windows + +package cli + +import ( + "context" + "io" + "os" + "path/filepath" + "strconv" + "strings" + "syscall" + "testing" + "time" + + zeroSandbox "github.com/Gitlawb/zero/internal/sandbox" +) + +// THE GRACEFUL PHASE HAS TO ACTUALLY HAPPEN. +// +// The shutdown comments described a graceful signal followed by escalation while +// the code did neither: Cancel called KillProcessTree, which is SIGKILL here, and +// WaitDelay only starts counting after Cancel returns so it could never postpone +// a kill that had already happened. A directed SIGTERM therefore reached the +// child as 137 rather than 143, with no chance to flush output, drop a lock file, +// or run a shutdown handler. +// +// Driven with a child that traps the signal and records that its handler ran, so +// the assertion is about the observed lifecycle rather than about a configured +// duration. +func TestACancelledCommandGetsAChanceToCleanUp(t *testing.T) { + marker := filepath.Join(t.TempDir(), "cleanup-ran") + // Exits 42 on the graceful signal: a distinctive code the child could only + // have chosen itself, so the status proves the handler ran rather than that + // something plausible happened. + script := "trap 'echo yes > " + marker + "; exit 42' TERM; while :; do sleep 0.05; done" + + ctx, cancel := context.WithCancel(context.Background()) + done := make(chan int, 1) + go func() { + done <- runSandboxPlannedCommand(ctx, zeroSandbox.CommandPlan{ + Name: "/bin/sh", + Args: []string{"-c", script}, + Dir: t.TempDir(), + }, io.Discard, io.Discard) + }() + + // The trap has to be installed before the signal, or this proves nothing. + time.Sleep(500 * time.Millisecond) + select { + case <-done: + t.Fatal("SETUP INVALID: the child exited on its own") + default: + } + + cancel() + var status int + select { + case status = <-done: + case <-time.After(sandboxExecShutdownGrace + 10*time.Second): + t.Fatal("the cancelled command never terminated") + } + + // The status the caller sees is the child's own, so a harness can tell an + // orderly shutdown from a kill. + if status != 42 { + t.Errorf("status = %d, want 42; the child chose that code in its TERM handler, so anything else means it never got to finish", status) + } + if _, err := os.Stat(marker); err != nil { + t.Fatalf("the child was killed outright instead of being asked to stop, so its cleanup handler never ran: %v", err) + } +} + +// And a child that ignores the request is still force-killed, after the bound +// rather than before it. Without this the fix could be "only ever ask nicely". +func TestAStubbornCancelledCommandIsStillKilled(t *testing.T) { + // Traps TERM and keeps running: only escalation can end this. + script := "trap '' TERM; while :; do sleep 0.05; done" + + ctx, cancel := context.WithCancel(context.Background()) + done := make(chan int, 1) + go func() { + done <- runSandboxPlannedCommand(ctx, zeroSandbox.CommandPlan{ + Name: "/bin/sh", + Args: []string{"-c", script}, + Dir: t.TempDir(), + }, io.Discard, io.Discard) + }() + time.Sleep(500 * time.Millisecond) + + started := time.Now() + cancel() + var status int + select { + case status = <-done: + case <-time.After(sandboxExecShutdownGrace + 15*time.Second): + t.Fatal("a child ignoring the graceful request was never force-killed") + } + // After the bound, not before: a kill that lands immediately is the defect + // this replaced. + if elapsed := time.Since(started); elapsed < sandboxExecShutdownGrace { + t.Fatalf("the child was killed after %s, before the %s grace it was owed", elapsed, sandboxExecShutdownGrace) + } + // And escalation really was a kill, so "ask nicely and give up" cannot pass. + if want := 128 + int(syscall.SIGKILL); status != want { + t.Errorf("status = %d, want %d after escalation", status, want) + } +} + +// AND THE STATUS HAS TO SAY WHICH SIGNAL ARRIVED. +// +// 137 rather than 143 was the reported symptom, and it is the only part of this +// a caller can see. A child that does not trap anything must come back as +// 128+SIGTERM: if the first signal were still SIGKILL, this would be 137 and the +// two cases above would both still pass, since one traps and the other ignores. +func TestACancelledCommandReportsTheGracefulSignal(t *testing.T) { + ctx, cancel := context.WithCancel(context.Background()) + done := make(chan int, 1) + go func() { + done <- runSandboxPlannedCommand(ctx, zeroSandbox.CommandPlan{ + Name: "/bin/sh", + Args: []string{"-c", "while :; do sleep 0.05; done"}, + Dir: t.TempDir(), + }, io.Discard, io.Discard) + }() + time.Sleep(500 * time.Millisecond) + cancel() + + var status int + select { + case status = <-done: + case <-time.After(sandboxExecShutdownGrace + 10*time.Second): + t.Fatal("the cancelled command never terminated") + } + if want := 128 + int(syscall.SIGTERM); status != want { + t.Fatalf("status = %d, want %d; %d means the child was SIGKILLed with no graceful phase", status, want, 128+int(syscall.SIGKILL)) + } +} + +// TERMINATION IS COMPLETE BEFORE THE FUNCTION RETURNS. +// +// `zero sandbox exec` releases the plan with a defer around this call, so plan +// cleanup is ordered after termination only if nothing survives the return. A +// future refactor that lets Wait come back while the tree is still up would move +// cleanup on top of a live child, deleting the policy-report file it is using. +// +// The stubborn child is the one that can outlive the call, since it ignores the +// graceful request and only escalation ends it. +func TestACancelledCommandIsFullyReapedBeforeCleanupCouldRun(t *testing.T) { + pidFile := filepath.Join(t.TempDir(), "child.pid") + script := "echo $$ > " + pidFile + "; trap '' TERM; while :; do sleep 0.05; done" + + ctx, cancel := context.WithCancel(context.Background()) + done := make(chan int, 1) + go func() { + done <- runSandboxPlannedCommand(ctx, zeroSandbox.CommandPlan{ + Name: "/bin/sh", + Args: []string{"-c", script}, + Dir: t.TempDir(), + }, io.Discard, io.Discard) + }() + time.Sleep(500 * time.Millisecond) + + raw, err := os.ReadFile(pidFile) + if err != nil { + t.Fatalf("SETUP INVALID: the child never recorded its pid, so there is nothing to check for survival: %v", err) + } + pid, err := strconv.Atoi(strings.TrimSpace(string(raw))) + if err != nil || pid <= 0 { + t.Fatalf("SETUP INVALID: unusable child pid %q: %v", raw, err) + } + + cancel() + select { + case <-done: + case <-time.After(sandboxExecShutdownGrace + 15*time.Second): + t.Fatal("a child ignoring the graceful request was never force-killed") + } + + // Signal 0 is the existence probe: no signal is sent, only the permission and + // liveness checks run, so ESRCH is the answer for a reaped process. + if err := syscall.Kill(pid, 0); err == nil { + t.Fatalf("the child (pid %d) was still alive when the run returned; plan cleanup is deferred around this call and would delete the policy-report file underneath it", pid) + } else if err != syscall.ESRCH { + t.Skipf("cannot probe pid %d here: %v", pid, err) + } +} diff --git a/internal/cli/sandbox_exec_signal_other.go b/internal/cli/sandbox_exec_signal_other.go new file mode 100644 index 000000000..c47442a41 --- /dev/null +++ b/internal/cli/sandbox_exec_signal_other.go @@ -0,0 +1,28 @@ +//go:build !windows + +package cli + +import ( + "os" + "syscall" +) + +// signaledExitStatus reports the conventional shell status for a child that was +// terminated by a signal, which exec.ExitError cannot represent. +// +// ExitCode() returns -1 for a signaled child, and the top level hands that to +// os.Exit, which truncates it to 255. A child that takes SIGTERM is then +// indistinguishable from one that exited 255 of its own accord, which breaks the +// documented status contract for a command whose whole job is to report the +// child's own status faithfully. 128+signal is what a shell reports and what a +// harness comparing statuses expects. +func signaledExitStatus(state *os.ProcessState) (int, bool) { + if state == nil { + return 0, false + } + status, ok := state.Sys().(syscall.WaitStatus) + if !ok || !status.Signaled() { + return 0, false + } + return 128 + int(status.Signal()), true +} diff --git a/internal/cli/sandbox_exec_signal_other_test.go b/internal/cli/sandbox_exec_signal_other_test.go new file mode 100644 index 000000000..bd838c934 --- /dev/null +++ b/internal/cli/sandbox_exec_signal_other_test.go @@ -0,0 +1,65 @@ +//go:build !windows + +package cli + +import ( + "errors" + "os/exec" + "syscall" + "testing" +) + +// A SIGNALED CHILD MUST NOT LOOK LIKE ONE THAT CHOSE TO EXIT 255. +// +// `sandbox exec` exists to report the child's own status faithfully, and +// exec.ExitError cannot represent signal termination: ExitCode() answers -1, and +// the top level hands that to os.Exit, which truncates it to 255. So a child +// killed by SIGTERM was indistinguishable from an ordinary exit of 255, and a +// harness comparing statuses could not tell a refusal from a kill. +// +// Driven with a real subprocess and a real signal rather than a synthetic +// WaitStatus, because the mapping is only correct if the OS agrees with it. +func TestSignaledChildReportsTheConventionalStatus(t *testing.T) { + for _, testCase := range []struct { + name string + script string + want int + }{ + {name: "SIGTERM", script: "kill -TERM $$; sleep 5", want: 128 + int(syscall.SIGTERM)}, + {name: "SIGINT", script: "kill -INT $$; sleep 5", want: 128 + int(syscall.SIGINT)}, + } { + t.Run(testCase.name, func(t *testing.T) { + err := exec.Command("/bin/sh", "-c", testCase.script).Run() + var exitErr *exec.ExitError + if !errors.As(err, &exitErr) { + t.Fatalf("SETUP INVALID: the child did not end with an ExitError: %v", err) + } + // The behaviour being corrected: the integer alone cannot say this. + if code := exitErr.ExitCode(); code != -1 { + t.Fatalf("SETUP INVALID: a signaled child reported exit code %d, expected -1 on this platform", code) + } + status, signaled := signaledExitStatus(exitErr.ProcessState) + if !signaled { + t.Fatal("a signaled child was not recognised as signaled, so it would be reported as exit 255") + } + if status != testCase.want { + t.Fatalf("status = %d, want %d (128 + signal), which is what a shell reports", status, testCase.want) + } + }) + } +} + +// An ordinary non-zero exit is untouched: it has a real code and must keep it. +func TestOrdinaryExitIsNotTreatedAsSignaled(t *testing.T) { + err := exec.Command("/bin/sh", "-c", "exit 3").Run() + var exitErr *exec.ExitError + if !errors.As(err, &exitErr) { + t.Fatalf("SETUP INVALID: the child did not end with an ExitError: %v", err) + } + if _, signaled := signaledExitStatus(exitErr.ProcessState); signaled { + t.Fatal("an ordinary exit was reported as signaled, which would rewrite its status") + } + if code := exitErr.ExitCode(); code != 3 { + t.Fatalf("exit code = %d, want the child's own 3", code) + } +} diff --git a/internal/cli/sandbox_exec_signal_windows.go b/internal/cli/sandbox_exec_signal_windows.go new file mode 100644 index 000000000..826400fdc --- /dev/null +++ b/internal/cli/sandbox_exec_signal_windows.go @@ -0,0 +1,12 @@ +//go:build windows + +package cli + +import "os" + +// signaledExitStatus has no Windows counterpart: a process there ends with an +// exit code, and there is no signal to fold into one. Windows keeps the exit +// code exec.ExitError already reports. +func signaledExitStatus(*os.ProcessState) (int, bool) { + return 0, false +} diff --git a/internal/cli/sandbox_exec_test.go b/internal/cli/sandbox_exec_test.go new file mode 100644 index 000000000..5eca61304 --- /dev/null +++ b/internal/cli/sandbox_exec_test.go @@ -0,0 +1,108 @@ +package cli + +import ( + "errors" + "strings" + "testing" +) + +// parseSandboxExecArgs sits on the boundary between Zero's flags and a child +// command's, and it had no tests at all: not for the reported defect, not even +// for the happy case. The contract it has to keep is that everything after the +// separator belongs to the child, help flags included, so the wrapper can never +// answer for a command it was supposed to be running. +func TestParseSandboxExecArgs(t *testing.T) { + for _, testCase := range []struct { + name string + args []string + command []string + help bool + usage bool + }{ + { + // The reported defect: --help after the separator is the CHILD's. + name: "help after the separator belongs to the child", + args: []string{"--", "cmd", "--help"}, + command: []string{"cmd", "--help"}, + }, + { + name: "a second separator is the child's too", + args: []string{"--", "cmd", "--", "inner"}, + command: []string{"cmd", "--", "inner"}, + }, + { + name: "a flag-looking command survives the separator", + args: []string{"--", "--weird-binary"}, + command: []string{"--weird-binary"}, + }, + { + // The separator-less form. The first token is the command, so its own + // flags are not ours to read either. + name: "help after a separatorless command belongs to the child", + args: []string{"cmd", "--help"}, + command: []string{"cmd", "--help"}, + }, + { + // A separator arriving AFTER the command has already started is part of + // the child's argv, not a delimiter we get a second go at. + name: "a separator after the command is the child's argument", + args: []string{"ls", "--", "foo"}, + command: []string{"ls", "--", "foo"}, + }, + {name: "bare help", args: []string{"--help"}, help: true}, + {name: "short help", args: []string{"-h"}, help: true}, + {name: "help subcommand", args: []string{"help"}, help: true}, + {name: "help in the prefix wins", args: []string{"--help", "--", "cmd"}, help: true}, + {name: "no arguments", args: nil, usage: true}, + {name: "separator with nothing after it", args: []string{"--"}, usage: true}, + } { + t.Run(testCase.name, func(t *testing.T) { + command, err := parseSandboxExecArgs(testCase.args) + switch { + case testCase.help: + if !errors.Is(err, errSandboxExecHelp) { + t.Fatalf("parseSandboxExecArgs(%q) = %q, %v; want the help sentinel", testCase.args, command, err) + } + case testCase.usage: + if err == nil || errors.Is(err, errSandboxExecHelp) { + t.Fatalf("parseSandboxExecArgs(%q) = %q, %v; want a usage error", testCase.args, command, err) + } + if !strings.Contains(err.Error(), "usage:") { + t.Errorf("a usage error must show the usage, got %q", err.Error()) + } + default: + if err != nil { + t.Fatalf("parseSandboxExecArgs(%q) returned %v", testCase.args, err) + } + if strings.Join(command, "\x00") != strings.Join(testCase.command, "\x00") { + t.Errorf("parseSandboxExecArgs(%q) = %q, want %q", testCase.args, command, testCase.command) + } + } + }) + } +} + +// The wrapper must never consume a token it then fails to hand over. Whatever it +// returns as the command has to be a trailing slice of what it was given, with +// at most a leading separator removed, or an argument has gone missing on its +// way to the child. +func TestParseSandboxExecArgsNeverDropsAChildArgument(t *testing.T) { + for _, args := range [][]string{ + {"--", "cmd", "--help", "-x", "--"}, + {"cmd", "--help"}, + {"ls", "--", "foo"}, + {"--", "--weird-binary", "arg"}, + } { + command, err := parseSandboxExecArgs(args) + if err != nil { + t.Fatalf("parseSandboxExecArgs(%q) returned %v", args, err) + } + if len(command) == 0 { + t.Fatalf("parseSandboxExecArgs(%q) returned an empty command", args) + } + suffix := strings.Join(args[len(args)-len(command):], "\x00") + if strings.Join(command, "\x00") != suffix { + t.Errorf("parseSandboxExecArgs(%q) = %q, which is not a trailing slice of the input; an argument was dropped or reordered", args, command) + } + } +} diff --git a/internal/doctor/hardening.go b/internal/doctor/hardening.go index baf21e04c..7df54c705 100644 --- a/internal/doctor/hardening.go +++ b/internal/doctor/hardening.go @@ -100,10 +100,17 @@ func windowsSandboxSetupCheck(goos string, backend sandbox.Backend, workspaceRoo } profile := sandbox.PermissionProfileFromPolicy(workspaceRoot, doctorSandboxPolicy(sandboxConfig), scope) setupConfig := sandbox.WindowsSandboxSetupConfig{ - SandboxHome: sandboxHome, - CommandCWD: workspaceRoot, - WorkspaceRoots: []string{workspaceRoot}, - PermissionProfile: profile, + SandboxHome: sandboxHome, + CommandCWD: workspaceRoot, + WorkspaceRoots: []string{workspaceRoot}, + // Same augmentation the setup args and the command plan apply, so doctor + // fingerprints what a real command fingerprints. Safe here: doctor runs in + // the operator's shell, not behind the sandbox's TEMP redirection. + PermissionProfile: sandbox.WindowsSandboxProfileWithRuntimeRoots(profile, []string{workspaceRoot}), + // Same opt-in a command would resolve, so doctor reports the principal + // mismatch as out-of-date setup instead of passing a check the next command + // will fail. + PrincipalOptIn: sandbox.WindowsSandboxPrincipalOptIn(nil), } if err := sandbox.ValidateWindowsSandboxSetupMarker(setupConfig); err != nil { result := check("sandbox.backend", "Sandbox backend", StatusWarn, fmt.Sprintf("Native sandbox backend %s is installed, but Windows sandbox setup is missing or out of date: %v.", backend.Name, err), map[string]any{ @@ -115,6 +122,53 @@ func windowsSandboxSetupCheck(goos string, backend sandbox.Backend, workspaceRoo }) return &result } + // Setup is valid and the opt-in is set, so report which principal a command + // will actually run as. + // + // This used to warn that the principal stood down under the DEFAULT + // network-deny policy, which was true while network denial could only be + // expressed on the restricted token: a principal token cannot carry the + // synthetic offline-marker SID the WFP filters matched. Dual-role provisioning + // removed that limit by giving each workspace an offline account whose real + // group membership the same filters match, so deny mode now runs on a + // principal exactly as allow mode does. Leaving the warning would have doctor + // telling operators their reads are unconfined while they are confined. + // + // Naming the role rather than just saying "active" is the useful part. The two + // accounts differ only in that membership, so an operator debugging network + // behaviour needs to know which one this workspace picks. + if setupConfig.PrincipalOptIn { + // Asked of the shared rule rather than re-derived here, so doctor and the + // runtime cannot disagree about which account a command runs as. Two + // copies drifting is what produced the wrong report this replaces. + role := sandbox.WindowsSandboxPrincipalRoleForNetwork(profile.Network.Mode) + // SETUP IS CURRENT IS NOT THE SAME AS THE PRINCIPAL IS LIVE, and this used + // to claim the second from the first. Marker validation compares serialized + // plans and hashes; it names no account. Delete the selected account or its + // secret, or drop the offline account out of ZeroSandboxOffline, and the + // marker still validates while the runtime either falls back to the weaker + // restricted-token path or fails outright. Doctor asserting read + // confinement in that state is worse than saying nothing, because it is the + // surface an operator checks precisely when they are unsure. + // + // Verifying liveness needs the account, a usable secret and (for offline) + // the group membership, all Windows-only live queries this package cannot + // make. So the claim is narrowed to what the marker actually proves, and + // the role is still named because that part IS derived rather than assumed + // and is what an operator debugging network behaviour needs. + result := check("sandbox.principal", "Sandbox principal", StatusPass, + fmt.Sprintf("Sandbox principal setup is current; a command in this workspace selects the %s principal. Setup state does not confirm the account is usable.", role), map[string]any{ + "backend": string(backend.Name), + "platform": goos, + "optIn": true, + // Deliberately not "active": nothing here checked the account, the + // secret or the group membership. + "setupCurrent": true, + "role": role, + "networkMode": string(profile.Network.Mode), + }) + return &result + } return nil } diff --git a/internal/doctor/hardening_principal_windows_test.go b/internal/doctor/hardening_principal_windows_test.go new file mode 100644 index 000000000..d1239343a --- /dev/null +++ b/internal/doctor/hardening_principal_windows_test.go @@ -0,0 +1,72 @@ +package doctor + +import ( + "strings" + "testing" + + "github.com/Gitlawb/zero/internal/config" + "github.com/Gitlawb/zero/internal/sandbox" +) + +// A VALID MARKER IS NOT A LIVE PRINCIPAL. Marker validation compares serialized +// plans and hashes and names no account, so it stays valid after the selected +// account is deleted, its secret is removed, or the offline account is dropped +// out of ZeroSandboxOffline. In each of those states the runtime falls back to +// the restricted-token path with no read confinement, or fails outright. +// +// Doctor is the surface an operator checks precisely when they are unsure, so +// claiming read confinement it has not verified is worse than reporting less. +// This pins the narrower claim: setup is current, the role is named because it +// is derived rather than assumed, and nothing asserts the account is usable. +func TestPrincipalCheckDoesNotClaimAnUnverifiedPrincipalIsActive(t *testing.T) { + sandboxHome := t.TempDir() + workspaceRoot := t.TempDir() + t.Setenv("ZERO_WINDOWS_SANDBOX_HOME", sandboxHome) + t.Setenv("ZERO_WINDOWS_SANDBOX_IDENTITY", "1") + + sandboxConfig := config.SandboxConfig{} + scope, err := sandbox.NewScope(workspaceRoot, sandboxConfig.AdditionalWriteRoots) + if err != nil { + t.Fatalf("NewScope: %v", err) + } + profile := sandbox.PermissionProfileFromPolicy(workspaceRoot, doctorSandboxPolicy(sandboxConfig), scope) + + // Mirrors what windowsSandboxSetupCheck derives, so the marker it writes is + // the one the check will validate against. Nothing here provisions an + // account, which is the whole point: the marker is valid and the principal + // does not exist. + if _, err := sandbox.WriteWindowsSandboxSetupMarker(sandbox.WindowsSandboxSetupConfig{ + SandboxHome: sandboxHome, + CommandCWD: workspaceRoot, + WorkspaceRoots: []string{workspaceRoot}, + PermissionProfile: sandbox.WindowsSandboxProfileWithRuntimeRoots(profile, []string{workspaceRoot}), + PrincipalOptIn: true, + }); err != nil { + t.Fatalf("WriteWindowsSandboxSetupMarker: %v", err) + } + + result := windowsSandboxSetupCheck("windows", sandbox.Backend{Name: sandbox.BackendWindowsRestrictedToken}, workspaceRoot, sandboxConfig) + if result == nil { + t.Fatal("expected a principal check once the marker validates and the opt-in is set") + } + if result.ID != "sandbox.principal" { + t.Fatalf("check = %q (%s): the marker did not validate, so this test is not exercising the principal branch", result.ID, result.Message) + } + + lowered := strings.ToLower(result.Message) + for _, claim := range []string{"is active", "commands run as"} { + if strings.Contains(lowered, claim) { + t.Errorf("doctor asserts %q about a principal it never checked: %q", claim, result.Message) + } + } + if active, ok := result.Details["active"]; ok { + t.Errorf("details carry active=%v, which nothing here verified", active) + } + + // The role is still worth reporting: it is derived from the network mode by + // the shared rule rather than assumed, and it is what an operator debugging + // network behaviour needs. + if role, ok := result.Details["role"]; !ok || strings.TrimSpace(role.(string)) == "" { + t.Errorf("the selected role should still be reported: %#v", result.Details) + } +} diff --git a/internal/peermsg/private_dir_other.go b/internal/peermsg/private_dir_other.go index 8ce11c413..be145e55b 100644 --- a/internal/peermsg/private_dir_other.go +++ b/internal/peermsg/private_dir_other.go @@ -10,7 +10,7 @@ import ( "strings" ) -func ensurePrivateDir(path string) error { +func EnsurePrivateDir(path string) error { abs, err := filepath.Abs(path) if err != nil { return err diff --git a/internal/peermsg/private_dir_unix.go b/internal/peermsg/private_dir_unix.go index 664550997..a19051141 100644 --- a/internal/peermsg/private_dir_unix.go +++ b/internal/peermsg/private_dir_unix.go @@ -12,7 +12,7 @@ import ( "golang.org/x/sys/unix" ) -func ensurePrivateDir(path string) error { +func EnsurePrivateDir(path string) error { abs, err := filepath.Abs(path) if err != nil { return err @@ -22,7 +22,13 @@ func ensurePrivateDir(path string) error { if err != nil { return err } - defer unix.Close(parentFD) + // THE DESCRIPTOR THIS FUNCTION OWNS CHANGES ON EVERY ITERATION. The walk + // closes parentFD and replaces it with the child it just opened, so a defer + // that captured the first number would close a descriptor this function no + // longer owns (or one the kernel has since reused) and leave the last one it + // does own open. Close whatever is current when we leave, exactly once, on + // every path. + defer func() { unix.Close(parentFD) }() for index, component := range components { if component == "" { continue diff --git a/internal/peermsg/private_dir_unix_test.go b/internal/peermsg/private_dir_unix_test.go index 3945c58ef..4fd088bf1 100644 --- a/internal/peermsg/private_dir_unix_test.go +++ b/internal/peermsg/private_dir_unix_test.go @@ -3,9 +3,12 @@ package peermsg import ( + "fmt" "os" "path/filepath" "testing" + + "golang.org/x/sys/unix" ) func TestEnsurePrivateDirRejectsSymlink(t *testing.T) { @@ -18,7 +21,7 @@ func TestEnsurePrivateDirRejectsSymlink(t *testing.T) { if err := os.Symlink(target, link); err != nil { t.Fatal(err) } - if err := ensurePrivateDir(link); err == nil { + if err := EnsurePrivateDir(link); err == nil { t.Fatal("expected symlink runtime directory to be rejected") } } @@ -33,7 +36,67 @@ func TestEnsurePrivateDirRejectsSymlinkedParent(t *testing.T) { if err := os.Symlink(target, link); err != nil { t.Fatal(err) } - if err := ensurePrivateDir(filepath.Join(link, "peers")); err == nil { + if err := EnsurePrivateDir(filepath.Join(link, "peers")); err == nil { t.Fatal("expected symlinked parent to be rejected") } } + +// THE DESCRIPTOR THE WALK ENDS ON HAS TO BE CLOSED TOO. +// +// EnsurePrivateDir walks the path one component at a time, closing each parent +// as it opens the child. A deferred close that captured the first descriptor +// closed a number this function no longer owned and left the last one open, so +// every fallback-runtime preparation leaked a descriptor. The lowest free +// descriptor is the watermark: dup(0) hands it back, and a leak per cycle walks +// it upward. Several depths, because the kernel reuses numbers and a single +// shallow fixture can hide the leak by accident. +func TestEnsurePrivateDirClosesTheDescriptorItOwns(t *testing.T) { + lowestFree := func() int { + fd, err := unix.Dup(0) + if err != nil { + t.Fatalf("dup: %v", err) + } + unix.Close(fd) + return fd + } + // The physical path: on macOS t.TempDir() sits under /var, which is itself + // a symlink to /private/var, and this walk refuses a symlink component by + // design. Production resolves its temp root the same way before calling. + base, err := filepath.EvalSymlinks(t.TempDir()) + if err != nil { + t.Fatalf("resolve temp dir: %v", err) + } + for _, depth := range []int{1, 3, 6} { + parts := []string{base} + for i := 0; i < depth; i++ { + parts = append(parts, fmt.Sprintf("level%d", i)) + } + target := filepath.Join(parts...) + + before := lowestFree() + for cycle := 0; cycle < 40; cycle++ { + if err := EnsurePrivateDir(target); err != nil { + t.Fatalf("depth %d cycle %d: %v", depth, cycle, err) + } + } + if after := lowestFree(); after != before { + t.Fatalf("depth %d: lowest free descriptor moved from %d to %d over 40 successful walks, so each one leaks", depth, before, after) + } + + // And the failing walk: a symlink component is refused partway down. + linked := filepath.Join(base, fmt.Sprintf("linked%d", depth)) + if err := os.Symlink(target, linked); err != nil { + t.Fatalf("symlink: %v", err) + } + failing := filepath.Join(linked, "below") + before = lowestFree() + for cycle := 0; cycle < 40; cycle++ { + if err := EnsurePrivateDir(failing); err == nil { + t.Fatalf("depth %d: a symlinked component was not refused", depth) + } + } + if after := lowestFree(); after != before { + t.Fatalf("depth %d: lowest free descriptor moved from %d to %d over 40 refused walks", depth, before, after) + } + } +} diff --git a/internal/peermsg/private_dir_windows.go b/internal/peermsg/private_dir_windows.go index 812b29808..27e961db6 100644 --- a/internal/peermsg/private_dir_windows.go +++ b/internal/peermsg/private_dir_windows.go @@ -12,7 +12,7 @@ import ( "golang.org/x/sys/windows" ) -func ensurePrivateDir(path string) (resultErr error) { +func EnsurePrivateDir(path string) (resultErr error) { abs, err := filepath.Abs(path) if err != nil { return err diff --git a/internal/peermsg/private_dir_windows_test.go b/internal/peermsg/private_dir_windows_test.go index 7e7f751b1..21ae2f13f 100644 --- a/internal/peermsg/private_dir_windows_test.go +++ b/internal/peermsg/private_dir_windows_test.go @@ -54,7 +54,7 @@ func TestEnsurePrivateDirAppliesOwnerOnlyProtectedDACL(t *testing.T) { if !windowsDirectoryDACLContains(t, path, worldSID) { t.Fatal("test setup did not grant the broad Everyone ACE") } - if err := ensurePrivateDir(path); err != nil { + if err := EnsurePrivateDir(path); err != nil { t.Fatal(err) } descriptor, err := windows.GetNamedSecurityInfo( @@ -149,7 +149,7 @@ func windowsDirectoryDACLContains(t *testing.T, path string, wanted *windows.SID func TestSecurePrivateDirectoryRejectsOwnerMismatch(t *testing.T) { path := filepath.Join(t.TempDir(), "private") - if err := ensurePrivateDir(path); err != nil { + if err := EnsurePrivateDir(path); err != nil { t.Fatal(err) } handle, err := openWindowsDirectory(path, windows.FILE_LIST_DIRECTORY|windows.FILE_TRAVERSE|windows.SYNCHRONIZE|windows.READ_CONTROL|windows.WRITE_DAC) @@ -172,7 +172,7 @@ func TestSecurePrivateDirectoryRejectsOwnerMismatch(t *testing.T) { func TestSecurePrivateDirectoryReportsDACLWriteFailure(t *testing.T) { path := filepath.Join(t.TempDir(), "private") - if err := ensurePrivateDir(path); err != nil { + if err := EnsurePrivateDir(path); err != nil { t.Fatal(err) } handle, err := openWindowsDirectory(path, windows.FILE_LIST_DIRECTORY|windows.FILE_TRAVERSE|windows.SYNCHRONIZE|windows.READ_CONTROL) @@ -206,7 +206,7 @@ func TestEnsurePrivateDirRejectsWindowsReparseParent(t *testing.T) { if err := os.Symlink(target, link); err != nil { t.Skipf("cannot create Windows directory symlink: %v", err) } - if err := ensurePrivateDir(filepath.Join(link, "peers")); err == nil { + if err := EnsurePrivateDir(filepath.Join(link, "peers")); err == nil { t.Fatal("expected reparse-point parent to be rejected") } } @@ -217,7 +217,7 @@ func TestEnsurePrivateDirRejectsWindowsFileComponent(t *testing.T) { if err := os.WriteFile(file, []byte("not a directory"), 0o600); err != nil { t.Fatal(err) } - if err := ensurePrivateDir(filepath.Join(file, "peers")); err == nil { + if err := EnsurePrivateDir(filepath.Join(file, "peers")); err == nil { t.Fatal("expected file path component to be rejected") } } diff --git a/internal/peermsg/service.go b/internal/peermsg/service.go index e32735c4a..f1afb8862 100644 --- a/internal/peermsg/service.go +++ b/internal/peermsg/service.go @@ -158,7 +158,7 @@ func New(options Options) (*Service, error) { } // canonicalRuntimePath normalizes aliases in the existing prefix. It does not -// establish a security boundary; ensurePrivateDir validates that boundary when +// establish a security boundary; EnsurePrivateDir validates that boundary when // the service starts. func canonicalRuntimePath(path string) (string, error) { missing := make([]string, 0, 4) @@ -219,10 +219,10 @@ func (service *Service) Start(handler Handler) error { if service.closed { return errors.New("peer messaging: service is closed") } - if err := ensurePrivateDir(service.root); err != nil { + if err := EnsurePrivateDir(service.root); err != nil { return fmt.Errorf("peer messaging: create runtime directory: %w", err) } - if err := ensurePrivateDir(service.registryDir()); err != nil { + if err := EnsurePrivateDir(service.registryDir()); err != nil { return fmt.Errorf("peer messaging: create registry: %w", err) } endpoint, err := service.transport.Endpoint(service.root, service.nonce, service.pid) diff --git a/internal/peermsg/transport_unix.go b/internal/peermsg/transport_unix.go index c32418dbb..651a85176 100644 --- a/internal/peermsg/transport_unix.go +++ b/internal/peermsg/transport_unix.go @@ -44,7 +44,7 @@ func canonicalPrivateDir(path string) (string, error) { if err != nil { return "", err } - if err := ensurePrivateDir(canonical); err != nil { + if err := EnsurePrivateDir(canonical); err != nil { return "", err } return canonical, nil diff --git a/internal/sandbox/analyzer.go b/internal/sandbox/analyzer.go index 7973fa951..d35c7865c 100644 --- a/internal/sandbox/analyzer.go +++ b/internal/sandbox/analyzer.go @@ -20,6 +20,10 @@ type AnalysisResult struct { TooComplex bool // Programs lists the distinct top-level command names found, for diagnostics. Programs []string + // GitInit is set when the script would create a git repository. It is a + // separate signal from Destructive and Network because the decision it feeds + // depends on WHERE the workspace sits, which the analyzer cannot see. + GitInit bool } // destructivePrograms are commands that can irrecoverably destroy data. @@ -183,6 +187,9 @@ func analyzeInto(script string, result *AnalysisResult, seen map[string]bool, de if commandUsesNetwork(prog, rest) { result.Network = true } + if commandCreatesGitRepository(call, prog, rest) { + result.GitInit = true + } if destructivePrograms[prog] || (prog == "rm" && hasRecursiveForce(rest)) || (powerShellRemoveItemPrograms[prog] && hasPowerShellRecursiveForce(rest)) || @@ -277,8 +284,77 @@ func packageManagerOffline(words []string) bool { return false } +// gitGlobalOptionsTakingValue are git's own options that consume the NEXT word. +// +// firstSubcommand skips words beginning with "-" but not the value that follows +// one, so "git -C sub clone " made it answer "sub" and the command was +// classified as touching no network at all. The "--opt=value" spelling is one +// token and needs no entry here. +// +// Keys are lowercase because the analyzer lowercases every word before it gets +// here, which also means -C and -c arrive identically. That is fine: both +// consume a value, and nothing below needs to tell them apart. +var gitGlobalOptionsTakingValue = map[string]bool{ + "-c": true, "-C": true, + "--git-dir": true, "--work-tree": true, "--namespace": true, + "--exec-path": true, "--config-env": true, "--super-prefix": true, + "--attr-source": true, +} + +// gitSubcommand returns git's subcommand, skipping global options AND the values +// they consume. Returns "" when the words carry no subcommand. +func gitSubcommand(words []string) string { + for index := 0; index < len(words); index++ { + word := words[index] + if word == "" { + continue + } + if !strings.HasPrefix(word, "-") { + return word + } + // "--opt=value" is one token and carries its own value. + if strings.HasPrefix(word, "--") && strings.Contains(word, "=") { + continue + } + if gitGlobalOptionsTakingValue[word] { + index++ + } + } + return "" +} + +// commandCreatesGitRepository reports whether this call would create a git +// repository. It shares gitSubcommand with the network gate deliberately, so the +// global options that already bypassed that gate cannot bypass this one either: +// "git -C sub init" and "git -c k=v init" both resolve to init. +// +// init-db is git's original spelling and still works today. clone creates a +// repository as surely as init does, with its config and hooks written by the +// fetch rather than by hand: "git clone ." into a workspace governed by an +// ancestor repository lands a root .git whose carveouts setup never planned, so +// it is the same nested-repository decision. The network prompt is a separate +// gate and granting it does not answer this one. Reported by @gnanam1990. +// +// The subcommand is the one git runs after alias resolution, which git_alias.go +// models: definitions from the command line and from the environment the call +// sets, expanded with git's own quoting and its in-alias option handling. A +// definition this analyzer cannot read makes the command unclassifiable, and +// under this guard that is a refusal. +func commandCreatesGitRepository(call *syntax.CallExpr, prog string, args []*syntax.Word) bool { + if prog != "git" { + return false + } + words := literalWordTexts(args) + aliases := gitAliasTable{} + if !gitEnvironmentAliases(aliases, callAssignments(call, args)) { + return true + } + gitInlineAliases(aliases, words) + return gitSubcommandCreatesRepository(gitSubcommand(words), aliases) +} + func gitUsesNetwork(words []string) bool { - switch firstSubcommand(words, nil) { + switch gitSubcommand(words) { case "clone", "fetch", "pull", "push", "ls-remote", "archive": return true default: diff --git a/internal/sandbox/engine.go b/internal/sandbox/engine.go index ed7440780..80a95c777 100644 --- a/internal/sandbox/engine.go +++ b/internal/sandbox/engine.go @@ -367,6 +367,19 @@ func (engine *Engine) Evaluate(ctx context.Context, request Request) Decision { return deny(request, risk, block.Code, block.Path, block.Reason, false) } } + // BEFORE THE NETWORK GATE, because this one cannot be approved away. + // + // `git clone` is both a network command and a repository-creating one. With + // the network gate first, a clone in a governed nested workspace under the + // default NetworkDeny came back as a network prompt: the operator was asked + // to approve network access for a command that approval could never make + // runnable, and the reason that actually decides it was never shown. The + // refusal below does not depend on policy, permission or mode, so it is the + // answer whatever the network gate would have said, and it is the one with a + // remedy the operator can act on. Reported by @jatmn. + if HasRiskCategory(risk, "nested_git_init") { + return deny(request, risk, BlockNestedGitInit, "", ReasonNestedGitInit, false) + } netMode := engine.effectiveNetworkMode(policy) if netMode == NetworkDeny && HasRiskCategory(risk, "network") && !engine.toolNetworkExempt(request) { if request.SideEffect == SideEffectShell && request.PermissionMode != PermissionUnsafe { diff --git a/internal/sandbox/git_alias.go b/internal/sandbox/git_alias.go new file mode 100644 index 000000000..92fd09660 --- /dev/null +++ b/internal/sandbox/git_alias.go @@ -0,0 +1,376 @@ +package sandbox + +import ( + "strconv" + "strings" + + "mvdan.cc/sh/v3/syntax" +) + +// THE NESTED-REPOSITORY GUARD CLASSIFIES THE COMMAND GIT RUNS, NOT THE ONE +// TYPED. Git rewrites its argument vector before choosing a builtin: an alias +// defined on the command line, in the environment or through --config-env is +// split with shell-style quoting, its own leading global options are applied, +// and the result is looked up again until a builtin is reached. Reading the +// literal token, or the first whitespace-separated field of an alias value, +// let `git -c alias.a='-c alias.b=init b' a .` and `git -c 'alias.a="init"' a .` +// create a repository the workspace profile carves nothing out for. What +// follows models that rewrite, and where the model runs out it refuses rather +// than guesses: under this guard the safe answer to "could this make a +// repository" is yes. Reported by @gnanam1990. + +// gitAliasTable maps a lowercased alias name to every definition the command +// supplies for it. Git keeps one value per key, the last one written, and the +// order in which the environment, -c and --config-env feed that list is not +// something this guard needs to reproduce: a definition git would discard but +// which reaches init here costs one refusal, while the opposite mistake costs +// a repository. So every definition is kept, and any of them reaching init +// refuses. For the same reason an alias that shadows a builtin, which git +// ignores, is followed here rather than skipped: this file carries no list of +// git's commands, and refusing `-c alias.status=init status` loses nothing. +type gitAliasTable map[string][]gitAliasDefinition + +// gitAliasDefinition is one alias value. unknown marks a definition whose text +// cannot be read statically, such as --config-env=alias.x=VAR or a +// GIT_CONFIG_VALUE_n written as $var: git will run whatever it holds, and +// this guard cannot say what that is. +type gitAliasDefinition struct { + text string + unknown bool +} + +func (table gitAliasTable) define(name, text string) { + name = strings.ToLower(strings.TrimSpace(name)) + if name == "" { + return + } + table[name] = append(table[name], gitAliasDefinition{text: strings.ToLower(text)}) +} + +func (table gitAliasTable) defineUnknown(name string) { + name = strings.ToLower(strings.TrimSpace(name)) + if name == "" { + return + } + table[name] = append(table[name], gitAliasDefinition{unknown: true}) +} + +// defineSetting records a `key=value` config setting when key names an alias. +// readable is false when the value is only known by reference, which makes the +// alias unknown rather than absent. A setting without "=" is a boolean in git +// (`alias.x` expands to "true") and defines nothing worth following. +func (table gitAliasTable) defineSetting(setting string, readable bool) { + name, value, found := strings.Cut(setting, "=") + if !found { + return + } + alias, ok := strings.CutPrefix(strings.ToLower(strings.TrimSpace(name)), "alias.") + if !ok { + return + } + if !readable { + table.defineUnknown(alias) + return + } + table.define(alias, value) +} + +// splitGitAliasExpansion tokenizes an alias value the way git's split_cmdline +// does: whitespace separates words, a single or double quote opens a quoted +// span anywhere inside a word, and a backslash outside single quotes escapes +// the next byte. An unclosed quote or a trailing backslash is a fatal error in +// git ("bad alias string"), so ok is false and nothing would have run. +func splitGitAliasExpansion(text string) (argv []string, ok bool) { + var current strings.Builder + var quoted byte + inWord := false + for index := 0; index < len(text); index++ { + c := text[index] + switch { + case quoted == 0 && isGitSpace(c): + if inWord { + argv = append(argv, current.String()) + current.Reset() + inWord = false + } + case quoted == 0 && (c == '\'' || c == '"'): + quoted = c + inWord = true + case c == quoted: + quoted = 0 + default: + if c == '\\' && quoted != '\'' { + index++ + if index >= len(text) { + return nil, false + } + c = text[index] + } + current.WriteByte(c) + inWord = true + } + } + if quoted != 0 { + return nil, false + } + if inWord { + argv = append(argv, current.String()) + } + return argv, true +} + +func isGitSpace(c byte) bool { + switch c { + case ' ', '\t', '\n', '\r', '\v', '\f': + return true + } + return false +} + +// gitAliasExpansionSubcommand runs a split expansion through git's own +// handling of global options inside an alias and returns the subcommand git +// would look up next. Inside an expansion git accepts -c and --config-env, +// which push settings the next lookup sees (that is how `-c alias.b=init b` +// reaches init), and the pager switches, which change nothing that runs. An +// option that changes the environment (-C, --git-dir, --bare, ...) is fatal +// inside an alias, and an expansion made only of options is an empty alias; +// both return ok false. So does an option this guard does not model, because +// "git would refuse this" and "this guard cannot read this" end the same way. +func gitAliasExpansionSubcommand(argv []string, aliases gitAliasTable) (subcommand string, ok bool) { + for index := 0; index < len(argv); index++ { + word := argv[index] + if word == "" { + return "", false + } + if !strings.HasPrefix(word, "-") { + return word, true + } + switch { + case word == "-c", word == "--config-env": + if index+1 >= len(argv) { + return "", false + } + index++ + aliases.defineSetting(argv[index], word == "-c") + case strings.HasPrefix(word, "--config-env="): + aliases.defineSetting(strings.TrimPrefix(word, "--config-env="), false) + case word == "-p", word == "--paginate", word == "-P", word == "--no-pager": + default: + return "", false + } + } + return "", false +} + +// gitSubcommandCreatesRepository reports whether subcommand, resolved through +// aliases the way git resolves it, is init, init-db or clone. The chain is +// followed to its end with the names on the current path kept, so a cycle is a +// refusal here the way "alias loop detected" is fatal in git. A shell alias +// runs a program this analyzer cannot classify; an expansion git cannot split, +// cannot run, or that leaves no subcommand behind is the same case, and so is +// a definition whose text is not readable. Every one of those refuses. +func gitSubcommandCreatesRepository(subcommand string, aliases gitAliasTable) bool { + return gitAliasPathCreatesRepository(subcommand, aliases, nil) +} + +func gitAliasPathCreatesRepository(subcommand string, aliases gitAliasTable, path []string) bool { + switch subcommand { + case "init", "init-db", "clone": + return true + } + definitions, ok := aliases[subcommand] + if !ok { + return false + } + for _, seen := range path { + if seen == subcommand { + return true + } + } + path = append(path, subcommand) + for _, definition := range definitions { + if definition.unknown || strings.HasPrefix(definition.text, "!") { + return true + } + argv, ok := splitGitAliasExpansion(definition.text) + if !ok { + return true + } + next, ok := gitAliasExpansionSubcommand(argv, aliases) + if !ok { + return true + } + if gitAliasPathCreatesRepository(next, aliases, path) { + return true + } + } + return false +} + +// gitInlineAliases records the alias definitions supplied ahead of the +// subcommand: `-c alias.NAME=EXPANSION`, and `--config-env alias.NAME=VAR` in +// both spellings, whose value lives in a variable this analyzer does not read. +// Git accepts no attached `-cNAME=VALUE` spelling ("unknown option"), so none +// is recognised here. Other global options that take a value are stepped over +// the way gitSubcommand steps over them, so a value that happens to look like +// a subcommand does not end the scan early. +func gitInlineAliases(aliases gitAliasTable, words []string) { + for index := 0; index < len(words); index++ { + word := words[index] + if word == "" { + continue + } + if !strings.HasPrefix(word, "-") { + return + } + switch { + case word == "-c", word == "--config-env": + if index+1 < len(words) { + index++ + aliases.defineSetting(words[index], word == "-c") + } + case strings.HasPrefix(word, "--config-env="): + aliases.defineSetting(strings.TrimPrefix(word, "--config-env="), false) + case strings.HasPrefix(word, "--") && strings.Contains(word, "="): + case gitGlobalOptionsTakingValue[word]: + index++ + } + } +} + +// shellAssignment is one NAME=value the command hands to git's environment. +// readable is false when the value is an expansion this analyzer cannot see. +type shellAssignment struct { + value string + readable bool +} + +// callAssignments collects the assignments in force for the program: the +// shell prefix (FOO=bar git ...) and the NAME=value words an env wrapper +// forwards (env FOO=bar git ...), which sit ahead of the program word. +func callAssignments(call *syntax.CallExpr, rest []*syntax.Word) map[string]shellAssignment { + assignments := map[string]shellAssignment{} + if call == nil { + return assignments + } + for _, assign := range call.Assigns { + if assign == nil || assign.Name == nil { + continue + } + readable := !assign.Append && assign.Array == nil && assign.Index == nil && + (assign.Value == nil || isLiteralWord(assign.Value)) + assignments[assign.Name.Value] = shellAssignment{value: wordText(assign.Value), readable: readable} + } + program := len(call.Args) - len(rest) - 1 + for index := 0; index < program && index < len(call.Args); index++ { + name, value, found := strings.Cut(wordText(call.Args[index]), "=") + if !found || !isEnvName(name) { + continue + } + assignments[name] = shellAssignment{value: value, readable: isLiteralWord(call.Args[index])} + } + return assignments +} + +func isEnvName(name string) bool { + if name == "" { + return false + } + for index, r := range name { + switch { + case r == '_', r >= 'a' && r <= 'z', r >= 'A' && r <= 'Z': + case r >= '0' && r <= '9' && index > 0: + default: + return false + } + } + return true +} + +// gitEnvironmentAliases records the alias definitions the command hands git +// through its environment, which git reads exactly as it reads -c: +// GIT_CONFIG_COUNT with GIT_CONFIG_KEY_n and GIT_CONFIG_VALUE_n, and +// GIT_CONFIG_PARAMETERS. It reports false when a definition this guard needs +// cannot be read: a count, a key or the parameter list written as an +// expansion, or the value of an alias key written that way. A count git would +// reject, or a numbered key with no value, is reported the same way; git dies +// on those, so refusing them costs nothing. +func gitEnvironmentAliases(aliases gitAliasTable, env map[string]shellAssignment) bool { + if parameters, ok := env["GIT_CONFIG_PARAMETERS"]; ok { + if !parameters.readable { + return false + } + settings, ok := splitGitConfigParameters(parameters.value) + if !ok { + return false + } + for _, setting := range settings { + aliases.defineSetting(setting, true) + } + } + count, ok := env["GIT_CONFIG_COUNT"] + if !ok { + return true + } + if !count.readable { + return false + } + n, err := strconv.Atoi(strings.TrimSpace(count.value)) + if err != nil || n < 0 || n > 1024 { + return false + } + for index := 0; index < n; index++ { + key, ok := env["GIT_CONFIG_KEY_"+strconv.Itoa(index)] + if !ok || !key.readable { + return false + } + value, ok := env["GIT_CONFIG_VALUE_"+strconv.Itoa(index)] + if !ok { + return false + } + aliases.defineSetting(key.value+"="+value.value, value.readable) + } + return true +} + +// splitGitConfigParameters dequotes GIT_CONFIG_PARAMETERS the way git does: +// whitespace-separated entries, each single-quoted, with `'\”` and `'\!'` +// standing for a quote or a bang inside one. Anything else is "bogus format" +// in git, a fatal error, so ok is false. +func splitGitConfigParameters(text string) (settings []string, ok bool) { + index := 0 + for { + for index < len(text) && isGitSpace(text[index]) { + index++ + } + if index >= len(text) { + return settings, true + } + if text[index] != '\'' { + return nil, false + } + index++ + var current strings.Builder + for { + if index >= len(text) { + return nil, false + } + c := text[index] + index++ + if c != '\'' { + current.WriteByte(c) + continue + } + if index+2 < len(text) && text[index] == '\\' && (text[index+1] == '\'' || text[index+1] == '!') && text[index+2] == '\'' { + current.WriteByte(text[index+1]) + index += 3 + continue + } + if index >= len(text) || isGitSpace(text[index]) { + settings = append(settings, current.String()) + break + } + return nil, false + } + } +} diff --git a/internal/sandbox/git_alias_test.go b/internal/sandbox/git_alias_test.go new file mode 100644 index 000000000..2393e5e4e --- /dev/null +++ b/internal/sandbox/git_alias_test.go @@ -0,0 +1,103 @@ +package sandbox + +import ( + "reflect" + "testing" +) + +// The tokenizer has to agree with git's split_cmdline on the forms that decide +// which subcommand runs: quotes anywhere in a word, backslash escapes outside +// single quotes, and the two errors git dies on. +func TestSplitGitAliasExpansionMatchesGitQuoting(t *testing.T) { + for _, tc := range []struct { + expansion string + want []string + ok bool + }{ + {"init", []string{"init"}, true}, + {`"init"`, []string{"init"}, true}, + {`in"it"`, []string{"init"}, true}, + {`in\it`, []string{"init"}, true}, + {`'in\it'`, []string{`in\it`}, true}, + {`"in\"it"`, []string{`in"it`}, true}, + {`-c alias.b=init b`, []string{"-c", "alias.b=init", "b"}, true}, + {`-c "alias.b=init x" b`, []string{"-c", "alias.b=init x", "b"}, true}, + {`'-c' 'alias.b=init' b`, []string{"-c", "alias.b=init", "b"}, true}, + {" init \t ", []string{"init"}, true}, + {"", nil, true}, + {`"init`, nil, false}, + {`'init`, nil, false}, + {`init\`, nil, false}, + } { + got, ok := splitGitAliasExpansion(tc.expansion) + if ok != tc.ok || !reflect.DeepEqual(got, tc.want) { + t.Errorf("splitGitAliasExpansion(%q) = %q, %v; want %q, %v", tc.expansion, got, ok, tc.want, tc.ok) + } + } +} + +// GIT_CONFIG_PARAMETERS is git's own single-quoted list; an entry git would +// call bogus makes the whole variable unreadable rather than partly read. +func TestSplitGitConfigParametersDequotesLikeGit(t *testing.T) { + for _, tc := range []struct { + text string + want []string + ok bool + }{ + {`'alias.a=init'`, []string{"alias.a=init"}, true}, + {`'alias.a=init' 'x=y'`, []string{"alias.a=init", "x=y"}, true}, + {` 'alias.a=in'\''it' `, []string{"alias.a=in'it"}, true}, + {`'alias.a=in'\!'it'`, []string{"alias.a=in!it"}, true}, + {"", nil, true}, + {`alias.a=init`, nil, false}, + {`'alias.a=init`, nil, false}, + {`'a'x`, nil, false}, + } { + got, ok := splitGitConfigParameters(tc.text) + if ok != tc.ok || !reflect.DeepEqual(got, tc.want) { + t.Errorf("splitGitConfigParameters(%q) = %q, %v; want %q, %v", tc.text, got, ok, tc.want, tc.ok) + } + } +} + +// Resolution applies git's in-alias option handling: a -c inside an expansion +// defines the alias the next lookup uses, pager switches are stepped over, an +// environment-changing option or an unreadable definition refuses, and only a +// chain that ends in a real repository-creating command is creation. +func TestGitAliasResolutionFollowsExpansionOptions(t *testing.T) { + for _, tc := range []struct { + name string + subcommand string + aliases map[string]string + unknown []string + creates bool + }{ + {name: "option inside expansion defines the next alias", subcommand: "a", aliases: map[string]string{"a": "-c alias.b=init b"}, creates: true}, + {name: "quoted expansion", subcommand: "a", aliases: map[string]string{"a": `"init"`}, creates: true}, + {name: "pager switch does not hide init", subcommand: "a", aliases: map[string]string{"a": "--no-pager init"}, creates: true}, + {name: "environment-changing option refuses", subcommand: "a", aliases: map[string]string{"a": "-C . init"}, creates: true}, + {name: "unclosed quote refuses", subcommand: "a", aliases: map[string]string{"a": `"in`}, creates: true}, + {name: "options only refuses", subcommand: "a", aliases: map[string]string{"a": "-c x=y"}, creates: true}, + {name: "config-env inside expansion is unreadable", subcommand: "a", aliases: map[string]string{"a": "--config-env=alias.b=var b"}, creates: true}, + {name: "redefinition loop refuses", subcommand: "a", aliases: map[string]string{"a": "-c alias.a=init a"}, creates: true}, + {name: "unknown definition refuses", subcommand: "a", unknown: []string{"a"}, creates: true}, + {name: "harmless option-bearing alias", subcommand: "st", aliases: map[string]string{"st": "-c color.ui=always status"}, creates: false}, + {name: "harmless quoted alias", subcommand: "st", aliases: map[string]string{"st": `"status"`}, creates: false}, + {name: "harmless pager alias", subcommand: "l", aliases: map[string]string{"l": "--no-pager log --oneline"}, creates: false}, + {name: "harmless chain through quotes", subcommand: "s", aliases: map[string]string{"s": "st", "st": `sta"tus"`}, creates: false}, + {name: "not an alias", subcommand: "status", creates: false}, + } { + t.Run(tc.name, func(t *testing.T) { + table := gitAliasTable{} + for name, expansion := range tc.aliases { + table.define(name, expansion) + } + for _, name := range tc.unknown { + table.defineUnknown(name) + } + if got := gitSubcommandCreatesRepository(tc.subcommand, table); got != tc.creates { + t.Fatalf("gitSubcommandCreatesRepository(%q) = %v, want %v", tc.subcommand, got, tc.creates) + } + }) + } +} diff --git a/internal/sandbox/git_global_options_test.go b/internal/sandbox/git_global_options_test.go new file mode 100644 index 000000000..28a40e03b --- /dev/null +++ b/internal/sandbox/git_global_options_test.go @@ -0,0 +1,58 @@ +package sandbox + +import "testing" + +// GIT'S GLOBAL OPTIONS CONSUME THE NEXT WORD, AND THE CLASSIFIER HAS TO KNOW. +// +// firstSubcommand skips words beginning with "-" but not the value that follows +// one. So "git -C sub clone " made it answer "sub", the command classified +// as touching no network at all, and the network gate never fired. Two spellings +// an ordinary user types by habit walked straight through it. +// +// Driven through AnalyzeCommand rather than the helper, because the gate reads +// the analysis and a helper-level assertion would not have caught this. +func TestGitGlobalOptionsDoNotHideTheSubcommand(t *testing.T) { + for _, testCase := range []struct { + script string + want bool + }{ + {script: "git clone https://example.com/x", want: true}, + {script: "git -C sub clone https://example.com/x", want: true}, + {script: "git -c user.name=x clone https://example.com/x", want: true}, + {script: "git -c a=b -C sub clone https://example.com/x", want: true}, + {script: "git --git-dir=/tmp/g clone https://example.com/x", want: true}, + {script: "git --namespace ns fetch origin", want: true}, + {script: "git -C sub push origin main", want: true}, + // Local-only work stays local, or the fix would be "call everything network". + {script: "git status", want: false}, + {script: "git -C sub status", want: false}, + {script: "git -c a=b commit -m x", want: false}, + {script: "git init", want: false}, + } { + if got := AnalyzeCommand(testCase.script).Network; got != testCase.want { + t.Errorf("AnalyzeCommand(%q).Network = %v, want %v", testCase.script, got, testCase.want) + } + } +} + +// The subcommand parser itself, including the option that consumes a value and +// the self-contained "--opt=value" spelling that does not. +func TestGitSubcommandSkipsOptionValues(t *testing.T) { + for _, testCase := range []struct { + words []string + want string + }{ + {words: []string{"clone", "url"}, want: "clone"}, + {words: []string{"-c", "sub", "clone"}, want: "clone"}, + {words: []string{"--git-dir=/tmp/g", "clone"}, want: "clone"}, + {words: []string{"--namespace", "ns", "fetch"}, want: "fetch"}, + {words: []string{"-c", "a=b", "-c", "sub", "push"}, want: "push"}, + {words: []string{"--no-pager", "status"}, want: "status"}, + {words: []string{}, want: ""}, + {words: []string{"-c", "a=b"}, want: ""}, + } { + if got := gitSubcommand(testCase.words); got != testCase.want { + t.Errorf("gitSubcommand(%v) = %q, want %q", testCase.words, got, testCase.want) + } + } +} diff --git a/internal/sandbox/git_nested_init_refusal_test.go b/internal/sandbox/git_nested_init_refusal_test.go new file mode 100644 index 000000000..e1742b591 --- /dev/null +++ b/internal/sandbox/git_nested_init_refusal_test.go @@ -0,0 +1,451 @@ +package sandbox + +import ( + "context" + "os" + "path/filepath" + "testing" +) + +// nestedGitWorkspace returns an ancestor repository and a workspace inside it +// that carries no git metadata of its own, which is the shape the whole refusal +// is about. +func nestedGitWorkspace(t *testing.T) (ancestor string, workspace string) { + t.Helper() + ancestor = t.TempDir() + if err := os.MkdirAll(filepath.Join(ancestor, ".git"), 0o700); err != nil { + t.Fatal(err) + } + workspace = filepath.Join(ancestor, "packages", "app") + if err := os.MkdirAll(workspace, 0o700); err != nil { + t.Fatal(err) + } + if _, err := os.Lstat(filepath.Join(workspace, ".git")); err == nil { + t.Fatal("SETUP INVALID: the workspace carries its own .git, which is the protected case, not this one") + } + return ancestor, workspace +} + +func gitCommandRequest(workspace string, command string) Request { + return Request{ + ToolName: "bash", + SideEffect: SideEffectShell, + Permission: PermissionPrompt, + PermissionMode: PermissionModeAsk, + WorkspaceRoot: workspace, + Args: map[string]any{"command": command}, + } +} + +func gitWorkspaceEngine(t *testing.T, workspace string) *Engine { + t.Helper() + return NewEngine(EngineOptions{ + WorkspaceRoot: workspace, + Policy: DefaultPolicy(), + Backend: nativeWrappingBackend, + }) +} + +// THE LIFECYCLE, END TO END: the workspace is protected at write time AND the +// command that would create a repository there is refused with an explanation. +// +// The setup guard is still load-bearing, but it now asserts the opposite of +// what it once did. The carveouts were restored for a nested workspace, because +// skipping them left nothing denying writes to /.git and a repository +// could be assembled there without running git at all. So the refusal is no +// longer the only thing standing between a sandboxed command and a writable +// config; it is the layer that tells an operator why, while the deny ACE is +// what stops it. If the carveouts disappear again this fails here, because the +// refusal on its own does not cover mkdir. +func TestNestedWorkspaceRefusesGitInit(t *testing.T) { + ancestor, workspace := nestedGitWorkspace(t) + + carved := map[string]bool{} + for _, carveout := range gitMetadataWriteCarveouts(workspace) { + carved[filepath.Base(carveout)] = true + } + if !carved["config"] || !carved["hooks"] { + t.Fatalf("SETUP INVALID: the nested workspace planned %v, so config and hooks are not denied and the refusal below is the only protection again", gitMetadataWriteCarveouts(workspace)) + } + + engine := gitWorkspaceEngine(t, workspace) + decision := engine.Evaluate(context.Background(), gitCommandRequest(workspace, "git init")) + if decision.Action != ActionDeny { + t.Fatalf("git init inside the repository at %s was %s, not denied; the repository it creates gets a writable config and hooks under the workspace grant", ancestor, decision.Action) + } + if decision.Block == nil || decision.Block.Code != BlockNestedGitInit { + t.Fatalf("block = %#v, want code %s so the operator gets the nested-repository explanation and not a generic denial", decision.Block, BlockNestedGitInit) + } +} + +// The refusal sits ahead of every allow path in Evaluate, so a tool already +// cleared by permissions does not walk through it. +func TestNestedGitInitRefusalOutranksAnAllowPermission(t *testing.T) { + _, workspace := nestedGitWorkspace(t) + engine := gitWorkspaceEngine(t, workspace) + + request := gitCommandRequest(workspace, "git init") + request.Permission = PermissionAllow + if decision := engine.Evaluate(context.Background(), request); decision.Action != ActionDeny { + t.Fatalf("an allow permission carried git init through: %#v", decision) + } + + request = gitCommandRequest(workspace, "git init") + request.PermissionMode = PermissionUnsafe + request.PermissionGranted = true + if decision := engine.Evaluate(context.Background(), request); decision.Action != ActionDeny { + t.Fatalf("unsafe mode carried git init through: %#v", decision) + } +} + +// The global options that bypassed the network gate must not bypass this one. +// These are the exact spellings that classified a clone as network=false before +// gitSubcommand learned which options take a value. +func TestNestedGitInitRefusalSurvivesGitGlobalOptions(t *testing.T) { + _, workspace := nestedGitWorkspace(t) + engine := gitWorkspaceEngine(t, workspace) + + for _, command := range []string{ + "git init", + "git init-db", + "git -C packages/app init", + "git -c core.hooksPath=/tmp/h init", + "git --git-dir=/tmp/g init", + "git --namespace ns init", + "sh -c \"git init\"", + "sudo git init", + "git status && git init", + } { + t.Run(command, func(t *testing.T) { + decision := engine.Evaluate(context.Background(), gitCommandRequest(workspace, command)) + if decision.Action != ActionDeny { + t.Fatalf("%q was %s, not denied", command, decision.Action) + } + }) + } +} + +// CONTROL: ordinary git in a nested workspace is untouched. +// +// Discovery still walks up to the ancestor and every command that uses it keeps +// working; only creation is refused. Without this, returning deny for anything +// containing the word git would pass the tests above. +func TestNestedWorkspaceStillRunsOrdinaryGitCommands(t *testing.T) { + _, workspace := nestedGitWorkspace(t) + engine := gitWorkspaceEngine(t, workspace) + + for _, command := range []string{ + "git status", + "git add .", + "git commit -m init", + "git log --oneline", + "git worktree list", + "git submodule update --init", + } { + t.Run(command, func(t *testing.T) { + decision := engine.Evaluate(context.Background(), gitCommandRequest(workspace, command)) + if decision.Action == ActionDeny { + t.Fatalf("%q was denied in a nested workspace: %s", command, decision.Reason) + } + }) + } +} + +// CONTROL: a standalone workspace still creates repositories. +// +// It keeps its materialized config and hooks carveouts, so the protection the +// refusal stands in for is actually there and the command is allowed. +func TestStandaloneWorkspaceStillAllowsGitInit(t *testing.T) { + workspace := t.TempDir() + if gitMetadataGovernedByAncestor(workspace) { + t.Skip("this temp directory sits inside a repository, so it is not the standalone case") + } + if len(gitMetadataWriteCarveouts(workspace)) == 0 { + t.Fatal("SETUP INVALID: a standalone workspace planned no carveouts, so it has no protection to allow git init against") + } + + engine := gitWorkspaceEngine(t, workspace) + decision := engine.Evaluate(context.Background(), gitCommandRequest(workspace, "git init")) + if decision.Action == ActionDeny { + t.Fatalf("git init was refused in a standalone workspace: %s", decision.Reason) + } +} + +// CONTROL: a linked worktree is a workspace whose .git is a FILE. +// +// It owns git metadata, keeps its pointer-file carveout, and is not somebody +// else's repository, so nothing here is refused. Zero's own development +// worktrees are this shape, so getting it wrong would refuse git init for the +// people writing this code. +func TestLinkedWorktreeWorkspaceStillAllowsGitInit(t *testing.T) { + ancestor := t.TempDir() + if err := os.MkdirAll(filepath.Join(ancestor, ".git", "worktrees", "w"), 0o700); err != nil { + t.Fatal(err) + } + workspace := filepath.Join(ancestor, "checkout") + if err := os.MkdirAll(workspace, 0o700); err != nil { + t.Fatal(err) + } + pointer := filepath.Join(workspace, ".git") + if err := os.WriteFile(pointer, []byte("gitdir: ../.git/worktrees/w\n"), 0o600); err != nil { + t.Fatal(err) + } + + // SETUP: the pointer carveout is intact, or this proves nothing about + // worktrees keeping their protection. + specs := gitMetadataWriteCarveoutSpecs(workspace) + if len(specs) != 1 || specs[0].Path != pointer || !specs[0].IsFile { + t.Fatalf("SETUP INVALID: linked worktree carveouts = %+v, want the file-shaped pointer %s", specs, pointer) + } + + engine := gitWorkspaceEngine(t, workspace) + if decision := engine.Evaluate(context.Background(), gitCommandRequest(workspace, "git init")); decision.Action == ActionDeny { + t.Fatalf("git init was refused in a linked worktree, which owns its git metadata: %s", decision.Reason) + } +} + +// And the workspace-level question is asked of the workspace itself, so a +// workspace that already carries .git is never treated as nested no matter what +// is above it. +func TestWorkspaceOwningGitIsNotTreatedAsNested(t *testing.T) { + ancestor := t.TempDir() + if err := os.MkdirAll(filepath.Join(ancestor, ".git"), 0o700); err != nil { + t.Fatal(err) + } + workspace := filepath.Join(ancestor, "vendored") + if err := os.MkdirAll(filepath.Join(workspace, ".git"), 0o700); err != nil { + t.Fatal(err) + } + if workspaceGovernedByAncestorRepository(workspace) { + t.Fatal("a workspace that owns .git was reported as governed by its ancestor, so its own protection would be refused instead of applied") + } +} + +// CLONE IS CREATION TOO, AND THE NETWORK GRANT DOES NOT ANSWER FOR IT. +// +// "git clone ." into a workspace governed by an ancestor repository lands +// a root .git whose config and hooks setup never planned carveouts for, exactly +// the state git init produces, but the guard only knew the init spelling. With +// network explicitly allowed and the shell grant already given, the clone came +// back as an ordinary allow. The repository-creation refusal is a separate +// decision from the network prompt and has to stand on its own. Reported by +// @gnanam1990. +func TestNestedWorkspaceRefusesGitCloneIntoTheRoot(t *testing.T) { + ancestor, workspace := nestedGitWorkspace(t) + policy := DefaultPolicy() + policy.Network = NetworkAllow + engine := NewEngine(EngineOptions{WorkspaceRoot: workspace, Policy: policy, Backend: nativeWrappingBackend}) + for _, command := range []string{ + "git clone https://example.invalid/repo.git .", + "git -C . clone https://example.invalid/repo.git .", + "git clone --depth 1 https://example.invalid/repo.git vendor/dep", + } { + request := gitCommandRequest(workspace, command) + request.Permission = PermissionAllow + request.PermissionGranted = true + decision := engine.Evaluate(context.Background(), request) + if decision.Action != ActionDeny { + t.Fatalf("%q inside the repository at %s was %s, not denied; the repository it creates gets a writable config and hooks under the workspace grant", command, ancestor, decision.Action) + } + if decision.Block == nil || decision.Block.Code != BlockNestedGitInit { + t.Fatalf("%q block = %#v, want code %s", command, decision.Block, BlockNestedGitInit) + } + } + // CONTROLS: ordinary git operations and the submodule init form stay clear + // of this guard, and a standalone workspace is not governed by anything. + for _, command := range []string{"git status", "git submodule update --init", "git fetch origin"} { + request := gitCommandRequest(workspace, command) + request.Permission = PermissionAllow + request.PermissionGranted = true + if decision := engine.Evaluate(context.Background(), request); decision.Action == ActionDeny && decision.Block != nil && decision.Block.Code == BlockNestedGitInit { + t.Fatalf("%q was refused as nested repository creation; it creates no repository", command) + } + } + standalone := t.TempDir() + if workspaceGovernedByAncestorRepository(standalone) { + t.Fatalf("SETUP INVALID: %s is governed by an ancestor repository", standalone) + } + request := gitCommandRequest(standalone, "git clone https://example.invalid/repo.git .") + request.Permission = PermissionAllow + request.PermissionGranted = true + if decision := gitWorkspaceEngine(t, standalone).Evaluate(context.Background(), request); decision.Block != nil && decision.Block.Code == BlockNestedGitInit { + t.Fatalf("clone into a standalone workspace was refused as nested: %#v", decision) + } +} + +// AN INLINE ALIAS IS STILL INIT. The refusal used to read the literal +// subcommand token, and `git -c alias.bootstrap=init bootstrap` passed it with +// shell and network granted and created a repository the profile carves nothing +// out for. The alias is resolved to the subcommand git runs; a chained alias +// follows the chain; a shell alias, which this analyzer cannot classify, counts +// as creation under this guard and nowhere else. +func TestNestedWorkspaceRefusesAnInlineGitAliasForInit(t *testing.T) { + ancestor, workspace := nestedGitWorkspace(t) + policy := DefaultPolicy() + policy.Network = NetworkAllow + engine := NewEngine(EngineOptions{WorkspaceRoot: workspace, Policy: policy, Backend: nativeWrappingBackend}) + for _, command := range []string{ + "git init", + "git -c alias.bootstrap=init bootstrap", + "git -c alias.mk=init-db mk", + "git -c alias.get=clone get https://example.invalid/repo.git .", + "git -c alias.two=bootstrap -c alias.bootstrap=init two", + "git -c alias.a=b -c alias.b=c -c alias.c=d -c alias.d=e -c alias.e=init a .", + "git -c alias.a=b -c alias.b=a a", + "git -c alias.dangling=nowhere -c alias.nowhere= dangling", + "git -C . -c alias.bootstrap=init bootstrap", + "git -c alias.sh=!sh -c 'git init' sh", + // THE EXPANSION IS PARSED THE WAY GIT PARSES IT. A -c inside the + // expansion defines the alias the next lookup uses, and quotes are + // stripped before the subcommand is read; both of the first two + // spellings created a repository while only the first whitespace + // field of the expansion was classified. + "git -c alias.a='-c alias.b=init b' a .", + "git -c 'alias.a=\"init\"' a .", + "git -c 'alias.a=in\"it\"' a", + "git -c 'alias.a=in\\it' a", + "git -c alias.a='--no-pager init' a", + "git -c alias.a='-c alias.a=init a' a", + // Git dies on these before running anything; refusing them costs + // nothing and keeps the rule that an expansion this analyzer cannot + // read is not evidence of safety. + "git -c 'alias.a=\"in' a", + "git -c 'alias.a=-C . init' a", + // An alias defined by reference, or through the environment the call + // sets, is the same command with its definition moved out of sight. + "git --config-env=alias.a=zz a .", + "git --config-env alias.a=zz a .", + "GIT_CONFIG_COUNT=1 GIT_CONFIG_KEY_0=alias.a GIT_CONFIG_VALUE_0=init git a .", + "env GIT_CONFIG_COUNT=1 GIT_CONFIG_KEY_0=alias.a GIT_CONFIG_VALUE_0=init git a .", + "GIT_CONFIG_PARAMETERS=\"'alias.a=init'\" git a .", + "GIT_CONFIG_PARAMETERS=$params git a .", + "GIT_CONFIG_COUNT=1 GIT_CONFIG_KEY_0=alias.a GIT_CONFIG_VALUE_0=$cmd git a .", + } { + request := gitCommandRequest(workspace, command) + request.Permission = PermissionAllow + request.PermissionGranted = true + decision := engine.Evaluate(context.Background(), request) + if decision.Action != ActionDeny || decision.Block == nil || decision.Block.Code != BlockNestedGitInit { + t.Fatalf("%q inside the repository at %s was not refused as nested repository creation: action=%s block=%#v", command, ancestor, decision.Action, decision.Block) + } + } + // CONTROLS: a harmless alias, an unrelated -c setting, ordinary use, and + // the quoted, option-bearing and environment-defined spellings of an alias + // that runs something harmless. + for _, command := range []string{ + "git -c alias.st=status st", + "git -c core.autocrlf=false status", + "git -c alias.bootstrap=init status", + "git -c alias.lg=log lg --oneline", + "git -c alias.a=b -c alias.b=c -c alias.c=d -c alias.d=e -c alias.e=status a", + "git -c alias.st='-c color.ui=always status' st", + "git -c 'alias.st=\"status\"' st", + "git -c 'alias.st=sta\"tus\"' st", + "git -c alias.l='--no-pager log --oneline' l", + "git -c alias.s=st -c 'alias.st=\"status\"' s", + "git --config-env=core.pager=p status", + "GIT_CONFIG_COUNT=1 GIT_CONFIG_KEY_0=alias.st GIT_CONFIG_VALUE_0=status git st", + "GIT_CONFIG_PARAMETERS=\"'alias.st=status'\" git st", + "GIT_CONFIG_COUNT=1 GIT_CONFIG_KEY_0=core.pager GIT_CONFIG_VALUE_0=$pager git log", + } { + request := gitCommandRequest(workspace, command) + request.Permission = PermissionAllow + request.PermissionGranted = true + if decision := engine.Evaluate(context.Background(), request); decision.Action == ActionDeny && decision.Block != nil && decision.Block.Code == BlockNestedGitInit { + t.Fatalf("%q was refused as nested repository creation; it creates no repository", command) + } + } + // And a standalone workspace is not governed at all, alias or not. + standalone := t.TempDir() + for _, command := range []string{ + "git -c alias.bootstrap=init bootstrap", + "git -c alias.a='-c alias.b=init b' a .", + "git -c 'alias.a=\"init\"' a .", + "GIT_CONFIG_COUNT=1 GIT_CONFIG_KEY_0=alias.a GIT_CONFIG_VALUE_0=init git a .", + } { + request := gitCommandRequest(standalone, command) + request.Permission = PermissionAllow + request.PermissionGranted = true + if decision := gitWorkspaceEngine(t, standalone).Evaluate(context.Background(), request); decision.Block != nil && decision.Block.Code == BlockNestedGitInit { + t.Fatalf("%q in a standalone workspace was refused as nested: %#v", command, decision) + } + } +} + +// THE REFUSAL THAT CANNOT BE APPROVED IS THE ONE THE OPERATOR HAS TO SEE. +// +// `git clone` carries both risks: it uses the network and it creates a +// repository. The clone test above sets NetworkAllow, which is the only policy +// under which the nested-repository branch was ever reached. Under the DEFAULT +// policy the network gate ran first and answered with a prompt, so the operator +// was asked to approve network access for a command no approval could make +// runnable, and ReasonNestedGitInit, the reason with a remedy in it, was never +// shown. Reported by @jatmn. +func TestNestedCloneIsRefusedAsNestedUnderTheDefaultNetworkPolicy(t *testing.T) { + _, workspace := nestedGitWorkspace(t) + policy := DefaultPolicy() + if policy.Network != NetworkDeny { + t.Fatalf("SETUP INVALID: the default network policy is %q, so this does not exercise the ordering", policy.Network) + } + engine := NewEngine(EngineOptions{WorkspaceRoot: workspace, Policy: policy, Backend: nativeWrappingBackend}) + + for _, granted := range []bool{false, true} { + request := gitCommandRequest(workspace, "git clone https://example.invalid/repo.git vendor/dep") + request.PermissionGranted = granted + decision := engine.Evaluate(context.Background(), request) + if decision.Action != ActionDeny || decision.Block == nil || decision.Block.Code != BlockNestedGitInit { + t.Fatalf("granted=%v: clone in a governed workspace was %s (block %#v, reason %q), want a %s denial", granted, decision.Action, decision.Block, decision.Reason, BlockNestedGitInit) + } + if decision.Reason != ReasonNestedGitInit { + t.Errorf("granted=%v: reason = %q, want the nested-repository reason and its remedy", granted, decision.Reason) + } + } + + // CONTROLS. A network command that creates no repository still gets the + // network answer, so the reordering did not swallow that gate, and a clone + // into a standalone workspace is a network question and nothing else. + fetch := engine.Evaluate(context.Background(), gitCommandRequest(workspace, "git fetch origin")) + if fetch.Action != ActionPrompt || fetch.Reason != ReasonNetworkBlocked { + t.Errorf("git fetch in the governed workspace was %s (%q), want the network prompt", fetch.Action, fetch.Reason) + } + standalone := t.TempDir() + if workspaceGovernedByAncestorRepository(standalone) { + t.Fatalf("SETUP INVALID: %s is governed by an ancestor repository", standalone) + } + standaloneEngine := NewEngine(EngineOptions{WorkspaceRoot: standalone, Policy: policy, Backend: nativeWrappingBackend}) + clone := standaloneEngine.Evaluate(context.Background(), gitCommandRequest(standalone, "git clone https://example.invalid/repo.git .")) + if clone.Action != ActionPrompt || clone.Reason != ReasonNetworkBlocked { + t.Errorf("clone into a standalone workspace was %s (%q), want the network prompt", clone.Action, clone.Reason) + } +} + +// THE LIMIT OF THIS GUARD, WRITTEN DOWN SO IT IS NOT MISTAKEN FOR COVERAGE. +// +// The nested-repository rule recognises git creating a repository. It does not +// recognise a repository assembled by other means, and it cannot: mkdir and a +// few writes, an archive, or a script in any language all produce the same +// directory, and none of them is a git command. These legs are allowed today. +// They are here so that anyone reading the rule as containment of the .git +// pathname finds a test saying otherwise, and so that a change which DOES close +// it at the decision layer has to come and update this deliberately. +// +// It is not a statement that the behaviour is acceptable. A repository built +// this way carries whatever config its author wrote, and a plain `git status` +// outside the sandbox honours it. That has to be closed where the carveouts +// live, at write time. Reported by @jatmn. +func TestHandAssembledRepositoryIsNotWhatThisGuardCatches(t *testing.T) { + _, workspace := nestedGitWorkspace(t) + engine := gitWorkspaceEngine(t, workspace) + for _, command := range []string{ + "mkdir .git", + "mkdir -p .git/objects .git/refs && printf 'ref: refs/heads/main\n' > .git/HEAD", + "cp -r ../template-repo/.git .git", + "python3 -c \"import os; os.makedirs('.git/refs')\"", + "tar -xf repo.tar", + } { + decision := engine.Evaluate(context.Background(), gitCommandRequest(workspace, command)) + if decision.Block != nil && decision.Block.Code == BlockNestedGitInit { + t.Errorf("%q is now refused as nested repository creation. If that is deliberate, this test and the WHAT THIS IS NOT note in risk.go both need to say what the rule covers now", command) + } + } +} diff --git a/internal/sandbox/git_nested_workspace_test.go b/internal/sandbox/git_nested_workspace_test.go new file mode 100644 index 000000000..8773d0ebf --- /dev/null +++ b/internal/sandbox/git_nested_workspace_test.go @@ -0,0 +1,263 @@ +package sandbox + +import ( + "os" + "os/exec" + "path/filepath" + "strings" + "testing" +) + +// A WORKSPACE INSIDE ANOTHER REPOSITORY KEEPS ITS CARVEOUTS. +// +// This branch skipped them for a while, so that Zero would not create a .git in +// a repository it does not own. The cost was worse than the thing avoided: +// nothing then denied writes to /.git, so a sandboxed command could build +// a repository there by hand and put core.fsmonitor in its config, and Zero runs +// git outside the sandbox against that workspace. The nested-repository refusal +// recognises git creating a repository; it cannot recognise mkdir. +// +// TestCarveoutsInsideAnAncestorRepositoryDoNotCaptureDiscovery measures the +// reason the skip existed and shows it does not hold. +func TestCarveoutsAreSynthesizedInsideAnAncestorRepository(t *testing.T) { + parent := t.TempDir() + if err := os.MkdirAll(filepath.Join(parent, ".git"), 0o700); err != nil { + t.Fatal(err) + } + workspace := filepath.Join(parent, "sub", "project") + if err := os.MkdirAll(workspace, 0o700); err != nil { + t.Fatal(err) + } + + // SETUP: the workspace itself has no .git, or this is a different case. + if _, err := os.Lstat(filepath.Join(workspace, ".git")); err == nil { + t.Fatal("SETUP INVALID: the workspace already carries .git") + } + if !gitMetadataGovernedByAncestor(workspace) { + t.Fatal("SETUP INVALID: the ancestor is not detected, so this proves nothing") + } + + specs := gitMetadataWriteCarveoutSpecs(workspace) + want := map[string]bool{ + filepath.Join(workspace, ".git", "config"): true, + filepath.Join(workspace, ".git", "hooks"): true, + } + for _, spec := range specs { + delete(want, spec.Path) + } + for path := range want { + t.Errorf("a workspace inside another repository gets no carveout for %s, so nothing denies writing it", path) + } +} + +// The reason the skip existed, measured rather than argued: materializing only +// config and hooks does not make the directory a repository, so git's discovery +// walk still reaches the ancestor and a config planted there is never read. +// If a future git changes that, this fails and the trade has to be reopened. +func TestCarveoutsInsideAnAncestorRepositoryDoNotCaptureDiscovery(t *testing.T) { + git, err := exec.LookPath("git") + if err != nil { + t.Skip("no git on PATH") + } + parent := t.TempDir() + run := func(dir string, args ...string) (string, error) { + command := exec.Command(git, args...) + command.Dir = dir + command.Env = append(os.Environ(), "GIT_CONFIG_NOSYSTEM=1", "HOME="+parent, "USERPROFILE="+parent) + out, err := command.CombinedOutput() + return strings.TrimSpace(string(out)), err + } + if out, err := run(parent, "init", "-q", "."); err != nil { + t.Skipf("cannot create the ancestor repository: %v: %s", err, out) + } + workspace := filepath.Join(parent, "sub", "project") + if err := os.MkdirAll(workspace, 0o700); err != nil { + t.Fatal(err) + } + ancestor, err := run(workspace, "rev-parse", "--show-toplevel") + if err != nil { + t.Skipf("cannot resolve the ancestor: %v: %s", err, ancestor) + } + + // Materialize exactly what the plan materializes, and make the config as + // hostile as a sandboxed command would if it could write one. + for _, spec := range gitMetadataWriteCarveoutSpecs(workspace) { + if spec.IsFile { + if err := os.MkdirAll(filepath.Dir(spec.Path), 0o700); err != nil { + t.Fatal(err) + } + if err := os.WriteFile(spec.Path, []byte("[core]\n\tfsmonitor = \"exit 9\"\n"), 0o600); err != nil { + t.Fatal(err) + } + continue + } + if err := os.MkdirAll(spec.Path, 0o700); err != nil { + t.Fatal(err) + } + } + + after, err := run(workspace, "rev-parse", "--show-toplevel") + if err != nil { + t.Fatalf("git stopped working in the workspace after the carveouts were materialized: %v: %s", err, after) + } + if after != ancestor { + t.Fatalf("discovery moved from %s to %s: the materialized carveouts captured the walk, which is the trade this test exists to watch", ancestor, after) + } + if out, err := run(workspace, "status", "--porcelain"); err != nil { + t.Fatalf("git status failed in the workspace, so the planted config was read: %v: %s", err, out) + } +} + +// And a standalone directory keeps the materialized carveouts, or the guard +// above would be satisfied by returning nothing everywhere and the protection +// would be gone for the case it was written for. +func TestCarveoutsStillCoverAStandaloneDirectory(t *testing.T) { + workspace := t.TempDir() + if gitMetadataGovernedByAncestor(workspace) { + t.Skip("this temp directory sits inside a repository, so it is not the standalone case") + } + + specs := gitMetadataWriteCarveoutSpecs(workspace) + wantConfig := filepath.Join(workspace, ".git", "config") + wantHooks := filepath.Join(workspace, ".git", "hooks") + var sawConfig, sawHooks bool + for _, spec := range specs { + switch spec.Path { + case wantConfig: + sawConfig = true + if !spec.IsFile { + t.Errorf("%s is planned as a directory; creating .git/config as a directory makes git init fail", spec.Path) + } + case wantHooks: + sawHooks = true + } + } + if !sawConfig || !sawHooks { + t.Fatalf("a standalone directory lost its carveouts: config=%t hooks=%t specs=%+v", sawConfig, sawHooks, specs) + } +} + +// A linked worktree is unchanged: the pointer file is denied and nothing beneath +// it is named, which is what keeps elevated setup from descending through a +// regular file. +func TestCarveoutsStillDenyTheLinkedWorktreePointer(t *testing.T) { + workspace := t.TempDir() + pointer := filepath.Join(workspace, ".git") + if err := os.WriteFile(pointer, []byte("gitdir: ../real/.git/worktrees/w\n"), 0o600); err != nil { + t.Fatal(err) + } + specs := gitMetadataWriteCarveoutSpecs(workspace) + if len(specs) != 1 || specs[0].Path != pointer || !specs[0].IsFile { + t.Fatalf("linked worktree carveouts = %+v, want exactly the file-shaped pointer %s", specs, pointer) + } +} + +// The ancestor test itself has to distinguish shapes, because a linked worktree +// or submodule ancestor carries .git as a FILE and still owns this directory. +func TestAncestorDetectionAcceptsBothGitShapes(t *testing.T) { + for _, shape := range []struct { + name string + make func(t *testing.T, parent string) + }{ + {name: "directory", make: func(t *testing.T, parent string) { + if err := os.MkdirAll(filepath.Join(parent, ".git"), 0o700); err != nil { + t.Fatal(err) + } + }}, + {name: "pointer file", make: func(t *testing.T, parent string) { + if err := os.WriteFile(filepath.Join(parent, ".git"), []byte("gitdir: elsewhere\n"), 0o600); err != nil { + t.Fatal(err) + } + }}, + } { + t.Run(shape.name, func(t *testing.T) { + parent := t.TempDir() + shape.make(t, parent) + workspace := filepath.Join(parent, "nested", "deep") + if err := os.MkdirAll(workspace, 0o700); err != nil { + t.Fatal(err) + } + if !gitMetadataGovernedByAncestor(workspace) { + t.Fatalf("an ancestor carrying .git as a %s was not recognised as governing %s", shape.name, workspace) + } + }) + } +} + +// A NESTED WORKSPACE THAT OWNS ITS OWN REPOSITORY KEEPS ITS OWN CARVEOUTS. +// +// The ancestor branch above exists so Zero does not synthesize a control +// directory inside somebody else's repository. It was asked unconditionally, +// so a workspace that already HAS its own .git lost the hooks and config +// carveouts too — and those are its own metadata, not the ancestor's. Nothing +// compensated: workspaceGovernedByAncestorRepository answers the same question +// by testing for a local .git, so it reported this workspace as not governed +// and the initialization refusal never fired. Commands then received plain +// workspace write access over a live .git/config and .git/hooks, which is +// exactly the configuration that decides what git executes. Reported by +// @jatmn. +// +// Asserted through the profile as well as the spec helper, because the spec +// list is only load-bearing if it reaches ReadOnlySubpaths. +func TestNestedWorkspaceOwningGitKeepsItsOwnCarveouts(t *testing.T) { + outer := t.TempDir() + if err := os.MkdirAll(filepath.Join(outer, ".git"), 0o700); err != nil { + t.Fatal(err) + } + workspace := filepath.Join(outer, "project") + if err := os.MkdirAll(filepath.Join(workspace, ".git"), 0o700); err != nil { + t.Fatal(err) + } + + wantHooks := filepath.Join(workspace, ".git", "hooks") + wantConfig := filepath.Join(workspace, ".git", "config") + specs := gitMetadataWriteCarveoutSpecs(workspace) + foundHooks, foundConfig := false, false + for _, spec := range specs { + switch { + case spec.Path == wantHooks && !spec.IsFile: + foundHooks = true + case spec.Path == wantConfig && spec.IsFile: + foundConfig = true + } + } + if !foundHooks || !foundConfig { + t.Fatalf("a workspace owning its own .git lost its carveouts: hooks=%v config=%v specs=%+v", + foundHooks, foundConfig, specs) + } + + // The specs only matter if they reach the profile the runner enforces. + profile := PermissionProfileFromPolicy(workspace, DefaultPolicy(), nil) + var root *WritableRoot + for index := range profile.FileSystem.WriteRoots { + if sameGitTestPath(profile.FileSystem.WriteRoots[index].Root, workspace) { + root = &profile.FileSystem.WriteRoots[index] + break + } + } + if root == nil { + t.Fatalf("no writable root for the workspace: %+v", profile.FileSystem.WriteRoots) + } + for _, want := range []string{wantHooks, wantConfig} { + covered := false + for _, sub := range root.ReadOnlySubpaths { + if sameGitTestPath(sub, want) { + covered = true + break + } + } + if !covered { + t.Errorf("%s is not read-only in the workspace write root: %+v", want, root.ReadOnlySubpaths) + } + } +} + +// Compared through the profile's own canonicalization. PermissionProfileFromPolicy +// stores normalizeProfilePath(root), which resolves symlinks, so a raw t.TempDir() +// path does not match it on a runner where the temp directory is reached through +// one: macOS /var -> /private/var, and the Windows short-name expansion of the +// runner's profile directory. Comparing cleaned strings passed locally and failed +// on both CI runners for that reason alone. +func sameGitTestPath(a, b string) bool { + return normalizeProfilePath(a) == normalizeProfilePath(b) +} diff --git a/internal/sandbox/git_pointer_shape_windows_test.go b/internal/sandbox/git_pointer_shape_windows_test.go new file mode 100644 index 000000000..4802e202a --- /dev/null +++ b/internal/sandbox/git_pointer_shape_windows_test.go @@ -0,0 +1,119 @@ +//go:build windows + +package sandbox + +import ( + "os" + "path/filepath" + "testing" +) + +func pointerWorkspace(t *testing.T) (workspace string, pointer string) { + t.Helper() + workspace = t.TempDir() + pointer = filepath.Join(workspace, ".git") + if err := os.WriteFile(pointer, []byte("gitdir: ../real/.git/worktrees/w\n"), 0o600); err != nil { + t.Fatal(err) + } + return workspace, pointer +} + +func pointerPlanConfig(t *testing.T, workspace string) WindowsSandboxCommandConfig { + t.Helper() + return WindowsSandboxCommandConfig{ + SandboxHome: t.TempDir(), + CommandCWD: workspace, + WorkspaceRoots: []string{workspace}, + PermissionProfile: PermissionProfile{ + FileSystem: FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{ + Root: workspace, + ReadOnlySubpaths: gitMetadataWriteCarveouts(workspace), + ProtectedMetadataNames: sandboxFullyProtectedMetadataNames, + }}, + }, + }, + } +} + +// THE SHAPE HAS TO SURVIVE THE STAGE BOUNDARY, NOT JUST BE CORRECT AT STAGE ONE. +// +// gitMetadataWriteCarveoutSpecs types a linked worktree's .git as a file. That +// typed result was flattened to a path before ACL planning, and the planner tried +// to recover the shape by deriving suffixes under a sentinel root whose .git never +// exists. The derivation therefore always took the directory branch, and a bare +// \.git could never match as a file. +// +// The window that turns a wrong plan into damage is the caller-to-elevated-helper +// gap: the profile is built in the user's shell and the plan is applied in a +// separately launched helper, across the UAC prompt. If the pointer goes away in +// between, the applier materializes a DIRECTORY at the pointer path and the +// worktree is broken, persistently, because the runner discards the rollback. +// +// Driven through BuildWindowsACLPlan and applyWindowsACLPlan with the pointer +// removed after planning, which is the only place this is visible. A stage-one +// assertion on specs[0].IsFile passed throughout the whole time the bug was live. +func TestALinkedWorktreePointerIsNeverMaterializedAsADirectory(t *testing.T) { + workspace, pointer := pointerWorkspace(t) + plan, err := BuildWindowsACLPlan(pointerPlanConfig(t, workspace)) + if err != nil { + t.Fatalf("BuildWindowsACLPlan: %v", err) + } + + // SETUP: the plan really does carry the pointer, typed as a file. + var found bool + for _, entry := range plan.Entries { + if entry.Path == pointer && entry.Action == WindowsACLDenyWrite { + found = true + if !entry.MaterializeFile { + t.Fatalf("the plan types %s as a directory; applying it would replace the worktree pointer with a directory", pointer) + } + } + } + if !found { + t.Fatalf("SETUP INVALID: no deny-write entry for the pointer %s: %+v", pointer, plan.Entries) + } + + // The window: the pointer goes away between planning and applying. + if err := os.Remove(pointer); err != nil { + t.Fatal(err) + } + if _, err := applyWindowsACLPlan(plan); err != nil { + t.Skipf("cannot apply an ACL plan here: %v", err) + } + + info, err := os.Lstat(pointer) + if err != nil { + // Nothing recreated is also acceptable: what must not happen is a directory. + return + } + if info.IsDir() { + t.Fatalf("a directory was created at the worktree pointer %s, which breaks the worktree", pointer) + } +} + +// Control: with the pointer present at apply, it stays a file and is not +// replaced. Without this, the assertion above could be satisfied by an apply +// that does nothing at all. +func TestALinkedWorktreePointerSurvivesAnApply(t *testing.T) { + workspace, pointer := pointerWorkspace(t) + plan, err := BuildWindowsACLPlan(pointerPlanConfig(t, workspace)) + if err != nil { + t.Fatalf("BuildWindowsACLPlan: %v", err) + } + if _, err := applyWindowsACLPlan(plan); err != nil { + t.Skipf("cannot apply an ACL plan here: %v", err) + } + info, err := os.Lstat(pointer) + if err != nil { + t.Fatalf("the pointer vanished across the apply: %v", err) + } + if info.IsDir() { + t.Fatalf("the pointer at %s became a directory", pointer) + } + body, err := os.ReadFile(pointer) + if err != nil || len(body) == 0 { + t.Fatalf("the pointer lost its contents: body=%q err=%v", body, err) + } +} diff --git a/internal/sandbox/profile.go b/internal/sandbox/profile.go index 5cd97a9bb..9b5e3d9f4 100644 --- a/internal/sandbox/profile.go +++ b/internal/sandbox/profile.go @@ -96,28 +96,209 @@ var protectedMetadataNames = []string{".git", ".zero", ".agents"} // gitMetadataWriteCarveouts below. var sandboxFullyProtectedMetadataNames = []string{".zero", ".agents"} -// gitMetadataWriteCarveouts returns the .git paths that stay write-denied under -// the OS-level sandbox even though the rest of .git is writable to git -// subprocesses. Worktrees and submodules store .git as a regular pointer file, -// so they protect that file itself: constructing .git/hooks below it makes -// bwrap fail before the command starts. All other entry and error states retain -// the legacy child paths. The pointer is deliberately not parsed here; -// repository-controlled metadata must not redirect sandbox rules elsewhere. +// sandboxRenameProtectedMetadataName is the metadata directory that cannot be +// fully write-protected but must still not be REPLACEABLE. +// +// It is deliberately not in the list above. That list denies write, and git has +// to write index, objects and refs. But the carveouts guarding it are attached +// to .git/config and .git/hooks as objects, so a principal that renames .git and +// recreates it gets fresh paths inheriting the workspace allow with no denies, +// which restores credential.helper and core.hooksPath. The Windows ACL plan +// therefore denies DELETE on this directory alone, uninherited. +const sandboxRenameProtectedMetadataName = ".git" + +// gitMetadataWriteCarveouts returns the .git subpaths that stay write-denied +// under the OS-level sandbox even though the rest of .git is writable to git +// subprocesses. Nonexistent paths are harmless no-ops in every backend's +// enforcement (seatbelt regex, bwrap ro-bind, Windows ACL deny entry). func gitMetadataWriteCarveouts(root string) []string { return gitMetadataWriteCarveoutsWithLstat(root, os.Lstat) } +// gitMetadataWriteCarveoutsWithLstat is the seam the worktree tests drive. +// Which carveouts apply is decided by the SHAPE of .git, and a test cannot +// create every shape it needs to cover: an lstat that fails is a real state +// (a racing checkout, a permission error) with its own required answer, and +// there is no filesystem it can be staged on. func gitMetadataWriteCarveoutsWithLstat(root string, lstat func(string) (os.FileInfo, error)) []string { + specs := gitMetadataWriteCarveoutSpecsWithLstat(root, lstat) + out := make([]string, 0, len(specs)) + for _, spec := range specs { + out = append(out, spec.Path) + } + return out +} + +// gitMetadataCarveout is a write-denied .git path together with the shape git +// expects it to have. The shape matters to exactly one backend: the Windows ACL +// plan creates a missing carveout so the deny ACE is in place before git first +// runs, and creating .git/config as a directory makes `git init` fail outright. +// Every other backend only ever names the path, so it can ignore IsFile. +type gitMetadataCarveout struct { + Path string + IsFile bool +} + +// gitMetadataWriteCarveoutSpecs is the single source of truth for the carveout +// set. gitMetadataWriteCarveouts derives its list from this so a path can never +// be added in one place and have its shape forgotten in the other. +func gitMetadataWriteCarveoutSpecs(root string) []gitMetadataCarveout { + return gitMetadataWriteCarveoutSpecsWithLstat(root, os.Lstat) +} + +func gitMetadataWriteCarveoutSpecsWithLstat(root string, lstat func(string) (os.FileInfo, error)) []gitMetadataCarveout { gitPath := filepath.Join(root, ".git") - children := []string{ - filepath.Join(gitPath, "hooks"), - filepath.Join(gitPath, "config"), + // A LINKED WORKTREE OR SUBMODULE HAS .git AS A FILE, NOT A DIRECTORY. + // + // It holds a `gitdir:` pointer to the real control directory, which lives + // outside this write root. Naming .git/config and .git/hooks there asks the + // Windows plan to materialize them by descending through .git as a directory, + // which cannot open a child beneath a regular file, so opted-in elevated setup + // aborts and the sandbox is unusable in any worktree. Zero's own development + // worktrees are exactly this shape. + // + // Deny the pointer file itself instead. It is the right protection rather than + // a lesser one: a principal that can rewrite `gitdir:` repoints the whole + // repository at a control directory it chooses, which subsumes editing config + // or dropping a hook. The real control directory is outside the write root, so + // the principal has no inherited access to it and needs no carveout there. + // + // Stat, not a lexical guess. Whether .git is a file is a property of the + // checkout, and the two layouts want different ACEs. A missing .git (git has + // not run yet) keeps the directory-shaped carveouts, which is what makes them + // materialize before git first runs. + // + // A REGULAR FILE, NOT MERELY A NON-DIRECTORY. The pointer this recognises is + // the `gitdir:` file git writes for a linked worktree or submodule. A symlink + // named .git, dangling or not, and a special entry such as a fifo are neither + // that nor a checkout git will use, so they keep the legacy child carveouts + // rather than being handed the stronger single-object protection on the + // strength of not being a directory. + gitInfo, gitErr := lstat(gitPath) + if gitErr == nil && gitInfo.Mode().IsRegular() { + return []gitMetadataCarveout{{Path: gitPath, IsFile: true}} + } + // EVERY WRITE ROOT GETS THE CARVEOUTS, INCLUDING ONE INSIDE ANOTHER REPOSITORY. + // + // The directory-shaped carveouts below are materialized by the Windows plan so + // the deny ACE is in place before git first runs. For a while this branch + // skipped them when an ancestor owned the git metadata, on the argument that + // creating .git/config and .git/hooks there synthesizes a control directory + // competing with the ancestor's for git's discovery walk. Measured against the + // git this repository builds against, it does not: a .git holding only config + // and hooks is not a valid repository, so `rev-parse --show-toplevel` still + // answers with the ancestor, `status` still works, and a core.fsmonitor + // written into that config is never executed. The test named below drives real + // git and fails if that ever stops being true. + // + // Skipping them cost much more than it saved. With no carveout on this root + // nothing denies writes to /.git at all, so a sandboxed command can + // assemble a repository there by hand, with mkdir and a few file writes and + // not one git command among them, and put core.fsmonitor in its config. Zero + // then runs `git -C status` OUTSIDE the sandbox (scratch_files.go, + // internal/repoinfo, internal/zerogit) and git executes it. The nested-repository + // refusal in risk.go does not cover that and cannot: no static reading of a + // shell command enumerates the ways to create a directory. It stays as the + // layer that explains itself to an operator, not as the boundary. + // + // So the boundary is here, where it is enforced at write time: config and + // hooks are denied on every write root, whatever owns the metadata above it. + return []gitMetadataCarveout{ + {Path: filepath.Join(root, ".git", "hooks")}, + {Path: filepath.Join(root, ".git", "config"), IsFile: true}, } - info, err := lstat(gitPath) - if err != nil || !info.Mode().IsRegular() { - return children +} + +// workspaceGovernedByAncestorRepository is the same question the carveouts ask, +// named for the caller that asks it about a whole workspace rather than about +// one write root. Sharing the rule is the point: the reason a workspace gets no +// carveouts is exactly the reason it may not create a repository. +func workspaceGovernedByAncestorRepository(workspaceRoot string) bool { + cleaned := strings.TrimSpace(workspaceRoot) + if cleaned == "" { + return false + } + if _, err := os.Lstat(filepath.Join(cleaned, ".git")); err == nil { + // It already owns git metadata, so setup protected it and nothing here is + // being created inside a repository that is not this workspace's. + return false + } + return gitMetadataGovernedByAncestor(cleaned) +} + +// gitMetadataGovernedByAncestor reports whether an ancestor of root carries git +// metadata, which is git's own discovery rule: the nearest ancestor with a .git +// entry owns this directory. +// +// Lstat rather than a resolved walk, because a .git of either shape answers the +// question: a directory is an ordinary repository root and a file is a linked +// worktree or submodule pointer. Both mean the metadata is not ours to create. +func gitMetadataGovernedByAncestor(root string) bool { + current := filepath.Clean(strings.TrimSpace(root)) + if current == "" || current == "." { + return false } - return []string{gitPath} + for { + parent := filepath.Dir(current) + if parent == current { + return false + } + if _, err := os.Lstat(filepath.Join(parent, ".git")); err == nil { + return true + } + current = parent + } +} + +// gitMetadataCarveoutSuffixBase is a sentinel root used only to recover the +// trailing segments of the carveout specs. It is never touched on disk. +const gitMetadataCarveoutSuffixBase = string(filepath.Separator) + "zero-carveout-base" + +// gitMetadataCarveoutIsFile reports whether path names a carveout git expects +// to be a file. +// +// It matches on the trailing segments rather than on a whole reconstructed +// path. The subpaths reaching the ACL plan are already normalized — resolved +// through EvalSymlinks where that succeeds — while a rebuilt spec path cannot +// be, because .git/config does not exist yet at setup and resolution falls back +// to a plain Clean. On a host where two spellings of the same path differ (an +// 8.3 short name, different casing) a whole-path equality check silently misses +// and the carveout is created as a directory again, which is the original bug +// reintroduced quietly. The suffix cannot drift from the spec list because it +// is derived from it. +func gitMetadataCarveoutIsFile(path string) bool { + candidate := strings.ToLower(filepath.Clean(strings.TrimSpace(path))) + if candidate == "" { + return false + } + // A BARE .git CARVEOUT IS THE LINKED-WORKTREE POINTER, AND ONLY THAT. + // + // The suffix derivation below runs specs against a sentinel root whose .git + // never exists, so it always takes the directory branch and yields exactly one + // file-shaped suffix, .gitconfig. A bare .git could therefore never + // match, and both planners emitted MaterializeFile:false for a path the profile + // stage had already typed correctly as a file. With the pointer absent at apply + // time, the plan then created a DIRECTORY at the pointer path and broke the + // worktree. + // + // The carveout SET already carries the answer. gitMetadataWriteCarveoutSpecs is + // the only producer of ReadOnlySubpaths, and it emits a bare .git in exactly one + // case: the linked-worktree or submodule pointer. The directory case emits + // .githooks and .gitconfig and never the parent. So the shape is not being + // guessed from a pathname here, it is being read off which carveout this is. + if strings.EqualFold(filepath.Base(candidate), ".git") { + return true + } + for _, spec := range gitMetadataWriteCarveoutSpecs(gitMetadataCarveoutSuffixBase) { + if !spec.IsFile { + continue + } + suffix := strings.ToLower(strings.TrimPrefix(spec.Path, gitMetadataCarveoutSuffixBase)) + if suffix != "" && strings.HasSuffix(candidate, suffix) { + return true + } + } + return false } func PermissionProfileFromPolicy(workspaceRoot string, policy Policy, scope *Scope) PermissionProfile { diff --git a/internal/sandbox/profile_gitfile_test.go b/internal/sandbox/profile_gitfile_test.go new file mode 100644 index 000000000..439df9063 --- /dev/null +++ b/internal/sandbox/profile_gitfile_test.go @@ -0,0 +1,79 @@ +package sandbox + +import ( + "os" + "path/filepath" + "strings" + "testing" +) + +// A LINKED WORKTREE HAS .git AS A FILE. +// +// The Windows principal plan materializes every carveout, and it does so by +// descending through .git as a directory. Against a `gitdir:` pointer file that +// walk cannot open a child, so opted-in elevated setup aborts and the sandbox +// cannot be used in a worktree at all. Zero's own development worktrees have +// exactly this layout, so this is the common case for anyone working on the +// sandbox itself. +func TestGitCarveoutsHandleAWorktreeGitfile(t *testing.T) { + root := t.TempDir() + gitFile := filepath.Join(root, ".git") + // The real shape git writes for a linked worktree. + if err := os.WriteFile(gitFile, []byte("gitdir: "+filepath.Join(root, "..", "main", ".git", "worktrees", "wt")+"\n"), 0o600); err != nil { + t.Fatalf("WriteFile: %v", err) + } + + specs := gitMetadataWriteCarveoutSpecs(root) + if len(specs) != 1 { + t.Fatalf("carveout specs = %#v, want only the .git pointer file", specs) + } + if filepath.Clean(specs[0].Path) != filepath.Clean(gitFile) { + t.Fatalf("carveout path = %s, want the .git pointer file %s", specs[0].Path, gitFile) + } + if !specs[0].IsFile { + t.Fatal("the .git pointer carveout is not marked as a file, so materialization would create a directory where git needs a file") + } + // Nothing may name a path BENEATH the pointer file, which is what forces the + // directory descent that aborts setup. + for _, path := range gitMetadataWriteCarveouts(root) { + rest := strings.TrimPrefix(filepath.Clean(path), filepath.Clean(gitFile)) + if rest != "" { + t.Errorf("carveout %s descends through the .git pointer file, which cannot have children", path) + } + } +} + +// The ordinary repository layout must keep the config and hooks carveouts. They +// are the whole reason the set exists: without them a principal reinstates +// credential.helper or core.hooksPath through inherited workspace write access. +func TestGitCarveoutsKeepConfigAndHooksForARealGitDir(t *testing.T) { + root := t.TempDir() + if err := os.MkdirAll(filepath.Join(root, ".git"), 0o700); err != nil { + t.Fatalf("MkdirAll: %v", err) + } + paths := gitMetadataWriteCarveouts(root) + for _, want := range []string{filepath.Join(root, ".git", "hooks"), filepath.Join(root, ".git", "config")} { + found := false + for _, got := range paths { + if filepath.Clean(got) == filepath.Clean(want) { + found = true + break + } + } + if !found { + t.Errorf("carveout %s missing for an ordinary .git directory: %#v", want, paths) + } + } +} + +// A workspace where git has never run keeps the directory-shaped carveouts. That +// is what lets the Windows plan create them before git first runs, so the deny +// ACEs are already in place rather than being applied to paths git will create +// later with inherited write access. +func TestGitCarveoutsKeepConfigAndHooksWhenGitIsAbsent(t *testing.T) { + root := t.TempDir() + paths := gitMetadataWriteCarveouts(root) + if len(paths) != 2 { + t.Fatalf("carveouts = %#v, want the config and hooks pair when .git is absent", paths) + } +} diff --git a/internal/sandbox/risk.go b/internal/sandbox/risk.go index faa4e09bc..dbbff459f 100644 --- a/internal/sandbox/risk.go +++ b/internal/sandbox/risk.go @@ -135,6 +135,45 @@ func classifyWithScope(request Request, scope *Scope) Risk { if analysis.Destructive { add("destructive", RiskCritical) } + // A REPOSITORY CREATED INSIDE SOMEBODY ELSE'S IS NOT PROTECTABLE LATER. + // + // gitMetadataWriteCarveoutSpecs returns NOTHING for a workspace governed by + // an ancestor repository, on every backend. That is deliberate: naming + // /.git/config and /.git/hooks makes the Windows plan create a + // control directory, and the bubblewrap helper mount one, that competes with + // the ancestor's for git's discovery walk inside a repository Zero does not + // own. So the workspace carries no config or hooks protection at all. + // + // A repository created here DURING the command therefore lands under the + // plain workspace write grant with nothing denying credential.helper or + // core.hooksPath, and on Windows nothing denying DELETE on .git either. The + // serialized plan never changed, so the cached setup marker stays valid and + // the next run will not notice. + // + // Refused rather than protected, because the protection would have to be + // established at a moment the sandbox is no longer at. + // + // WHAT THIS IS NOT. It refuses GIT creating a repository, by every spelling + // git offers. It is not containment of the .git pathname. A command that + // assembles a repository by other means (mkdir and a few file writes, an + // archive, a script in any language) is not recognised here, and no static + // reading of a shell command can recognise all of them, so this rule must + // not be described or relied on as the boundary. A repository assembled + // that way carries whatever config its author wrote, and an ordinary + // `git status` run outside the sandbox honours it, core.fsmonitor included. + // Closing that is the carveouts' job, at write time, not this rule's. + // TestHandAssembledRepositoryIsNotWhatThisGuardCatches pins the limit so it + // cannot be mistaken for coverage. Reported by @jatmn. + // + // The condition is the WORKSPACE, not the directory the command names. + // Resolving that directory would mean tracking -C and cwd through the + // script, which is precisely the option-parsing surface that let + // "git -C sub clone" past the network gate. Refusing a git init outside the + // workspace too is the conservative side of that trade, and the reason text + // says so. + if analysis.GitInit && workspaceGovernedByAncestorRepository(request.WorkspaceRoot) { + add("nested_git_init", RiskCritical) + } if analysis.TooComplex { add("unparseable_command", RiskHigh) } diff --git a/internal/sandbox/runner_windows_integration_test.go b/internal/sandbox/runner_windows_integration_test.go index cd6d0388e..485ae0dd0 100644 --- a/internal/sandbox/runner_windows_integration_test.go +++ b/internal/sandbox/runner_windows_integration_test.go @@ -36,10 +36,14 @@ func TestWindowsRestrictedTokenRealSandboxSmoke(t *testing.T) { Network: NetworkPolicy{Mode: NetworkDeny}, } config := WindowsSandboxCommandArgsOptions{ - SandboxHome: sandboxHome, - CommandCWD: root, - WorkspaceRoots: []string{root}, - PermissionProfile: profile, + SandboxHome: sandboxHome, + CommandCWD: root, + WorkspaceRoots: []string{root}, + // The setup builder folds the runtime roots into its profile before the + // helper sees it; the command builder does not. Passing the bare profile + // here made every command disagree with the marker setup wrote, nine + // entries against five, before the first write probe ran. + PermissionProfile: WindowsSandboxProfileWithRuntimeRoots(profile, []string{root}), SandboxLevel: WindowsSandboxLevelRestrictedToken, } runWindowsRealSmokeSetup(t, setupExe, WindowsSandboxSetupArgsOptions{ @@ -393,6 +397,15 @@ func realSmokeExecutable(t *testing.T, envKey string, fallbackName string) strin func runWindowsRealSmokeSetup(t *testing.T, setupExe string, options WindowsSandboxSetupArgsOptions) { t.Helper() + // options.PrincipalOptIn is deliberately left nil by both call sites, which + // makes BuildWindowsSandboxSetupArgs resolve the opt-in from this process's + // environment — the same value the command half resolves, since the smoke + // WindowsSandboxCommandArgsOptions carries no explicit entry either. Do not + // "fix" this by setting it to false: anyone running this suite with + // ZERO_WINDOWS_SANDBOX_IDENTITY=1 (the only way to exercise the principal + // backend) would then serialize `--sandbox-principal 0`, disagree with the + // command half, and fail every command at marker validation instead of + // testing the sandbox. args, err := BuildWindowsSandboxSetupArgs(options) if err != nil { t.Fatalf("BuildWindowsSandboxSetupArgs: %v", err) diff --git a/internal/sandbox/runtime_fallback_anchor_other.go b/internal/sandbox/runtime_fallback_anchor_other.go new file mode 100644 index 000000000..3e1c4da87 --- /dev/null +++ b/internal/sandbox/runtime_fallback_anchor_other.go @@ -0,0 +1,33 @@ +//go:build !windows + +package sandbox + +import ( + "os" + "path/filepath" +) + +// physicalTempDir returns os.TempDir with every symlink in its ancestry +// resolved, so the fallback anchor's parent chain is physical and only the +// anchor itself is a new component. +// +// peermsg.EnsurePrivateDir walks from the root with O_NOFOLLOW and refuses any +// component that is a link. That is right for the owned tail and wrong for the +// ancestors above it: on macOS os.TempDir lives under /var, which is a symlink +// to /private/var, so the first version of this fix refused every fallback on +// every Mac with "refusing non-directory or symlink runtime path component +// var". The constraint jatmn stated for #901 applies here identically: +// redirected cache and TEMP locations above the owned tail are the operator's +// business and must keep working; the restriction is on what Zero owns. +// +// EvalSymlinks is the correct resolver off Windows, where the only reparse +// shape is a symlink and it traverses them. +func physicalTempDir() (string, error) { + return physicalDir(os.TempDir()) +} + +// physicalDir resolves any directory the same way, for the other places that +// have to hand EnsurePrivateDir a physical parent. +func physicalDir(path string) (string, error) { + return filepath.EvalSymlinks(path) +} diff --git a/internal/sandbox/runtime_fallback_anchor_windows.go b/internal/sandbox/runtime_fallback_anchor_windows.go new file mode 100644 index 000000000..7bcc73f16 --- /dev/null +++ b/internal/sandbox/runtime_fallback_anchor_windows.go @@ -0,0 +1,63 @@ +//go:build windows + +package sandbox + +import ( + "fmt" + "os" + + "golang.org/x/sys/windows" +) + +// physicalTempDir returns os.TempDir as the object it actually names, so the +// fallback anchor's parent chain is physical and only the anchor itself is a +// new component. +// +// peermsg.EnsurePrivateDir walks from the volume root with OBJ_DONT_REPARSE and +// refuses any component that is a reparse point. That is right for the owned +// tail and wrong for the ancestors above it: a redirected %LOCALAPPDATA% is an +// ordinary Windows configuration, and TEMP sits beneath it. The constraint +// jatmn stated for #901 applies here identically: redirected cache and TEMP +// locations above the owned tail must keep working; the restriction is on what +// Zero owns. +// +// GetFinalPathNameByHandle rather than filepath.EvalSymlinks, because +// EvalSymlinks does not traverse a junction on Windows: it returns the +// junction's own path and the walker then refuses it. The handle answers where +// the directory actually is. Same recipe as verifyWindowsACLTargetNotRedirected, +// which is why the flag constants and the prefix trim are shared. +func physicalTempDir() (string, error) { + return physicalDir(os.TempDir()) +} + +// physicalDir resolves any directory through a handle, for the other places +// that have to hand EnsurePrivateDir a physical parent. GetFinalPathNameByHandle +// rather than EvalSymlinks, because EvalSymlinks does not traverse a junction. +func physicalDir(path string) (string, error) { + utf16Path, err := windows.UTF16PtrFromString(path) + if err != nil { + return "", fmt.Errorf("encode directory %s: %w", path, err) + } + handle, err := windows.CreateFile( + utf16Path, + windows.FILE_READ_ATTRIBUTES|windows.SYNCHRONIZE, + windows.FILE_SHARE_READ|windows.FILE_SHARE_WRITE|windows.FILE_SHARE_DELETE, + nil, + windows.OPEN_EXISTING, + windows.FILE_FLAG_BACKUP_SEMANTICS, + 0, + ) + if err != nil { + return "", fmt.Errorf("open directory %s: %w", path, err) + } + defer windows.CloseHandle(handle) + buffer := make([]uint16, windows.MAX_LONG_PATH) + n, err := windows.GetFinalPathNameByHandle(handle, &buffer[0], uint32(len(buffer)), windowsFileNameNormalized|windowsVolumeNameDOS) + if err != nil { + return "", fmt.Errorf("resolve directory %s: %w", path, err) + } + if int(n) < len(buffer) { + buffer = buffer[:n] + } + return trimWindowsExtendedPathPrefix(windows.UTF16ToString(buffer)), nil +} diff --git a/internal/sandbox/runtime_fallback_anchor_windows_test.go b/internal/sandbox/runtime_fallback_anchor_windows_test.go new file mode 100644 index 000000000..bb1152e9b --- /dev/null +++ b/internal/sandbox/runtime_fallback_anchor_windows_test.go @@ -0,0 +1,157 @@ +//go:build windows + +package sandbox + +import ( + "os" + "os/exec" + "path/filepath" + "strings" + "testing" +) + +// DETERMINISM IS NOT OWNERSHIP. +// +// The fallback runtime root moved from a fresh os.MkdirTemp parent to a +// predictable name so that elevated setup and the later command agree on it. +// That is required. But a predictable path beneath the shared system temp is +// not a private allocation: whatever another local user, or a stray earlier +// process, has already placed at that name is what preparation then leases, +// creates, chmods and cleans through. The anchor beneath which the tree lives +// must therefore be proven to be this user's private directory before anything +// is built under it, on BOTH paths that can hand back a fallback root. +// +// Junctions need no privilege on Windows, so the "attacker-precreated parent" +// jatmn described is modelled as a junction at the anchor pointing into a +// directory the test owns and watches. Either trigger must refuse, and the +// watched directory must stay empty. + +// fallbackAnchorFixture redirects TEMP so the anchor lands inside a directory +// this test owns, plants a junction at the anchor into a watched target, and +// returns the target. The junction is the shape that defeats a pathname walk: +// it is not a symlink to os.ModeSymlink, and EvalSymlinks will not traverse it. +func fallbackAnchorFixture(t *testing.T) (target string) { + t.Helper() + tempRoot := t.TempDir() + target = t.TempDir() + t.Setenv("TMP", tempRoot) + t.Setenv("TEMP", tempRoot) + + // The anchor's parent is the PHYSICAL temp dir, so compare against the + // resolved form of the redirect rather than its spelling: t.TempDir can + // come back as an 8.3 short name or in a different case from what the + // handle reports, and a prefix check on the raw string would fail for a + // correct anchor. + anchor := fallbackRuntimeAnchor() + resolvedTempRoot, err := physicalTempDir() + if err != nil { + t.Fatalf("SETUP INVALID: cannot resolve the redirected TEMP %s: %v", tempRoot, err) + } + if !pathWithinRoot(resolvedTempRoot, anchor) { + t.Fatalf("SETUP INVALID: anchor %s is not under the redirected TEMP %s (resolved %s)", anchor, tempRoot, resolvedTempRoot) + } + if out, err := exec.Command("cmd", "/c", "mklink", "/J", anchor, target).CombinedOutput(); err != nil { + t.Fatalf("mklink /J: %v\n%s", err, out) + } + return target +} + +func assertNothingUnder(t *testing.T, target string) { + t.Helper() + entries, err := os.ReadDir(target) + if err != nil { + t.Fatal(err) + } + if len(entries) != 0 { + names := make([]string, 0, len(entries)) + for _, entry := range entries { + names = append(names, entry.Name()) + } + t.Errorf("preparation built %v beneath the junction target, which is somebody else's directory", names) + } +} + +// Trigger one: the user cache resolves INSIDE the workspace, so +// sandboxRuntimeRootFor itself returns the fallback root and the first lease +// attempt is already against it. +func TestFallbackAnchorIsRefusedWhenTheCacheLandsInsideTheWorkspace(t *testing.T) { + isolateSandboxRuntimeRoots(t) + target := fallbackAnchorFixture(t) + workspace := t.TempDir() + + previous := sandboxUserCacheDir + t.Cleanup(func() { sandboxUserCacheDir = previous }) + sandboxUserCacheDir = func() (string, error) { return filepath.Join(workspace, ".cache"), nil } + + _, release, err := prepareSandboxRuntime(workspace) + if err == nil { + release() + t.Fatal("preparation went through a junction at the fallback anchor and reported success") + } + if !strings.Contains(err.Error(), "fallback anchor") { + t.Errorf("the refusal does not name the anchor, so the operator cannot act on it: %v", err) + } + assertNothingUnder(t, target) +} + +// Trigger two: the cache root cannot be leased (a FILE sits where the cache +// runtime tree would go), so preparation falls back explicitly. +func TestFallbackAnchorIsRefusedWhenTheCacheRootCannotBeLeased(t *testing.T) { + isolateSandboxRuntimeRoots(t) + target := fallbackAnchorFixture(t) + workspace := t.TempDir() + + // Make the cache-derived root unleasable at its ROOT: the user cache is a + // regular file, so nothing beneath it can be created and the first lease + // fails, which is the path that asks for the fallback explicitly. The + // first draft planted the file at the derived leaf instead, and the + // preparation failed on that mkdir before ever reaching the fallback + // branch, so the trigger under test never fired. + cache := filepath.Join(t.TempDir(), "cache-is-a-file") + if err := os.WriteFile(cache, []byte("not a directory"), 0o600); err != nil { + t.Fatal(err) + } + + previous := sandboxUserCacheDir + t.Cleanup(func() { sandboxUserCacheDir = previous }) + sandboxUserCacheDir = func() (string, error) { return cache, nil } + + _, release, err := prepareSandboxRuntime(workspace) + if err == nil { + release() + t.Fatal("preparation fell back through a junction at the anchor and reported success") + } + if !strings.Contains(err.Error(), "fallback anchor") { + t.Errorf("the refusal does not name the anchor: %v", err) + } + assertNothingUnder(t, target) +} + +// And the honest control: with nothing planted, the fallback prepares, the +// anchor is a real directory this test can see, and the runtime tree sits +// beneath it. +func TestFallbackAnchorIsCreatedPrivatelyWhenNothingIsInTheWay(t *testing.T) { + tempRoot := t.TempDir() + t.Setenv("TMP", tempRoot) + t.Setenv("TEMP", tempRoot) + workspace := t.TempDir() + + previous := sandboxUserCacheDir + t.Cleanup(func() { sandboxUserCacheDir = previous }) + sandboxUserCacheDir = func() (string, error) { return filepath.Join(workspace, ".cache"), nil } + + runtimeState, release, err := prepareSandboxRuntime(workspace) + if err != nil { + t.Fatalf("a clean fallback was refused: %v", err) + } + defer release() + + anchor := fallbackRuntimeAnchor() + info, err := os.Lstat(anchor) + if err != nil || !info.IsDir() { + t.Fatalf("the anchor was not created as a directory: err=%v", err) + } + if !pathWithinRoot(anchor, runtimeState.Root) { + t.Errorf("runtime root %s is not beneath the anchor %s", runtimeState.Root, anchor) + } +} diff --git a/internal/sandbox/runtime_fallback_test.go b/internal/sandbox/runtime_fallback_test.go new file mode 100644 index 000000000..2c5852d3f --- /dev/null +++ b/internal/sandbox/runtime_fallback_test.go @@ -0,0 +1,85 @@ +package sandbox + +import ( + "path/filepath" + "strings" + "testing" +) + +// SETUP AND EVERY COMMAND RUNNER MUST DERIVE THE SAME FALLBACK ROOT. +// +// They are separate processes. The fallback used to call os.MkdirTemp and cache +// the answer in a process-global map, so elevated setup granted the sandbox +// principal write access to the directory IT made, and each later +// __windows-command-runner made a different one and pointed TMP, GOCACHE and npm +// there. Those directories are created by the calling user and carry no ACE for +// the principal, so ordinary cache writes failed with a bare ACCESS_DENIED. +// +// A repeated call is what a second process looks like from here: no shared +// state, same inputs. The old implementation could not pass this, because the +// only thing that made repeat calls agree was the in-process map. +func TestFallbackRuntimeRootIsTheSameForEveryProcess(t *testing.T) { + workspace := filepath.Join(t.TempDir(), "ws") + + first, err := fallbackSandboxRuntimeRoot(workspace) + if err != nil { + t.Fatalf("fallbackSandboxRuntimeRoot: %v", err) + } + second, err := fallbackSandboxRuntimeRoot(workspace) + if err != nil { + t.Fatalf("fallbackSandboxRuntimeRoot (second process): %v", err) + } + if first != second { + t.Fatalf("two processes derived different runtime roots, so setup grants one and commands write to another:\n %s\n %s", first, second) + } + if first == "" { + t.Fatal("no runtime root derived") + } +} + +// Different workspaces must not share a runtime tree, or one workspace's +// principal would hold an ACE on another's caches. +func TestFallbackRuntimeRootIsPerWorkspace(t *testing.T) { + base := t.TempDir() + first, err := fallbackSandboxRuntimeRoot(filepath.Join(base, "alpha")) + if err != nil { + t.Fatal(err) + } + second, err := fallbackSandboxRuntimeRoot(filepath.Join(base, "beta")) + if err != nil { + t.Fatal(err) + } + if first == second { + t.Fatalf("two workspaces share the runtime root %q", first) + } +} + +// It must NAME the tree without creating it. Teardown asks for this path on its +// way out, and materializing a directory there would leave one behind on every +// teardown. +func TestFallbackRuntimeRootCreatesNothing(t *testing.T) { + workspace := filepath.Join(t.TempDir(), "ws") + root, err := fallbackSandboxRuntimeRoot(workspace) + if err != nil { + t.Fatal(err) + } + if pathExists(root) { + t.Errorf("naming the runtime root created %s", root) + } +} + +// The root must land outside the workspace, which is the entire reason this +// branch exists. +func TestFallbackRuntimeRootStaysOutsideTheWorkspace(t *testing.T) { + workspace := filepath.Join(t.TempDir(), "ws") + root, err := fallbackSandboxRuntimeRoot(workspace) + if err != nil { + t.Fatal(err) + } + if pathWithinRoot(workspace, root) { + t.Errorf("runtime root %q is inside workspace %q", root, workspace) + } + if !strings.Contains(root, "zero") { + t.Errorf("runtime root %q does not look like a zero-owned tree", root) + } +} diff --git a/internal/sandbox/runtime_lease.go b/internal/sandbox/runtime_lease.go index b05530fae..f7f68c708 100644 --- a/internal/sandbox/runtime_lease.go +++ b/internal/sandbox/runtime_lease.go @@ -24,6 +24,18 @@ func acquireSandboxRuntimeLease(root string) (*sandboxRuntimeLease, error) { return &sandboxRuntimeLease{handle: handle}, nil } +// acquireSandboxRuntimeLeaseIn takes the shared lease named name inside parent, +// which the caller has just validated: the directory is opened without +// following a link at its final component and the lease file is created or +// opened relative to that handle, never by a fresh pathname walk. +func acquireSandboxRuntimeLeaseIn(parent, name string) (*sandboxRuntimeLease, error) { + handle, err := acquireSharedRuntimeLeaseIn(parent, name) + if err != nil { + return nil, fmt.Errorf("acquire sandbox runtime lease: %w", err) + } + return &sandboxRuntimeLease{handle: handle}, nil +} + func tryAcquireSandboxRuntimeCleanupLease(root string) (*sandboxRuntimeLease, bool, error) { handle, inUse, err := tryAcquireExclusiveRuntimeLease(sandboxRuntimeLeasePath(root)) if err != nil || inUse { diff --git a/internal/sandbox/runtime_lease_redirect_unix_test.go b/internal/sandbox/runtime_lease_redirect_unix_test.go new file mode 100644 index 000000000..53ca3d3ba --- /dev/null +++ b/internal/sandbox/runtime_lease_redirect_unix_test.go @@ -0,0 +1,87 @@ +//go:build !windows + +package sandbox + +import ( + "os" + "path/filepath" + "strings" + "testing" +) + +// THE REUSABLE CHILD OF THE FALLBACK ANCHOR IS VALIDATED BEFORE THE LEASE. +// +// The anchor was checked, but its reusable "v1" child was created by pathname +// and the lease opened beside it, so a symlink planted at v1 by an earlier +// same-user sandboxed command sent the .lease into another directory before +// validation ever ran. With the fallback root now predictable and persistent, +// that state carried across processes. Preparation must refuse without any +// effect in the redirected target. +func TestPrepareSandboxRuntimeRefusesARedirectedFallbackChild(t *testing.T) { + _, tempRoot := isolateSandboxRuntimeRoots(t) + workspace := filepath.Join(t.TempDir(), "ws") + if err := os.MkdirAll(workspace, 0o700); err != nil { + t.Fatal(err) + } + // Force the fallback: the cache root inside the workspace is unusable. + original := sandboxUserCacheDir + sandboxUserCacheDir = func() (string, error) { return filepath.Join(workspace, ".cache"), nil } + t.Cleanup(func() { sandboxUserCacheDir = original }) + + anchor := fallbackRuntimeAnchor() + // Compared physically: the anchor is derived from the resolved temp root, + // and on macOS the fixture temp under /var resolves to /private/var. + physicalTemp, err := filepath.EvalSymlinks(tempRoot) + if err != nil { + t.Fatalf("resolve fixture temp: %v", err) + } + if rel, err := filepath.Rel(physicalTemp, anchor); err != nil || rel == ".." || filepath.IsAbs(rel) || strings.HasPrefix(rel, "..") { + t.Fatalf("SETUP INVALID: fallback anchor %s is not under the fixture temp %s", anchor, physicalTemp) + } + if err := os.MkdirAll(anchor, 0o700); err != nil { + t.Fatal(err) + } + elsewhere := filepath.Join(t.TempDir(), "elsewhere") + if err := os.MkdirAll(elsewhere, 0o700); err != nil { + t.Fatal(err) + } + if err := os.Symlink(elsewhere, filepath.Join(anchor, "v1")); err != nil { + t.Fatalf("symlink: %v", err) + } + + _, release, err := prepareSandboxRuntime(workspace) + if release != nil { + release() + } + if err == nil { + t.Fatal("preparation followed a redirected v1 instead of refusing") + } + entries, readErr := os.ReadDir(elsewhere) + if readErr != nil { + t.Fatal(readErr) + } + if len(entries) != 0 { + names := make([]string, 0, len(entries)) + for _, entry := range entries { + names = append(names, entry.Name()) + } + t.Fatalf("preparation left %v in the redirected target before refusing", names) + } + + // CONTROL: with v1 an ordinary directory, or absent, preparation works and + // its lease lands beside the root it selected. + if err := os.Remove(filepath.Join(anchor, "v1")); err != nil { + t.Fatal(err) + } + state, release, err := prepareSandboxRuntime(workspace) + if err != nil { + t.Fatalf("ordinary preparation: %v", err) + } + defer release() + if filepath.Dir(state.Root) != filepath.Join(anchor, "v1") { + t.Fatalf("selected root %s is not under the fallback anchor's v1", state.Root) + } + if _, err := os.Lstat(sandboxRuntimeLeasePath(state.Root)); err != nil { + t.Fatalf("lease was not created beside the selected root: %v", err) + } +} diff --git a/internal/sandbox/runtime_lease_redirect_windows_test.go b/internal/sandbox/runtime_lease_redirect_windows_test.go new file mode 100644 index 000000000..cfb892482 --- /dev/null +++ b/internal/sandbox/runtime_lease_redirect_windows_test.go @@ -0,0 +1,81 @@ +//go:build windows + +package sandbox + +import ( + "os" + "path/filepath" + "strings" + "testing" +) + +// THE REUSABLE CHILD OF THE FALLBACK ANCHOR IS VALIDATED BEFORE THE LEASE. +// +// The anchor was checked, but its reusable "v1" child was created by pathname +// and the lease opened beside it, so a junction planted at v1 sent the .lease +// into another directory before validation ever ran. Preparation must refuse +// without any effect in the redirected target. A junction, not a symlink, +// because that is what an unprivileged account on this platform can create. +func TestPrepareSandboxRuntimeRefusesARedirectedFallbackChild(t *testing.T) { + _, tempRoot := isolateSandboxRuntimeRoots(t) + workspace := filepath.Join(t.TempDir(), "ws") + if err := os.MkdirAll(workspace, 0o700); err != nil { + t.Fatal(err) + } + // Force the fallback: the cache root inside the workspace is unusable. + original := sandboxUserCacheDir + sandboxUserCacheDir = func() (string, error) { return filepath.Join(workspace, ".cache"), nil } + t.Cleanup(func() { sandboxUserCacheDir = original }) + + anchor := fallbackRuntimeAnchor() + if !strings.HasPrefix(strings.ToLower(anchor), strings.ToLower(tempRoot)) { + physical, _ := physicalDir(tempRoot) + if physical == "" || !strings.HasPrefix(strings.ToLower(anchor), strings.ToLower(physical)) { + t.Fatalf("SETUP INVALID: fallback anchor %s is not under the fixture temp %s", anchor, tempRoot) + } + } + if err := os.MkdirAll(anchor, 0o700); err != nil { + t.Fatal(err) + } + elsewhere := filepath.Join(t.TempDir(), "elsewhere") + if err := os.MkdirAll(elsewhere, 0o700); err != nil { + t.Fatal(err) + } + makeJunction(t, filepath.Join(anchor, "v1"), elsewhere) + + _, release, err := prepareSandboxRuntime(workspace) + if release != nil { + release() + } + if err == nil { + t.Fatal("preparation followed a redirected v1 instead of refusing") + } + entries, readErr := os.ReadDir(elsewhere) + if readErr != nil { + t.Fatal(readErr) + } + if len(entries) != 0 { + names := make([]string, 0, len(entries)) + for _, entry := range entries { + names = append(names, entry.Name()) + } + t.Fatalf("preparation left %v in the redirected target before refusing", names) + } + + // CONTROL: with the junction gone, preparation works and its lease lands + // beside the root it selected. + if err := os.Remove(filepath.Join(anchor, "v1")); err != nil { + t.Fatal(err) + } + state, release, err := prepareSandboxRuntime(workspace) + if err != nil { + t.Fatalf("ordinary preparation: %v", err) + } + defer release() + if !strings.EqualFold(filepath.Dir(state.Root), filepath.Join(anchor, "v1")) { + t.Fatalf("selected root %s is not under the fallback anchor's v1", state.Root) + } + if _, err := os.Lstat(sandboxRuntimeLeasePath(state.Root)); err != nil { + t.Fatalf("lease was not created beside the selected root: %v", err) + } +} diff --git a/internal/sandbox/runtime_lease_unix.go b/internal/sandbox/runtime_lease_unix.go index 3f51f6d90..855085419 100644 --- a/internal/sandbox/runtime_lease_unix.go +++ b/internal/sandbox/runtime_lease_unix.go @@ -4,7 +4,9 @@ package sandbox import ( "errors" + "fmt" "os" + "path/filepath" "golang.org/x/sys/unix" ) @@ -13,6 +15,24 @@ type runtimeLeaseHandle struct { file *os.File } +func acquireSharedRuntimeLeaseIn(parent, name string) (runtimeLeaseHandle, error) { + dirFD, err := unix.Open(parent, unix.O_RDONLY|unix.O_DIRECTORY|unix.O_NOFOLLOW|unix.O_CLOEXEC, 0) + if err != nil { + return runtimeLeaseHandle{}, fmt.Errorf("open sandbox runtime parent %s: %w", parent, err) + } + defer unix.Close(dirFD) + fd, err := unix.Openat(dirFD, name, unix.O_CREAT|unix.O_RDWR|unix.O_NOFOLLOW|unix.O_CLOEXEC, 0o600) + if err != nil { + return runtimeLeaseHandle{}, fmt.Errorf("open sandbox runtime lease %s: %w", name, err) + } + file := os.NewFile(uintptr(fd), filepath.Join(parent, name)) + if err := unix.Flock(int(file.Fd()), unix.LOCK_SH); err != nil { + _ = file.Close() + return runtimeLeaseHandle{}, err + } + return runtimeLeaseHandle{file: file}, nil +} + func acquireSharedRuntimeLease(path string) (runtimeLeaseHandle, error) { file, err := os.OpenFile(path, os.O_CREATE|os.O_RDWR, 0o600) if err != nil { diff --git a/internal/sandbox/runtime_lease_windows.go b/internal/sandbox/runtime_lease_windows.go index a27594728..2b601d0ef 100644 --- a/internal/sandbox/runtime_lease_windows.go +++ b/internal/sandbox/runtime_lease_windows.go @@ -4,7 +4,10 @@ package sandbox import ( "errors" + "fmt" "os" + "path/filepath" + "unsafe" "golang.org/x/sys/windows" ) @@ -14,6 +17,58 @@ type runtimeLeaseHandle struct { overlapped windows.Overlapped } +// acquireSharedRuntimeLeaseIn opens or creates the lease relative to a handle on +// its parent that refuses a reparse point at the final component, so the +// pathname the caller just validated cannot be swapped underneath the open. +func acquireSharedRuntimeLeaseIn(parent, name string) (runtimeLeaseHandle, error) { + dir, err := openWindowsACLDirectoryNoFollow(parent) + if err != nil { + return runtimeLeaseHandle{}, fmt.Errorf("open sandbox runtime parent: %w", err) + } + defer func() { _ = windows.CloseHandle(dir) }() + if err := validateWindowsACLComponent(name); err != nil { + return runtimeLeaseHandle{}, err + } + objectName, err := windows.NewNTUnicodeString(name) + if err != nil { + return runtimeLeaseHandle{}, fmt.Errorf("encode sandbox runtime lease %s: %w", name, err) + } + attributes := windows.OBJECT_ATTRIBUTES{ + RootDirectory: dir, + ObjectName: objectName, + Attributes: windows.OBJ_CASE_INSENSITIVE, + } + attributes.Length = uint32(unsafe.Sizeof(attributes)) + var handle windows.Handle + var status windows.IO_STATUS_BLOCK + if err := windows.NtCreateFile( + &handle, + windows.GENERIC_READ|windows.GENERIC_WRITE|windows.SYNCHRONIZE, + &attributes, + &status, + nil, + windows.FILE_ATTRIBUTE_NORMAL, + windows.FILE_SHARE_READ|windows.FILE_SHARE_WRITE|windows.FILE_SHARE_DELETE, + windows.FILE_OPEN_IF, + windows.FILE_NON_DIRECTORY_FILE|windows.FILE_OPEN_REPARSE_POINT|windows.FILE_SYNCHRONOUS_IO_NONALERT, + 0, + 0, + ); err != nil { + return runtimeLeaseHandle{}, fmt.Errorf("open sandbox runtime lease %s: %w", name, err) + } + if err := rejectWindowsACLReparseHandle(handle, name); err != nil { + _ = windows.CloseHandle(handle) + return runtimeLeaseHandle{}, err + } + file := os.NewFile(uintptr(handle), filepath.Join(parent, name)) + lease := runtimeLeaseHandle{file: file} + if err := windows.LockFileEx(windows.Handle(file.Fd()), 0, 0, 1, 0, &lease.overlapped); err != nil { + _ = file.Close() + return runtimeLeaseHandle{}, err + } + return lease, nil +} + func acquireSharedRuntimeLease(path string) (runtimeLeaseHandle, error) { file, err := os.OpenFile(path, os.O_CREATE|os.O_RDWR, 0o600) if err != nil { diff --git a/internal/sandbox/runtime_state.go b/internal/sandbox/runtime_state.go index 4a5fdfc9a..2907db5f3 100644 --- a/internal/sandbox/runtime_state.go +++ b/internal/sandbox/runtime_state.go @@ -7,10 +7,14 @@ import ( "fmt" "os" "path/filepath" + "runtime" "sort" + "strconv" "strings" "sync" "time" + + "github.com/Gitlawb/zero/internal/peermsg" ) var sandboxUserCacheDir = os.UserCacheDir @@ -21,11 +25,6 @@ const ( sandboxRuntimeMaxRoots = 64 ) -var fallbackSandboxRuntimes = struct { - sync.Mutex - roots map[string]string -}{roots: make(map[string]string)} - type SandboxRuntime struct { Root string `json:"root,omitempty"` Cache string `json:"cache,omitempty"` @@ -33,8 +32,37 @@ type SandboxRuntime struct { Temp string `json:"temp,omitempty"` } +// sandboxRuntimeRootFor derives the per-workspace runtime root. It is separated +// from prepareSandboxRuntime because the elevated Windows setup path needs the +// same answer WITHOUT taking a lease or creating anything: a sandbox principal +// is a separate account with no inherited rights under the user cache, so setup +// has to grant it write access to this tree before any command runs. +// +// Both callers must agree exactly. If they ever drift, setup grants the ACE on +// one directory while commands write to another, and the failure is a bare +// ACCESS_DENIED from npm or go build with nothing pointing at the sandbox. +func sandboxRuntimeRootFor(workspaceRoot string, cacheRoot string) (string, error) { + if root, ok := deterministicSandboxRuntimeRoot(workspaceRoot, cacheRoot); ok { + return root, nil + } + return fallbackSandboxRuntimeRoot(workspaceRoot) +} + +// deterministicSandboxRuntimeRoot returns the cache-derived runtime root and +// whether it is usable, meaning it lands outside the workspace. +// +// Neither this nor the fallback creates anything now, so a caller that only +// needs to NAME the tree can safely go through sandboxRuntimeRootFor and get the +// answer commands will actually use. This remains separate for callers that need +// to distinguish the cache-derived root from the temp-derived one. +func deterministicSandboxRuntimeRoot(workspaceRoot string, cacheRoot string) (string, bool) { + digest := sha256.Sum256([]byte(workspaceRoot)) + root := filepath.Join(cacheRoot, "zero", "runtime", "v1", hex.EncodeToString(digest[:8])) + return root, !pathWithinRoot(workspaceRoot, root) +} + func prepareSandboxRuntime(workspaceRoot string) (SandboxRuntime, func(), error) { - workspaceRoot = filepath.Clean(strings.TrimSpace(workspaceRoot)) + workspaceRoot = canonicalSandboxWorkspaceRoot(workspaceRoot) if workspaceRoot == "" || workspaceRoot == "." { return SandboxRuntime{}, nil, errors.New("sandbox runtime requires a workspace root") } @@ -42,17 +70,30 @@ func prepareSandboxRuntime(workspaceRoot string) (SandboxRuntime, func(), error) if err != nil { return SandboxRuntime{}, nil, fmt.Errorf("resolve user cache directory: %w", err) } - cacheRoot = filepath.Clean(strings.TrimSpace(cacheRoot)) + // Canonicalized the SAME way as the workspace root, because + // sandboxRuntimeRootFor compares the two: it falls back to a private temp + // tree when the derived runtime root would land inside the workspace. + // Normalizing only one side made that comparison run on two different + // spellings of the same path — /var vs /private/var on macOS, an 8.3 short + // name vs its long form on Windows — so the containment check missed and the + // fallback never fired. + cacheRoot = canonicalSandboxWorkspaceRoot(cacheRoot) if cacheRoot == "" || cacheRoot == "." { return SandboxRuntime{}, nil, errors.New("user cache directory is unavailable") } - digest := sha256.Sum256([]byte(workspaceRoot)) - root := filepath.Join(cacheRoot, "zero", "runtime", "v1", hex.EncodeToString(digest[:8])) - if pathWithinRoot(workspaceRoot, root) { - root, err = fallbackSandboxRuntimeRoot(workspaceRoot) - if err != nil { - return SandboxRuntime{}, nil, err - } + root, err := sandboxRuntimeRootFor(workspaceRoot, cacheRoot) + if err != nil { + return SandboxRuntime{}, nil, err + } + // THE ANCHOR IS PROVEN BEFORE THE TREE IS BUILT, on both paths that can + // hand back a fallback root: sandboxRuntimeRootFor returns one itself when + // the cache lands inside the workspace, and the branch below asks for one + // when the cache root cannot be leased. A deterministic name under the + // shared temp is only safe beneath a directory this user owns; if the anchor + // is a link, or exists and belongs to someone else, this refuses rather than + // leasing, creating and chmod-ing through it. + if err := ensureFallbackRuntimeAnchor(root); err != nil { + return SandboxRuntime{}, nil, err } lease, err := prepareSandboxRuntimeLease(root) if err != nil { @@ -60,6 +101,9 @@ func prepareSandboxRuntime(workspaceRoot string) (SandboxRuntime, func(), error) if err != nil { return SandboxRuntime{}, nil, err } + if err := ensureFallbackRuntimeAnchor(root); err != nil { + return SandboxRuntime{}, nil, err + } lease, err = prepareSandboxRuntimeLease(root) if err != nil { return SandboxRuntime{}, nil, err @@ -90,13 +134,13 @@ func prepareSandboxRuntime(workspaceRoot string) (SandboxRuntime, func(), error) filepath.Join(runtimeState.Data, "go-mod"), filepath.Join(runtimeState.Data, "cargo"), } - for _, directory := range directories { - if err := os.MkdirAll(directory, 0o700); err != nil { - return SandboxRuntime{}, nil, fmt.Errorf("create sandbox runtime directory %s: %w", directory, err) - } - if err := os.Chmod(directory, 0o700); err != nil { - return SandboxRuntime{}, nil, fmt.Errorf("secure sandbox runtime directory %s: %w", directory, err) - } + // Handle-relative and no-follow below the operator-owned base. The tree is + // persistent input from the previous sandboxed command, which is granted write + // access to cache, data and tmp and can leave a link in place of one of them. + // os.MkdirAll and os.Chmod both follow, so the host process created package + // caches wherever that command pointed them. See ensureRuntimeTreeDirs. + if err := ensureRuntimeTreeDirs(runtimeState.Root, directories); err != nil { + return SandboxRuntime{}, nil, err } now := sandboxRuntimeNow() if err := os.Chtimes(runtimeState.Root, now, now); err != nil { @@ -108,10 +152,20 @@ func prepareSandboxRuntime(workspaceRoot string) (SandboxRuntime, func(), error) } func prepareSandboxRuntimeLease(root string) (*sandboxRuntimeLease, error) { - if err := os.MkdirAll(filepath.Dir(root), 0o700); err != nil { - return nil, fmt.Errorf("create sandbox runtime parent: %w", err) + // THE OWNED TAIL IS VALIDATED BEFORE THE FIRST EFFECT, AND THE LEASE IS + // OPENED THROUGH THAT VALIDATION. The anchor check above this covers the + // anchor; the reusable "v1" child beneath it is the runtime's too, and + // creating it by pathname followed whatever was sitting at that name. + // Redirect v1 into another directory and the old sequence created a .lease + // there and only then failed validation. EnsurePrivateDir walks each + // component without following links and creates what is missing in place; + // the lease is then opened relative to a no-follow handle on that parent, so + // a component swapped in after the check cannot move the open either. + parent := filepath.Dir(root) + if err := peermsg.EnsurePrivateDir(parent); err != nil { + return nil, fmt.Errorf("sandbox runtime parent %s is not a private directory owned by this user: %w", parent, err) } - return acquireSandboxRuntimeLease(root) + return acquireSandboxRuntimeLeaseIn(parent, filepath.Base(root)+sandboxRuntimeLeaseSuffix) } // cleanupSandboxRuntimeRoots applies a conservative age/count policy. Cleanup @@ -174,22 +228,109 @@ func combineSandboxCleanups(cleanups ...func()) func() { } } -func fallbackSandboxRuntimeRoot(workspaceRoot string) (string, error) { - fallbackSandboxRuntimes.Lock() - defer fallbackSandboxRuntimes.Unlock() - if root := fallbackSandboxRuntimes.roots[workspaceRoot]; root != "" { - return root, nil +// fallbackRuntimeAnchor is the per-user directory the deterministic fallback +// runtime tree lives beneath. +// +// DETERMINISM IS NOT OWNERSHIP. The fallback moved from a fresh, private +// os.MkdirTemp parent to a predictable name so that elevated setup and the +// later command process agree on the runtime root. That agreement is required +// and must stay. But a predictable path directly under the shared system temp +// is not a private allocation: on an ordinary shared /tmp another local user +// can create /tmp/zero with mode 0700 first, and every affected user then fails +// before the sandbox command starts; a pre-existing redirected component makes +// host-side preparation operate somewhere other than the derived tree. +// +// So the tree is rooted beneath an anchor that names THIS user, and the anchor +// is created and validated by peermsg.EnsurePrivateDir before anything is +// built under it: handle-relative descent, no-follow, refused if any component +// is a link or the leaf is not owned by the current user, mode 0700. On Unix +// the uid is in the name because /tmp is shared between users; on Windows +// os.TempDir is already per-user, and the private descriptor plus the owner +// check do the rest. +func fallbackRuntimeAnchor() string { + tag := "user" + if runtime.GOOS != "windows" { + tag = strconv.Itoa(os.Getuid()) } - parent, err := os.MkdirTemp("", "zero-runtime-") - if err != nil { - return "", fmt.Errorf("create fallback sandbox runtime: %w", err) + // PHYSICAL PARENT, OWNED LEAF. The validator walks every component no-follow + // and refuses links, which is right for the anchor Zero owns and wrong for + // the operator's ancestors above it: macOS puts TempDir under /var, a + // symlink to /private/var, and a redirected %LOCALAPPDATA% is ordinary on + // Windows. The first version handed the validator the unresolved path and + // refused every fallback on every Mac. Resolving the parent first means the + // only new component the validator sees is the anchor itself. + // + // On a resolve error the unresolved path is used, deliberately: the + // validator is still the fail-closed check, and refusing there names the + // real component rather than hiding it behind a resolver failure. + base := os.TempDir() + if physical, err := physicalTempDir(); err == nil && physical != "" { + base = physical } - root := filepath.Join(parent, "runtime") + return filepath.Join(base, "zero-runtime-"+tag) +} + +// isFallbackRuntimeRoot reports whether root was derived by +// fallbackSandboxRuntimeRoot, so the caller knows the anchor must be validated +// before the tree beneath it is prepared. Both candidates a command derives can +// be fallback roots: sandboxRuntimeRootFor returns one itself when the cache +// lands inside the workspace, and prepareSandboxRuntime asks for one when the +// cache root cannot be leased. +func isFallbackRuntimeRoot(root string) bool { + // pathWithinRoot(parent, child): is root beneath the anchor. The first + // draft had these reversed, which asked whether the anchor was beneath the + // root, answered false every time, and skipped the validation entirely + // while every test that depended on it reported success through a + // junction. Argument order is the whole function. + return pathWithinRoot(fallbackRuntimeAnchor(), root) +} + +// ensureFallbackRuntimeAnchor validates the anchor beneath which a fallback +// runtime tree is about to be built, and fails closed. Called only when the +// root is a fallback root; the cache-derived root lives under the user cache, +// which is already the user's own directory. +func ensureFallbackRuntimeAnchor(root string) error { + if !isFallbackRuntimeRoot(root) { + return nil + } + anchor := fallbackRuntimeAnchor() + if err := peermsg.EnsurePrivateDir(anchor); err != nil { + return fmt.Errorf("sandbox runtime fallback anchor %s is not a private directory owned by this user: %w", anchor, err) + } + return nil +} + +// fallbackSandboxRuntimeRoot returns the runtime root for a workspace whose +// cache-derived root would land inside itself. +// +// DERIVED, not minted, and that is the whole of the fix. It used to call +// os.MkdirTemp and remember the answer in a process-global map, which made the +// result private to whichever process asked first. Elevated Windows setup +// granted the sandbox principal write access to the directory IT created, then +// every later __windows-command-runner process created a DIFFERENT one and +// pointed TMP, GOCACHE, npm and the rest at it. Those directories are created by +// the calling user and carry no ACE for the principal, so ordinary cache and +// temp writes failed with a bare ACCESS_DENIED and nothing naming the sandbox. +// +// sandboxRuntimeRootFor already documents that both callers must agree exactly. +// Hashing the workspace, the same way the cache-derived root does, is what makes +// that true for this branch too: every process reaches the same path without +// having to share any state. +// +// It creates nothing, so deterministicSandboxRuntimeRoot's promise about naming +// a tree without materializing it now holds for the fallback as well. +func fallbackSandboxRuntimeRoot(workspaceRoot string) (string, error) { + digest := sha256.Sum256([]byte(workspaceRoot)) + root := filepath.Join(fallbackRuntimeAnchor(), "v1", hex.EncodeToString(digest[:8])) if pathWithinRoot(workspaceRoot, root) { - _ = os.RemoveAll(parent) - return "", fmt.Errorf("fallback sandbox runtime root %q must be outside workspace %q", root, workspaceRoot) + // Both candidates land inside the workspace, so there is nowhere left to + // put a runtime tree the workspace's own policy does not govern. Refused + // rather than pointed somewhere arbitrary: a runtime root inside the + // workspace makes the sandbox's own cache writes indistinguishable from + // the work it is meant to be confining. + return "", fmt.Errorf("sandbox runtime root %q would fall inside workspace %q; "+ + "open the workspace somewhere other than the cache or temp directory", root, workspaceRoot) } - fallbackSandboxRuntimes.roots[workspaceRoot] = root return root, nil } @@ -226,3 +367,59 @@ func permissionProfileWithRuntime(profile PermissionProfile, runtimeState Sandbo profile.FileSystem.WriteRoots = append(profile.FileSystem.WriteRoots, WritableRoot{Root: runtimeState.Root}) return profile } + +// canonicalSandboxWorkspaceRoot normalizes a workspace root the way +// Engine.resolveCommandDir already does — clean, absolutize, then resolve +// symlinks — so every derivation keyed to a workspace agrees on the string. +// +// The runtime root is a hash of this, and the elevated Windows setup grants the +// principal that tree while commands derive it again. Cleaning alone was not +// enough for the two to agree, and it does not take a symlink for them to +// differ: a path opened in different casing, or through an 8.3 short name (what +// a Windows CI runner's TEMP looks like), resolves to a different spelling. +// Setup then granted one tree and every command used another, so the grant that +// makes npm/go/pip caches writable landed where nothing reads and surfaced as a +// bare ACCESS_DENIED. +// +// Resolution failing is not an error: an unresolvable root still needs a stable +// key, and falling back to the cleaned absolute path is what the command path +// does too. +func canonicalSandboxWorkspaceRoot(root string) string { + cleaned := filepath.Clean(strings.TrimSpace(root)) + if cleaned == "" || cleaned == "." { + return "" + } + if !filepath.IsAbs(cleaned) { + if absolute, err := filepath.Abs(cleaned); err == nil { + cleaned = absolute + } + } + // EvalSymlinks fails outright when the LEAF does not exist, which is the + // normal case for a cache or runtime root that has not been created yet. A + // plain call therefore resolved an existing workspace while leaving a + // not-yet-created cache root unresolved, and the two were compared against + // each other — the containment check that decides whether the runtime tree + // must move out of the workspace then ran on /private/var/... versus + // /var/..., missed, and left the tree inside the workspace. + // + // Resolve the longest existing ancestor and re-append the rest, so a path + // normalizes the same way whether or not its final segments exist yet. + remainder := "" + current := cleaned + for { + if resolved, err := filepath.EvalSymlinks(current); err == nil { + if remainder == "" { + return resolved + } + return filepath.Join(resolved, remainder) + } + parent := filepath.Dir(current) + if parent == current { + // Nothing along the path resolved; the cleaned absolute form is the + // best stable key available. + return cleaned + } + remainder = filepath.Join(filepath.Base(current), remainder) + current = parent + } +} diff --git a/internal/sandbox/runtime_state_test.go b/internal/sandbox/runtime_state_test.go index b707f67f2..5bc51808a 100644 --- a/internal/sandbox/runtime_state_test.go +++ b/internal/sandbox/runtime_state_test.go @@ -10,6 +10,7 @@ import ( ) func TestPrepareSandboxRuntimeStaysOutsideWorkspace(t *testing.T) { + isolateSandboxRuntimeRoots(t) workspace := t.TempDir() cacheRoot := t.TempDir() original := sandboxUserCacheDir @@ -39,6 +40,7 @@ func TestPrepareSandboxRuntimeStaysOutsideWorkspace(t *testing.T) { } func TestPrepareSandboxRuntimeCleansExpiredSibling(t *testing.T) { + isolateSandboxRuntimeRoots(t) workspace := t.TempDir() cacheRoot := t.TempDir() now := time.Date(2026, 7, 20, 12, 0, 0, 0, time.UTC) @@ -70,6 +72,7 @@ func TestPrepareSandboxRuntimeCleansExpiredSibling(t *testing.T) { } func TestPrepareSandboxRuntimeFallsBackWhenUserCacheIsInsideWorkspace(t *testing.T) { + isolateSandboxRuntimeRoots(t) workspace := t.TempDir() original := sandboxUserCacheDir sandboxUserCacheDir = func() (string, error) { return filepath.Join(workspace, ".cache"), nil } @@ -89,6 +92,7 @@ func TestPrepareSandboxRuntimeFallsBackWhenUserCacheIsInsideWorkspace(t *testing } func TestCleanupSandboxRuntimeSkipsActiveLease(t *testing.T) { + isolateSandboxRuntimeRoots(t) workspace := t.TempDir() cacheRoot := t.TempDir() now := time.Date(2026, 7, 20, 12, 0, 0, 0, time.UTC) @@ -329,3 +333,106 @@ func TestEngineCommandPlanCarriesManagedRuntime(t *testing.T) { } cleanupLease.release() } + +// sandboxRuntimeRootFor compares the workspace root against the derived runtime +// root to decide whether to fall back to a private temp tree. Both sides +// therefore have to be the same spelling of the same path. +// +// Canonicalizing only the workspace root broke this on CI: the workspace +// resolved (/var to /private/var on macOS, an 8.3 short name to its long form +// on Windows) while the cache root kept its original spelling, so the +// containment check compared two different strings, the fallback never fired, +// and the runtime tree was placed inside the workspace it exists to stay out of. +// +// A symlink is the portable way to produce a spelling that only resolution +// reconciles — Clean cannot see through one. Windows refuses to create symlinks +// without privilege, so this skips there; the platforms that CI caught the bug +// on are the ones that run it. +func TestPrepareSandboxRuntimeNormalizesTheCacheRootBeforeComparingIt(t *testing.T) { + isolateSandboxRuntimeRoots(t) + workspace := t.TempDir() + link := filepath.Join(t.TempDir(), "workspace-link") + if err := os.Symlink(workspace, link); err != nil { + t.Skipf("cannot create a symlink here: %v", err) + } + // The cache root reaches us spelled through the symlink; the workspace does + // not. Resolved, it is plainly inside the workspace and the fallback must + // fire. Unresolved, the two strings share no prefix and it does not. + original := sandboxUserCacheDir + sandboxUserCacheDir = func() (string, error) { return filepath.Join(link, ".cache"), nil } + t.Cleanup(func() { sandboxUserCacheDir = original }) + + runtimeState, release, err := prepareSandboxRuntime(workspace) + if err != nil { + t.Fatalf("prepareSandboxRuntime: %v", err) + } + if release != nil { + defer release() + } + // Resolve before comparing. Spelled through the link the runtime root shares + // no textual prefix with the workspace, so an unresolved comparison would + // call it "outside" while it sits physically inside — the test would pass + // against the very bug it exists for. + resolved := runtimeState.Root + if actual, err := filepath.EvalSymlinks(runtimeState.Root); err == nil { + resolved = actual + } + if pathWithinRoot(workspace, resolved) { + t.Fatalf("runtime root %q resolves to %q, inside workspace %q; the containment check did not see through the cache root's spelling", runtimeState.Root, resolved, workspace) + } +} + +// A path whose final segments do not exist yet must still normalize the same +// way as one that does. This is the shape macOS CI hit: t.TempDir() sits under +// /var, a symlink to /private/var, and the cache root it derives has not been +// created when it is first normalized. Resolving only the workspace left the +// two sides of the containment check spelled differently. +func TestCanonicalSandboxWorkspaceRootResolvesThroughAMissingLeaf(t *testing.T) { + real := t.TempDir() + link := filepath.Join(t.TempDir(), "link") + if err := os.Symlink(real, link); err != nil { + t.Skipf("cannot create a symlink here: %v", err) + } + + existing := canonicalSandboxWorkspaceRoot(link) + if existing != canonicalSandboxWorkspaceRoot(real) { + t.Fatalf("an existing symlinked dir did not resolve: %q vs %q", existing, canonicalSandboxWorkspaceRoot(real)) + } + + // The leaf, and its parent, do not exist. + missing := filepath.Join(link, ".cache", "zero") + got := canonicalSandboxWorkspaceRoot(missing) + want := filepath.Join(existing, ".cache", "zero") + if got != want { + t.Errorf("missing leaf normalized to %q, want %q — the ancestor was not resolved", got, want) + } + if !pathWithinRoot(existing, got) { + t.Errorf("%q should be inside %q once both are canonical", got, existing) + } +} + +// isolateSandboxRuntimeRoots gives a test its own cache root AND its own temp +// root, so every runtime tree the production code can select, the preferred +// cache-derived one and the temp-derived fallback, lands under the fixture. +// Workspace isolation is not runtime-storage isolation: those paths come from +// separate resolver inputs, and a test that redirected only the workspace ran +// real setup, preparation and reclamation against the developer's own cache +// and temp. Both roots are returned for assertions on where things landed. +func isolateSandboxRuntimeRoots(t *testing.T) (cacheRoot, tempRoot string) { + t.Helper() + cacheRoot = filepath.Join(t.TempDir(), "cache") + tempRoot = filepath.Join(t.TempDir(), "temp") + for _, dir := range []string{cacheRoot, tempRoot} { + if err := os.MkdirAll(dir, 0o700); err != nil { + t.Fatal(err) + } + } + previous := sandboxUserCacheDir + sandboxUserCacheDir = func() (string, error) { return cacheRoot, nil } + t.Cleanup(func() { sandboxUserCacheDir = previous }) + // os.TempDir reads these at call time on every platform this runs on. + t.Setenv("TMP", tempRoot) + t.Setenv("TEMP", tempRoot) + t.Setenv("TMPDIR", tempRoot) + return cacheRoot, tempRoot +} diff --git a/internal/sandbox/runtime_tree.go b/internal/sandbox/runtime_tree.go new file mode 100644 index 000000000..4df3da27c --- /dev/null +++ b/internal/sandbox/runtime_tree.go @@ -0,0 +1,88 @@ +package sandbox + +import ( + "fmt" + "path/filepath" + "strings" +) + +// ensureRuntimeTreeDirs creates every directory of the runtime tree without ever +// following a link below the operator-owned base. +// +// THE TREE IS PERSISTENT INPUT FROM THE PREVIOUS SANDBOXED COMMAND. +// +// The runtime root is deterministic so setup and later runners agree on one +// path, and the sandboxed command is granted write access to cache, data and +// tmp. So the command can replace one of them with a symlink or a Windows +// junction on its way out. The next preparation ran os.MkdirAll and os.Chmod on +// raw pathnames, which follow, and the HOST Zero process, the ordinary user +// rather than the confined principal, then created package-cache directories +// inside a target the previous command chose. +// +// The anchor check is not enough on its own: it proves the per-user anchor and +// says nothing about the reusable root or any descendant beneath it. +// +// NOT peermsg.EnsurePrivateDir. That ends in a protected-DACL write, which on +// Windows strips the sandbox principal's grant that elevated setup installed on +// the runtime tree and brings back the bare access denials from npm and go that +// the grant exists to prevent. This descent validates and creates; it does not +// re-secure. +func ensureRuntimeTreeDirs(root string, directories []string) error { + // Canonicalized the same way the rest of the runtime state is, because + // runtimeCandidateBase decides ownership with a containment test and that + // test runs on spellings. An 8.3 short name (C:UsersRUNNER~1) or a + // symlinked temp compares unequal to the long form of the same directory, + // which reads as "no operator-owned base" for a perfectly ordinary tree. + if canonical := canonicalSandboxWorkspaceRoot(root); canonical != "" { + root = canonical + } + base, ok := runtimeCandidateBase(root) + if !ok || strings.TrimSpace(base) == "" { + return fmt.Errorf("sandbox runtime root %s has no operator-owned base to descend from", root) + } + // Physical, so a redirected cache or TEMP above the tree stays the operator's + // business while everything Zero owns below it is addressed no-follow. + physical, err := physicalDir(base) + if err != nil { + return fmt.Errorf("resolve the sandbox runtime base %s: %w", base, err) + } + for _, directory := range directories { + tail, err := runtimeTreeTail(physical, base, directory) + if err != nil { + return err + } + if err := ensureRuntimeTreeDir(physical, tail); err != nil { + return err + } + } + return nil +} + +// runtimeTreeTail returns the components of directory below the base, which are +// the ones Zero owns and the ones that must not be links. +func runtimeTreeTail(physicalBase string, base string, directory string) ([]string, error) { + cleaned := filepath.Clean(strings.TrimSpace(directory)) + relative, err := filepath.Rel(base, cleaned) + if err != nil || relative == "." || strings.HasPrefix(relative, "..") { + // Try the physical spelling too: the caller's path may already be physical. + relative, err = filepath.Rel(physicalBase, cleaned) + if err != nil || relative == "." || strings.HasPrefix(relative, "..") { + return nil, fmt.Errorf("sandbox runtime directory %s does not sit under its base %s", directory, base) + } + } + parts := strings.Split(relative, string(filepath.Separator)) + out := make([]string, 0, len(parts)) + for _, part := range parts { + if part == "" || part == "." { + continue + } + if part == ".." { + return nil, fmt.Errorf("sandbox runtime directory %s escapes its base %s", directory, base) + } + out = append(out, part) + } + if len(out) == 0 { + return nil, fmt.Errorf("sandbox runtime directory %s resolves to its own base", directory) + } + return out, nil +} diff --git a/internal/sandbox/runtime_tree_other.go b/internal/sandbox/runtime_tree_other.go new file mode 100644 index 000000000..e0a3f4439 --- /dev/null +++ b/internal/sandbox/runtime_tree_other.go @@ -0,0 +1,49 @@ +//go:build !windows + +package sandbox + +import ( + "fmt" + "os" + "path/filepath" + + "golang.org/x/sys/unix" +) + +// ensureRuntimeTreeDir creates the owned tail beneath a physical base using +// openat and mkdirat with O_NOFOLLOW, so a symlink planted at any component is +// refused rather than followed. +// +// fchmod on the retained descriptor rather than chmod on the pathname: chmod(2) +// follows symlinks, which is the half of this defect that actually applies here. +// Every owned component gets the mode, not only the leaf, because every one of +// them is ours. +func ensureRuntimeTreeDir(physicalBase string, tail []string) error { + parent, err := unix.Open(physicalBase, unix.O_RDONLY|unix.O_DIRECTORY|unix.O_CLOEXEC, 0) + if err != nil { + return fmt.Errorf("open the sandbox runtime base %s: %w", physicalBase, err) + } + defer func() { _ = unix.Close(parent) }() + + current := physicalBase + for _, name := range tail { + current = filepath.Join(current, name) + if err := unix.Mkdirat(parent, name, 0o700); err != nil && err != unix.EEXIST { + return fmt.Errorf("create sandbox runtime directory %s: %w", current, err) + } + child, openErr := unix.Openat(parent, name, unix.O_RDONLY|unix.O_DIRECTORY|unix.O_NOFOLLOW|unix.O_CLOEXEC, 0) + if openErr != nil { + if openErr == unix.ELOOP || openErr == unix.ENOTDIR { + return fmt.Errorf("sandbox runtime component %s is a link or not a directory, so a previous sandboxed command may have redirected it: %w", current, os.ErrInvalid) + } + return fmt.Errorf("open sandbox runtime directory %s: %w", current, openErr) + } + if err := unix.Fchmod(child, 0o700); err != nil { + _ = unix.Close(child) + return fmt.Errorf("secure sandbox runtime directory %s: %w", current, err) + } + _ = unix.Close(parent) + parent = child + } + return nil +} diff --git a/internal/sandbox/runtime_tree_redirect_windows_test.go b/internal/sandbox/runtime_tree_redirect_windows_test.go new file mode 100644 index 000000000..dc4a333e1 --- /dev/null +++ b/internal/sandbox/runtime_tree_redirect_windows_test.go @@ -0,0 +1,193 @@ +//go:build windows + +package sandbox + +import ( + "os" + "os/exec" + "path/filepath" + "testing" +) + +// A PREVIOUS SANDBOXED COMMAND CAN LEAVE A REDIRECT BEHIND. +// +// The runtime root is deterministic so setup and later runners agree on one +// path, and the sandboxed command is granted write access to cache, data and +// tmp. So it can replace one of them with a junction on its way out. The next +// preparation used os.MkdirAll and os.Chmod on raw pathnames, which follow, and +// the HOST Zero process, the ordinary user rather than the confined principal, +// then created package-cache directories inside a target the previous command +// chose. +// +// The anchor check does not cover this: it proves the per-user anchor and says +// nothing about the reusable root or anything beneath it. +// +// A junction needs no privilege, so the previous command's half runs here on an +// ordinary unelevated box. +func TestRuntimeTreePreparationRefusesARedirectedDescendant(t *testing.T) { + base := runtimeTreeTestBase(t) + target := t.TempDir() + root := filepath.Join(base, "zero", "runtime", "v1", "abcdef0123456789") + cache := filepath.Join(root, "cache") + if err := os.MkdirAll(root, 0o700); err != nil { + t.Fatal(err) + } + // The previous lifetime leaves a junction where cache used to be. + if out, err := exec.Command("cmd", "/c", "mklink", "/J", cache, target).CombinedOutput(); err != nil { + t.Skipf("mklink /J unavailable here: %v\n%s", err, out) + } + // SETUP: the junction really redirects, or a clean target proves nothing. + probe := filepath.Join(cache, "probe.txt") + if err := os.WriteFile(probe, []byte("x"), 0o600); err != nil { + t.Skipf("the junction does not accept writes here: %v", err) + } + if _, err := os.Stat(filepath.Join(target, "probe.txt")); err != nil { + t.Skipf("SETUP: the junction does not redirect on this filesystem: %v", err) + } + if err := os.Remove(probe); err != nil { + t.Fatal(err) + } + + err := ensureRuntimeTreeDirs(root, []string{root, cache, filepath.Join(cache, "npm")}) + if err == nil { + t.Fatal("preparation descended through a junction the previous sandboxed command planted") + } + + entries, readErr := os.ReadDir(target) + if readErr != nil { + t.Fatal(readErr) + } + if len(entries) != 0 { + names := make([]string, 0, len(entries)) + for _, entry := range entries { + names = append(names, entry.Name()) + } + t.Fatalf("preparation created %v beneath the redirected target", names) + } +} + +// And an ordinary tree is still created, or the refusal above would be satisfied +// by a preparation that refuses everything. +func TestRuntimeTreePreparationStillCreatesAnOrdinaryTree(t *testing.T) { + base := runtimeTreeTestBase(t) + root := filepath.Join(base, "zero", "runtime", "v1", "abcdef0123456789") + cache := filepath.Join(root, "cache") + npm := filepath.Join(cache, "npm") + + if err := ensureRuntimeTreeDirs(root, []string{root, cache, npm}); err != nil { + t.Fatalf("an ordinary runtime tree was refused: %v", err) + } + for _, dir := range []string{root, cache, npm} { + info, err := os.Stat(dir) + if err != nil || !info.IsDir() { + t.Fatalf("%s was not created: err=%v", dir, err) + } + } +} + +// Reusing an existing tree is the common case and must not be refused. +func TestRuntimeTreePreparationIsIdempotent(t *testing.T) { + base := runtimeTreeTestBase(t) + root := filepath.Join(base, "zero", "runtime", "v1", "abcdef0123456789") + dirs := []string{root, filepath.Join(root, "cache"), filepath.Join(root, "cache", "npm")} + if err := ensureRuntimeTreeDirs(root, dirs); err != nil { + t.Fatalf("first preparation: %v", err) + } + if err := ensureRuntimeTreeDirs(root, dirs); err != nil { + t.Fatalf("second preparation over an existing tree was refused: %v", err) + } +} + +// AND THE PRODUCTION ENTRY POINT USES IT, WHICH THE HELPER TESTS CANNOT SEE. +// +// The three tests above call ensureRuntimeTreeDirs directly, so reverting +// prepareSandboxRuntime to the old pathname loop leaves every one of them green. +// The defect was in what the entry point called, so this drives the entry point. +func TestPrepareSandboxRuntimeRefusesARedirectedDescendant(t *testing.T) { + isolateSandboxRuntimeRoots(t) + cacheRoot := t.TempDir() + target := t.TempDir() + previous := sandboxUserCacheDir + t.Cleanup(func() { sandboxUserCacheDir = previous }) + sandboxUserCacheDir = func() (string, error) { return cacheRoot, nil } + + workspace := t.TempDir() + root, ok := deterministicSandboxRuntimeRoot(canonicalSandboxWorkspaceRoot(workspace), canonicalSandboxWorkspaceRoot(cacheRoot)) + if !ok { + t.Fatalf("SETUP INVALID: no deterministic runtime root for %s", workspace) + } + cache := filepath.Join(root, "cache") + if err := os.MkdirAll(root, 0o700); err != nil { + t.Fatal(err) + } + if out, err := exec.Command("cmd", "/c", "mklink", "/J", cache, target).CombinedOutput(); err != nil { + t.Skipf("mklink /J unavailable here: %v\n%s", err, out) + } + probe := filepath.Join(cache, "probe.txt") + if err := os.WriteFile(probe, []byte("x"), 0o600); err != nil { + t.Skipf("the junction does not accept writes here: %v", err) + } + if _, err := os.Stat(filepath.Join(target, "probe.txt")); err != nil { + t.Skipf("SETUP: the junction does not redirect on this filesystem: %v", err) + } + if err := os.Remove(probe); err != nil { + t.Fatal(err) + } + + _, release, err := prepareSandboxRuntime(workspace) + if release != nil { + release() + } + if err == nil { + t.Fatal("prepareSandboxRuntime descended through a junction the previous sandboxed command planted") + } + + entries, readErr := os.ReadDir(target) + if readErr != nil { + t.Fatal(readErr) + } + if len(entries) != 0 { + names := make([]string, 0, len(entries)) + for _, entry := range entries { + names = append(names, entry.Name()) + } + t.Fatalf("prepareSandboxRuntime created %v beneath the redirected target", names) + } +} + +// And an ordinary workspace still prepares through the entry point. +func TestPrepareSandboxRuntimeStillPreparesAnOrdinaryTree(t *testing.T) { + isolateSandboxRuntimeRoots(t) + cacheRoot := t.TempDir() + previous := sandboxUserCacheDir + t.Cleanup(func() { sandboxUserCacheDir = previous }) + sandboxUserCacheDir = func() (string, error) { return cacheRoot, nil } + + runtimeState, release, err := prepareSandboxRuntime(t.TempDir()) + if err != nil { + t.Fatalf("an ordinary workspace was refused: %v", err) + } + defer release() + for _, dir := range []string{runtimeState.Root, runtimeState.Cache, runtimeState.Data, runtimeState.Temp} { + info, statErr := os.Stat(dir) + if statErr != nil || !info.IsDir() { + t.Fatalf("%s was not created: err=%v", dir, statErr) + } + } +} + +// runtimeTreeTestBase makes the operator-owned base deterministic. +// +// runtimeCandidateBase answers by containment against the user cache directory, +// so a bare t.TempDir() only works where temp happens to sit under it. On a CI +// runner it does not, or its 8.3 short name compares unequal to the long form, +// and the tree is reported as having no owned base. Stubbing the cache root is +// what makes these tests about the descent rather than about where temp lives. +func runtimeTreeTestBase(t *testing.T) string { + t.Helper() + base := canonicalSandboxWorkspaceRoot(t.TempDir()) + previous := sandboxUserCacheDir + t.Cleanup(func() { sandboxUserCacheDir = previous }) + sandboxUserCacheDir = func() (string, error) { return base, nil } + return base +} diff --git a/internal/sandbox/runtime_tree_windows.go b/internal/sandbox/runtime_tree_windows.go new file mode 100644 index 000000000..bf10ae9b1 --- /dev/null +++ b/internal/sandbox/runtime_tree_windows.go @@ -0,0 +1,44 @@ +//go:build windows + +package sandbox + +import ( + "fmt" + + "golang.org/x/sys/windows" +) + +// ensureRuntimeTreeDir creates the owned tail beneath a physical base, one +// component at a time, from retained handles. +// +// Every open and create goes through the NtCreateFile helpers with +// OBJ_DONT_REPARSE and FILE_OPEN_REPARSE_POINT, so a junction planted at any +// component is seen as the link it is and refused rather than descended. No +// component name is resolved twice, so there is no interval for a swap to land +// in. +// +// No chmod. On Windows os.Chmod only toggles READONLY, and against a junction it +// lands on the link rather than the target, so it bought nothing here. The +// directory's protection is the ACL that elevated setup applied, which this +// descent deliberately does not touch. +func ensureRuntimeTreeDir(physicalBase string, tail []string) error { + parent, err := openWindowsACLDirectoryNoFollow(physicalBase) + if err != nil { + return fmt.Errorf("open the sandbox runtime base %s: %w", physicalBase, err) + } + defer func() { _ = windows.CloseHandle(parent) }() + + for _, name := range tail { + child, _, createErr := createWindowsACLChildDirectory(parent, name) + if createErr != nil { + existing, openErr := openWindowsACLChildDirectory(parent, name) + if openErr != nil { + return fmt.Errorf("prepare the sandbox runtime component %s beneath %s: %w", name, physicalBase, createErr) + } + child = existing + } + _ = windows.CloseHandle(parent) + parent = child + } + return nil +} diff --git a/internal/sandbox/types.go b/internal/sandbox/types.go index 897254546..7de3fd6cb 100644 --- a/internal/sandbox/types.go +++ b/internal/sandbox/types.go @@ -104,12 +104,21 @@ const ( BlockNetwork BlockCode = "network" BlockDestructiveCommand BlockCode = "destructive_command" BlockPersistentDeny BlockCode = "persistent_deny" + BlockNestedGitInit BlockCode = "nested_git_init" // BlockDenied is the catch-all for a denied decision that carries no more // specific block code. BlockDenied BlockCode = "denied" ) const ReasonNetworkBlocked = "network access requires approval" + +// ReasonNestedGitInit explains a refusal the operator can act on: the remedy is +// to run git init outside the sandbox, or to open the repository itself as the +// workspace rather than a subdirectory of it. +const ReasonNestedGitInit = "this workspace sits inside an existing git repository, so the sandbox has no git protection here to extend to a new one: " + + "a repository created now would carry a writable config and hooks. " + + "Every git init in the command is refused, because the sandbox does not resolve which directory each one targets. " + + "Run git init outside the sandbox, or open the repository root as the workspace" const ReasonEscalatedSandboxRequired = "unsandboxed shell command requires approval" const ( diff --git a/internal/sandbox/windows_acl.go b/internal/sandbox/windows_acl.go index c7d47956e..bb84860c2 100644 --- a/internal/sandbox/windows_acl.go +++ b/internal/sandbox/windows_acl.go @@ -7,12 +7,72 @@ import ( "strings" ) +// isWindowsVolumeRoot reports whether a cleaned path is the top of a volume, +// with nothing above it: `C:\`, a bare separator, or a UNC share root. +// +// Detected structurally rather than by pattern matching drive letters, because +// filepath.Dir of a root is that same root and of anything else is strictly +// shorter. That holds for drive-qualified paths, for the separator alone, and +// for UNC roots, on either build host. +func isWindowsVolumeRoot(path string) bool { + cleaned := filepath.Clean(strings.TrimSpace(path)) + if cleaned == "" || cleaned == "." { + return false + } + return filepath.Dir(cleaned) == cleaned +} + +// validateWindowsACLComponent rejects anything that is not a single path +// component. +// +// This is load-bearing in two places, which is why it lives in the portable file +// rather than beside either of them. +// +// At apply time NtCreateFile happily resolves a RELATIVE name containing +// separators, and it resolves it the ordinary way, so an intermediate junction +// inside that name is followed and the object lands outside the pinned parent. +// A name with a separator reopens exactly the hole the parent handle exists to +// close. +// +// At plan time the same shape escapes the write root: a name is joined onto the +// root to place a deny ACE, so ".." or a separator puts that ACE on a directory +// outside the workspace entirely. +// +// The separators are checked explicitly rather than via filepath.Base, because +// these are Windows paths whatever the build host is, and on Linux +// filepath.Base leaves a backslash-joined name untouched and would wave it +// through. A colon is rejected too: it names an alternate data stream or a +// drive, neither of which is a child. +func validateWindowsACLComponent(name string) error { + switch { + case name == "": + return errors.New("windows ACL path component is empty") + case name == "." || name == "..": + return fmt.Errorf("windows ACL path component %q is a relative reference, not a child", name) + case strings.ContainsAny(name, `\/`): + return fmt.Errorf("windows ACL path component %q contains a separator, so it would resolve through intermediate directories instead of staying a child", name) + case strings.Contains(name, ":"): + return fmt.Errorf("windows ACL path component %q contains a colon, which names a stream or a drive rather than a child", name) + } + return nil +} + type WindowsACLAction string const ( WindowsACLAllowWrite WindowsACLAction = "allow-write" WindowsACLDenyRead WindowsACLAction = "deny-read" WindowsACLDenyWrite WindowsACLAction = "deny-write" + // WindowsACLDenyDelete denies removing or renaming the object it names, + // WITHOUT denying writes to it or inside it, and without inheriting. + // + // It exists for .git. The write-denied carveouts live on .git/config and + // .git/hooks as objects, so replacing the .git directory discards them: the + // recreated config and hooks inherit the workspace allow with no deny, which + // restores credential.helper and core.hooksPath. .git cannot simply join + // sandboxFullyProtectedMetadataNames, because DenyWrite's mask includes + // FILE_GENERIC_WRITE and git must write index, objects and refs. + WindowsACLDenyDelete WindowsACLAction = "deny-delete" // WindowsACLRevokeCapability removes any existing ACE (allow or deny) for // Capability at Path, without itself granting or denying anything (applied // via SetEntriesInAclW's SET_ACCESS mode with a zero mask, not @@ -41,6 +101,11 @@ type WindowsACLEntry struct { // sandbox. Empty for an operator-named path, which has no owned tail. Anchor string `json:"anchor,omitempty"` Materialize bool `json:"materialize,omitempty"` + // MaterializeFile makes Materialize create an empty FILE instead of a + // directory. Only meaningful with Materialize. .git/config is the case that + // forces the distinction: created as a directory it does not merely carry + // the wrong ACL, it makes `git init` fail outright. + MaterializeFile bool `json:"materializeFile,omitempty"` } type WindowsACLPlan struct { @@ -62,6 +127,34 @@ func BuildWindowsACLPlan(config WindowsSandboxCommandConfig) (WindowsACLPlan, er Path: capability.Root, Capability: capability.SID, }) + // The grant above carries DELETE, so the same object guard the principal + // plan has must come with it: not DenyWrite (git writes index, objects and + // refs), not materialized (git creates .git, and an empty one breaks + // git init), and not inherited, so everything underneath stays writable. + entries = append(entries, WindowsACLEntry{ + Action: WindowsACLDenyDelete, + Path: windowsRenameProtectedObject(capability.Root), + Capability: capability.SID, + }) + // MATERIALIZED, LIKE THE PRINCIPAL PLAN'S, AND FOR TWO REASONS. + // + // First, a guard attached to an object that does not exist is not applied. + // On a workspace that had no .git when setup ran, the first pass skipped all + // three of these as missing and the deferred pass skipped them again, + // because nothing else in this plan created them. Setup still recorded + // success. A later git init then created .git, config and hooks beneath the + // already-granted workspace, where they inherit the allow, DELETE included, + // with no object-specific deny of their own: the sandboxed command could + // rename .git aside, recreate it, and get credential.helper and + // core.hooksPath back. + // + // Second, materializing these is what lets the applier's deferred pass land + // the deny-delete on .git above, since creating .gitconfig and .githooks + // creates .git as their parent. + // + // The principal planner has done this from the start. Sharing the shape + // rather than only the pathname is the point: this tier is the DEFAULT + // backend, so the weaker of the two rules was the one almost everyone got. for _, path := range capability.ProtectedWriteDenyPaths { // ANCHORED ONLY WHERE THE ANCHOR IS TRUE. Most of these are derived // from the root and sit under it, and holding those inside it is the @@ -75,10 +168,58 @@ func BuildWindowsACLPlan(config WindowsSandboxCommandConfig) (WindowsACLPlan, er anchor = capability.Root } entries = append(entries, WindowsACLEntry{ - Action: WindowsACLDenyWrite, + Action: WindowsACLDenyWrite, + Path: path, + Capability: capability.SID, + Anchor: anchor, + Materialize: true, + MaterializeFile: gitMetadataCarveoutIsFile(path), + }) + } + } + // Read roots, granted to the read-capability SID. + // + // A profile carrying DenyRead runs the command on a strict token rather than a + // WRITE_RESTRICTED one, and the strict token applies the restricted-SID check + // to reads. Read roots reached that check granted only to the principal's own + // account SID, which the write jail keeps out of the restricting set on + // purpose, so every read failed it — including opening the executable. The + // principal plan still grants the account SID; this is the other half of the + // same grant, so the allow-list and the restriction now come from one place. + // AND IT REFUSES TO EXPRESS "READ EVERYWHERE" AS AN ACE ON THE VOLUME ROOT. + // + // Production profiles seed ReadRoots with the bare filesystem root, so this + // loop used to add an inheritable allow-read for the synthetic read SID at + // "C:\". SetSecurityInfo propagates inheritable ACEs onto existing children, + // so that is not one sandbox-owned object being changed: it walks and rewrites + // DACL inheritance across unrelated system, application and user trees on the + // drive, and a locked or exclusively opened descendant makes the result depend + // on ambient filesystem state. + // + // It is not even complete after paying that price. A bare root resolves on one + // volume, while the executables, DLLs and tool installations a command needs + // can sit on another without appearing as their own read roots, so the strict + // token still fails before its executable starts. Measured: with the read SID + // granted nowhere the strict token cannot open C:WindowsSystem32cmd.exe. + // + // So the profile that needs this is refused, at both setup tiers, rather than + // half-served by a persistent volume-wide ACL edit. That leaves denyRead + // unavailable on Windows until there is a bounded authorization model covering + // the real platform, runtime and executable dependencies on every relevant + // volume, which is what #869 tracks. + readSID, err := windowsReadAllowCapabilitySID(config) + if err != nil { + return WindowsACLPlan{}, err + } + if readSID != "" { + for _, path := range config.PermissionProfile.FileSystem.ReadRoots { + if path = strings.TrimSpace(path); path == "" { + continue + } + entries = append(entries, WindowsACLEntry{ + Action: WindowsACLAllowRead, Path: path, - Capability: capability.SID, - Anchor: anchor, + Capability: readSID, }) } } @@ -114,6 +255,102 @@ func BuildWindowsACLPlan(config WindowsSandboxCommandConfig) (WindowsACLPlan, er return WindowsACLPlan{Entries: dedupeWindowsACLEntries(entries)}, nil } +// WindowsACLPlanReadGrantRefusal reports why a plan must not be applied, or "". +// +// ONE ANSWER FOR EVERY TIER THAT APPLIES A PLAN. This started inside the +// unelevated tier, because that is where it was first observed: an ordinary user +// cannot write the volume root's DACL, so every command failed there with a +// generic diagnosis. Elevated setup CAN write it, and that is worse rather than +// better. SetSecurityInfo propagates inheritable ACEs to existing children, so +// applying it rewrites DACL inheritance across unrelated trees, and it still does +// not cover a second volume, so the strict token can fail to open its own +// executable after all that. +// +// KEYED ON THE READ GRANT, NOT ON THE VOLUME ROOT. Checking for a volume root +// tested a symptom of the production profile rather than the thing that is +// unsafe. permissionProfileReadRoots happens to seed the bare filesystem root, so +// that check covered production by coincidence; a profile with a narrowed read +// list and a denyRead still put an inheritable ACE on C:Windows and passed. The +// grant itself is what cannot be applied safely to a directory Zero does not own, +// and the grant exists only for a denyRead profile, so that is what is refused. +func WindowsACLPlanReadGrantRefusal(plan WindowsACLPlan) string { + root := windowsPlanReadGrantTarget(plan) + if root == "" { + return "" + } + return "this sandbox profile configures denyRead, which selects a fully restricted token, and that token applies its restricted-SID check to reads as well as writes. " + + "Serving it needs the read capability granted on directories Zero does not own, starting at " + root + ", which cannot be done safely: the grant is inheritable, so applying it rewrites permissions across unrelated system, application and user directories, " + + "and it still would not cover executables or libraries on another volume. " + + "denyRead is therefore not available on Windows yet (see #869). Remove denyRead from the sandbox configuration to run on a write-restricted token, which keeps the workspace write jail intact" +} + +// windowsPlanReadGrantTarget returns the broadest path the plan would grant the +// read capability on that Zero does not already own, preferring a volume root so +// the diagnostic names the worst of them. +// +// A read grant on a write root is Zero's own directory and is not the problem; +// the objection is to writing an inheritable ACE onto somebody else's tree. So +// the write roots are excluded, and an empty answer means every read grant lands +// on a directory this sandbox already governs, which needs no refusal. +func windowsPlanReadGrantTarget(plan WindowsACLPlan) string { + owned := make(map[string]struct{}) + for _, entry := range plan.Entries { + if entry.Action == WindowsACLAllowWrite { + owned[strings.ToLower(normalizeProfilePath(entry.Path))] = struct{}{} + } + } + first := "" + for _, entry := range plan.Entries { + if entry.Action != WindowsACLAllowRead { + continue + } + normalized := normalizeProfilePath(entry.Path) + if _, ours := owned[strings.ToLower(normalized)]; ours { + continue + } + if isWindowsVolumeRoot(normalized) { + return entry.Path + } + if first == "" { + first = entry.Path + } + } + return first +} + +// windowsPlanVolumeRootGrant returns the first volume root the plan would have +// to change permissions on, or empty when it needs none. +// +// Only elevated setup can write a volume-root DACL, so this is the one entry +// that decides whether a plan is applicable by the unelevated tier at all. The +// check is on the PLAN rather than on the profile because the plan is what gets +// applied: a future entry that lands at a volume root for some other reason is +// caught by the same test. +func windowsPlanVolumeRootGrant(plan WindowsACLPlan) string { + for _, entry := range plan.Entries { + if isWindowsVolumeRoot(normalizeProfilePath(entry.Path)) { + return entry.Path + } + } + return "" +} + +// windowsRenameProtectedObject names the object whose DELETE must be denied on +// a write root, whatever trustee holds the grant. +// +// ONE DERIVATION, BECAUSE TWO PLANNERS CONSUME IT. The allow-write mask both +// backends share includes DELETE, and it inherits from the write root onto .git. +// The carveouts that actually protect git live on .git/config and .git/hooks as +// OBJECTS, so renaming .git aside and recreating it discards them: the fresh +// config and hooks inherit the workspace allow with no deny of their own, which +// hands back credential.helper and core.hooksPath. The principal planner denied +// DELETE here and the capability planner did not, so the default restricted-token +// backend was missing the guard entirely. Deriving it in one place is what stops +// a change to the shared mask from updating one consumer and not the other. +func windowsRenameProtectedObject(root string) string { + return filepath.Join(root, sandboxRenameProtectedMetadataName) +} + type windowsWriteRootCapability struct { Root string SID string @@ -175,17 +412,32 @@ func windowsWriteCapabilitySIDs(capabilities []windowsWriteRootCapability) []str } func windowsReadDenyCapabilitySIDs(config WindowsSandboxCommandConfig, writeSIDs []string) ([]string, error) { - if len(writeSIDs) > 0 { - return writeSIDs, nil - } if len(config.PermissionProfile.FileSystem.DenyRead) == 0 { return nil, nil } - caps, err := LoadOrCreateWindowsCapabilitySIDs(config.SandboxHome) + // The read-capability SID must be denied here as well as granted above. It + // holds read on the read roots, and those start at the filesystem root, so a + // carveout sitting inside one of them would stay readable through that grant + // and the deny list would quietly stop meaning anything. + // Every SID the token can carry has to be denied, not just the newest one. + // With no write roots the token's only capability is ReadOnly, so dropping it + // here would leave the deny naming a SID the command never holds. + denySIDs := append([]string{}, writeSIDs...) + if len(denySIDs) == 0 { + caps, err := LoadOrCreateWindowsCapabilitySIDs(config.SandboxHome) + if err != nil { + return nil, err + } + denySIDs = append(denySIDs, caps.ReadOnly) + } + readSID, err := windowsReadAllowCapabilitySID(config) if err != nil { return nil, err } - return []string{caps.ReadOnly}, nil + if readSID != "" { + denySIDs = append(denySIDs, readSID) + } + return denySIDs, nil } func planWindowsDenyReadPaths(paths []string) []string { @@ -232,3 +484,19 @@ func dedupeWindowsACLEntries(entries []WindowsACLEntry) []WindowsACLEntry { } return out } + +// windowsReadAllowCapabilitySID returns the read-capability SID when this profile +// will run on a strict token, and "" otherwise. +// +// Gated on DenyRead because that is exactly what selects the strict token in the +// runner: with no DenyRead the token stays WRITE_RESTRICTED, reads skip the +// restricted-SID check entirely, and granting read on roots as broad as the +// filesystem root would add ACEs that buy nothing. Derived purely from the +// profile, so the setup half and the command half reach the same answer without +// consulting anything ambient. +func windowsReadAllowCapabilitySID(config WindowsSandboxCommandConfig) (string, error) { + if len(config.PermissionProfile.FileSystem.DenyRead) == 0 { + return "", nil + } + return WindowsReadAllowSID(config.SandboxHome) +} diff --git a/internal/sandbox/windows_acl_apply_windows.go b/internal/sandbox/windows_acl_apply_windows.go index e8ecb008c..b4bac3a4f 100644 --- a/internal/sandbox/windows_acl_apply_windows.go +++ b/internal/sandbox/windows_acl_apply_windows.go @@ -6,6 +6,7 @@ import ( "errors" "fmt" "os" + "path/filepath" "sort" "strings" "unsafe" @@ -26,28 +27,128 @@ type windowsACLPathGroup struct { Path string // Anchor is the write root Path was derived from, empty when the operator // named the path. See verifyWindowsACLHandleUnderAnchor. - Anchor string - Entries []WindowsACLEntry - Materialize bool + Anchor string + Entries []WindowsACLEntry + Materialize bool + MaterializeFile bool } +// windowsACLChainStep is one directory component beneath the anchor, and +// whether THIS run created it. The flag comes from the kernel rather than from +// "it was missing when we looked", because something else can win the gap +// between the probe and the create, and removing a directory the sandbox did not +// make is how a rollback deletes a user's data. +type windowsACLChainStep struct { + Name string + Made bool + // ID is the identity of the directory this run created at Name, read from + // the handle that created it. Rollback removes only an object with this + // identity: a created name says the parent is unchanged, not that the + // child has not been renamed aside and replaced since. + ID windowsFileIdentity +} + +// windowsACLMaterialization is exactly what materialization created, recorded in +// the shape rollback needs to undo it without resolving a single pathname below +// the anchor. +// +// The anchor is the deepest directory that already existed, and it is the ONLY +// pathname rollback re-resolves. Everything under it is a list of single +// components walked one handle at a time, because a name containing a separator +// is resolved the ordinary way by the kernel and would follow an intermediate +// junction straight out of the approved tree. +type windowsACLMaterialization struct { + AnchorPath string + AnchorID windowsFileIdentity + // Chain is every component between the anchor and the target, shallow to + // deep. All of them are needed to descend at rollback time; only the ones + // with Made set are removed. + Chain []windowsACLChainStep + // File is the leaf file component created inside the deepest Chain entry, + // for the .git/config carveout. Empty when the target is a directory. + File string + FileMade bool + // FileID identifies the leaf file this run created, for the same reason. + FileID windowsFileIdentity +} + +func (materialization windowsACLMaterialization) createdAnything() bool { + if materialization.FileMade { + return true + } + for _, step := range materialization.Chain { + if step.Made { + return true + } + } + return false +} + +// windowsACLRestoreHook observes restore order. Nil in production; the ordering +// it exposes is load-bearing and was previously pinned by nothing at all, +// despite a comment claiming otherwise. +var windowsACLRestoreHook func(path string) + type windowsACLSnapshot struct { - Path string - Descriptor *windows.SECURITY_DESCRIPTOR - Materialized bool + Path string + Descriptor *windows.SECURITY_DESCRIPTOR + Created windowsACLMaterialization + // TargetID is the object the DACL was read from, so the restore can prove it + // is writing back to that same object rather than to whatever the pathname + // resolves to at rollback time. + // + // A no-follow re-open catches a REPARSE POINT swapped in since apply, and + // that is what the restore relied on. It cannot catch the other + // substitution: rename the target aside and put an ordinary directory of the + // same name in its place. Nothing about that decoy is a reparse point, so the + // re-open succeeds and the old DACL is written onto the attacker's object + // while the real one keeps the aborted setup's ACEs. + // + // Volume serial plus file index is the same identity the materialization + // unwind already anchors on. See reopenWindowsACLDirectoryAsIdentity. + TargetID windowsFileIdentity } func applyWindowsACLPlan(plan WindowsACLPlan) (func() error, error) { groups := groupWindowsACLPlanByPath(plan) snapshots := make([]windowsACLSnapshot, 0, len(groups)) + abort := func(err error) (func() error, error) { + if rollbackErr := rollbackWindowsACLSnapshots(snapshots); rollbackErr != nil { + return nil, fmt.Errorf("%w; rollback failed: %v", err, rollbackErr) + } + return nil, err + } + + var deferred []windowsACLPathGroup for _, group := range groups { snapshot, applied, err := applyWindowsACLPathGroup(group) if err != nil { - rollbackErr := rollbackWindowsACLSnapshots(snapshots) - if rollbackErr != nil { - return nil, fmt.Errorf("%w; rollback failed: %v", err, rollbackErr) - } - return nil, err + return abort(err) + } + if applied { + snapshots = append(snapshots, snapshot) + continue + } + deferred = append(deferred, group) + } + + // A group that applied to nothing was skipped because its target did not + // exist and the group does not materialize one. A LATER group can still + // create it, so the skip has to be retried rather than treated as final. + // + // The .git rename guard is exactly this shape and was silently absent + // because of it. .git must NOT be materialized, since an empty one breaks + // `git init`, so its deny-delete group carries no Materialize. But .git does + // get created, as the parent chain of the .git\config carveout, and that + // group sorts AFTER it. So on every workspace that did not already have a + // .git, the guard was skipped, the directory appeared moments later, and + // nothing went back for it: the principal could then rename .git aside and + // shed the config and hooks carveouts, which is the escape the guard exists + // to stop. Groups that are genuinely absent simply skip again here. + for _, group := range deferred { + snapshot, applied, err := applyWindowsACLPathGroup(group) + if err != nil { + return abort(err) } if applied { snapshots = append(snapshots, snapshot) @@ -72,6 +173,7 @@ func groupWindowsACLPlanByPath(plan WindowsACLPlan) []windowsACLPathGroup { } group.Entries = append(group.Entries, entry) group.Materialize = group.Materialize || entry.Materialize + group.MaterializeFile = group.MaterializeFile || entry.MaterializeFile // One path is derived from at most one root, so the first anchor seen is // the anchor. Taking it rather than overwriting keeps a later // operator-named duplicate of the same path from clearing it. @@ -101,8 +203,42 @@ func applyWindowsACLPathGroup(group windowsACLPathGroup) (windowsACLSnapshot, bo // elevated setup a lower-privileged local user could swap the target for a // symlink/junction between operations and redirect the ACL change onto a // system object it never validated (issue #728, a TOCTOU privilege boundary). - materialized := false - handle, isDir, err := openWindowsACLTarget(path) + // Reject a malformed group BEFORE touching the filesystem. This validation + // used to run after materialization, so a bad SID or an unknown action + // created a directory chain and only then failed, leaving the error path to + // unwind work that never needed doing. Nothing here depends on isDir: that + // argument only selects the inheritance flag, while the errors come from the + // action lookup and the SID parse. + if _, _, err := prepareWindowsACLPathGroupEntries(group.Entries, false, nil); err != nil { + return windowsACLSnapshot{}, false, err + } + + var created windowsACLMaterialization + var handle windows.Handle + // Every exit from here goes through one closure, because there are now two + // things to undo rather than one: the open handle, and whatever + // materialization created. Both are captured by reference and both start + // zero, so calling this before either is set is safe and does nothing. + // + // The unwind is handle-relative. It must never fall back to a pathname + // delete: the failure being cleaned up here can BE the path swap, and + // os.RemoveAll on a swapped ancestor is precisely the recursive elevated + // delete outside the workspace that this cleanup is supposed to prevent. + fail := func(err error) (windowsACLSnapshot, bool, error) { + if handle != 0 { + _ = windows.CloseHandle(handle) + } + // The apply itself is failing, so nothing has been left with a DACL to + // restore; only the removal outcome matters here. + if _, unwindErr := rollbackWindowsACLMaterialization(created); unwindErr != nil { + return windowsACLSnapshot{}, false, fmt.Errorf("%w; cleanup failed: %v", err, unwindErr) + } + return windowsACLSnapshot{}, false, err + } + + var isDir bool + var err error + handle, isDir, err = openWindowsACLTargetFinalComponent(path) if err != nil { if !errors.Is(err, os.ErrNotExist) { return windowsACLSnapshot{}, false, err @@ -118,26 +254,20 @@ func applyWindowsACLPathGroup(group windowsACLPathGroup) (windowsACLSnapshot, bo // elevated setup create the directory outside the write root and only the // containment check below would notice, after the fact. if err := verifyWindowsACLPathUnderAnchor(group.Anchor, path); err != nil { - return windowsACLSnapshot{}, false, err + return fail(err) } - if err := os.MkdirAll(path, 0o700); err != nil { - return windowsACLSnapshot{}, false, fmt.Errorf("materialize windows ACL target %s: %w", path, err) - } - materialized = true - handle, isDir, err = openWindowsACLTarget(path) + // created is assigned even on failure: materialization reports what it + // managed to make before it stopped, and fail() unwinds exactly that. + created, err = materializeWindowsACLTarget(path, group.MaterializeFile) if err != nil { - _ = os.RemoveAll(path) - return windowsACLSnapshot{}, false, fmt.Errorf("open materialized windows ACL target %s: %w", path, err) + return fail(fmt.Errorf("materialize windows ACL target %s: %w", path, err)) } - } - // From here the handle is open; every early return must close it first (and - // remove a freshly materialized target) so a failure leaks neither. - fail := func(err error) (windowsACLSnapshot, bool, error) { - _ = windows.CloseHandle(handle) - if materialized { - _ = os.RemoveAll(path) + handle, isDir, err = openWindowsACLTargetFinalComponent(path) + if err != nil { + // This is the branch that fires when the post-create verify catches a + // swap, so it is the single most important cleanup in the file. + return fail(fmt.Errorf("open materialized windows ACL target %s: %w", path, err)) } - return windowsACLSnapshot{}, false, err } // THE OBJECT HAS TO BE WHERE THE PLAN SAID IT WOULD BE. The open above // refuses a reparse point at the final component and resolves every one @@ -176,8 +306,14 @@ func applyWindowsACLPathGroup(group windowsACLPathGroup) (windowsACLSnapshot, bo // The handle has served its purpose (read+write bound to one object) and is // closed now — rollback re-opens no-follow rather than holding a handle for // the whole sandbox lifetime, since one caller discards the rollback closure. + // Captured BEFORE the handle closes, because that handle is the only thing + // that names the object rather than the path. + targetID, err := windowsIdentityOfHandle(handle) + if err != nil { + return fail(fmt.Errorf("read windows ACL target identity for %s: %w", path, err)) + } _ = windows.CloseHandle(handle) - return windowsACLSnapshot{Path: path, Descriptor: descriptor, Materialized: materialized}, true, nil + return windowsACLSnapshot{Path: path, Descriptor: descriptor, Created: created, TargetID: targetID}, true, nil } // openWindowsACLTarget opens path for reading and rewriting its DACL without @@ -189,6 +325,30 @@ func applyWindowsACLPathGroup(group windowsACLPathGroup) (windowsACLSnapshot, bo // is exactly the redirection this guard exists to prevent. A missing target is // surfaced as os.ErrNotExist so the caller's materialize path still fires. func openWindowsACLTarget(path string) (windows.Handle, bool, error) { + handle, isDir, err := openWindowsACLTargetFinalComponent(path) + if err != nil { + return 0, false, err + } + // Ancestors are resolved by CreateFile even with FILE_FLAG_OPEN_REPARSE_POINT, + // so the final-component check is not enough on its own here: this opener + // serves the rollback re-open and the runtime ACE, which know no anchor and + // must not be redirected by anything on the way down. + if err := verifyWindowsACLTargetNotRedirected(handle, path); err != nil { + _ = windows.CloseHandle(handle) + return 0, false, err + } + return handle, isDir, nil +} + +// openWindowsACLTargetFinalComponent is the open without the ancestor check. +// The plan apply uses it: a reparse point ABOVE a target is the operator's +// business (a workspace on a junction or a dev drive is a configuration that +// works), and a reparse point on a tail this package derived from a write root +// is caught by verifyWindowsACLHandleUnderAnchor, which knows the anchor and +// tolerates exactly the aliases above it. That is main's rule from #1040 and +// the two containment tests pin it; the blanket refusal above is kept for the +// callers that have no anchor to reason with. +func openWindowsACLTargetFinalComponent(path string) (windows.Handle, bool, error) { utf16Path, err := windows.UTF16PtrFromString(path) if err != nil { return 0, false, fmt.Errorf("encode windows ACL target %s: %w", path, err) @@ -270,13 +430,27 @@ func prepareWindowsACLPathGroupEntries(entries []WindowsACLEntry, isDir bool, ol return nil, nil, err } inheritance := uint32(0) - if isDir && !entry.NoInherit { + // DenyDelete governs the object it names and nothing beneath it. Inherited + // onto .git's children it would deny DELETE on every file inside, so git + // could not remove a lock file, a ref, or anything else it rewrites, and + // the guard would read as a broken repository rather than as a blocked + // rename. Decided here, not in the planner, so no plan can make it inherit. + if isDir && !entry.NoInherit && entry.Action != WindowsACLDenyDelete { inheritance = windows.SUB_CONTAINERS_AND_OBJECTS_INHERIT } + entryInheritance := inheritance + // DenyDelete governs the object it names and nothing beneath it. Inherited + // onto .git's children it would deny DELETE on every file inside, so git + // could not remove a lock file, a ref, or anything else it rewrites, and + // the guard would read as a broken repository rather than as a blocked + // rename. + if entry.Action == WindowsACLDenyDelete { + entryInheritance = windows.NO_INHERITANCE + } out = append(out, windows.EXPLICIT_ACCESS{ AccessPermissions: permissions, AccessMode: accessMode, - Inheritance: inheritance, + Inheritance: entryInheritance, Trustee: windows.TRUSTEE{ TrusteeForm: windows.TRUSTEE_IS_SID, TrusteeType: windows.TRUSTEE_IS_GROUP, @@ -585,13 +759,61 @@ func windowsAceSID(ace *windows.ACCESS_ALLOWED_ACE) (sid *windows.SID, ok bool) func windowsACLAccess(action WindowsACLAction) (windows.ACCESS_MODE, windows.ACCESS_MASK, error) { switch action { case WindowsACLAllowWrite: - return windows.GRANT_ACCESS, windows.FILE_GENERIC_READ | windows.FILE_GENERIC_WRITE | windows.FILE_GENERIC_EXECUTE, nil + // DELETE is part of the grant, not an extra. FILE_GENERIC_WRITE covers + // creating and modifying but not removing or renaming, and a rename needs + // delete access on the source. Under the old same-user token that gap was + // invisible, because the caller already held inherited rights on its own + // tree; a sandbox principal is a separate account with no such + // inheritance, so without DELETE it can write a file it can never delete. + // Ordinary editing and most git operations rewrite files by replacing + // them, so the omission fails normal work rather than an edge case. + // + // FILE_DELETE_CHILD is deliberately NOT granted, for the same reason + // WRITE_DAC and WRITE_OWNER are not. On a parent it authorises deleting a + // child whatever the child's own DACL says, so granting it on a write root + // hands back the write-denied carve-outs underneath it: a principal could + // delete .git/config and recreate it, and the replacement inherits this + // grant with no deny of its own — restoring exactly the credential.helper + // and core.hooksPath control the carve-out exists to prevent. + // + // It was granted here originally to keep the mask symmetric with + // WindowsACLDenyWrite, which does treat FILE_DELETE_CHILD as part of + // write. Symmetry is the wrong goal: denying a capability is not a reason + // to grant it. DELETE alone covers removing and renaming files the + // principal owns inside its roots, which is what the grant is for. + return windows.GRANT_ACCESS, windows.FILE_GENERIC_READ | windows.FILE_GENERIC_WRITE | windows.FILE_GENERIC_EXECUTE | windows.DELETE, nil + case WindowsACLAllowRead: + // Read and traverse without write. A sandbox principal is a separate + // account with no inherent access to the caller's tree, so a read-only + // root has to be granted rather than assumed. Deliberately omits + // FILE_GENERIC_WRITE, DELETE and WRITE_DAC. + return windows.GRANT_ACCESS, windows.FILE_GENERIC_READ | windows.FILE_GENERIC_EXECUTE, nil + case windowsACLRevoke: + // REVOKE_ACCESS drops every ACE naming the trustee regardless of the mask, + // so the mask is ignored here. Used to retire a principal without having + // to remember which access each path was granted. + return windows.REVOKE_ACCESS, 0, nil case WindowsACLDenyRead: return windows.DENY_ACCESS, windows.FILE_GENERIC_READ | windows.FILE_GENERIC_EXECUTE, nil case WindowsACLDenyWrite: return windows.DENY_ACCESS, (windows.FILE_GENERIC_WRITE | windows.DELETE | windowsFileDeleteChild | windows.WRITE_DAC | windows.WRITE_OWNER) &^ windows.SYNCHRONIZE, nil + case WindowsACLDenyDelete: + // Deny removing or RENAMING the object itself, nothing more. Renaming a + // directory needs DELETE on that directory, so denying DELETE is what + // stops .git being moved aside and recreated without its carveouts. + // + // WRITE_DAC and WRITE_OWNER come along because a guard the principal can + // rewrite, or take ownership of and then rewrite, is not a guard. + // + // FILE_GENERIC_WRITE is deliberately absent: git writes index, objects and + // refs constantly, and denying it would break every commit rather than the + // rename. FILE_DELETE_CHILD is absent for the same reason one level down, + // since git deletes its own lock files and refs. Neither is needed here: + // this ACE does not inherit (see prepareWindowsACLPathGroupEntries), so it + // governs the .git directory object alone. + return windows.DENY_ACCESS, windows.DELETE | windows.WRITE_DAC | windows.WRITE_OWNER, nil case WindowsACLRevokeCapability: - // Handled specially in windowsExplicitAccessEntries (preserve DenyRead). + // Handled specially in prepareWindowsACLPathGroupEntries (preserve DenyRead). return windows.SET_ACCESS, 0, nil default: return 0, 0, fmt.Errorf("unsupported windows ACL action %q", action) @@ -600,29 +822,76 @@ func windowsACLAccess(action WindowsACLAction) (windows.ACCESS_MODE, windows.ACC func rollbackWindowsACLSnapshots(snapshots []windowsACLSnapshot) error { var errs []error + // Reverse order is load-bearing twice over. Groups are sorted ascending by + // path key and an ancestor key is always a proper prefix of its descendants, + // so walking backwards unwinds descendant before ancestor: a materialized + // directory is therefore empty by the time its own removal is attempted. + // Restoring ACLs in the same order is independently right, because + // SetSecurityInfo propagates inheritable ACEs down, so the ancestor must go + // last. TestRollbackUnwindsDescendantsBeforeAncestors pins it. for index := len(snapshots) - 1; index >= 0; index-- { snapshot := snapshots[index] - if snapshot.Materialized { - if err := os.RemoveAll(snapshot.Path); err != nil { - errs = append(errs, fmt.Errorf("remove materialized windows ACL target %s: %w", snapshot.Path, err)) + if snapshot.Created.createdAnything() { + removed, err := rollbackWindowsACLMaterialization(snapshot.Created) + if err != nil { + errs = append(errs, err) + } + // Only skip the restore when the object carrying the DACL is actually + // gone. A racer can win the leaf while this run made its parent, and + // then the aggregate says "ours" about a file that is not: the parent + // cannot be removed because it is non-empty, and skipping here left a + // pre-existing file wearing an aborted setup's ACL. Falling through + // reaches the same no-follow, TargetID-guarded restore as any other + // snapshot, which refuses on a mismatch rather than forcing. + if removed { + continue } - continue } dacl, _, err := snapshot.Descriptor.DACL() if err != nil { errs = append(errs, fmt.Errorf("read rollback windows DACL for %s: %w", snapshot.Path, err)) continue } - // Re-open no-follow rather than restoring by pathname: the restore must - // land on the real object, not a reparse point swapped in since apply. The - // residual window is small because the target is ACL-restricted by now, but - // a handle keeps the restore honest. On a materialized-target rollback we - // remove it above, so only the restore-existing path opens here. - handle, _, err := openWindowsACLTarget(snapshot.Path) + // Re-open no-follow AND prove it is the same object. The no-follow open + // alone only rules out a reparse point swapped in since apply; a target + // renamed aside and replaced by an ordinary directory of the same name + // passes it, and the old DACL would land on the decoy while the real + // object kept the aborted setup's ACEs. See windowsACLSnapshot.TargetID. + handle, _, err := openWindowsACLTargetFinalComponent(snapshot.Path) if err != nil { errs = append(errs, fmt.Errorf("re-open windows ACL target %s for rollback: %w", snapshot.Path, err)) continue } + // An identity to compare against makes the pathname route irrelevant: the + // object either is the one the DACL was read from or it is not. Without + // one, the route is all there is, so keep the blanket redirection check. + if snapshot.TargetID.empty() { + if err := verifyWindowsACLTargetNotRedirected(handle, snapshot.Path); err != nil { + _ = windows.CloseHandle(handle) + errs = append(errs, fmt.Errorf("re-open windows ACL target %s for rollback: %w", snapshot.Path, err)) + continue + } + } + if !snapshot.TargetID.empty() { + got, identityErr := windowsIdentityOfHandle(handle) + if identityErr != nil { + _ = windows.CloseHandle(handle) + errs = append(errs, fmt.Errorf("read rollback identity for %s: %w", snapshot.Path, identityErr)) + continue + } + if got != snapshot.TargetID { + // REFUSED, not forced through. Leaving the real object with the + // aborted setup's ACEs is the safe direction: the caller is failing + // anyway, and writing a stale DACL onto an object setup never + // touched is the one outcome rollback must not produce. + _ = windows.CloseHandle(handle) + errs = append(errs, fmt.Errorf("refusing to restore windows ACL for %s: it is no longer the object setup applied to", snapshot.Path)) + continue + } + } + if windowsACLRestoreHook != nil { + windowsACLRestoreHook(snapshot.Path) + } if err := windows.SetSecurityInfo(handle, windows.SE_FILE_OBJECT, windows.DACL_SECURITY_INFORMATION, nil, nil, dacl, nil); err != nil { errs = append(errs, fmt.Errorf("rollback windows ACL for %s: %w", snapshot.Path, err)) } @@ -630,3 +899,280 @@ func rollbackWindowsACLSnapshots(snapshots []windowsACLSnapshot) error { } return errors.Join(errs...) } + +// materializeWindowsACLTarget creates a missing ACL target with the shape the +// owning tool expects. A directory target is created whole; a file target gets +// its parent chain created and then an empty file, because creating it as a +// directory would break the tool that owns it rather than just mis-ACL it. +// The returned record is meaningful even when the error is non-nil: a chain that +// got three levels deep and then failed still has three levels to unwind. +func materializeWindowsACLTarget(path string, asFile bool) (windowsACLMaterialization, error) { + directory := path + leaf := "" + if asFile { + directory = filepath.Dir(path) + leaf = filepath.Base(path) + } + created, parent, err := makeWindowsACLDirChainNoFollow(directory) + if err != nil { + return created, err + } + defer func() { _ = windows.CloseHandle(parent) }() + if !asFile { + return created, nil + } + // windowsACLRacedLeafHook fires at the ONE instant a racer can win the leaf + // while this run already owns the chain above it: the directories exist and + // the file does not yet. That mixed ownership is what made the old rollback + // skip its restore, and it is unreachable from the anchor-stage hook, which + // fires before the chain exists at all. Always nil in production. + if windowsACLRacedLeafHook != nil { + windowsACLRacedLeafHook(path) + } + // A racing creator winning is still fine: the target exists, which is all + // materialization needed. createWindowsACLChildFile reports that as + // created=false, so rollback will not delete a file the sandbox did not make. + created.File = leaf + madeFile, fileID, err := createWindowsACLChildFile(parent, leaf) + created.FileMade = madeFile + if madeFile { + created.FileID = fileID + } + return created, err +} + +// rollbackWindowsACLMaterialization removes exactly what materialization +// created, deepest first, without resolving any pathname below the anchor. +// +// This is the other half of the pathname problem. The old cleanup called +// os.RemoveAll on the target pathname, which re-resolves every ancestor at the +// moment it runs, so an ancestor swapped to a junction after the object was +// created sent a recursive elevated delete into an unrelated tree. It also only +// ever removed the final component, leaving every intermediate directory the +// chain had created behind. +// +// The anchor is the one pathname that has to be resolved again, and it is +// checked by file identity rather than by name: replacing a directory with +// another REAL directory of the same name needs no reparse point at all and +// would otherwise pass every no-follow check there is. +// +// Residue is preferable to over-deletion throughout. When something cannot be +// removed safely this reports it and leaves it, and never falls back to a +// pathname delete. +// rollbackWindowsACLMaterialization removes what this run created and reports +// whether the ACL-BEARING TARGET is among what it removed. +// +// AGGREGATE OWNERSHIP IS NOT THE TARGET'S DISPOSITION. The caller used to treat +// createdAnything() as "this run owns the object it applied a DACL to", which is +// an OR across the whole materialization record. When this run creates the +// parent chain and a racer wins the leaf, that is true while the ACL-bearing +// file belongs to somebody else. The caller then skipped the identity-validated +// restore and left a pre-existing file carrying an aborted setup's ACL. +// +// The target is the file leaf when FileMade is set, and the deepest chain entry +// otherwise. Anything this cannot prove it removed reports false, which routes +// the caller to a restore that is already no-follow and TargetID-guarded, so the +// conservative direction is the default. +func rollbackWindowsACLMaterialization(materialization windowsACLMaterialization) (targetRemoved bool, err error) { + if !materialization.createdAnything() { + return false, nil + } + anchor, anchorErr := reopenWindowsACLDirectoryAsIdentity(materialization.AnchorPath, materialization.AnchorID) + if anchorErr != nil { + if errors.Is(anchorErr, os.ErrNotExist) { + // The anchor is gone, so everything created beneath it is gone too, + // the target included. Nothing to undo, and no way to undo it. + return true, nil + } + return false, fmt.Errorf("unwind windows ACL materialization under %s: %w", materialization.AnchorPath, anchorErr) + } + + // One handle per level: handles[i] is the parent of Chain[i], which is what + // deleting Chain[i] relative to a pinned parent requires. + handles := []windows.Handle{anchor} + defer func() { + for _, handle := range handles { + if handle != 0 { + _ = windows.CloseHandle(handle) + } + } + }() + + // Descend only as far as is actually required. Removing a directory needs its + // PARENT's handle, not its own, so the deepest component is opened only when + // a file leaf lives inside it. This is not just economy: the deepest + // component is usually the ACL target itself, so it may already carry the + // deny-read ACE this rollback is undoing, and opening it would be refused by + // the very ACL being unwound. + needed := len(materialization.Chain) + if !materialization.FileMade && needed > 0 { + needed-- + } + depth := 0 + for ; depth < needed; depth++ { + child, openErr := openWindowsACLChildDirectory(handles[depth], materialization.Chain[depth].Name) + if openErr != nil { + // Already removed by something else. Stop descending; whatever is + // below it is gone with it. + if isWindowsNotExist(openErr) { + break + } + return false, fmt.Errorf("unwind windows ACL materialization under %s: %w", materialization.AnchorPath, openErr) + } + // A component this run created has to still be the object it created + // before anything beneath it is touched: an ordinary directory renamed + // into its place beneath the unchanged anchor is not ours to descend + // into, let alone delete. + if step := materialization.Chain[depth]; step.Made { + got, identityErr := windowsIdentityOfHandle(child) + if identityErr != nil { + _ = windows.CloseHandle(child) + return false, fmt.Errorf("unwind windows ACL materialization under %s: identify %s: %w", materialization.AnchorPath, step.Name, identityErr) + } + if got != step.ID { + _ = windows.CloseHandle(child) + return false, fmt.Errorf("refusing to unwind windows ACL materialization under %s: %s is no longer the directory setup created", materialization.AnchorPath, step.Name) + } + } + handles = append(handles, child) + } + + var errs []error + // An ancestor vanished mid-descent, so the target went with it. + targetRemoved = depth < needed + // The file leaf lives inside the deepest chain directory, so it goes first + // and only if the descent actually reached that far. + if materialization.FileMade && depth == needed { + if err := deleteWindowsACLChildFileAs(handles[len(handles)-1], materialization.File, materialization.FileID); err != nil { + errs = append(errs, fmt.Errorf("remove materialized windows ACL file %s: %w", materialization.File, err)) + } else { + targetRemoved = true + } + } + // Chain[i] is removed through handles[i], so the deepest one that can be + // removed is bounded by how far the descent actually got. + deepest := len(handles) - 1 + if last := len(materialization.Chain) - 1; deepest > last { + deepest = last + } + for index := deepest; index >= 0; index-- { + // Close the child's own handle, if the descent opened one, before asking + // its parent to remove it. A directory with a live handle open still + // counts as present, so the parent's delete would come back as non-empty. + if index+1 < len(handles) && handles[index+1] != 0 { + _ = windows.CloseHandle(handles[index+1]) + handles[index+1] = 0 + } + if !materialization.Chain[index].Made { + continue + } + if err := deleteWindowsACLChildDirectoryAs(handles[index], materialization.Chain[index].Name, materialization.Chain[index].ID); err != nil { + errs = append(errs, fmt.Errorf("remove materialized windows ACL directory %s: %w", materialization.Chain[index].Name, err)) + continue + } + // The deepest chain entry IS the ACL target for a directory + // materialization, and it is the leaf's parent when a racer made the leaf, + // so removing it means the target is gone either way. + if !materialization.FileMade && index == len(materialization.Chain)-1 { + targetRemoved = true + } + } + return targetRemoved, errors.Join(errs...) +} + +// windowsACLMaterializeSwapHook is a test seam and nothing else. It fires inside +// makeWindowsACLDirChainNoFollow at the exact instant the race used to be +// exploitable: the anchor is verified and pinned, and nothing has been created +// yet. A race reproducible only by luck is not a regression test, so the instant +// is made addressable rather than hoped for. Always nil in production. +var windowsACLMaterializeSwapHook func(anchor string) + +// windowsACLRacedLeafHook fires after the directory chain is created and before +// the file leaf is. See materializeWindowsACLTarget. Always nil in production. +var windowsACLRacedLeafHook func(path string) + +// makeWindowsACLDirChainNoFollow is an os.MkdirAll that never resolves a +// pathname below its anchor. +// +// It walks UP to the deepest ancestor that already exists and opens it +// no-follow. Because GetFinalPathNameByHandle answers for the whole resolved +// path, that single check clears every ancestor above it too. Then it walks back +// DOWN, creating one component at a time relative to the HANDLE of the level +// above, so the tree it descends is pinned to objects rather than named by +// strings. +// +// That is the difference that matters. This used to verify a component by +// pathname, close the handle, and then hand the same string to os.Mkdir: two +// independent kernel resolutions with a gap between them. A workspace owner who +// swapped the verified ancestor for a junction inside that gap got the component +// created outside the approved tree, as Administrator, and verifying again +// afterwards cannot un-create it. Junctions need no privilege, so this was +// reachable by exactly the unprivileged user the sandbox exists to contain. +// +// It deliberately does NOT re-verify each created component by pathname. A child +// created relative to a pinned parent is in the right place by construction, so +// comparing pathnames afterwards would add nothing and would reject correct +// creates whenever the tree was legitimately renamed mid-setup. +// +// Returns what it created, plus an open handle to the deepest directory which +// the caller must close. +func makeWindowsACLDirChainNoFollow(dir string) (windowsACLMaterialization, windows.Handle, error) { + cleaned := filepath.Clean(strings.TrimSpace(dir)) + if cleaned == "" || cleaned == "." { + return windowsACLMaterialization{}, 0, fmt.Errorf("materialize windows ACL target: empty directory path %q", dir) + } + + // Walk up to the deepest ancestor that exists, collecting the component + // NAMES that are missing. Names, not paths: everything below the anchor is + // addressed relative to a handle from here on. + var missing []string + current := cleaned + var anchor windows.Handle + var anchorID windowsFileIdentity + for { + handle, identity, err := openWindowsACLDirectoryNoFollowWithIdentity(current) + if err == nil { + anchor, anchorID = handle, identity + break + } + if !errors.Is(err, os.ErrNotExist) { + return windowsACLMaterialization{}, 0, err + } + parent := filepath.Dir(current) + if parent == current { + return windowsACLMaterialization{}, 0, fmt.Errorf("materialize windows ACL target %s: no existing ancestor to anchor on", dir) + } + missing = append(missing, filepath.Base(current)) + current = parent + } + + created := windowsACLMaterialization{AnchorPath: current, AnchorID: anchorID} + + if hook := windowsACLMaterializeSwapHook; hook != nil { + hook(current) + } + + // Walk back down, one component per handle. The anchor handle is released as + // soon as its child is open, so at most two levels are held at once. + parent := anchor + for index := len(missing) - 1; index >= 0; index-- { + name := missing[index] + child, madeNow, err := createWindowsACLChildDirectory(parent, name) + if err != nil { + _ = windows.CloseHandle(parent) + return created, 0, err + } + step := windowsACLChainStep{Name: name, Made: madeNow} + if madeNow { + if step.ID, err = windowsIdentityOfHandle(child); err != nil { + _ = windows.CloseHandle(parent) + _ = windows.CloseHandle(child) + return created, 0, fmt.Errorf("identify materialized windows ACL directory %s: %w", name, err) + } + } + created.Chain = append(created.Chain, step) + _ = windows.CloseHandle(parent) + parent = child + } + return created, parent, nil +} diff --git a/internal/sandbox/windows_acl_apply_windows_test.go b/internal/sandbox/windows_acl_apply_windows_test.go index 1d57b2350..5f9d8a03c 100644 --- a/internal/sandbox/windows_acl_apply_windows_test.go +++ b/internal/sandbox/windows_acl_apply_windows_test.go @@ -37,8 +37,15 @@ func TestApplyWindowsACLPathGroupHandleBasedRoundTrip(t *testing.T) { if !applied { t.Fatal("applied = false, want true for an existing directory target") } - if snapshot.Path != dir || snapshot.Materialized { - t.Fatalf("snapshot = %#v, want Path=%q Materialized=false", snapshot, dir) + if snapshot.Path != dir { + t.Fatalf("snapshot.Path = %q, want %q", snapshot.Path, dir) + } + // The target already existed, so nothing was created and rollback must have + // nothing to remove. Asserting the chain rather than a bool matters: a + // rewiring that recorded the walked components instead of only the created + // ones would make rollback delete a directory the sandbox never made. + if snapshot.Created.createdAnything() { + t.Fatalf("snapshot recorded %#v as created for a target that already existed", snapshot.Created) } if snapshot.Descriptor == nil { t.Fatal("snapshot has no captured descriptor to roll back to") @@ -110,8 +117,21 @@ func TestApplyWindowsACLPathGroupMaterializes(t *testing.T) { if err != nil { t.Fatalf("applyWindowsACLPathGroup: %v", err) } - if !applied || !snapshot.Materialized { - t.Fatalf("applied=%v materialized=%v, want both true", applied, snapshot.Materialized) + if !applied { + t.Fatal("applied = false, want true for a materialized target") + } + // Exactly one component was missing, so exactly one must be recorded as + // created, and it must be the leaf's own name rather than a path, carrying + // the identity of the directory it made so rollback can prove it is still + // deleting that object. + if len(snapshot.Created.Chain) != 1 || snapshot.Created.Chain[0].Name != "created" || !snapshot.Created.Chain[0].Made { + t.Fatalf("created chain = %#v, want one step {created true}", snapshot.Created.Chain) + } + if snapshot.Created.Chain[0].ID.empty() { + t.Fatalf("created chain step carries no identity: %#v", snapshot.Created.Chain[0]) + } + if snapshot.Created.AnchorPath != filepath.Dir(target) { + t.Fatalf("anchor = %q, want the existing parent %q", snapshot.Created.AnchorPath, filepath.Dir(target)) } if _, err := os.Stat(target); err != nil { t.Fatalf("materialized target not created: %v", err) diff --git a/internal/sandbox/windows_acl_git_guard_windows_test.go b/internal/sandbox/windows_acl_git_guard_windows_test.go new file mode 100644 index 000000000..b72c1cfce --- /dev/null +++ b/internal/sandbox/windows_acl_git_guard_windows_test.go @@ -0,0 +1,122 @@ +//go:build windows + +package sandbox + +import ( + "os" + "path/filepath" + "testing" + "unsafe" + + "golang.org/x/sys/windows" +) + +// gitGuardDenyDeleteApplied reports whether the applied DACL on path carries a +// DENY ace granting DELETE to sid. Read back from the object rather than from +// the plan, because the plan is what was already wrong. +func gitGuardDenyDeleteApplied(t *testing.T, path string, sid string) bool { + t.Helper() + handle, _, err := openWindowsACLTarget(path) + if err != nil { + t.Fatalf("open %s: %v", path, err) + } + defer func() { _ = windows.CloseHandle(handle) }() + descriptor, err := windows.GetSecurityInfo(handle, windows.SE_FILE_OBJECT, windows.DACL_SECURITY_INFORMATION) + if err != nil { + t.Fatalf("read the DACL of %s: %v", path, err) + } + dacl, _, err := descriptor.DACL() + if err != nil || dacl == nil { + t.Fatalf("read the DACL of %s: %v", path, err) + } + wanted, err := windows.StringToSid(sid) + if err != nil { + t.Fatalf("parse the capability SID %q: %v", sid, err) + } + // GetAce hands back a generic header that is reinterpreted here. Sound only + // for the fixed-layout ACE types: an object ACE carries Flags and two GUIDs + // ahead of the trustee, so SidStart would land mid-structure. Allowed and + // denied ACEs share that fixed layout, and nothing under test builds an + // object ACE, so a type this does not recognise is skipped rather than + // decoded into a nonsense SID. + for index := uint32(0); index < uint32(dacl.AceCount); index++ { + var ace *windows.ACCESS_ALLOWED_ACE + if err := windows.GetAce(dacl, index, &ace); err != nil { + continue + } + if ace.Header.AceType != windows.ACCESS_DENIED_ACE_TYPE { + continue + } + aceSID := (*windows.SID)(unsafe.Pointer(&ace.SidStart)) + if !aceSID.Equals(wanted) { + continue + } + if ace.Mask&windows.DELETE != 0 { + return true + } + } + return false +} + +// THE GUARD BELONGS TO THE GRANT, NOT TO ONE PLANNER. +// +// The allow-write mask both backends share includes DELETE, and it inherits from +// the write root onto .git. What protects git is attached to .git/config and +// .git/hooks as OBJECTS, so renaming .git aside and recreating it discards those +// carveouts: the fresh config and hooks inherit the workspace allow with no deny +// of their own, which hands back credential.helper and core.hooksPath. +// +// The principal planner denied DELETE on .git; the capability planner did not, +// and the capability backend is the default. So on the reachable path the +// caller's own token and the capability restricting SID could both authorise the +// rename. +// +// Read back from the applied object, not asserted on the plan, because the plan +// is exactly what was wrong. DACL edits on a user-owned directory need no +// Administrator rights, so this runs unelevated. +func TestCapabilityPlanDeniesDeleteOnGit(t *testing.T) { + workspace := t.TempDir() + gitDir := filepath.Join(workspace, ".git") + if err := os.MkdirAll(gitDir, 0o700); err != nil { + t.Fatal(err) + } + + config := WindowsSandboxCommandConfig{ + SandboxHome: t.TempDir(), + CommandCWD: workspace, + WorkspaceRoots: []string{workspace}, + PermissionProfile: PermissionProfile{ + FileSystem: FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{Root: workspace}}, + }, + }, + } + plan, err := BuildWindowsACLPlan(config) + if err != nil { + t.Fatalf("BuildWindowsACLPlan: %v", err) + } + sid, err := windowsCapabilitySIDForWriteRoot(config, workspace) + if err != nil { + t.Fatalf("resolve the workspace capability SID: %v", err) + } + // SETUP: the plan must actually grant this root, or the guard below would be + // vacuously satisfied by a plan that grants nothing. + granted := false + for _, entry := range plan.Entries { + if entry.Action == WindowsACLAllowWrite && entry.Path == workspace { + granted = true + } + } + if !granted { + t.Fatalf("SETUP INVALID: the capability plan does not grant write on %s", workspace) + } + + if _, err := applyWindowsACLPlan(plan); err != nil { + t.Skipf("cannot apply an ACL plan here: %v", err) + } + + if !gitGuardDenyDeleteApplied(t, gitDir, sid) { + t.Fatalf(".git carries no deny-DELETE ace for the capability SID %s, so the sandboxed token can rename it and recreate it without the config and hooks carveouts", sid) + } +} diff --git a/internal/sandbox/windows_acl_junction_ancestor_windows_test.go b/internal/sandbox/windows_acl_junction_ancestor_windows_test.go new file mode 100644 index 000000000..24817ee94 --- /dev/null +++ b/internal/sandbox/windows_acl_junction_ancestor_windows_test.go @@ -0,0 +1,94 @@ +//go:build windows + +package sandbox + +import ( + "os" + "os/exec" + "path/filepath" + "strings" + "testing" +) + +// makeJunction points link at target, skipping the test when the environment +// refuses to create one. A junction needs no privilege, which is exactly why +// this attack is reachable by an ordinary workspace owner. +func makeJunction(t *testing.T, link, target string) { + t.Helper() + if out, err := exec.Command("cmd", "/c", "mklink", "/J", link, target).CombinedOutput(); err != nil { + t.Skipf("cannot create a junction here: %v: %s", err, out) + } + // Assert it actually redirects. A junction that silently did nothing would + // green this test while proving nothing. + probe := filepath.Join(link, "redirect-probe") + if err := os.WriteFile(probe, []byte("x"), 0o600); err != nil { + t.Fatalf("write through junction: %v", err) + } + if _, err := os.Stat(filepath.Join(target, "redirect-probe")); err != nil { + t.Fatalf("junction does not redirect, so this test would prove nothing: %v", err) + } + if err := os.Remove(probe); err != nil { + t.Fatalf("clean probe: %v", err) + } +} + +// Elevated setup must not create anything through a reparse-point ancestor. +// +// FILE_FLAG_OPEN_REPARSE_POINT only stops the FINAL component being followed. +// materializeWindowsACLTarget used os.MkdirAll/os.OpenFile on the pathname, both +// of which resolve ancestors, so a workspace owner who turned .git into a +// junction before setup ran got objects created at a location of their choosing +// — as Administrator — and the no-follow check only rejected it afterwards, far +// too late to un-create them. +func TestMaterializeRefusesAncestorJunctionBeforeCreating(t *testing.T) { + for name, asFile := range map[string]bool{"file target": true, "directory target": false} { + t.Run(name, func(t *testing.T) { + external := t.TempDir() + workspace := t.TempDir() + gitDir := filepath.Join(workspace, ".git") + makeJunction(t, gitDir, external) + + target := filepath.Join(gitDir, "hooks", "config") + _, err := materializeWindowsACLTarget(target, asFile) + if err == nil { + t.Fatalf("materialized %s through a junction ancestor instead of refusing", target) + } + if !strings.Contains(err.Error(), "reparse") { + t.Fatalf("refused for the wrong reason: %v", err) + } + // Nothing may survive on the other side of the junction. The old code + // left every intermediate directory MkdirAll had created. + leaked, lerr := os.ReadDir(external) + if lerr != nil { + t.Fatalf("read external dir: %v", lerr) + } + if len(leaked) != 0 { + names := make([]string, 0, len(leaked)) + for _, entry := range leaked { + names = append(names, entry.Name()) + } + t.Fatalf("created %v outside the workspace through the junction", names) + } + }) + } +} + +// The ordinary path still works: no reparse point anywhere, target gets made. +func TestMaterializeStillCreatesOrdinaryTargets(t *testing.T) { + root := t.TempDir() + for name, asFile := range map[string]bool{"file target": true, "directory target": false} { + t.Run(name, func(t *testing.T) { + target := filepath.Join(root, name, "nested", "deeper", "target") + if _, err := materializeWindowsACLTarget(target, asFile); err != nil { + t.Fatalf("materializeWindowsACLTarget: %v", err) + } + info, err := os.Stat(target) + if err != nil { + t.Fatalf("target was not created: %v", err) + } + if info.IsDir() == asFile { + t.Fatalf("target isDir=%v, wanted file=%v", info.IsDir(), asFile) + } + }) + } +} diff --git a/internal/sandbox/windows_acl_materialize_swap_windows_test.go b/internal/sandbox/windows_acl_materialize_swap_windows_test.go new file mode 100644 index 000000000..5a4654c93 --- /dev/null +++ b/internal/sandbox/windows_acl_materialize_swap_windows_test.go @@ -0,0 +1,453 @@ +//go:build windows + +package sandbox + +import ( + "errors" + "os" + "path/filepath" + "strings" + "testing" + + "golang.org/x/sys/windows" +) + +// swapAncestorAside moves approved out of the way and leaves a junction wearing +// its name, pointing at elsewhere. This is the whole attack in three lines, and +// it needs no privilege: junctions are creatable by any user, which is exactly +// why an unprivileged workspace owner can aim elevated setup wherever they like. +// +// Returns the path the real directory now lives at. +func swapAncestorAside(t *testing.T, approved, elsewhere string) string { + t.Helper() + moved := approved + "-moved" + if err := os.Rename(approved, moved); err != nil { + t.Skipf("cannot rename a directory with an open handle on this filesystem: %v", err) + } + makeJunction(t, approved, elsewhere) + return moved +} + +// requireNothingEscaped fails when anything at all was created on the far side +// of the junction. +func requireNothingEscaped(t *testing.T, elsewhere string) { + t.Helper() + leaked, err := os.ReadDir(elsewhere) + if err != nil { + t.Fatalf("read the decoy directory: %v", err) + } + if len(leaked) == 0 { + return + } + names := make([]string, 0, len(leaked)) + for _, entry := range leaked { + names = append(names, entry.Name()) + } + t.Fatalf("ESCAPED: created %v outside the approved tree, as Administrator", names) +} + +// THE MATERIALIZATION RACE, ON THE PRODUCTION CALL PATH. +// +// The existing junction test plants its junction before the walk even starts, so +// the very first check sees it and refuses. That proves the easy half. The half +// the reviewer filed is the swap that happens AFTER a component has been +// verified and BEFORE it is used, and no test reached it: a race nobody can +// trigger on demand is not a regression test, so makeWindowsACLDirChainNoFollow +// carries a seam that fires at exactly that instant. +// +// The control arm matters as much as the fixed one. It performs the identical +// swap and then does what this code used to do, creating by pathname, and +// asserts that the object DOES escape. Without it, the fixed arm passing proves +// only that some code ran, not that the hole it closes was ever open. +func TestMaterializeSurvivesAnAncestorSwappedMidWalk(t *testing.T) { + t.Run("control: creating by pathname escapes", func(t *testing.T) { + root := t.TempDir() + approved := filepath.Join(root, "ws") + elsewhere := filepath.Join(root, "OUTSIDE") + for _, dir := range []string{approved, elsewhere} { + if err := os.Mkdir(dir, 0o700); err != nil { + t.Fatalf("seed %s: %v", dir, err) + } + } + // Verified, exactly as the walk verifies its anchor. + if err := verifyWindowsACLPathComponentNotRedirected(approved); err != nil { + t.Fatalf("anchor did not verify before the swap: %v", err) + } + moved := swapAncestorAside(t, approved, elsewhere) + + // The old create: a pathname, re-resolved by the kernel right now. + if err := os.MkdirAll(filepath.Join(approved, "a", "b"), 0o700); err != nil { + t.Fatalf("pathname create: %v", err) + } + if _, err := os.Stat(filepath.Join(elsewhere, "a", "b")); err != nil { + t.Fatalf("the control arm did not reproduce the escape, so the fixed arm below proves nothing: %v", err) + } + if _, err := os.Stat(filepath.Join(moved, "a", "b")); err == nil { + t.Error("the control arm created inside the verified directory, which is not the behaviour being contrasted") + } + }) + + t.Run("fixed: creating through the pinned handle stays put", func(t *testing.T) { + root := t.TempDir() + approved := filepath.Join(root, "ws") + elsewhere := filepath.Join(root, "OUTSIDE") + for _, dir := range []string{approved, elsewhere} { + if err := os.Mkdir(dir, 0o700); err != nil { + t.Fatalf("seed %s: %v", dir, err) + } + } + + var moved string + swapped := false + windowsACLMaterializeSwapHook = func(anchor string) { + if swapped { + return + } + swapped = true + if anchor != approved { + t.Errorf("anchored on %q, want the deepest existing ancestor %q", anchor, approved) + } + moved = swapAncestorAside(t, approved, elsewhere) + } + t.Cleanup(func() { windowsACLMaterializeSwapHook = nil }) + + created, err := materializeWindowsACLTarget(filepath.Join(approved, "a", "b"), false) + if err != nil { + t.Fatalf("materializeWindowsACLTarget: %v", err) + } + if !swapped { + t.Fatal("the seam never fired, so the swap never happened and this test proved nothing") + } + requireNothingEscaped(t, elsewhere) + if _, err := os.Stat(filepath.Join(moved, "a", "b")); err != nil { + t.Errorf("the target did not land in the directory that was verified: %v", err) + } + // And the record must describe what to unwind, in components rather than + // paths, shallowest first. + if len(created.Chain) != 2 || created.Chain[0].Name != "a" || created.Chain[1].Name != "b" { + t.Fatalf("created chain = %#v, want [a b] shallow to deep", created.Chain) + } + for _, step := range created.Chain { + if !step.Made { + t.Errorf("component %q was not recorded as created, so rollback would leave it behind", step.Name) + } + } + }) +} + +// The FILE target has the same race, and it is the one that matters most: +// .git/config is materialized as a file on every stock setup, and it is the file +// whose credential.helper is worth stealing. +func TestMaterializeFileSurvivesAnAncestorSwappedMidWalk(t *testing.T) { + root := t.TempDir() + approved := filepath.Join(root, "ws") + elsewhere := filepath.Join(root, "OUTSIDE") + for _, dir := range []string{approved, elsewhere} { + if err := os.Mkdir(dir, 0o700); err != nil { + t.Fatalf("seed %s: %v", dir, err) + } + } + + var moved string + swapped := false + windowsACLMaterializeSwapHook = func(string) { + if swapped { + return + } + swapped = true + moved = swapAncestorAside(t, approved, elsewhere) + } + t.Cleanup(func() { windowsACLMaterializeSwapHook = nil }) + + created, err := materializeWindowsACLTarget(filepath.Join(approved, ".git", "config"), true) + if err != nil { + t.Fatalf("materializeWindowsACLTarget: %v", err) + } + if !swapped { + t.Fatal("the seam never fired, so this test proved nothing") + } + requireNothingEscaped(t, elsewhere) + + landed := filepath.Join(moved, ".git", "config") + info, err := os.Stat(landed) + if err != nil { + t.Fatalf("the file did not land in the directory that was verified: %v", err) + } + if info.IsDir() { + t.Error("materialized .git/config as a directory, which breaks git init") + } + if !created.FileMade || created.File != "config" { + t.Errorf("file record = %q made=%v, want config/true", created.File, created.FileMade) + } +} + +// THE ROLLBACK RACE. The ancestor is swapped AFTER the target was created, which +// is the window the teardown path lives in: minutes or hours, not microseconds. +// +// The bystander is the point. If the unwind resolves by pathname it walks into +// the decoy and deletes what it finds there, recursively and elevated. Its +// survival is the only thing that proves the unwind did not. +func TestRollbackDoesNotFollowAnAncestorSwappedAfterCreation(t *testing.T) { + root := t.TempDir() + approved := filepath.Join(root, "ws") + elsewhere := filepath.Join(root, "OUTSIDE") + for _, dir := range []string{approved, elsewhere} { + if err := os.Mkdir(dir, 0o700); err != nil { + t.Fatalf("seed %s: %v", dir, err) + } + } + + created, err := materializeWindowsACLTarget(filepath.Join(approved, "a", "b"), false) + if err != nil { + t.Fatalf("materializeWindowsACLTarget: %v", err) + } + + // A bystander tree under the decoy, shaped exactly like what we created, so a + // pathname unwind would find something to destroy at every level. + bystander := filepath.Join(elsewhere, "a", "b") + if err := os.MkdirAll(bystander, 0o700); err != nil { + t.Fatalf("seed bystander: %v", err) + } + witness := filepath.Join(bystander, "irreplaceable.txt") + if err := os.WriteFile(witness, []byte("not yours to delete"), 0o600); err != nil { + t.Fatalf("seed witness: %v", err) + } + + moved := swapAncestorAside(t, approved, elsewhere) + + // The anchor pathname now names the decoy, and the decoy is a junction, so + // the unwind must refuse rather than proceed. Either way it must not delete. + _, err = rollbackWindowsACLMaterialization(created) + + if _, statErr := os.Stat(witness); statErr != nil { + t.Fatalf("DESTRUCTIVE: rollback followed the junction and deleted a tree outside the approved directory: %v", statErr) + } + if _, statErr := os.Stat(bystander); statErr != nil { + t.Fatalf("DESTRUCTIVE: rollback removed the bystander directory outside the approved directory: %v", statErr) + } + if err == nil { + t.Error("rollback reported success while unwinding through a swapped ancestor; it must say it could not") + } + // Residue inside the real tree is the accepted price: leaving it is strictly + // better than a recursive delete through a path someone else controls. + if _, statErr := os.Stat(filepath.Join(moved, "a", "b")); statErr != nil { + t.Logf("note: the real tree was also unwound (%v); leaving it would be acceptable too", statErr) + } +} + +// A real directory wearing the anchor's name is a swap with NO reparse point +// anywhere, so every no-follow check in this package passes it. Only the file +// identity notices. +func TestRollbackRefusesAnAnchorReplacedByARealDirectory(t *testing.T) { + root := t.TempDir() + approved := filepath.Join(root, "ws") + if err := os.Mkdir(approved, 0o700); err != nil { + t.Fatalf("seed: %v", err) + } + created, err := materializeWindowsACLTarget(filepath.Join(approved, "a"), false) + if err != nil { + t.Fatalf("materializeWindowsACLTarget: %v", err) + } + + if err := os.Rename(approved, approved+"-moved"); err != nil { + t.Skipf("cannot rename here: %v", err) + } + // An ordinary directory. Nothing is a link; nothing is a reparse point. + if err := os.Mkdir(approved, 0o700); err != nil { + t.Fatalf("plant the replacement: %v", err) + } + decoy := filepath.Join(approved, "a") + if err := os.Mkdir(decoy, 0o700); err != nil { + t.Fatalf("plant the decoy child: %v", err) + } + + _, err = rollbackWindowsACLMaterialization(created) + if err == nil { + t.Error("rollback accepted a different directory wearing the anchor's name") + } else if !strings.Contains(err.Error(), "no longer the directory") { + t.Errorf("refused for the wrong reason: %v", err) + } + if _, statErr := os.Stat(decoy); statErr != nil { + t.Errorf("rollback deleted a directory it never created: %v", statErr) + } +} + +// Rollback removes ONLY what this run created. A pre-existing ancestor is walked +// through and left alone. +func TestRollbackLeavesDirectoriesItDidNotCreate(t *testing.T) { + root := t.TempDir() + existing := filepath.Join(root, "ws", "already-here") + if err := os.MkdirAll(existing, 0o700); err != nil { + t.Fatalf("seed: %v", err) + } + target := filepath.Join(existing, "made", "deeper") + + created, err := materializeWindowsACLTarget(target, false) + if err != nil { + t.Fatalf("materializeWindowsACLTarget: %v", err) + } + if created.AnchorPath != existing { + t.Fatalf("anchor = %q, want the deepest pre-existing directory %q", created.AnchorPath, existing) + } + if _, err := rollbackWindowsACLMaterialization(created); err != nil { + t.Fatalf("rollbackWindowsACLMaterialization: %v", err) + } + if _, err := os.Stat(filepath.Join(existing, "made")); !errors.Is(err, os.ErrNotExist) { + t.Errorf("created directory survived rollback: stat err = %v", err) + } + if _, err := os.Stat(existing); err != nil { + t.Errorf("rollback removed a directory that already existed: %v", err) + } +} + +// The whole apply path, through the closure callers actually hold, rather than +// through rollbackWindowsACLSnapshots directly. Every other rollback test in +// this package calls the unwind by hand, which cannot catch applyWindowsACLPlan +// failing to carry the created record into the snapshots it hands over. +func TestApplyWindowsACLPlanClosureRemovesWhatItMaterialized(t *testing.T) { + root := t.TempDir() + directoryTarget := filepath.Join(root, "ws", "hooks") + fileTarget := filepath.Join(root, "ws", "config") + + plan := WindowsACLPlan{Entries: []WindowsACLEntry{ + {Action: WindowsACLDenyWrite, Path: directoryTarget, Capability: "S-1-1-0", Materialize: true}, + {Action: WindowsACLDenyWrite, Path: fileTarget, Capability: "S-1-1-0", Materialize: true, MaterializeFile: true}, + }} + + rollback, err := applyWindowsACLPlan(plan) + if err != nil { + t.Fatalf("applyWindowsACLPlan: %v", err) + } + for _, path := range []string{directoryTarget, fileTarget} { + if _, err := os.Stat(path); err != nil { + t.Fatalf("%s was not materialized: %v", path, err) + } + } + if err := rollback(); err != nil { + t.Fatalf("rollback closure: %v", err) + } + for _, path := range []string{directoryTarget, fileTarget} { + if _, err := os.Stat(path); !errors.Is(err, os.ErrNotExist) { + t.Errorf("%s survived the rollback closure: stat err = %v", path, err) + } + } + // The shared prefix both targets needed must go too, and it is created by + // whichever group runs first rather than being owned by both. + if _, err := os.Stat(filepath.Join(root, "ws")); !errors.Is(err, os.ErrNotExist) { + t.Errorf("the shared parent survived: stat err = %v", err) + } +} + +// A rollback that cannot remove something must SAY so. This is the regression +// guard for the trap a naive handle-relative port walks straight into: +// FILE_DELETE_ON_CLOSE reports success on a non-empty directory and leaves it +// there, which turns a loud failure into a silent lie. The directory being +// populated is not adversarial; .git/hooks fills up the moment git runs. +func TestRollbackReportsWhatItCouldNotRemove(t *testing.T) { + root := t.TempDir() + target := filepath.Join(root, "ws", "hooks") + created, err := materializeWindowsACLTarget(target, false) + if err != nil { + t.Fatalf("materializeWindowsACLTarget: %v", err) + } + if err := os.WriteFile(filepath.Join(target, "pre-commit"), []byte("#!/bin/sh\n"), 0o600); err != nil { + t.Fatalf("populate: %v", err) + } + + _, err = rollbackWindowsACLMaterialization(created) + if err == nil { + t.Fatal("rollback reported success on a directory it could not empty, so callers cannot tell teardown failed") + } + if !strings.Contains(strings.ToLower(err.Error()), "hooks") { + t.Errorf("the error does not name what was left behind: %v", err) + } + // Left in place deliberately. Removing it would mean recursing, and recursion + // through a path the workspace owner controls is the thing being avoided. + if _, statErr := os.Stat(target); statErr != nil { + t.Errorf("rollback recursed into a populated directory instead of reporting it: %v", statErr) + } +} + +// The primitives take a single component and the walk relies on that. A joined +// name is resolved the ordinary way by the kernel, so an intermediate junction +// inside it is followed and the object lands outside the anchor: the pinned +// parent buys nothing if the name itself walks. +func TestChildOperationsRefuseNamesThatAreNotSingleComponents(t *testing.T) { + root := t.TempDir() + parent, err := openWindowsACLDirectoryNoFollow(root) + if err != nil { + t.Fatalf("open root: %v", err) + } + defer func() { _ = windows.CloseHandle(parent) }() + + for _, name := range []string{`sub\child`, "sub/child", "..", ".", "", `C:\absolute`, "stream:name"} { + t.Run("create dir "+name, func(t *testing.T) { + handle, _, err := createWindowsACLChildDirectory(parent, name) + if err == nil { + _ = windows.CloseHandle(handle) + t.Fatalf("accepted %q, which the kernel would resolve through intermediate directories", name) + } + }) + t.Run("delete dir "+name, func(t *testing.T) { + if err := deleteWindowsACLChildDirectory(parent, name); err == nil { + t.Fatalf("accepted %q for deletion", name) + } + }) + t.Run("create file "+name, func(t *testing.T) { + if _, _, err := createWindowsACLChildFile(parent, name); err == nil { + t.Fatalf("accepted %q for file creation", name) + } + }) + } +} + +// A junction sitting where a chain component should be must be refused when it +// is OPENED, not merely when it is created. FILE_OPEN_REPARSE_POINT hands back a +// handle to the junction itself, and using that as the next parent puts every +// deeper create on the far side of it. +func TestChildOperationsRefuseAnExistingJunction(t *testing.T) { + root := t.TempDir() + elsewhere := t.TempDir() + makeJunction(t, filepath.Join(root, "hop"), elsewhere) + + parent, err := openWindowsACLDirectoryNoFollow(root) + if err != nil { + t.Fatalf("open root: %v", err) + } + defer func() { _ = windows.CloseHandle(parent) }() + + handle, _, err := createWindowsACLChildDirectory(parent, "hop") + if err == nil { + _ = windows.CloseHandle(handle) + t.Error("createWindowsACLChildDirectory returned a junction as the next parent in the walk") + } + opened, err := openWindowsACLChildDirectory(parent, "hop") + if err == nil { + _ = windows.CloseHandle(opened) + t.Error("openWindowsACLChildDirectory returned a junction to descend through") + } +} + +// Rollback descends; it must never create. If a component was removed by +// something else in the meantime, re-making it and then deleting it would remove +// a directory the sandbox never made. +func TestRollbackDescentNeverCreatesAMissingComponent(t *testing.T) { + root := t.TempDir() + parent, err := openWindowsACLDirectoryNoFollow(root) + if err != nil { + t.Fatalf("open root: %v", err) + } + defer func() { _ = windows.CloseHandle(parent) }() + + handle, err := openWindowsACLChildDirectory(parent, "never-existed") + if err == nil { + _ = windows.CloseHandle(handle) + t.Fatal("the descent open created a directory that did not exist") + } + if !isWindowsNotExist(err) { + t.Errorf("a missing component reported %v, which rollback cannot distinguish from a real failure", err) + } + if _, statErr := os.Stat(filepath.Join(root, "never-existed")); statErr == nil { + t.Error("a directory appeared on disk from an open that should never create") + } +} diff --git a/internal/sandbox/windows_acl_raced_leaf_windows_test.go b/internal/sandbox/windows_acl_raced_leaf_windows_test.go new file mode 100644 index 000000000..6929d96e8 --- /dev/null +++ b/internal/sandbox/windows_acl_raced_leaf_windows_test.go @@ -0,0 +1,128 @@ +//go:build windows + +package sandbox + +import ( + "os" + "path/filepath" + "strings" + "testing" + "unsafe" + + "golang.org/x/sys/windows" +) + +// racedLeafDenyMask reads the DENY mask a path carries for sid. +func racedLeafDenyMask(t *testing.T, path string, sid string) uint32 { + t.Helper() + handle, _, err := openWindowsACLTarget(path) + if err != nil { + t.Fatalf("open %s: %v", path, err) + } + defer func() { _ = windows.CloseHandle(handle) }() + descriptor, err := windows.GetSecurityInfo(handle, windows.SE_FILE_OBJECT, windows.DACL_SECURITY_INFORMATION) + if err != nil { + t.Fatalf("read the DACL of %s: %v", path, err) + } + dacl, _, err := descriptor.DACL() + if err != nil || dacl == nil { + t.Fatalf("read the DACL of %s: %v", path, err) + } + wanted, err := windows.StringToSid(sid) + if err != nil { + t.Fatalf("parse %q: %v", sid, err) + } + var mask uint32 + for index := uint32(0); index < uint32(dacl.AceCount); index++ { + var ace *windows.ACCESS_ALLOWED_ACE + if windows.GetAce(dacl, index, &ace) != nil { + continue + } + if ace.Header.AceType != windows.ACCESS_DENIED_ACE_TYPE { + continue + } + if !(*windows.SID)(unsafe.Pointer(&ace.SidStart)).Equals(wanted) { + continue + } + mask |= uint32(ace.Mask) + } + return mask +} + +// OWNING THE PARENT IS NOT OWNING THE TARGET. +// +// rollbackWindowsACLSnapshots branched on createdAnything(), an OR across the +// whole materialization record. When this run creates the parent chain and a +// racer wins the leaf, that is true while the ACL-bearing file belongs to +// somebody else: Chain carries Made:true and FileMade is false. +// +// The rollback then could not remove the parent, because it is not empty, and +// the unconditional skip meant the existing no-follow, identity-validated +// restore never ran either. A file that existed before setup started was left +// wearing an aborted setup's DACL. +// +// Driven through applyWindowsACLPlan with the leaf created inside the swap hook, +// which fires after the anchor is pinned and before anything is made, so the +// race is deterministic rather than hoped for. +func TestAbortRestoresALeafARacerCreated(t *testing.T) { + workspace := t.TempDir() + parent := filepath.Join(workspace, "materialized") + leaf := filepath.Join(parent, "config") + + previous := windowsACLRacedLeafHook + t.Cleanup(func() { windowsACLRacedLeafHook = previous }) + raced := false + windowsACLRacedLeafHook = func(string) { + // The chain above already belongs to this run; the racer wins only the + // file. That mixed ownership is the whole case. + if raced { + return + } + raced = true + _ = os.WriteFile(leaf, []byte("pre-existing content"), 0o600) + } + + sid := testPrincipalSID + + plan := WindowsACLPlan{Entries: []WindowsACLEntry{{ + Action: WindowsACLDenyWrite, + Path: leaf, + Capability: sid, + Materialize: true, + MaterializeFile: true, + }}} + rollback, err := applyWindowsACLPlan(plan) + if err != nil { + t.Skipf("cannot apply an ACL plan here: %v", err) + } + + if !raced { + t.Fatal("SETUP INVALID: the leaf hook never fired, so the mixed-ownership case was not reproduced") + } + // SETUP: the racer really won, and the apply really landed on its file. + if _, statErr := os.Stat(leaf); statErr != nil { + t.Skipf("SETUP: the racer's leaf is not present, so this is not the mixed-ownership case: %v", statErr) + } + applied := racedLeafDenyMask(t, leaf, sid) + if applied == 0 { + t.Skipf("SETUP: the apply did not put a deny ACE on the raced leaf, so there is nothing to restore") + } + + // The setup aborts. + if rollback != nil { + _ = rollback() + } + + // The racer's file must not be left wearing this run's ACL. + if got := racedLeafDenyMask(t, leaf, sid); got != 0 { + t.Fatalf("a file the racer created still carries the aborted setup's deny ACE (mask=%#x); rollback skipped its restore because an ancestor happened to be ours", got) + } + // And it must still exist: a raced leaf is never ours to delete. + body, readErr := os.ReadFile(leaf) + if readErr != nil { + t.Fatalf("rollback removed a file it did not create: %v", readErr) + } + if !strings.Contains(string(body), "pre-existing") { + t.Fatalf("the raced leaf's contents changed: %q", body) + } +} diff --git a/internal/sandbox/windows_acl_relative_windows.go b/internal/sandbox/windows_acl_relative_windows.go new file mode 100644 index 000000000..77fd6d6a8 --- /dev/null +++ b/internal/sandbox/windows_acl_relative_windows.go @@ -0,0 +1,496 @@ +//go:build windows + +package sandbox + +import ( + "errors" + "fmt" + "os" + "unsafe" + + "golang.org/x/sys/windows" +) + +// Handle-relative directory operations. +// +// WHY THESE EXIST. Every pathname-based call re-resolves the whole path inside +// the kernel at the moment it runs. So verifying a component and then creating +// through it are two separate resolutions of the same string, and a workspace +// owner can swap an ancestor for a junction in the gap between them: setup +// verifies, the attacker swaps, setup creates, and the object lands outside the +// approved tree. Checking again afterwards is too late, because the thing has +// already been created somewhere it should not be. +// +// A HANDLE pins the object rather than the name. Once a directory is open, that +// handle keeps referring to the same directory however the path is later +// rearranged, so creating a child relative to it cannot be redirected. os.Mkdir, +// os.OpenFile and os.RemoveAll are pathname-based by construction with no +// relative form on Windows, which is why this drops to NtCreateFile with +// OBJECT_ATTRIBUTES.RootDirectory. + +// IO_STATUS_BLOCK.Information values for a create/open, named because "2 means +// it was created" is not something a reader should have to look up. +const ( + windowsFileOpened uintptr = 1 + windowsFileCreated uintptr = 2 +) + +// windowsACLDirectoryShare is the share mode every open here uses. A sandbox +// tree is live, so refusing to share would fail on any directory something else +// happens to have open: a denial of service on ourselves rather than a security +// property. +const windowsACLDirectoryShare = windows.FILE_SHARE_READ | windows.FILE_SHARE_WRITE | windows.FILE_SHARE_DELETE + +// windowsFileIdentity is the kernel's own answer to "is this the same object", +// independent of what it is currently called. +// +// It exists because rollback cannot hold the anchor handle open for the whole +// sandbox lifetime (see rollbackWindowsACLSnapshots), so it has to re-open the +// anchor by pathname, and a pathname can be made to name a different object. +// Crucially that substitution needs NO reparse point: rename the real directory +// aside and create an ordinary directory wearing its name, and every no-follow +// check still passes because nothing anywhere is a link. Comparing the volume +// and file index catches it, because those identify the object the kernel +// actually opened. +type windowsFileIdentity struct { + Volume uint32 + IndexHigh uint32 + IndexLow uint32 +} + +func (identity windowsFileIdentity) empty() bool { + return identity == windowsFileIdentity{} +} + +// windowsIdentityOfHandle reads the identity of an already-open object. +func windowsIdentityOfHandle(handle windows.Handle) (windowsFileIdentity, error) { + var info windows.ByHandleFileInformation + if err := windows.GetFileInformationByHandle(handle, &info); err != nil { + return windowsFileIdentity{}, fmt.Errorf("read windows file identity: %w", err) + } + return windowsFileIdentity{ + Volume: info.VolumeSerialNumber, + IndexHigh: info.FileIndexHigh, + IndexLow: info.FileIndexLow, + }, nil +} + +// openWindowsACLDirectoryNoFollow opens an existing directory by pathname, +// refusing to traverse or land on a reparse point. +// +// This is the ANCHOR for a handle-relative walk: the one pathname resolution +// that has to happen, with everything below it relative to the handle it +// returns. FILE_FLAG_OPEN_REPARSE_POINT stops the final component being +// followed, and verifyWindowsACLTargetNotRedirected then confirms no ancestor +// redirected either, because GetFinalPathNameByHandle answers for the whole +// resolved path. +func openWindowsACLDirectoryNoFollow(path string) (windows.Handle, error) { + utf16Path, err := windows.UTF16PtrFromString(path) + if err != nil { + return 0, fmt.Errorf("encode windows ACL directory %s: %w", path, err) + } + handle, err := windows.CreateFile( + utf16Path, + windows.FILE_LIST_DIRECTORY|windows.FILE_READ_ATTRIBUTES|windows.SYNCHRONIZE, + windowsACLDirectoryShare, + nil, + windows.OPEN_EXISTING, + windows.FILE_FLAG_BACKUP_SEMANTICS|windows.FILE_FLAG_OPEN_REPARSE_POINT, + 0, + ) + if err != nil { + // Errno.Is maps the not-found codes to os.ErrNotExist, so a caller + // walking up to find the deepest existing ancestor keeps working. + return 0, fmt.Errorf("open windows ACL directory %s: %w", path, err) + } + if err := verifyWindowsACLHandleIsCleanDirectory(handle, path); err != nil { + _ = windows.CloseHandle(handle) + return 0, err + } + return handle, nil +} + +// openWindowsACLDirectoryNoFollowWithIdentity is the anchor open plus the +// identity a later rollback needs in order to prove it re-opened the same +// object. +func openWindowsACLDirectoryNoFollowWithIdentity(path string) (windows.Handle, windowsFileIdentity, error) { + handle, err := openWindowsACLDirectoryNoFollow(path) + if err != nil { + return 0, windowsFileIdentity{}, err + } + identity, err := windowsIdentityOfHandle(handle) + if err != nil { + _ = windows.CloseHandle(handle) + return 0, windowsFileIdentity{}, fmt.Errorf("identify windows ACL directory %s: %w", path, err) + } + return handle, identity, nil +} + +// reopenWindowsACLDirectoryAsIdentity re-opens an anchor by pathname and refuses +// it unless the kernel says it is the object that was opened before. +// +// Used only by rollback. See windowsFileIdentity for why the pathname alone is +// not enough, and rollbackWindowsACLSnapshots for why a handle cannot simply be +// held instead. +func reopenWindowsACLDirectoryAsIdentity(path string, want windowsFileIdentity) (windows.Handle, error) { + handle, err := openWindowsACLDirectoryNoFollow(path) + if err != nil { + return 0, err + } + if want.empty() { + return handle, nil + } + got, err := windowsIdentityOfHandle(handle) + if err != nil { + _ = windows.CloseHandle(handle) + return 0, fmt.Errorf("identify windows ACL directory %s: %w", path, err) + } + if got != want { + _ = windows.CloseHandle(handle) + return 0, fmt.Errorf("refusing to unwind under %s: it is no longer the directory setup created into, so something replaced it since (possible path swap during elevated setup)", path) + } + return handle, nil +} + +// createWindowsACLChildDirectory creates one directory directly beneath parent, +// or opens it when it already exists, and reports which happened. +// +// name must be a single component; see validateWindowsACLComponent for why that +// is enforced rather than assumed. The kernel resolves it relative to the parent +// HANDLE, so nothing above it is consulted and nothing above it can be swapped +// underneath us. +// +// created is true only when this call made the directory, which the rollback +// needs: removing one that already existed would delete a user's data over a +// failure that had nothing to do with it. +func createWindowsACLChildDirectory(parent windows.Handle, name string) (handle windows.Handle, created bool, err error) { + if err := validateWindowsACLComponent(name); err != nil { + return 0, false, err + } + objectName, err := windows.NewNTUnicodeString(name) + if err != nil { + return 0, false, fmt.Errorf("encode windows ACL directory component %s: %w", name, err) + } + attributes := windows.OBJECT_ATTRIBUTES{ + RootDirectory: parent, + ObjectName: objectName, + Attributes: windows.OBJ_CASE_INSENSITIVE, + } + attributes.Length = uint32(unsafe.Sizeof(attributes)) + + var status windows.IO_STATUS_BLOCK + if err := windows.NtCreateFile( + &handle, + windows.FILE_LIST_DIRECTORY|windows.FILE_READ_ATTRIBUTES|windows.SYNCHRONIZE|windows.DELETE, + &attributes, + &status, + nil, + windows.FILE_ATTRIBUTE_DIRECTORY, + windowsACLDirectoryShare, + windows.FILE_OPEN_IF, + windows.FILE_DIRECTORY_FILE|windows.FILE_OPEN_REPARSE_POINT|windows.FILE_SYNCHRONOUS_IO_NONALERT, + 0, + 0, + ); err != nil { + return 0, false, fmt.Errorf("create windows ACL directory component %s: %w", name, err) + } + // FILE_OPEN_IF means an EXISTING child is opened rather than created, and + // FILE_OPEN_REPARSE_POINT means a junction is opened AS the junction. Without + // this check that junction becomes the parent of the next level and every + // create beneath it lands wherever it points, which is the mid-walk swap this + // whole file exists to stop. + if err := rejectWindowsACLReparseHandle(handle, name); err != nil { + _ = windows.CloseHandle(handle) + return 0, false, err + } + return handle, status.Information == windowsFileCreated, nil +} + +// openWindowsACLChildDirectory opens an EXISTING directory beneath parent and +// never creates one. +// +// Rollback walks back down the chain it created, and it must not conjure a +// component that has since been removed: FILE_OPEN_IF would recreate it, and +// then the unwind would delete a directory setup never made. FILE_OPEN is the +// whole difference from createWindowsACLChildDirectory. +func openWindowsACLChildDirectory(parent windows.Handle, name string) (handle windows.Handle, err error) { + if err := validateWindowsACLComponent(name); err != nil { + return 0, err + } + objectName, err := windows.NewNTUnicodeString(name) + if err != nil { + return 0, fmt.Errorf("encode windows ACL directory component %s: %w", name, err) + } + attributes := windows.OBJECT_ATTRIBUTES{ + RootDirectory: parent, + ObjectName: objectName, + Attributes: windows.OBJ_CASE_INSENSITIVE, + } + attributes.Length = uint32(unsafe.Sizeof(attributes)) + + // Ask for the least that still allows passing through and checking the + // reparse attribute. This runs during rollback, on directories whose ACEs + // have already been applied, so every extra right is another way for the + // unwind to be refused by the very ACL it is unwinding. Notably absent: + // SYNCHRONIZE, and with it FILE_SYNCHRONOUS_IO_NONALERT, since nothing is + // read or written through this handle. + var status windows.IO_STATUS_BLOCK + if err := windows.NtCreateFile( + &handle, + windows.FILE_TRAVERSE|windows.FILE_READ_ATTRIBUTES, + &attributes, + &status, + nil, + windows.FILE_ATTRIBUTE_DIRECTORY, + windowsACLDirectoryShare, + windows.FILE_OPEN, + windows.FILE_DIRECTORY_FILE|windows.FILE_OPEN_REPARSE_POINT, + 0, + 0, + ); err != nil { + return 0, fmt.Errorf("open windows ACL directory component %s: %w", name, err) + } + if err := rejectWindowsACLReparseHandle(handle, name); err != nil { + _ = windows.CloseHandle(handle) + return 0, err + } + return handle, nil +} + +// createWindowsACLChildFile creates an empty file directly beneath parent, or +// opens it when it already exists, and reports which happened. +// +// The counterpart to createWindowsACLChildDirectory for the one materialized +// target that must be a FILE: .git/config, where creating a directory instead +// would break `git init` outright rather than merely mis-ACL it. +// +// FILE_OPEN_IF rather than FILE_CREATE deliberately. The pathname version this +// replaces used O_CREATE|O_EXCL and then tolerated os.ErrExist, so a racing +// creator winning was fine. FILE_CREATE's collision status is +// STATUS_OBJECT_NAME_COLLISION, which errors.Is(err, os.ErrExist) does NOT +// match, so porting it literally would have turned that tolerated race into a +// hard failure. FILE_OPEN_IF keeps the old behaviour and reports the truth in +// created, which rollback needs so it never deletes a file it did not make. +func createWindowsACLChildFile(parent windows.Handle, name string) (created bool, identity windowsFileIdentity, err error) { + if err := validateWindowsACLComponent(name); err != nil { + return false, windowsFileIdentity{}, err + } + objectName, err := windows.NewNTUnicodeString(name) + if err != nil { + return false, windowsFileIdentity{}, fmt.Errorf("encode windows ACL file component %s: %w", name, err) + } + attributes := windows.OBJECT_ATTRIBUTES{ + RootDirectory: parent, + ObjectName: objectName, + Attributes: windows.OBJ_CASE_INSENSITIVE, + } + attributes.Length = uint32(unsafe.Sizeof(attributes)) + + var handle windows.Handle + var status windows.IO_STATUS_BLOCK + if err := windows.NtCreateFile( + &handle, + windows.FILE_READ_ATTRIBUTES|windows.SYNCHRONIZE, + &attributes, + &status, + nil, + windows.FILE_ATTRIBUTE_NORMAL, + windowsACLDirectoryShare, + windows.FILE_OPEN_IF, + windows.FILE_NON_DIRECTORY_FILE|windows.FILE_OPEN_REPARSE_POINT|windows.FILE_SYNCHRONOUS_IO_NONALERT, + 0, + 0, + ); err != nil { + return false, windowsFileIdentity{}, fmt.Errorf("create windows ACL file component %s: %w", name, err) + } + if err := rejectWindowsACLReparseHandle(handle, name); err != nil { + _ = windows.CloseHandle(handle) + return false, windowsFileIdentity{}, err + } + createdNow := status.Information == windowsFileCreated + identity, err = windowsIdentityOfHandle(handle) + if err != nil { + _ = windows.CloseHandle(handle) + return createdNow, windowsFileIdentity{}, fmt.Errorf("identify windows ACL file component %s: %w", name, err) + } + if err := windows.CloseHandle(handle); err != nil { + return createdNow, identity, fmt.Errorf("close windows ACL file component %s: %w", name, err) + } + return createdNow, identity, nil +} + +// deleteWindowsACLChildDirectory removes one directory directly beneath parent. +// +// The counterpart to the create above, and the reason rollback cannot use +// os.RemoveAll: that takes a pathname, so an ancestor swapped to a junction +// AFTER the object was created sends the recursive delete somewhere else and +// takes unrelated trees with it. Resolving relative to the parent handle makes +// that impossible. +// +// A missing child is not an error: rollback runs on failure paths where the +// object may never have been created. +func deleteWindowsACLChildDirectory(parent windows.Handle, name string) error { + return deleteWindowsACLChild(parent, name, true, windowsFileIdentity{}) +} + +// deleteWindowsACLChildDirectoryAs and deleteWindowsACLChildFileAs remove the +// child only if the object at name is still want. The identity is read from +// the very handle the delete disposition is set on, so a replacement renamed +// in between a check and the unlink cannot be the thing removed. +func deleteWindowsACLChildDirectoryAs(parent windows.Handle, name string, want windowsFileIdentity) error { + return deleteWindowsACLChild(parent, name, true, want) +} + +func deleteWindowsACLChildFileAs(parent windows.Handle, name string, want windowsFileIdentity) error { + return deleteWindowsACLChild(parent, name, false, want) +} + +// deleteWindowsACLChildFile removes one file directly beneath parent, for the +// materialized .git/config carveout. Rollback picks between this and the +// directory form from the shape recorded at materialization time rather than by +// stat-ing the pathname, because a stat is another pathname resolution and this +// whole file exists to avoid those. +func deleteWindowsACLChildFile(parent windows.Handle, name string) error { + return deleteWindowsACLChild(parent, name, false, windowsFileIdentity{}) +} + +// deleteWindowsACLChild opens one child relative to parent and deletes it by +// SETTING ITS DISPOSITION, not with FILE_DELETE_ON_CLOSE. +// +// That distinction is the whole point, and it was measured rather than assumed. +// FILE_DELETE_ON_CLOSE defers the removal to cleanup, where a non-empty +// directory makes it fail with nothing to report it to: NtCreateFile returns +// success, CloseHandle returns success, and the directory is still there. A +// rollback built on it would report success while leaving materialized state on +// disk, which is strictly worse than the os.RemoveAll it replaces, because +// os.RemoveAll at least removed it. +// +// NtSetInformationFile answers synchronously and to the caller, so a non-empty +// directory comes back as STATUS_DIRECTORY_NOT_EMPTY. Leaving residue is +// acceptable, since the alternative is a recursive delete through a pathname an +// attacker may control; lying about having removed it is not. +// +// FILE_DIRECTORY_FILE / FILE_NON_DIRECTORY_FILE also make the open refuse an +// object of the wrong shape rather than deleting it. +func deleteWindowsACLChild(parent windows.Handle, name string, directory bool, want windowsFileIdentity) error { + if err := validateWindowsACLComponent(name); err != nil { + return err + } + objectName, err := windows.NewNTUnicodeString(name) + if err != nil { + return fmt.Errorf("encode windows ACL component %s: %w", name, err) + } + attributes := windows.OBJECT_ATTRIBUTES{ + RootDirectory: parent, + ObjectName: objectName, + Attributes: windows.OBJ_CASE_INSENSITIVE, + } + attributes.Length = uint32(unsafe.Sizeof(attributes)) + + shapeOption := uint32(windows.FILE_NON_DIRECTORY_FILE) + shapeAttribute := uint32(windows.FILE_ATTRIBUTE_NORMAL) + if directory { + shapeOption = windows.FILE_DIRECTORY_FILE + shapeAttribute = windows.FILE_ATTRIBUTE_DIRECTORY + } + + // DELETE alone. Asking for SYNCHRONIZE as well would make this fail on any + // object already carrying a deny-read ACE, because FILE_GENERIC_READ and + // FILE_GENERIC_EXECUTE both include SYNCHRONIZE, and rollback exists + // precisely to undo objects that have just been ACL'd. Nothing is read or + // written through this handle, so synchronous IO is not needed either. + var handle windows.Handle + var status windows.IO_STATUS_BLOCK + if err := windows.NtCreateFile( + &handle, + windows.DELETE|windows.FILE_READ_ATTRIBUTES, + &attributes, + &status, + nil, + shapeAttribute, + windowsACLDirectoryShare, + windows.FILE_OPEN, + shapeOption|windows.FILE_OPEN_REPARSE_POINT, + 0, + 0, + ); err != nil { + if isWindowsNotExist(err) { + return nil + } + return fmt.Errorf("open windows ACL component %s for delete: %w", name, err) + } + defer func() { _ = windows.CloseHandle(handle) }() + if !want.empty() { + got, err := windowsIdentityOfHandle(handle) + if err != nil { + return fmt.Errorf("identify windows ACL component %s before delete: %w", name, err) + } + if got != want { + return fmt.Errorf("refusing to remove %s: it is no longer the object setup created", name) + } + } + + // One BOOLEAN: FILE_DISPOSITION_INFORMATION.DeleteFile = TRUE. + disposition := byte(1) + var setStatus windows.IO_STATUS_BLOCK + if err := windows.NtSetInformationFile( + handle, + &setStatus, + &disposition, + uint32(unsafe.Sizeof(disposition)), + windows.FileDispositionInformation, + ); err != nil { + return fmt.Errorf("delete windows ACL component %s: %w", name, err) + } + return nil +} + +// rejectWindowsACLReparseHandle refuses a handle that landed on a reparse point. +// +// Deliberately NOT verifyWindowsACLTargetNotRedirected: that one compares the +// handle's resolved path against an expected pathname, which is exactly the +// pathname dependency the handle-relative walk removes. A child opened relative +// to a pinned parent is in the right place by construction even when the +// pathname no longer leads there, so comparing paths would reject correct, safe +// creates whenever the tree was legitimately renamed. The attribute is the only +// thing worth checking here. +func rejectWindowsACLReparseHandle(handle windows.Handle, name string) error { + var info windows.ByHandleFileInformation + if err := windows.GetFileInformationByHandle(handle, &info); err != nil { + return fmt.Errorf("inspect windows ACL component %s: %w", name, err) + } + if info.FileAttributes&windows.FILE_ATTRIBUTE_REPARSE_POINT != 0 { + return fmt.Errorf("refusing to work through reparse-point component %s: possible path swap during elevated setup", name) + } + return nil +} + +// verifyWindowsACLHandleIsCleanDirectory rejects a handle that landed on a +// reparse point, or on an object other than the path asked for. +func verifyWindowsACLHandleIsCleanDirectory(handle windows.Handle, path string) error { + var info windows.ByHandleFileInformation + if err := windows.GetFileInformationByHandle(handle, &info); err != nil { + return fmt.Errorf("inspect windows ACL directory %s: %w", path, err) + } + if info.FileAttributes&windows.FILE_ATTRIBUTE_REPARSE_POINT != 0 { + return fmt.Errorf("refusing to materialize under reparse-point path component %s: possible path swap during elevated setup", path) + } + return verifyWindowsACLTargetNotRedirected(handle, path) +} + +// isWindowsNotExist reports a missing-object error from either side of the API. +// NtCreateFile returns NTSTATUS values, which do not map to os.ErrNotExist the +// way the Win32 error codes do. +func isWindowsNotExist(err error) bool { + if err == nil { + return false + } + if os.IsNotExist(err) { + return true + } + var status windows.NTStatus + if errors.As(err, &status) { + return status == windows.STATUS_OBJECT_NAME_NOT_FOUND || status == windows.STATUS_OBJECT_PATH_NOT_FOUND + } + return false +} diff --git a/internal/sandbox/windows_acl_relative_windows_test.go b/internal/sandbox/windows_acl_relative_windows_test.go new file mode 100644 index 000000000..4f52a83d5 --- /dev/null +++ b/internal/sandbox/windows_acl_relative_windows_test.go @@ -0,0 +1,230 @@ +//go:build windows + +package sandbox + +import ( + "os" + "path/filepath" + "testing" + + "golang.org/x/sys/windows" +) + +// THE SWAP THAT PATHNAMES CANNOT SURVIVE. +// +// This is the race behind the materialization P1. A pathname-based create +// resolves the whole path again at the moment it runs, so an ancestor replaced +// between the verification and the create sends the new directory somewhere else +// entirely. Verifying afterwards is too late: the object already exists in the +// wrong place. +// +// A handle pins the OBJECT, not the name. This test performs the swap for real, +// in the window that used to be exploitable, and asserts the child still lands +// in the directory that was verified. +// +// Worth noting for anyone extending this: the fix is what makes the race +// testable at all. Against the old code the swap had to be threaded into the +// middle of a function; here it is three ordinary lines between an open and a +// create, because the handle is held across them. +func TestChildCreationFollowsTheHandleNotThePath(t *testing.T) { + root := t.TempDir() + approved := filepath.Join(root, "approved") + elsewhere := filepath.Join(root, "elsewhere") + for _, dir := range []string{approved, elsewhere} { + if err := os.Mkdir(dir, 0o700); err != nil { + t.Fatalf("seed %s: %v", dir, err) + } + } + + // Verified once, exactly as setup does before it materializes anything. + parent, err := openWindowsACLDirectoryNoFollow(approved) + if err != nil { + t.Fatalf("open approved directory: %v", err) + } + defer func() { _ = windows.CloseHandle(parent) }() + + // THE SWAP, in the window that used to be exploitable: move the verified + // directory aside and leave a junction to somewhere else wearing its name. + moved := filepath.Join(root, "approved-moved") + if err := os.Rename(approved, moved); err != nil { + t.Skipf("cannot rename a directory with an open handle on this filesystem: %v", err) + } + makeJunction(t, approved, elsewhere) + + // The create resolves against the handle, so it must ignore the junction now + // sitting at the original pathname. + child, created, err := createWindowsACLChildDirectory(parent, "materialized") + if err != nil { + t.Fatalf("create child relative to the pinned handle: %v", err) + } + defer func() { _ = windows.CloseHandle(child) }() + if !created { + t.Error("created = false for a directory that did not exist") + } + + if _, err := os.Stat(filepath.Join(moved, "materialized")); err != nil { + t.Errorf("the child did not land in the verified directory: %v", err) + } + if _, err := os.Stat(filepath.Join(elsewhere, "materialized")); err == nil { + t.Fatal("ESCAPED: the child was created through the junction, outside the approved tree") + } +} + +// Materialization runs on trees that may already be half-built, so creating an +// existing directory has to be a no-op rather than a failure. created must still +// report the truth, because rollback deletes only what this call made. +func TestCreatingAnExistingChildOpensItInstead(t *testing.T) { + root := t.TempDir() + if err := os.Mkdir(filepath.Join(root, "already"), 0o700); err != nil { + t.Fatalf("seed: %v", err) + } + parent, err := openWindowsACLDirectoryNoFollow(root) + if err != nil { + t.Fatalf("open root: %v", err) + } + defer func() { _ = windows.CloseHandle(parent) }() + + handle, created, err := createWindowsACLChildDirectory(parent, "already") + if err != nil { + t.Fatalf("open existing child: %v", err) + } + _ = windows.CloseHandle(handle) + if created { + t.Error("created = true for a directory that already existed; rollback would delete a user's data") + } +} + +// The rollback counterpart. os.RemoveAll on a pathname whose ancestor has since +// become a junction recurses outside the workspace and deletes unrelated trees, +// which is the second P1. Resolving relative to the parent handle cannot. +func TestDeleteFollowsTheHandleNotThePath(t *testing.T) { + root := t.TempDir() + approved := filepath.Join(root, "approved") + elsewhere := filepath.Join(root, "elsewhere") + for _, dir := range []string{approved, elsewhere} { + if err := os.Mkdir(dir, 0o700); err != nil { + t.Fatalf("seed %s: %v", dir, err) + } + } + // A bystander under the decoy: if the delete ever resolves by pathname it is + // reachable, and its survival is what proves the delete did not. + if err := os.Mkdir(filepath.Join(elsewhere, "victim"), 0o700); err != nil { + t.Fatalf("seed victim: %v", err) + } + if err := os.Mkdir(filepath.Join(approved, "victim"), 0o700); err != nil { + t.Fatalf("seed target: %v", err) + } + + parent, err := openWindowsACLDirectoryNoFollow(approved) + if err != nil { + t.Fatalf("open approved directory: %v", err) + } + defer func() { _ = windows.CloseHandle(parent) }() + + moved := filepath.Join(root, "approved-moved") + if err := os.Rename(approved, moved); err != nil { + t.Skipf("cannot rename a directory with an open handle on this filesystem: %v", err) + } + makeJunction(t, approved, elsewhere) + + if err := deleteWindowsACLChildDirectory(parent, "victim"); err != nil { + t.Fatalf("delete relative to the pinned handle: %v", err) + } + if _, err := os.Stat(filepath.Join(elsewhere, "victim")); err != nil { + t.Fatal("DESTRUCTIVE: rollback followed the junction and deleted a directory outside the approved tree") + } + if _, err := os.Stat(filepath.Join(moved, "victim")); err == nil { + t.Error("the directory inside the approved tree was not removed") + } +} + +// A delete that cannot happen must SAY so. +// +// This is the gap that let a silent bug ship in the first version of this file. +// It deleted with FILE_DELETE_ON_CLOSE, which defers the removal to cleanup, +// where a non-empty directory makes it fail with nowhere to report it: the open +// returned success, the close returned success, and the directory was still +// there. The only test covering deletion used an EMPTY directory, so it passed +// throughout. Rollback built on that would have reported success while leaving +// materialized state on disk, which is worse than the os.RemoveAll it replaced, +// because os.RemoveAll actually removed it. +func TestDeletingANonEmptyDirectoryIsReported(t *testing.T) { + root := t.TempDir() + populated := filepath.Join(root, "populated") + if err := os.Mkdir(populated, 0o700); err != nil { + t.Fatalf("seed: %v", err) + } + if err := os.WriteFile(filepath.Join(populated, "occupant"), []byte("x"), 0o600); err != nil { + t.Fatalf("seed occupant: %v", err) + } + parent, err := openWindowsACLDirectoryNoFollow(root) + if err != nil { + t.Fatalf("open root: %v", err) + } + defer func() { _ = windows.CloseHandle(parent) }() + + err = deleteWindowsACLChildDirectory(parent, "populated") + if _, statErr := os.Stat(populated); statErr != nil { + t.Fatalf("the directory was removed with its contents, which this delete must never do: %v", statErr) + } + if err == nil { + t.Fatal("reported success while leaving the directory in place") + } +} + +// The directory form must refuse a file rather than delete it, and the file form +// must handle the one materialized target that is a file. +func TestDeleteDistinguishesFilesFromDirectories(t *testing.T) { + root := t.TempDir() + if err := os.WriteFile(filepath.Join(root, "config"), []byte("x"), 0o600); err != nil { + t.Fatalf("seed: %v", err) + } + parent, err := openWindowsACLDirectoryNoFollow(root) + if err != nil { + t.Fatalf("open root: %v", err) + } + defer func() { _ = windows.CloseHandle(parent) }() + + if err := deleteWindowsACLChildDirectory(parent, "config"); err == nil { + t.Error("the directory delete accepted a file") + } + if err := deleteWindowsACLChildFile(parent, "config"); err != nil { + t.Fatalf("deleteWindowsACLChildFile: %v", err) + } + if _, err := os.Stat(filepath.Join(root, "config")); err == nil { + t.Error("the file survived its delete") + } +} + +// Rollback runs on failure paths where the object may never have been created, +// so a missing child is success rather than an error to report. +func TestDeletingAMissingChildIsNotAnError(t *testing.T) { + root := t.TempDir() + parent, err := openWindowsACLDirectoryNoFollow(root) + if err != nil { + t.Fatalf("open root: %v", err) + } + defer func() { _ = windows.CloseHandle(parent) }() + + if err := deleteWindowsACLChildDirectory(parent, "never-existed"); err != nil { + t.Fatalf("deleting a missing child reported an error: %v", err) + } +} + +// The anchor open is the one pathname resolution in the walk, so it has to +// refuse a junction itself rather than leaving it to a later check. +func TestOpeningAJunctionAnchorIsRefused(t *testing.T) { + root := t.TempDir() + real := filepath.Join(root, "real") + if err := os.Mkdir(real, 0o700); err != nil { + t.Fatalf("seed: %v", err) + } + link := filepath.Join(root, "link") + makeJunction(t, link, real) + + handle, err := openWindowsACLDirectoryNoFollow(link) + if err == nil { + _ = windows.CloseHandle(handle) + t.Fatal("opened a junction as the materialization anchor; every create beneath it would land outside the approved tree") + } +} diff --git a/internal/sandbox/windows_acl_reparse_windows.go b/internal/sandbox/windows_acl_reparse_windows.go new file mode 100644 index 000000000..71c6e206c --- /dev/null +++ b/internal/sandbox/windows_acl_reparse_windows.go @@ -0,0 +1,161 @@ +//go:build windows + +package sandbox + +import ( + "fmt" + "path/filepath" + "strings" + + "golang.org/x/sys/windows" +) + +// GetFinalPathNameByHandle flags. x/sys/windows does not export these. +const ( + windowsFileNameNormalized uint32 = 0x0 + windowsVolumeNameDOS uint32 = 0x0 +) + +// verifyWindowsACLTargetNotRedirected fails when an opened handle resolved +// somewhere other than the requested path, which means a component along the way +// is a reparse point. +// +// openWindowsACLTarget's FILE_FLAG_OPEN_REPARSE_POINT check covers the FINAL +// component only; CreateFile still resolves ANCESTORS. A user who controls the +// workspace can turn an ancestor — .git, say — into a junction before elevated +// setup runs, and setup would then apply its DACL change to an object outside +// the approved tree while the final-component check still passed. Junctions need +// no privilege to create, unlike symlinks, so this is reachable by exactly the +// unprivileged user the sandbox exists to contain. +// +// GetFinalPathNameByHandle answers where the handle actually landed, covering +// every component in one call rather than walking the path and re-checking each +// component (which would also race between the checks). +// +// The comparison basis must NEVER follow a reparse point, and getting that wrong +// is what made this guard mostly ornamental. +// +// It used to compare against canonicalSandboxWorkspaceRoot(path), which runs +// filepath.EvalSymlinks — the same resolution the kernel just performed. So for a +// directory SYMLINK the two sides agreed precisely BECAUSE the redirect happened: +// GetFinalPathNameByHandle followed the link to produce one side, EvalSymlinks +// followed the same link to produce the other, and elevated setup went on to +// rewrite the DACL of an object outside the workspace. +// +// Junctions were rejected, but by accident rather than by design. Go reports a +// junction as ModeIrregular rather than ModeSymlink, so EvalSymlinks refuses it, +// canonicalSandboxWorkspaceRoot falls back to the lexical path, and the mismatch +// surfaces. That is a property of the standard library's mode bits, not of this +// check: a Go release that resolved mount points would silently disarm the one +// case this was known to catch. +// +// So the expected side is now normalized WITHOUT resolving anything. Casing and +// 8.3 short names still have to compare equal, since a legitimate target can be +// spelled either way and still be the same object — EqualFold covers the first +// and GetLongPathName the second, and neither follows a link. +// +// This does mean a target deliberately spelled through a symlink is now refused. +// That is the intended direction: the whole question here is whether the object +// the handle landed on is the object that was named, and a caller that wants a +// link's target should name the target. +func verifyWindowsACLTargetNotRedirected(handle windows.Handle, path string) error { + buffer := make([]uint16, windows.MAX_LONG_PATH) + n, err := windows.GetFinalPathNameByHandle(handle, &buffer[0], uint32(len(buffer)), windowsFileNameNormalized|windowsVolumeNameDOS) + if err != nil { + return fmt.Errorf("resolve windows ACL target %s: %w", path, err) + } + if int(n) < len(buffer) { + buffer = buffer[:n] + } + actual := trimWindowsExtendedPathPrefix(windows.UTF16ToString(buffer)) + expected := trimWindowsExtendedPathPrefix(windowsACLComparablePath(path)) + if !strings.EqualFold(filepath.Clean(actual), filepath.Clean(expected)) { + return fmt.Errorf("refusing to apply ACL to %s: it resolves to %s, so a parent directory is a reparse point (possible path swap during elevated setup)", path, actual) + } + return nil +} + +// windowsACLComparablePath normalizes a path for comparison against where a +// handle actually landed, WITHOUT resolving any reparse point. +// +// The two things that must still compare equal are casing and 8.3 short names, +// because either can spell the same object. GetLongPathName expands the short +// form by reading directory entries; it does not follow a symlink or a junction +// to its target, which is the entire reason it is used here instead of +// EvalSymlinks. Casing is left to the caller's EqualFold. +// +// Every failure degrades to the lexically cleaned path rather than to a resolved +// one. A path this cannot expand is compared as written, which can only produce +// a spurious MISMATCH — a refusal — and never a spurious match. That is the +// direction a redirect check has to fail in. +func windowsACLComparablePath(path string) string { + cleaned := filepath.Clean(strings.TrimSpace(path)) + if cleaned == "" || cleaned == "." { + return cleaned + } + if absolute, err := filepath.Abs(cleaned); err == nil { + cleaned = absolute + } + wide, err := windows.UTF16PtrFromString(cleaned) + if err != nil { + return cleaned + } + buffer := make([]uint16, windows.MAX_LONG_PATH) + n, err := windows.GetLongPathName(wide, &buffer[0], uint32(len(buffer))) + if err != nil || n == 0 || int(n) >= len(buffer) { + return cleaned + } + return filepath.Clean(windows.UTF16ToString(buffer[:n])) +} + +// trimWindowsExtendedPathPrefix strips the \?\ form GetFinalPathNameByHandle +// returns so it can be compared with an ordinary path. +func trimWindowsExtendedPathPrefix(path string) string { + // Built from filepath.Separator rather than written as literals so the + // backslashes cannot be miscounted by whatever writes this file. + sep := string(filepath.Separator) + devicePrefix := sep + sep + "?" + sep + uncPrefix := devicePrefix + "UNC" + sep + if strings.HasPrefix(path, uncPrefix) { + return sep + sep + strings.TrimPrefix(path, uncPrefix) + } + return strings.TrimPrefix(path, devicePrefix) +} + +// verifyWindowsACLPathComponentNotRedirected opens one path component no-follow +// and refuses it if it is a reparse point or resolves anywhere other than its own +// pathname. Because GetFinalPathNameByHandle answers for the WHOLE resolved path, +// verifying a single existing component also clears every ancestor above it. +// +// A missing component surfaces as os.ErrNotExist so the caller can walk further +// up. Only FILE_READ_ATTRIBUTES is requested: this inspects, it never writes, and +// asking for more would fail on ancestors the setup process has no rights on. +func verifyWindowsACLPathComponentNotRedirected(path string) error { + utf16Path, err := windows.UTF16PtrFromString(path) + if err != nil { + return fmt.Errorf("encode windows ACL path component %s: %w", path, err) + } + handle, err := windows.CreateFile( + utf16Path, + windows.FILE_READ_ATTRIBUTES, + windows.FILE_SHARE_READ|windows.FILE_SHARE_WRITE|windows.FILE_SHARE_DELETE, + nil, + windows.OPEN_EXISTING, + windows.FILE_FLAG_BACKUP_SEMANTICS|windows.FILE_FLAG_OPEN_REPARSE_POINT, + 0, + ) + if err != nil { + // syscall.Errno.Is maps ERROR_FILE_NOT_FOUND/ERROR_PATH_NOT_FOUND to + // os.ErrNotExist, so the caller's errors.Is check keeps working. + return fmt.Errorf("open windows ACL path component %s: %w", path, err) + } + defer func() { _ = windows.CloseHandle(handle) }() + var info windows.ByHandleFileInformation + if err := windows.GetFileInformationByHandle(handle, &info); err != nil { + return fmt.Errorf("inspect windows ACL path component %s: %w", path, err) + } + if info.FileAttributes&windows.FILE_ATTRIBUTE_REPARSE_POINT != 0 { + return fmt.Errorf("refusing to materialize under reparse-point path component %s: possible path swap during elevated setup", path) + } + return verifyWindowsACLTargetNotRedirected(handle, path) +} diff --git a/internal/sandbox/windows_acl_reparse_windows_test.go b/internal/sandbox/windows_acl_reparse_windows_test.go new file mode 100644 index 000000000..c529f461b --- /dev/null +++ b/internal/sandbox/windows_acl_reparse_windows_test.go @@ -0,0 +1,79 @@ +//go:build windows + +package sandbox + +import ( + "os" + "os/exec" + "path/filepath" + "strings" + "testing" + + "golang.org/x/sys/windows" +) + +// An unprivileged user who controls the workspace can turn an ancestor of a +// configured ACL target into a junction before elevated setup runs. CreateFile +// resolves ancestors even with FILE_FLAG_OPEN_REPARSE_POINT, so the +// final-component check passes and setup would rewrite the DACL of an object +// outside the approved tree. +// +// Junctions, unlike symlinks, need no privilege — which is what makes this +// reachable by exactly the user the sandbox is containing. +func TestOpenWindowsACLTargetRefusesAJunctionAncestor(t *testing.T) { + base := t.TempDir() + workspace := filepath.Join(base, "workspace") + outside := filepath.Join(base, "OUTSIDE") + for _, dir := range []string{workspace, outside} { + if err := os.MkdirAll(dir, 0o700); err != nil { + t.Fatalf("mkdir %s: %v", dir, err) + } + } + // The object the attacker wants setup to touch. + victim := filepath.Join(outside, "hooks") + if err := os.MkdirAll(victim, 0o700); err != nil { + t.Fatalf("mkdir victim: %v", err) + } + + // .git is the ancestor, and it is a junction to OUTSIDE. + gitDir := filepath.Join(workspace, ".git") + if out, err := exec.Command("cmd", "/c", "mklink", "/J", gitDir, outside).CombinedOutput(); err != nil { + t.Skipf("cannot create a junction here: %v %s", err, out) + } + + // This is the path setup would be configured with. + target := filepath.Join(gitDir, "hooks") + if _, err := os.Stat(target); err != nil { + t.Fatalf("precondition: the junction should make %s reachable: %v", target, err) + } + + handle, _, err := openWindowsACLTarget(target) + if err == nil { + _ = windows.CloseHandle(handle) + t.Fatal("opened an ACL target through a junction ancestor; setup would have rewritten a DACL outside the workspace") + } + if !strings.Contains(err.Error(), "reparse point") { + t.Errorf("error = %v, want it to name the reparse point", err) + } + t.Logf("refused as expected: %v", err) +} + +// The guard must not reject ordinary targets, including ones spelled +// non-canonically — a differently-cased path is the same object, not a redirect. +func TestOpenWindowsACLTargetAcceptsAnOrdinaryTarget(t *testing.T) { + dir := t.TempDir() + nested := filepath.Join(dir, "Nested") + if err := os.MkdirAll(nested, 0o700); err != nil { + t.Fatalf("mkdir: %v", err) + } + for _, spelling := range []string{nested, strings.ToLower(nested)} { + handle, isDir, err := openWindowsACLTarget(spelling) + if err != nil { + t.Fatalf("openWindowsACLTarget(%q): %v", spelling, err) + } + if !isDir { + t.Errorf("%q reported as not a directory", spelling) + } + _ = windows.CloseHandle(handle) + } +} diff --git a/internal/sandbox/windows_acl_restore_identity_windows_test.go b/internal/sandbox/windows_acl_restore_identity_windows_test.go new file mode 100644 index 000000000..c3c04a8ff --- /dev/null +++ b/internal/sandbox/windows_acl_restore_identity_windows_test.go @@ -0,0 +1,148 @@ +//go:build windows + +package sandbox + +import ( + "os" + "path/filepath" + "strings" + "testing" + + "golang.org/x/sys/windows" +) + +// THE RESTORE MUST LAND ON THE OBJECT THE DACL CAME FROM. +// +// Creation and the materialization unwind were both bound to handles, and the +// restore branch was left re-resolving the target by pathname. Its comment +// justified that with a no-follow re-open, which rules out a REPARSE POINT +// swapped in since apply and nothing else. The other substitution passes it +// untouched: rename the target aside and put an ordinary directory of the same +// name in its place. Nothing in that decoy is a link, so the no-follow open +// succeeds, the pre-apply DACL is written onto the attacker's object, and the +// real one keeps the ACEs from the setup that just aborted. +func TestRollbackRefusesToRestoreOntoAReplacedTarget(t *testing.T) { + root := t.TempDir() + approved := filepath.Join(root, "ws") + target := filepath.Join(approved, "target") + if err := os.MkdirAll(target, 0o700); err != nil { + t.Fatalf("seed the approved tree: %v", err) + } + + handle, _, err := openWindowsACLTarget(target) + if err != nil { + t.Fatalf("open the target: %v", err) + } + descriptor, err := windows.GetSecurityInfo(handle, windows.SE_FILE_OBJECT, windows.DACL_SECURITY_INFORMATION) + if err != nil { + _ = windows.CloseHandle(handle) + t.Fatalf("read the target DACL: %v", err) + } + identity, err := windowsIdentityOfHandle(handle) + if err != nil { + _ = windows.CloseHandle(handle) + t.Fatalf("read the target identity: %v", err) + } + _ = windows.CloseHandle(handle) + + // The swap, exactly as it would happen between apply and rollback. Ordinary + // directories throughout: no links, no reparse points. + if err := os.Rename(target, target+"-moved"); err != nil { + t.Skipf("cannot rename here: %v", err) + } + if err := os.Mkdir(target, 0o700); err != nil { + t.Fatalf("plant the replacement: %v", err) + } + + snapshot := windowsACLSnapshot{Path: target, Descriptor: descriptor, TargetID: identity} + err = rollbackWindowsACLSnapshots([]windowsACLSnapshot{snapshot}) + if err == nil { + t.Fatal("rollback wrote the pre-apply DACL onto a different object wearing the target's name") + } + if !strings.Contains(err.Error(), "no longer the object") { + t.Errorf("refused for the wrong reason: %v", err) + } +} + +// And an untouched target still restores, or the guard above would be satisfied +// by a rollback that refuses everything. +func TestRollbackRestoresAnUnchangedTarget(t *testing.T) { + target := filepath.Join(t.TempDir(), "target") + if err := os.Mkdir(target, 0o700); err != nil { + t.Fatalf("seed the target: %v", err) + } + + handle, _, err := openWindowsACLTarget(target) + if err != nil { + t.Fatalf("open the target: %v", err) + } + descriptor, err := windows.GetSecurityInfo(handle, windows.SE_FILE_OBJECT, windows.DACL_SECURITY_INFORMATION) + if err != nil { + _ = windows.CloseHandle(handle) + t.Fatalf("read the target DACL: %v", err) + } + identity, err := windowsIdentityOfHandle(handle) + if err != nil { + _ = windows.CloseHandle(handle) + t.Fatalf("read the target identity: %v", err) + } + _ = windows.CloseHandle(handle) + + snapshot := windowsACLSnapshot{Path: target, Descriptor: descriptor, TargetID: identity} + if err := rollbackWindowsACLSnapshots([]windowsACLSnapshot{snapshot}); err != nil { + t.Fatalf("an unchanged target was refused: %v", err) + } +} + +// THE COMMENT AT rollbackWindowsACLSnapshots CLAIMED A TEST THAT DID NOT EXIST. +// +// It said "TestRollbackUnwindsDescendantsBeforeAncestors pins it" about the +// reverse iteration order, and grepping the repo found only that sentence. The +// ordering is load-bearing twice over: a materialized directory must be empty +// before its own removal is attempted, and SetSecurityInfo propagates +// inheritable ACEs downward so the ancestor has to be restored last. Neither +// was pinned by anything. +func TestRollbackUnwindsDescendantsBeforeAncestors(t *testing.T) { + root := t.TempDir() + ancestor := filepath.Join(root, "ws") + descendant := filepath.Join(ancestor, "child") + if err := os.MkdirAll(descendant, 0o700); err != nil { + t.Fatalf("seed the tree: %v", err) + } + + snapshotFor := func(path string) windowsACLSnapshot { + t.Helper() + handle, _, err := openWindowsACLTarget(path) + if err != nil { + t.Fatalf("open %s: %v", path, err) + } + defer windows.CloseHandle(handle) + descriptor, err := windows.GetSecurityInfo(handle, windows.SE_FILE_OBJECT, windows.DACL_SECURITY_INFORMATION) + if err != nil { + t.Fatalf("read the DACL for %s: %v", path, err) + } + identity, err := windowsIdentityOfHandle(handle) + if err != nil { + t.Fatalf("read the identity for %s: %v", path, err) + } + return windowsACLSnapshot{Path: path, Descriptor: descriptor, TargetID: identity} + } + + // Ascending by path key, the order applyWindowsACLPlan produces. + snapshots := []windowsACLSnapshot{snapshotFor(ancestor), snapshotFor(descendant)} + + var order []string + restore := windowsACLRestoreHook + windowsACLRestoreHook = func(path string) { order = append(order, path) } + t.Cleanup(func() { windowsACLRestoreHook = restore }) + + if err := rollbackWindowsACLSnapshots(snapshots); err != nil { + t.Fatalf("rollbackWindowsACLSnapshots: %v", err) + } + if len(order) != 2 { + t.Fatalf("restored %d targets, want 2: %v", len(order), order) + } + if order[0] != descendant || order[1] != ancestor { + t.Errorf("restore order was %v; the descendant must be restored before its ancestor, because SetSecurityInfo propagates inheritable ACEs downward", order) + } +} diff --git a/internal/sandbox/windows_acl_rollback_identity_windows_test.go b/internal/sandbox/windows_acl_rollback_identity_windows_test.go new file mode 100644 index 000000000..765b211e8 --- /dev/null +++ b/internal/sandbox/windows_acl_rollback_identity_windows_test.go @@ -0,0 +1,102 @@ +//go:build windows + +package sandbox + +import ( + "os" + "path/filepath" + "testing" +) + +// ROLLBACK DELETES THE OBJECT IT CREATED, NOT WHATEVER NOW WEARS ITS NAME. +// +// The anchor's identity was verified, but the created children were recorded +// by name only, so after setup materialized .git/config another process could +// move it aside and put an ordinary file at that name beneath the unchanged +// anchor, and a later setup failure deleted the replacement. Each created +// component now carries the identity read from the handle that created it, +// and the delete disposition is set on a handle whose identity was checked. +func TestRollbackRefusesAReplacedMaterializedChild(t *testing.T) { + for name, asFile := range map[string]bool{"file target": true, "directory target": false} { + t.Run(name, func(t *testing.T) { + workspace := t.TempDir() + target := filepath.Join(workspace, ".git", "hooks") + if asFile { + target = filepath.Join(workspace, ".git", "config") + } + created, err := materializeWindowsACLTarget(target, asFile) + if err != nil { + t.Fatalf("materialize: %v", err) + } + if !created.createdAnything() { + t.Fatal("SETUP INVALID: nothing was materialized") + } + + // Another process moves the created object aside and puts an + // ordinary one of the same shape at its name. The anchor above is + // untouched, so the anchor identity check cannot see this. + aside := target + ".aside" + if err := os.Rename(target, aside); err != nil { + t.Fatalf("rename aside: %v", err) + } + if asFile { + if err := os.WriteFile(target, []byte("theirs"), 0o600); err != nil { + t.Fatal(err) + } + } else { + if err := os.Mkdir(target, 0o700); err != nil { + t.Fatal(err) + } + if err := os.WriteFile(filepath.Join(target, "keep"), []byte("theirs"), 0o600); err != nil { + t.Fatal(err) + } + } + + removed, err := rollbackWindowsACLMaterialization(created) + // The replacement is the whole point: it has to survive, whatever the + // rollback reported about itself. + if _, statErr := os.Lstat(target); statErr != nil { + t.Fatalf("rollback deleted the replacement at %s, which it never created: %v", target, statErr) + } + if !asFile { + if _, statErr := os.Lstat(filepath.Join(target, "keep")); statErr != nil { + t.Fatalf("the replacement directory's contents are gone: %v", statErr) + } + } + if err == nil { + t.Fatal("rollback left a replacement it never created without saying why") + } + if removed { + t.Fatal("rollback reported the target removed after refusing the replacement") + } + }) + } +} + +// CONTROL: the unreplaced case still unwinds completely, so the identity check +// cannot pass by refusing everything. +func TestRollbackRemovesTheChildrenItActuallyCreated(t *testing.T) { + for name, asFile := range map[string]bool{"file target": true, "directory target": false} { + t.Run(name, func(t *testing.T) { + workspace := t.TempDir() + target := filepath.Join(workspace, ".git", "hooks") + if asFile { + target = filepath.Join(workspace, ".git", "config") + } + created, err := materializeWindowsACLTarget(target, asFile) + if err != nil { + t.Fatalf("materialize: %v", err) + } + removed, err := rollbackWindowsACLMaterialization(created) + if err != nil { + t.Fatalf("rollback: %v", err) + } + if !removed { + t.Fatal("rollback did not report the target removed") + } + if _, statErr := os.Lstat(filepath.Join(workspace, ".git")); !os.IsNotExist(statErr) { + t.Fatalf(".git survived a rollback of what created it: %v", statErr) + } + }) + } +} diff --git a/internal/sandbox/windows_acl_symlink_ancestor_windows_test.go b/internal/sandbox/windows_acl_symlink_ancestor_windows_test.go new file mode 100644 index 000000000..6794a1aeb --- /dev/null +++ b/internal/sandbox/windows_acl_symlink_ancestor_windows_test.go @@ -0,0 +1,113 @@ +//go:build windows + +package sandbox + +import ( + "os" + "os/exec" + "path/filepath" + "strings" + "testing" + + "golang.org/x/sys/windows" +) + +// THE COMPARISON BASIS MUST NOT FOLLOW WHAT IT IS CHECKING FOR. +// +// verifyWindowsACLTargetNotRedirected asks GetFinalPathNameByHandle where the +// handle landed and compares that against the requested path. It used to build +// the second side with filepath.EvalSymlinks, which is the same resolution the +// kernel had just performed, so for a directory symlink the two sides agreed +// BECAUSE the redirect happened, and elevated setup rewrote the DACL of an +// object outside the workspace. +// +// Junctions were caught, but only because Go reports one as ModeIrregular and +// EvalSymlinks declines it, leaving the lexical path to disagree. That is a +// standard-library detail, not a property of the guard. + +// The property, testable with no privilege at all: the normalizer must not +// resolve a reparse point. +// +// HONEST LIMIT, because it would be easy to read more into this than it proves. +// A junction is used because creating one needs no privilege, and EvalSymlinks +// does not resolve a junction either — so this test passes against BOTH the old +// basis and the new one, and reverting the fix does not fail it. It is an +// invariant test, not a regression for this change: it pins that normalization +// leaves a reparse point alone, which is what would break first if a future Go +// started resolving mount points and silently disarmed the junction case. +// +// The case that actually separates the two bases is the directory symlink below, +// and that one skips without privilege. On this machine it skipped. +func TestACLComparablePathDoesNotResolveAReparsePoint(t *testing.T) { + base := t.TempDir() + outside := filepath.Join(base, "outside") + if err := os.MkdirAll(filepath.Join(outside, "hooks"), 0o700); err != nil { + t.Fatalf("mkdir outside: %v", err) + } + link := filepath.Join(base, "link") + if out, err := exec.Command("cmd", "/c", "mklink", "/J", link, outside).CombinedOutput(); err != nil { + t.Skipf("cannot create a junction here: %v (%s)", err, out) + } + + through := filepath.Join(link, "hooks") + comparable := windowsACLComparablePath(through) + + // Compared against the normalizer's OWN answer for each path, never against + // the raw input. GetLongPathName legitimately rewrites the string: on a CI + // runner the temp directory is an 8.3 short name, so expanding RUNNER~1 to + // runneradmin makes the result differ from what was passed in while naming + // exactly the same object. An earlier version of this test asserted equality + // with the input and failed on Windows CI for that reason, which was the test + // being wrong rather than the code. + target := windowsACLComparablePath(filepath.Join(outside, "hooks")) + + // The whole property: normalizing a path THROUGH the reparse point must not + // produce the target's normalization. If it did, the guard would be comparing + // a redirected handle against a redirected expectation and matching itself. + if strings.EqualFold(filepath.Clean(comparable), filepath.Clean(target)) { + t.Errorf("windowsACLComparablePath resolved through the reparse point:\n through %s\n target %s", comparable, target) + } + // And it still has to name the link component, or it resolved to something + // else entirely rather than leaving the path alone. + if !strings.Contains(strings.ToLower(comparable), "link") { + t.Errorf("windowsACLComparablePath lost the link component: %s", comparable) + } + + // The old basis is checked alongside for contrast: where it resolves, it + // could not have been the comparison basis. + if resolved := canonicalSandboxWorkspaceRoot(through); strings.EqualFold(filepath.Clean(resolved), filepath.Clean(filepath.Join(outside, "hooks"))) { + t.Logf("confirmed: canonicalSandboxWorkspaceRoot resolves %s to %s, which is why it cannot be the basis", through, resolved) + } +} + +// The end-to-end symlink case the finding names. Directory symlinks need either +// Developer Mode or SeCreateSymbolicLinkPrivilege, so this skips where it cannot +// build the fixture rather than passing vacuously. +func TestOpenWindowsACLTargetRefusesASymlinkAncestor(t *testing.T) { + base := t.TempDir() + outside := filepath.Join(base, "outside") + if err := os.MkdirAll(filepath.Join(outside, "hooks"), 0o700); err != nil { + t.Fatalf("mkdir outside: %v", err) + } + workspace := filepath.Join(base, "workspace") + if err := os.MkdirAll(workspace, 0o700); err != nil { + t.Fatalf("mkdir workspace: %v", err) + } + + // A DIRECTORY symlink, not a junction: os.Symlink maps to CreateSymbolicLinkW + // and the directory flag is inferred from the existing target. + link := filepath.Join(workspace, ".git") + if err := os.Symlink(outside, link); err != nil { + t.Skipf("cannot create a directory symlink here (needs Developer Mode or SeCreateSymbolicLinkPrivilege): %v", err) + } + + target := filepath.Join(link, "hooks") + handle, _, err := openWindowsACLTarget(target) + if err == nil { + _ = windows.CloseHandle(handle) + t.Fatal("elevated setup opened a target through a symlinked ancestor; its DACL change would land outside the workspace") + } + if !strings.Contains(err.Error(), "reparse point") { + t.Errorf("refusal does not name the cause: %v", err) + } +} diff --git a/internal/sandbox/windows_acl_test.go b/internal/sandbox/windows_acl_test.go index bcabedbdb..1e6568238 100644 --- a/internal/sandbox/windows_acl_test.go +++ b/internal/sandbox/windows_acl_test.go @@ -46,9 +46,14 @@ func TestBuildWindowsACLPlanForWorkspaceWriteProfile(t *testing.T) { assertWindowsACLEntry(t, plan, WindowsACLAllowWrite, `C:\workspace`, workspaceSID, false) assertWindowsACLEntry(t, plan, WindowsACLAllowWrite, `D:\cache`, cacheSID, false) - assertWindowsACLEntry(t, plan, WindowsACLDenyWrite, `C:\workspace\vendor`, workspaceSID, false) - assertWindowsACLEntry(t, plan, WindowsACLDenyWrite, `C:\workspace\.git`, workspaceSID, false) - assertWindowsACLEntry(t, plan, WindowsACLDenyWrite, `C:\workspace\.zero`, workspaceSID, false) + // Materialized, matching the principal plan. A guard attached to an object + // that does not exist yet is never applied: on a workspace with no .git at + // setup time both passes skipped these while setup still recorded success. + // Creating them is also what lets the deferred pass reach the deny-delete on + // .git, since config and hooks create .git as their parent. + assertWindowsACLEntry(t, plan, WindowsACLDenyWrite, `C:\workspace\vendor`, workspaceSID, true) + assertWindowsACLEntry(t, plan, WindowsACLDenyWrite, `C:\workspace\.git`, workspaceSID, true) + assertWindowsACLEntry(t, plan, WindowsACLDenyWrite, `C:\workspace\.zero`, workspaceSID, true) assertWindowsACLEntry(t, plan, WindowsACLDenyWrite, `C:\workspace\secret-write`, workspaceSID, false) assertWindowsACLEntry(t, plan, WindowsACLDenyWrite, `C:\workspace\secret-write`, cacheSID, false) assertWindowsACLEntry(t, plan, WindowsACLDenyRead, `C:\workspace\secret-read`, workspaceSID, true) @@ -354,3 +359,37 @@ func TestBuildWindowsACLPlanLeavesAnOutOfRootSubpathUnanchored(t *testing.T) { t.Errorf("out-of-root subpath carries anchor %q, want none: the apply would refuse a config that works today", outsideAnchor) } } + +func TestBuildWindowsACLPlanUsesReadOnlySIDWithoutWriteRoots(t *testing.T) { + home := t.TempDir() + caps, err := LoadOrCreateWindowsCapabilitySIDs(home) + if err != nil { + t.Fatalf("LoadOrCreateWindowsCapabilitySIDs: %v", err) + } + plan, err := BuildWindowsACLPlan(WindowsSandboxCommandConfig{ + SandboxHome: home, + PermissionProfile: PermissionProfile{ + FileSystem: FileSystemPolicy{ + Kind: FileSystemRestricted, + DenyRead: []string{`C:\workspace\secret-read`}, + }, + Network: NetworkPolicy{Mode: NetworkDeny}, + }, + }) + if err != nil { + t.Fatalf("BuildWindowsACLPlan: %v", err) + } + // Two deny entries, one per SID the token can carry here. ReadOnly is the + // only capability a profile with no write roots gets, and ReadAllow is the + // read grant a DenyRead profile's strict token carries: denying just one of + // them leaves the carveout readable through the other. + if len(plan.Entries) != 2 { + t.Fatalf("ACL entries = %#v, want a deny-read entry per capability SID", plan.Entries) + } + assertWindowsACLEntry(t, plan, WindowsACLDenyRead, `C:\workspace\secret-read`, caps.ReadOnly, true) + readSID, err := WindowsReadAllowSID(home) + if err != nil { + t.Fatalf("WindowsReadAllowSID: %v", err) + } + assertWindowsACLEntry(t, plan, WindowsACLDenyRead, `C:\workspace\secret-read`, readSID, true) +} diff --git a/internal/sandbox/windows_candidate_creation_windows_test.go b/internal/sandbox/windows_candidate_creation_windows_test.go new file mode 100644 index 000000000..a59655df1 --- /dev/null +++ b/internal/sandbox/windows_candidate_creation_windows_test.go @@ -0,0 +1,87 @@ +//go:build windows + +package sandbox + +import ( + "os" + "os/exec" + "path/filepath" + "strings" + "testing" +) + +// ELEVATED SETUP MUST NOT CREATE THROUGH A JUNCTION THE CALLER PLANTED. +// +// ensureWindowsSandboxRuntimeCandidates runs as Administrator, and os.MkdirAll +// follows links. Both candidate ancestries are prepared by the invoking user, so +// a junction at a not-yet-created component below the cache directory — or at the +// fallback anchor — made elevated setup build the runtime tree beneath a +// redirected, Administrator-writable target. The ACL applier's no-follow check +// runs later and spots the redirection, but the privileged create has already +// happened by then, outside its rollback boundary. +// +// A junction needs no privilege on Windows, so this reproduces the caller's half +// of the race on an ordinary unelevated box. +func TestRuntimeCandidateCreationRefusesAJunctionedComponent(t *testing.T) { + cacheRoot := t.TempDir() + target := t.TempDir() + + previous := sandboxUserCacheDir + t.Cleanup(func() { sandboxUserCacheDir = previous }) + sandboxUserCacheDir = func() (string, error) { return cacheRoot, nil } + + workspace := t.TempDir() + candidate, ok := deterministicSandboxRuntimeRoot(canonicalSandboxWorkspaceRoot(workspace), canonicalSandboxWorkspaceRoot(cacheRoot)) + if !ok { + t.Fatalf("SETUP INVALID: no deterministic candidate for %s under %s", workspace, cacheRoot) + } + // The component the caller controls: the first one Zero owns, below the + // operator's cache directory. + owned := filepath.Join(canonicalSandboxWorkspaceRoot(cacheRoot), "zero") + if !strings.HasPrefix(strings.ToLower(candidate), strings.ToLower(owned)) { + t.Fatalf("SETUP INVALID: candidate %s does not sit under the owned component %s", candidate, owned) + } + if out, err := exec.Command("cmd", "/c", "mklink", "/J", owned, target).CombinedOutput(); err != nil { + t.Fatalf("SETUP INVALID: mklink /J: %v\n%s", err, out) + } + + err := ensureRuntimeCandidateDir(candidate) + if err == nil { + t.Fatal("elevated setup created the runtime tree through a junction the caller planted") + } + + // The whole point: nothing may appear beneath the redirected target. + entries, readErr := os.ReadDir(target) + if readErr != nil { + t.Fatal(readErr) + } + if len(entries) != 0 { + names := make([]string, 0, len(entries)) + for _, entry := range entries { + names = append(names, entry.Name()) + } + t.Fatalf("setup created %v beneath the redirected target", names) + } +} + +// And an ordinary cache directory still gets its candidate, or the refusal above +// would be satisfied by a creator that refuses everything. +func TestRuntimeCandidateCreationStillCreatesAnOrdinaryTree(t *testing.T) { + cacheRoot := t.TempDir() + previous := sandboxUserCacheDir + t.Cleanup(func() { sandboxUserCacheDir = previous }) + sandboxUserCacheDir = func() (string, error) { return cacheRoot, nil } + + workspace := t.TempDir() + candidate, ok := deterministicSandboxRuntimeRoot(canonicalSandboxWorkspaceRoot(workspace), canonicalSandboxWorkspaceRoot(cacheRoot)) + if !ok { + t.Fatalf("SETUP INVALID: no deterministic candidate for %s", workspace) + } + if err := ensureRuntimeCandidateDir(candidate); err != nil { + t.Fatalf("an ordinary cache tree was refused: %v", err) + } + info, err := os.Stat(candidate) + if err != nil || !info.IsDir() { + t.Fatalf("the candidate was not created: err=%v", err) + } +} diff --git a/internal/sandbox/windows_command_runner_windows.go b/internal/sandbox/windows_command_runner_windows.go index de8eebfea..e7373965d 100644 --- a/internal/sandbox/windows_command_runner_windows.go +++ b/internal/sandbox/windows_command_runner_windows.go @@ -3,8 +3,11 @@ package sandbox import ( + "errors" "fmt" "io" + "os" + "strings" ) func runWindowsSandboxCommand(config WindowsSandboxCommandConfig, stderr io.Writer) int { @@ -23,6 +26,15 @@ func runWindowsSandboxCommand(config WindowsSandboxCommandConfig, stderr io.Writ fmt.Fprintln(stderr, WindowsSandboxCommandRunnerName+": "+err.Error()) return 1 } + // The marker attests to a PLAN. This attests to the OBJECT that plan was + // applied to, which cleanup can reclaim and an ordinary run then recreates + // without the capability ACE. Checked before the token is minted, so the + // operator is told to rerun setup instead of watching every sandboxed write + // fail with a bare access-denied. + if err := verifyWindowsRuntimeRootCapability(config); err != nil { + fmt.Fprintln(stderr, WindowsSandboxCommandRunnerName+": "+err.Error()) + return 1 + } case WindowsSandboxLevelUnelevated: if err := ensureWindowsUnelevatedSetup(config); err != nil { fmt.Fprintln(stderr, WindowsSandboxCommandRunnerName+": "+err.Error()) @@ -84,6 +96,115 @@ func runWindowsSandboxCommand(config WindowsSandboxCommandConfig, stderr io.Writ // DenyRead is rejected above for both runner levels rather than launching a // fully restricted narrow-SID token that cannot execute normal tools. writeRestricted := len(config.PermissionProfile.FileSystem.DenyRead) == 0 + + // A provisioned sandbox principal replaces the restricted token entirely: it + // is a separate account, so reads outside its granted roots are denied by the + // filesystem rather than left open the way a same-user restricted token has + // to leave them (#662). Absent, unprovisioned or opted-out, ok is false and + // the restricted-token backend below runs exactly as before. + principalToken, ok, err := windowsSandboxPrincipalToken(config) + if err != nil { + // The one path here that does not fall back, because a provisioned but + // unusable principal means the sandbox is broken rather than absent. Say + // how to get out of it, since the whole backend is opt-in. + fmt.Fprintf(stderr, "%s: sandbox principal is provisioned but unusable: %v. Re-run `zero sandbox setup` from an elevated terminal, or unset %s to fall back to the restricted-token sandbox.\n", + WindowsSandboxCommandRunnerName, err, windowsSandboxIdentityEnv) + return 1 + } + if ok { + defer principalToken.Close() + // The principal gets its own identity AND the write jail, not one or the + // other. Its ACEs confine reads; without the restricted token it would + // still hold every write its ambient memberships grant, so a profile + // permitting writes only to the workspace could still write anywhere + // BATCH or BUILTIN\Users may — C:\Users\Public\Documents, for one. + // + // The principal's own SID is deliberately NOT a restricting SID, and that + // is the whole of the write jail on this path. + // + // WRITE_RESTRICTED allows a write only when BOTH the normal token and the + // restricting-SID list allow it. The account SID is already enabled in the + // normal token, so listing it as a restricting SID makes the second check + // a formality for anything granted to that account: every path carrying a + // direct principal ACE passes both halves and is writable wherever it + // sits. The principal's own profile directory, which Windows creates on + // first logon with exactly such an ACE, is outside every configured write + // root and was writable for that reason. + // + // Same defect as the World SID in the restricted list (#865), with the + // account SID in place of Everyone: a SID already carried by the normal + // token cannot also serve as the restriction on it. + // + // Confining to the capability SIDs alone restores the intersection the + // jail is supposed to be. A configured write root carries BOTH a principal + // ACE, from the principal plan, satisfying the normal token, AND a + // capability ACE, from BuildWindowsACLPlan which setup applies on every + // path, satisfying the restriction. So the principal keeps the tree it + // owns, while anything holding only a principal ACE now fails the + // restricted check. + principalUser, err := principalToken.GetTokenUser() + if err != nil { + fmt.Fprintln(stderr, WindowsSandboxCommandRunnerName+": read sandbox principal SID: "+err.Error()) + return 1 + } + jailSIDs := windowsPrincipalJailSIDs(tokenSIDs, principalUser.User.Sid.String()) + // A strict token restricts READS too, and the account SID that carries the + // read roots was just removed above. Without the read capability the jail + // stops being a write jail and becomes a total one: no read root, no + // executable, nothing. BuildWindowsACLPlan grants this same SID on every + // read root and denies it on every DenyRead path, so the allow-list and the + // restriction stay one decision. + if !writeRestricted { + readSID, err := WindowsReadAllowSID(config.SandboxHome) + if err != nil { + fmt.Fprintln(stderr, WindowsSandboxCommandRunnerName+": resolve sandbox read capability: "+err.Error()) + return 1 + } + jailSIDs = append(jailSIDs, readSID) + } + jailedToken, err := restrictWindowsTokenForCapabilitySIDs(principalToken, jailSIDs, writeRestricted) + if err != nil { + fmt.Fprintln(stderr, WindowsSandboxCommandRunnerName+": "+err.Error()) + return 1 + } + defer jailedToken.Close() + // CreateProcessAsUser below is about to be handed a token for a DIFFERENT + // account, which forfeits the own-token exemption the ordinary restricted + // path relies on. Enable what that needs, and refuse with a specific reason + // when this process cannot, rather than letting an unelevated run surface a + // bare "Access is denied" that reads as the command being rejected. + // + // Refusing rather than falling back: an operator who set the opt-in to + // confine reads must not be handed the same-user restricted token while + // believing otherwise. + if err := enableWindowsPrincipalLaunchPrivileges(); err != nil { + fmt.Fprintln(stderr, WindowsSandboxCommandRunnerName+": "+err.Error()) + return 1 + } + // Stop the child describing the CALLER. Everything identifying the account + // survived from the invoking user's environment, so a principal command + // resolved its per-user state through paths inside a profile it cannot + // open. Applied here, on the principal path only, because this is the first + // point that knows which account the command is about to run as. + // The role has to be the one the token above was minted for, or the child + // is told it is the other account and resolves its per-user state under a + // profile it cannot open. + config.Env = windowsPrincipalIdentityEnvironment( + config.Env, + windowsSandboxUserName( + windowsSandboxPrincipalKey(config), + windowsSandboxRoleForNetwork(config.PermissionProfile.Network.Mode), + ), + config.PermissionProfile.Runtime, + ) + exitCode, err := runWindowsCommandAsUser(jailedToken, config) + if err != nil { + fmt.Fprintln(stderr, WindowsSandboxCommandRunnerName+": "+err.Error()) + return 1 + } + return exitCode + } + token, err := createWindowsRestrictedTokenForCapabilitySIDs(tokenSIDs, writeRestricted) if err != nil { fmt.Fprintln(stderr, WindowsSandboxCommandRunnerName+": "+err.Error()) @@ -117,12 +238,127 @@ func ensureWindowsUnelevatedSetup(config WindowsSandboxCommandConfig) error { if err != nil { return err } - if marker.contains(applied) { + // THE MARKER ATTESTS TO A PLAN. THE COMMAND CONSUMES AN OBJECT. + // + // The marker fingerprints pathnames and actions, and cleanup is allowed to + // reclaim the deterministic runtime directory that plan was realized on. An + // ordinary later run recreates the same pathname with the caller-private DACL + // and no capability ACE, and the serialized plan is unchanged, so this fast + // path returned before anything looked at the directory. The restricted child + // still carried the capability SID; the new object simply did not grant it, and + // every temp, package-cache and build-cache write failed after launch with a + // bare ACCESS_DENIED and nothing pointing at setup. + // + // This tier owns its plan and needs no elevation, so the answer here is to + // reapply rather than to refuse. Refusing would print advice to run elevated + // setup, which is unnecessary on this tier and unfollowable for a user with no + // Administrator account. The restricted-token tier keeps its refusal, because + // an ordinary user cannot restore principal provisioning. + if marker.contains(applied) && verifyWindowsRuntimeRootCapability(config) == nil { return nil } + // THE TIER BOUNDARY, STATED BEFORE ANY MUTATION IS ATTEMPTED. + // + // One plan is consumed by two tiers with different authority. A profile that + // carries DenyRead runs on a strict token, and the strict token applies the + // restricted-SID check to READS, so the read capability has to be granted + // wherever the command reads from -- including the volume root that + // permissionProfileReadRoots seeds. Elevated setup can write that DACL. An + // ordinary user cannot, and the common opener asks for WRITE_DAC on every + // entry, so this tier fails on that one root every time. + // + // Dropping the root ACE instead is not an option: the strict token would then + // fail its own read check for the executable and every ambient read. So the + // tier is refused explicitly, naming the root and the reason, rather than + // discovered as an ACCESS_DENIED after the fact. The real smoke test misses + // this because it substitutes a user-owned temporary directory for the + // production read root. + if refusal := WindowsACLPlanReadGrantRefusal(plan); refusal != "" { + return errors.New("unelevated sandbox setup cannot apply this plan: " + refusal) + } if _, err := applyWindowsACLPlan(plan); err != nil { + // Refusing to run is right: without these ACEs the write jail does not + // exist, so continuing would run the command believing it is sandboxed + // when it is not. What was wrong was the diagnosis. Every failure got the + // same "the workspace may be on a filesystem you do not own" guess, and + // the suggested remedy was elevated setup, which does not help at all + // when the real problem is one root in the plan that nobody can ACL. + // + // Being precise matters because this failure repeats: the success marker + // is only recorded on success, so the same plan fails identically on + // every later command until the offending root leaves it. A reader who + // cannot tell which root is at fault has no way out of that. + // Every remedy named below is one the reader can actually carry out. + // Both messages used to offer `--sandbox forbid`, which is not an option + // at all: SandboxPreferenceForbid is an internal engine state with no flag + // behind it, so acting on it produced an unknown option and left the reader + // stuck on the failure they had just been told how to clear. Advice that + // does not work costs more than none, since finding that out takes time. + if denied := windowsACLPlanDeniedPath(err); denied != "" { + return fmt.Errorf("apply unelevated workspace ACLs: %w; %s cannot have its permissions changed by this user, "+ + "so the sandbox cannot enforce a write boundary there and will not run the command. "+ + "That path is one of this workspace's sandbox roots, usually a system directory that arrived via TEMP or TMP. "+ + "Check those, or turn the sandbox off in your user config with "+ + `"sandbox": {"enabled": false}. `+ + "Running `zero sandbox setup` elevated will NOT fix this", err, denied) + } return fmt.Errorf("apply unelevated workspace ACLs: %w — the workspace may be on a filesystem the current user does not own; "+ - "run `zero sandbox setup` from an elevated (Administrator) terminal, or re-run with `--sandbox forbid` to skip OS sandboxing", err) + "run `zero sandbox setup` from an elevated (Administrator) terminal, "+ + `or turn the sandbox off in your user config with "sandbox": {"enabled": false}`, err) } return recordWindowsUnelevatedAppliedPlan(config.SandboxHome, applied) } + +// windowsACLPlanDeniedPath pulls the target path out of an apply failure that +// was an access denial, and returns "" for anything else. +// +// applyWindowsACLPathGroup already wraps the path into its error, so this reads +// the message rather than threading a typed error through four layers for one +// diagnostic. The string it matches is produced in the same package by +// openWindowsACLTarget, and a test pins the pairing so the two cannot drift +// apart silently. +func windowsACLPlanDeniedPath(err error) string { + if err == nil || !errors.Is(err, os.ErrPermission) { + return "" + } + const marker = "open windows ACL target " + message := err.Error() + start := strings.Index(message, marker) + if start < 0 { + return "" + } + // Colon-SPACE, not colon. The wrapper is "...target %s: %w", and on Windows + // the path itself starts with a drive colon, so splitting on the first colon + // returns "C". A drive colon is always followed by a separator, never a + // space, which makes ": " the only unambiguous boundary here. + rest := message[start+len(marker):] + end := strings.Index(rest, ": ") + if end <= 0 { + return "" + } + return strings.TrimSpace(rest[:end]) +} + +// windowsPrincipalJailSIDs returns the restricting SIDs for a principal command. +// +// A named function with its own test rather than an inline slice literal, +// because the defect it encodes was invisible: the production composition added +// the account's own SID while the only jail test granted a GROUP, so the +// configuration that shipped was never the configuration under test. +// +// accountSID is taken and DELIBERATELY EXCLUDED rather than simply not passed. +// Naming it here makes the exclusion the function's stated contract instead of an +// omission a later edit could undo without noticing, and it also removes the SID +// should it ever arrive through capabilitySIDs, which is the same route the World +// SID took in #865. +func windowsPrincipalJailSIDs(capabilitySIDs []string, accountSID string) []string { + account := strings.TrimSpace(accountSID) + jail := make([]string, 0, len(capabilitySIDs)) + for _, sid := range capabilitySIDs { + if account != "" && strings.EqualFold(strings.TrimSpace(sid), account) { + continue + } + jail = append(jail, sid) + } + return jail +} diff --git a/internal/sandbox/windows_dualrole_rollback_windows_test.go b/internal/sandbox/windows_dualrole_rollback_windows_test.go new file mode 100644 index 000000000..cb07a95ce --- /dev/null +++ b/internal/sandbox/windows_dualrole_rollback_windows_test.go @@ -0,0 +1,236 @@ +//go:build windows + +package sandbox + +import ( + "errors" + "os" + "path/filepath" + "testing" + "time" + + "golang.org/x/sys/windows" +) + +// The outer rollback must only delete principals this run created. +// +// Dual-role setup provisions offline then online. If the offline role adopts an +// account that already existed and anything afterwards fails, the outer rollback +// used to delete it anyway, turning a partial re-setup failure into the loss of +// a principal that was working before the run started. Two roles make that the +// common case rather than an unlucky one: the first role usually succeeds, so +// there is nearly always something for a later failure to destroy. +func TestDualRoleSetupRollbackSparesAdoptedPrincipals(t *testing.T) { + for name, testCase := range map[string]struct { + createdByRole map[windowsSandboxRole]bool + wantRemoved []windowsSandboxRole + }{ + "offline adopted, online created": { + createdByRole: map[windowsSandboxRole]bool{ + windowsSandboxRoleOffline: false, + windowsSandboxRoleOnline: true, + }, + // Only the one this run made. Deleting the adopted offline principal + // is the data loss this guards against. + wantRemoved: []windowsSandboxRole{windowsSandboxRoleOnline}, + }, + "both adopted": { + createdByRole: map[windowsSandboxRole]bool{ + windowsSandboxRoleOffline: false, + windowsSandboxRoleOnline: false, + }, + wantRemoved: nil, + }, + "both created": { + createdByRole: map[windowsSandboxRole]bool{ + windowsSandboxRoleOffline: true, + windowsSandboxRoleOnline: true, + }, + wantRemoved: []windowsSandboxRole{windowsSandboxRoleOffline, windowsSandboxRoleOnline}, + }, + } { + t.Run(name, func(t *testing.T) { + var removed []windowsSandboxRole + restoreDualRoleSeams(t, testCase.createdByRole, &removed) + + // Fail on the SECOND role, not the first. ACL application happens per + // role inside the loop, so failing the first aborts before the second + // is provisioned at all and the rollback never has more than one + // principal to consider. The case worth testing is the one review + // described: the first role succeeds, the second fails, and the + // question is whether the first gets destroyed on the way out. + // Count GRANT plans, not ACL calls. applyWindowsPrincipalACLs revokes + // this trustee's existing ACEs before applying the new plan, so there + // are two calls per role now. The contract under test is "the second + // ROLE fails"; counting raw calls would fail the first role's grant + // instead and prove something else entirely. + grants := 0 + applyWindowsACLPlanFn = func(plan WindowsACLPlan) (func() error, error) { + if len(plan.Entries) > 0 && plan.Entries[0].Action == windowsACLRevoke { + return func() error { return nil }, nil + } + grants++ + if grants < 2 { + return func() error { return nil }, nil + } + return nil, errors.New("ACL apply refused") + } + + if _, err := setupWindowsSandboxPrincipal(windowsSandboxTestConfig(t)); err == nil { + t.Fatal("setup reported success despite an injected ACL failure") + } + assertRolesEqual(t, removed, testCase.wantRemoved) + }) + } +} + +// The same contract on the success path: the rollback handed back to the caller +// for a LATER setup step to invoke must be just as reluctant. +func TestDualRoleSetupReturnedRollbackSparesAdoptedPrincipals(t *testing.T) { + var removed []windowsSandboxRole + restoreDualRoleSeams(t, map[windowsSandboxRole]bool{ + windowsSandboxRoleOffline: false, + windowsSandboxRoleOnline: true, + }, &removed) + + rollback, err := setupWindowsSandboxPrincipal(windowsSandboxTestConfig(t)) + if err != nil { + t.Fatalf("setup: %v", err) + } + if len(removed) != 0 { + t.Fatalf("setup removed principals before anything failed: %v", removed) + } + if err := rollback(); err != nil { + t.Fatalf("rollback: %v", err) + } + assertRolesEqual(t, removed, []windowsSandboxRole{windowsSandboxRoleOnline}) +} + +func restoreDualRoleSeams(t *testing.T, createdByRole map[windowsSandboxRole]bool, removed *[]windowsSandboxRole) { + t.Helper() + prevProvision := provisionWindowsSandboxPrincipalForSetupFn + prevApply := applyWindowsACLPlanFn + prevRemove := removeWindowsSandboxPrincipalForSetupFn + t.Cleanup(func() { + provisionWindowsSandboxPrincipalForSetupFn = prevProvision + applyWindowsACLPlanFn = prevApply + removeWindowsSandboxPrincipalForSetupFn = prevRemove + }) + + provisionWindowsSandboxPrincipalForSetupFn = func(_ WindowsSandboxCommandConfig, role windowsSandboxRole) (windowsSandboxIdentity, bool, error) { + sid, err := windows.CreateWellKnownSid(windows.WinLocalSystemSid) + if err != nil { + return windowsSandboxIdentity{}, false, err + } + return windowsSandboxIdentity{Username: "zero-sbx-test", SID: sid}, createdByRole[role], nil + } + applyWindowsACLPlanFn = func(WindowsACLPlan) (func() error, error) { + return func() error { return nil }, nil + } + removeWindowsSandboxPrincipalForSetupFn = func(_ WindowsSandboxCommandConfig, role windowsSandboxRole) error { + *removed = append(*removed, role) + return nil + } +} + +// windowsSandboxTestConfig hands the production setup path test-owned roots and +// nothing else. The principal and ACL seams intercept the identity work, not the +// filesystem work: setupWindowsSandboxPrincipal materializes runtime candidates +// under the user cache and TEMP, and applyWindowsPrincipalACLs writes ledgers +// below the sandbox home, so the fixed C:\sandboxhome and C:\ws these tests used +// to name meant an ordinary run wrote to the live cache, failed on drive-root +// permissions, or collided with host state. Everything setup touches is +// redirected to t.TempDir, and windowsSandboxTestRootsUntouched proves the +// redirect took. Reported by @gnanam1990. +func windowsSandboxTestConfig(t *testing.T) WindowsSandboxCommandConfig { + t.Helper() + home := filepath.Join(t.TempDir(), "sandboxhome") + workspace := filepath.Join(t.TempDir(), "ws") + if err := os.MkdirAll(workspace, 0o700); err != nil { + t.Fatal(err) + } + cache := t.TempDir() + prevCache := sandboxUserCacheDir + sandboxUserCacheDir = func() (string, error) { return cache, nil } + t.Cleanup(func() { sandboxUserCacheDir = prevCache }) + temp := t.TempDir() + t.Setenv("TMP", temp) + t.Setenv("TEMP", temp) + before := windowsFixedTestRootsState() + t.Cleanup(func() { windowsSandboxTestRootsUntouched(t, cache, before) }) + return WindowsSandboxCommandConfig{ + SandboxHome: home, + CommandCWD: workspace, + WorkspaceRoots: []string{workspace}, + PermissionProfile: PermissionProfile{ + FileSystem: FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{Root: workspace}}, + }, + }, + } +} + +// windowsFixedRootState is what a run may not change about a path outside the +// test-owned roots: whether it exists, and if it does, when its entries last +// changed. A directory's modification time moves whenever something is created +// or removed inside it, which is exactly what setup would do to a home it was +// still writing ledgers into. +type windowsFixedRootState struct { + exists bool + modTime time.Time +} + +var windowsFixedTestRoots = []string{`C:\sandboxhome`, `C:\ws`} + +func windowsFixedTestRootsState() map[string]windowsFixedRootState { + states := make(map[string]windowsFixedRootState, len(windowsFixedTestRoots)) + for _, fixed := range windowsFixedTestRoots { + info, err := os.Stat(fixed) + if err != nil { + states[fixed] = windowsFixedRootState{} + continue + } + states[fixed] = windowsFixedRootState{exists: true, modTime: info.ModTime()} + } + return states +} + +// windowsSandboxTestRootsUntouched asserts the production setup path stayed +// inside the test-owned roots. Not "the fixed paths are absent": a box that ran +// the old form of these tests still has C:\sandboxhome from them, which is the +// host state the finding was about, so the assertion is that THIS run neither +// created nor changed them. And the cache the run was redirected to must not be +// the live one, or the redirect proved nothing. +func windowsSandboxTestRootsUntouched(t *testing.T, cache string, before map[string]windowsFixedRootState) { + t.Helper() + after := windowsFixedTestRootsState() + for _, fixed := range windowsFixedTestRoots { + was, now := before[fixed], after[fixed] + switch { + case !was.exists && now.exists: + t.Errorf("setup under test created %s, a path outside the test-owned roots", fixed) + case was.exists && now.exists && !now.modTime.Equal(was.modTime): + t.Errorf("setup under test wrote into %s, a path outside the test-owned roots", fixed) + } + } + if live, err := os.UserCacheDir(); err == nil && live == cache { + t.Fatalf("SETUP INVALID: the redirected cache %s is the live one", cache) + } +} + +func assertRolesEqual(t *testing.T, got []windowsSandboxRole, want []windowsSandboxRole) { + t.Helper() + if len(got) != len(want) { + t.Fatalf("removed %v, want %v", got, want) + } + seen := map[windowsSandboxRole]bool{} + for _, role := range got { + seen[role] = true + } + for _, role := range want { + if !seen[role] { + t.Fatalf("removed %v, want %v", got, want) + } + } +} diff --git a/internal/sandbox/windows_fresh_workspace_git_guard_windows_test.go b/internal/sandbox/windows_fresh_workspace_git_guard_windows_test.go new file mode 100644 index 000000000..c9e1fdf35 --- /dev/null +++ b/internal/sandbox/windows_fresh_workspace_git_guard_windows_test.go @@ -0,0 +1,172 @@ +//go:build windows + +package sandbox + +import ( + "os" + "os/exec" + "path/filepath" + "testing" + "unsafe" + + "golang.org/x/sys/windows" +) + +// appliedDenyMask returns the DENY mask an applied DACL carries for sid, read +// back off the object rather than asserted on the plan, because the plan is what +// was wrong. +func appliedDenyMask(t *testing.T, path string, sid string) uint32 { + t.Helper() + handle, _, err := openWindowsACLTarget(path) + if err != nil { + t.Fatalf("open %s: %v", path, err) + } + defer func() { _ = windows.CloseHandle(handle) }() + descriptor, err := windows.GetSecurityInfo(handle, windows.SE_FILE_OBJECT, windows.DACL_SECURITY_INFORMATION) + if err != nil { + t.Fatalf("read the DACL of %s: %v", path, err) + } + dacl, _, err := descriptor.DACL() + if err != nil || dacl == nil { + t.Fatalf("read the DACL of %s: %v", path, err) + } + wanted, err := windows.StringToSid(sid) + if err != nil { + t.Fatalf("parse %q: %v", sid, err) + } + var mask uint32 + // Fixed-layout ACE types only; an object ACE puts GUIDs ahead of the trustee + // so SidStart would land mid-structure. Nothing here builds one, and an + // unrecognised type is skipped rather than decoded into a nonsense SID. + for index := uint32(0); index < uint32(dacl.AceCount); index++ { + var ace *windows.ACCESS_ALLOWED_ACE + if err := windows.GetAce(dacl, index, &ace); err != nil { + continue + } + if ace.Header.AceType != windows.ACCESS_DENIED_ACE_TYPE { + continue + } + if !(*windows.SID)(unsafe.Pointer(&ace.SidStart)).Equals(wanted) { + continue + } + mask |= uint32(ace.Mask) + } + return mask +} + +// A GUARD ON AN OBJECT THAT DOES NOT EXIST YET IS NOT A GUARD. +// +// The capability planner emitted the deny-delete on .git and the deny-writes on +// .git\config and .git\hooks without Materialize. On a workspace that had no +// .git when setup ran, the first apply pass skipped all three as missing and the +// deferred pass skipped them again, because nothing else in the capability plan +// created them. Setup recorded success anyway. +// +// A later git init then created .git, config and hooks beneath the +// already-granted workspace. They inherit the workspace allow, which carries +// DELETE, with no object-specific deny of their own, so the sandboxed command +// could rename .git aside, recreate it, and get credential.helper and +// core.hooksPath back. +// +// This is the DEFAULT backend, so the weaker of the two planners' rules was the +// one almost every Windows user got. The principal plan was never affected: its +// carveouts are materialized, which both applies them and creates .git as their +// parent so the deferred retry lands the deny-delete. +// +// Read back off the applied object, and driven through the same +// BuildWindowsACLPlan + applyWindowsACLPlan pair both setup tiers use. DACL edits +// on a user-owned directory need no Administrator rights. +func TestCapabilityGitGuardReachesAWorkspaceThatGetsGitLater(t *testing.T) { + workspace := t.TempDir() + gitDir := filepath.Join(workspace, ".git") + + // SETUP: no .git at setup time. That is the entire case. + if _, err := os.Stat(gitDir); err == nil { + t.Fatal("SETUP INVALID: .git already exists, so this is the covered case, not the uncovered one") + } + + config := WindowsSandboxCommandConfig{ + SandboxHome: t.TempDir(), + CommandCWD: workspace, + WorkspaceRoots: []string{workspace}, + PermissionProfile: PermissionProfile{ + FileSystem: FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{ + Root: workspace, + ReadOnlySubpaths: gitMetadataWriteCarveouts(workspace), + ProtectedMetadataNames: sandboxFullyProtectedMetadataNames, + }}, + }, + }, + } + plan, err := BuildWindowsACLPlan(config) + if err != nil { + t.Fatalf("BuildWindowsACLPlan: %v", err) + } + sid, err := windowsCapabilitySIDForWriteRoot(config, workspace) + if err != nil { + t.Fatalf("resolve the workspace capability SID: %v", err) + } + if _, err := applyWindowsACLPlan(plan); err != nil { + t.Skipf("cannot apply an ACL plan here: %v", err) + } + + // The workspace grant really landed, so a missing deny below is a missing + // deny and not an apply that did nothing. + if appliedAllowMask(t, workspace, sid) == 0 { + t.Fatal("SETUP INVALID: the workspace carries no allow ACE for the capability SID, so the plan did not apply") + } + + // Now git arrives, the way it does for a scaffold or a clone after setup. + if out, err := exec.Command("git", "init", workspace).CombinedOutput(); err != nil { + t.Skipf("git init unavailable here: %v\n%s", err, out) + } + + if mask := appliedDenyMask(t, gitDir, sid); mask&uint32(windows.DELETE) == 0 { + t.Fatalf(".git carries no deny-DELETE for the capability SID after git init (mask=%#x); the sandboxed token can rename it aside and recreate it without the config and hooks carveouts", mask) + } + hooks := filepath.Join(gitDir, "hooks") + if _, err := os.Stat(hooks); err == nil { + if mask := appliedDenyMask(t, hooks, sid); mask == 0 { + t.Error(".git\\hooks carries no deny for the capability SID, so core.hooksPath is writable from inside the sandbox") + } + } +} + +// appliedAllowMask is the allow-side companion, used only as a setup guard. +func appliedAllowMask(t *testing.T, path string, sid string) uint32 { + t.Helper() + handle, _, err := openWindowsACLTarget(path) + if err != nil { + t.Fatalf("open %s: %v", path, err) + } + defer func() { _ = windows.CloseHandle(handle) }() + descriptor, err := windows.GetSecurityInfo(handle, windows.SE_FILE_OBJECT, windows.DACL_SECURITY_INFORMATION) + if err != nil { + t.Fatalf("read the DACL of %s: %v", path, err) + } + dacl, _, err := descriptor.DACL() + if err != nil || dacl == nil { + t.Fatalf("read the DACL of %s: %v", path, err) + } + wanted, err := windows.StringToSid(sid) + if err != nil { + t.Fatalf("parse %q: %v", sid, err) + } + var mask uint32 + for index := uint32(0); index < uint32(dacl.AceCount); index++ { + var ace *windows.ACCESS_ALLOWED_ACE + if err := windows.GetAce(dacl, index, &ace); err != nil { + continue + } + if ace.Header.AceType != windows.ACCESS_ALLOWED_ACE_TYPE { + continue + } + if !(*windows.SID)(unsafe.Pointer(&ace.SidStart)).Equals(wanted) { + continue + } + mask |= uint32(ace.Mask) + } + return mask +} diff --git a/internal/sandbox/windows_git_carveout_windows_test.go b/internal/sandbox/windows_git_carveout_windows_test.go new file mode 100644 index 000000000..2f444f06c --- /dev/null +++ b/internal/sandbox/windows_git_carveout_windows_test.go @@ -0,0 +1,66 @@ +//go:build windows + +package sandbox + +import ( + "os" + "os/exec" + "path/filepath" + "testing" +) + +// The git control-plane carveouts are two different shapes: .git/hooks is a +// directory, .git/config is a FILE. Materialization has to respect that. +// +// On a fresh workspace neither exists yet, which is precisely the case +// Materialize was added for — so this is the common path, not a corner. Creating +// .git/config as a directory does not just mis-ACL it: it makes the workspace +// permanently unusable, because git refuses to initialise over a directory +// where its config file belongs. +func TestPrincipalACLPlanMaterializesGitConfigAsFile(t *testing.T) { + workspace := t.TempDir() + + plan, err := buildWindowsPrincipalACLPlan(windowsPrincipalACLInput{ + // Guests, deliberately: the deny-write ACE must land on the sandbox + // principal, not on whoever runs the test. With Everyone (S-1-1-0) the + // ACE denies the test process too and `git init` fails with "Permission + // denied" for a reason that has nothing to do with the shape bug. + PrincipalSID: "S-1-5-32-546", + WriteRoots: []WritableRoot{{ + Root: workspace, + ReadOnlySubpaths: gitMetadataWriteCarveouts(workspace), + }}, + }) + if err != nil { + t.Fatalf("buildWindowsPrincipalACLPlan: %v", err) + } + + rollback, err := applyWindowsACLPlan(plan) + if err != nil { + t.Fatalf("applyWindowsACLPlan: %v", err) + } + t.Cleanup(func() { + if rollback != nil { + _ = rollback() + } + }) + + configPath := filepath.Join(workspace, ".git", "config") + if info, err := os.Stat(configPath); err == nil && info.IsDir() { + t.Errorf(".git/config was materialized as a directory; git requires a file") + } + hooksPath := filepath.Join(workspace, ".git", "hooks") + if info, err := os.Stat(hooksPath); err == nil && !info.IsDir() { + t.Errorf(".git/hooks was materialized as a file; git requires a directory") + } + + // The failure users would actually hit. + if _, err := exec.LookPath("git"); err != nil { + t.Skip("git not on PATH; shape assertions above still ran") + } + cmd := exec.Command("git", "init") + cmd.Dir = workspace + if out, err := cmd.CombinedOutput(); err != nil { + t.Errorf("git init failed on a workspace after sandbox setup: %v\n%s", err, out) + } +} diff --git a/internal/sandbox/windows_git_rename_guard_apply_windows_test.go b/internal/sandbox/windows_git_rename_guard_apply_windows_test.go new file mode 100644 index 000000000..3dcfd6202 --- /dev/null +++ b/internal/sandbox/windows_git_rename_guard_apply_windows_test.go @@ -0,0 +1,145 @@ +//go:build windows + +package sandbox + +import ( + "os" + "path/filepath" + "strconv" + "strings" + "testing" + + "golang.org/x/sys/windows" +) + +// windowsPathDeniesDelete reports whether the object's real DACL carries a deny +// ACE covering DELETE for the given SID. +// +// The plan-shape tests assert what the planner emits. This reads what actually +// landed on disk, which is the gap that let the guard go missing: the plan was +// right the whole time and the apply silently dropped it. +func windowsPathDeniesDelete(t *testing.T, path, sid string) bool { + t.Helper() + handle, _, err := openWindowsACLTarget(path) + if err != nil { + t.Fatalf("open %s: %v", path, err) + } + defer func() { _ = windows.CloseHandle(handle) }() + + descriptor, err := windows.GetSecurityInfo(handle, windows.SE_FILE_OBJECT, windows.DACL_SECURITY_INFORMATION) + if err != nil { + t.Fatalf("read DACL for %s: %v", path, err) + } + // SDDL rather than walking the ACE buffer by hand: this x/sys does not export + // an ACE enumerator, and unsafe pointer arithmetic in a test that exists to + // catch a security regression is its own hazard. + // + // An ACE renders as (type;flags;rights;guid;inherit_guid;sid), so the deny + // entries are the ones whose type field is D. Rights come back as SDDL + // abbreviations when they fit and as a hex mask when they do not, so both are + // accepted; SD is the abbreviation for DELETE. + for _, ace := range strings.Split(descriptor.String(), "(") { + fields := strings.Split(strings.TrimSuffix(strings.TrimSpace(ace), ")"), ";") + if len(fields) != 6 || fields[0] != "D" || !strings.EqualFold(fields[5], sid) { + continue + } + rights := fields[2] + if strings.Contains(rights, "SD") { + return true + } + if mask, err := strconv.ParseUint(strings.TrimPrefix(strings.ToLower(rights), "0x"), 16, 32); err == nil { + if uint32(mask)&uint32(windows.DELETE) != 0 { + return true + } + } + } + return false +} + +// THE GUARD MUST REACH DISK ON A WORKSPACE THAT HAD NO .git. +// +// This is the case the whole plan is built around and it was the one case where +// the guard was absent. .git deliberately carries no Materialize, because an +// empty .git breaks `git init`. Groups are applied in ascending path order, so +// the .git group ran while .git did not yet exist, was skipped as a +// non-materializing missing target, and nothing revisited it once the +// .git\config carveout created .git as its parent chain moments later. +// +// Four tests already covered the deny-delete mask and the plan emitting it. +// None of them applied the plan and looked at the object, which is exactly why +// a green suite said nothing. +func TestGitRenameGuardReachesDiskOnAWorkspaceWithoutGit(t *testing.T) { + const principal = testPrincipalSID // Unaliased, so it renders literally in SDDL. + workspace := t.TempDir() + gitDir := filepath.Join(workspace, ".git") + if _, err := os.Stat(gitDir); err == nil { + t.Fatal("the workspace already has .git, so this proves nothing") + } + + plan, err := buildWindowsPrincipalACLPlan(windowsPrincipalACLInput{ + PrincipalSID: principal, + WriteRoots: []WritableRoot{{ + Root: workspace, + ProtectedMetadataNames: sandboxFullyProtectedMetadataNames, + ReadOnlySubpaths: gitMetadataWriteCarveouts(workspace), + }}, + }) + if err != nil { + t.Fatalf("buildWindowsPrincipalACLPlan: %v", err) + } + + rollback, err := applyWindowsACLPlan(plan) + if err != nil { + t.Fatalf("applyWindowsACLPlan: %v", err) + } + t.Cleanup(func() { + if rollback != nil { + _ = rollback() + } + }) + + // The carveouts create .git on the way to .git\config. + if _, err := os.Stat(gitDir); err != nil { + t.Fatalf(".git was never created by the carveout materialization: %v", err) + } + if !windowsPathDeniesDelete(t, gitDir, principal) { + t.Fatal("no deny-delete ACE on .git after applying the plan: the principal can rename it aside and recreate it without the config and hooks carveouts") + } +} + +// The same guard on a workspace that already had .git must keep working. This +// case was already correct, and it is kept so a fix aimed at the fresh +// workspace cannot quietly trade one for the other. +func TestGitRenameGuardStillReachesDiskWhenGitAlreadyExists(t *testing.T) { + const principal = testPrincipalSID + workspace := t.TempDir() + gitDir := filepath.Join(workspace, ".git") + if err := os.Mkdir(gitDir, 0o700); err != nil { + t.Fatalf("seed .git: %v", err) + } + + plan, err := buildWindowsPrincipalACLPlan(windowsPrincipalACLInput{ + PrincipalSID: principal, + WriteRoots: []WritableRoot{{ + Root: workspace, + ProtectedMetadataNames: sandboxFullyProtectedMetadataNames, + ReadOnlySubpaths: gitMetadataWriteCarveouts(workspace), + }}, + }) + if err != nil { + t.Fatalf("buildWindowsPrincipalACLPlan: %v", err) + } + rollback, err := applyWindowsACLPlan(plan) + if err != nil { + t.Fatalf("applyWindowsACLPlan: %v", err) + } + t.Cleanup(func() { + if rollback != nil { + _ = rollback() + } + }) + + if !windowsPathDeniesDelete(t, gitDir, principal) { + t.Fatal("no deny-delete ACE on a .git that already existed") + } +} diff --git a/internal/sandbox/windows_git_rename_guard_test.go b/internal/sandbox/windows_git_rename_guard_test.go new file mode 100644 index 000000000..4d82f243b --- /dev/null +++ b/internal/sandbox/windows_git_rename_guard_test.go @@ -0,0 +1,98 @@ +package sandbox + +import ( + "path/filepath" + "testing" +) + +// A sandbox principal must not be able to REPLACE .git. +// +// The write-denied carveouts are attached to .git/config and .git/hooks as +// objects. Rename .git aside, recreate it, and those objects are gone: the fresh +// config and hooks inherit the workspace allow with no deny of their own, which +// hands back credential.helper and core.hooksPath, and with them arbitrary code +// execution on the next git command. +// +// .git cannot join sandboxFullyProtectedMetadataNames to fix this, because that +// list emits DenyWrite, whose mask includes FILE_GENERIC_WRITE. Git has to write +// index, objects and refs. So the directory needs DELETE denied on itself while +// staying writable underneath. +func TestThePrincipalCannotRenameTheGitDirectory(t *testing.T) { + root := filepath.FromSlash("/ws/repo") + plan, err := buildWindowsPrincipalACLPlan(windowsPrincipalACLInput{ + PrincipalSID: "S-1-5-21-1-2-3-1001", + WriteRoots: []WritableRoot{{ + Root: root, + ProtectedMetadataNames: sandboxFullyProtectedMetadataNames, + ReadOnlySubpaths: gitMetadataWriteCarveouts(root), + }}, + }) + if err != nil { + t.Fatalf("buildWindowsPrincipalACLPlan: %v", err) + } + + // The plan normalizes every write root before it names an ACE, so the + // expected path has to be normalized too or this compares two spellings of + // the same directory. Hardcoding a drive letter instead would pass on + // Windows and fail everywhere else, since the builder is portable code. + gitDir := filepath.Join(normalizeProfilePath(root), ".git") + var denyDelete *WindowsACLEntry + for index := range plan.Entries { + if plan.Entries[index].Action == WindowsACLDenyDelete && plan.Entries[index].Path == gitDir { + denyDelete = &plan.Entries[index] + break + } + } + if denyDelete == nil { + t.Fatalf("no deny-delete entry for %s, so the principal can rename .git and recreate it without the carveouts:\n%#v", gitDir, plan.Entries) + } + if denyDelete.Capability != "S-1-5-21-1-2-3-1001" { + t.Errorf("deny-delete names %q, want the principal SID", denyDelete.Capability) + } +} + +// The guard must not become a write ban. Git writes constantly inside .git, so a +// deny that reached the children would break every commit rather than just the +// rename. It also must not be materialized into existence: .git is git's to +// create, and an empty .git directory made by setup breaks `git init`. +func TestTheGitRenameGuardDoesNotBlockGitsOwnWrites(t *testing.T) { + root := filepath.FromSlash("/ws/repo") + plan, err := buildWindowsPrincipalACLPlan(windowsPrincipalACLInput{ + PrincipalSID: "S-1-5-21-1-2-3-1001", + WriteRoots: []WritableRoot{{ + Root: root, + ProtectedMetadataNames: sandboxFullyProtectedMetadataNames, + ReadOnlySubpaths: gitMetadataWriteCarveouts(root), + }}, + }) + if err != nil { + t.Fatalf("buildWindowsPrincipalACLPlan: %v", err) + } + + gitDir := filepath.Join(normalizeProfilePath(root), ".git") + for _, entry := range plan.Entries { + if entry.Path != gitDir { + continue + } + if entry.Action == WindowsACLDenyWrite { + t.Errorf("deny-write on %s would stop git writing index/objects/refs", gitDir) + } + if entry.Action == WindowsACLDenyDelete && entry.Materialize { + t.Errorf("the rename guard materializes %s; git must create it, an empty .git breaks git init", gitDir) + } + } + + // The existing carveouts must survive unchanged. + for _, want := range []string{filepath.Join(gitDir, "config"), filepath.Join(gitDir, "hooks")} { + found := false + for _, entry := range plan.Entries { + if entry.Action == WindowsACLDenyWrite && entry.Path == want { + found = true + break + } + } + if !found { + t.Errorf("the write-deny carveout for %s disappeared", want) + } + } +} diff --git a/internal/sandbox/windows_git_rename_guard_windows_test.go b/internal/sandbox/windows_git_rename_guard_windows_test.go new file mode 100644 index 000000000..973d30c5d --- /dev/null +++ b/internal/sandbox/windows_git_rename_guard_windows_test.go @@ -0,0 +1,81 @@ +//go:build windows + +package sandbox + +import ( + "testing" + + "golang.org/x/sys/windows" +) + +// The mask is the whole point of the action, so it is asserted bit by bit. +// +// Renaming a directory needs DELETE on the directory itself, so denying DELETE +// is what stops .git being replaced. Everything else in the mask is there to +// stop the principal removing the guard: WRITE_DAC would let it rewrite the +// DACL, WRITE_OWNER would let it take ownership and then rewrite the DACL. +// +// What must NOT be in it matters just as much. FILE_GENERIC_WRITE would stop git +// writing index, objects and refs. FILE_DELETE_CHILD would stop git deleting its +// own lock files and refs. Either one turns a rename guard into a broken repo. +func TestDenyDeleteMaskStopsRenameWithoutStoppingGit(t *testing.T) { + mode, mask, err := windowsACLAccess(WindowsACLDenyDelete) + if err != nil { + t.Fatalf("windowsACLAccess(deny-delete): %v", err) + } + if mode != windows.DENY_ACCESS { + t.Fatalf("access mode = %v, want DENY_ACCESS", mode) + } + + for _, required := range []struct { + name string + bit windows.ACCESS_MASK + }{ + {"DELETE", windows.DELETE}, + {"WRITE_DAC", windows.WRITE_DAC}, + {"WRITE_OWNER", windows.WRITE_OWNER}, + } { + if mask&required.bit == 0 { + t.Errorf("mask %#x is missing %s, so the guard can be removed or bypassed", mask, required.name) + } + } + for _, forbidden := range []struct { + name string + bit windows.ACCESS_MASK + breaks string + }{ + {"FILE_GENERIC_WRITE", windows.FILE_GENERIC_WRITE, "git writing index/objects/refs"}, + {"FILE_DELETE_CHILD", windowsFileDeleteChild, "git deleting its own lock files and refs"}, + } { + if mask&forbidden.bit != 0 { + t.Errorf("mask %#x includes %s, which breaks %s", mask, forbidden.name, forbidden.breaks) + } + } +} + +// Inheritance is the other half. An inherited deny would reach every file inside +// .git and stop git deleting anything at all, so this ACE has to apply to the +// directory object alone while the other actions keep inheriting as before. +func TestDenyDeleteDoesNotInheritWhileOtherActionsStillDo(t *testing.T) { + entries := []WindowsACLEntry{ + {Action: WindowsACLDenyDelete, Path: `C:\work\repo\.git`, Capability: "S-1-5-32-9999"}, + {Action: WindowsACLDenyWrite, Path: `C:\work\repo\.zero`, Capability: "S-1-5-32-9999"}, + {Action: WindowsACLAllowWrite, Path: `C:\work\repo`, Capability: "S-1-5-32-9999"}, + } + + _, access, err := prepareWindowsACLPathGroupEntries(entries, true, nil) + if err != nil { + t.Fatalf("prepareWindowsACLPathGroupEntries: %v", err) + } + if len(access) != len(entries) { + t.Fatalf("got %d access entries, want %d", len(access), len(entries)) + } + if access[0].Inheritance != windows.NO_INHERITANCE { + t.Errorf("deny-delete inheritance = %#x, want NO_INHERITANCE; an inherited deny would stop git deleting inside .git", access[0].Inheritance) + } + for index, entry := range entries[1:] { + if got := access[index+1].Inheritance; got != windows.SUB_CONTAINERS_AND_OBJECTS_INHERIT { + t.Errorf("%s inheritance = %#x, want the directory default to be unchanged", entry.Action, got) + } + } +} diff --git a/internal/sandbox/windows_group_adoption_windows_test.go b/internal/sandbox/windows_group_adoption_windows_test.go new file mode 100644 index 000000000..f1652d8f2 --- /dev/null +++ b/internal/sandbox/windows_group_adoption_windows_test.go @@ -0,0 +1,86 @@ +//go:build windows + +package sandbox + +import ( + "errors" + "strings" + "testing" +) + +// A NAME IS NOT PROOF OF PROVENANCE, for the USERS group specifically. +// +// ensureWindowsLocalGroup covers both managed groups, and the sibling test in +// windows_offline_group_ownership_windows_test.go pins the offline one. This +// pins ZeroSandboxUsers, which is the group anandh8x reported: it holds every +// sandbox principal, so adopting a same-named group created by an installer or +// by policy would hand the principal whatever that group already grants. +// +// Driven through the seams so it needs no Administrator and leaves no real local +// group on the machine running the suite. +func TestAdoptingAForeignUsersGroupIsRefused(t *testing.T) { + for _, status := range []uintptr{nerrGroupExists, errorAliasExists} { + prevAdd, prevOwned := addWindowsLocalGroupFn, windowsLocalGroupOwnedByZeroFn + t.Cleanup(func() { addWindowsLocalGroupFn, windowsLocalGroupOwnedByZeroFn = prevAdd, prevOwned }) + + addWindowsLocalGroupFn = func(string, string) (uintptr, error) { return status, nil } + windowsLocalGroupOwnedByZeroFn = func(string, string) (bool, error) { return false, nil } + + err := ensureWindowsSandboxGroup() + if err == nil { + t.Fatalf("status %d adopted a group Zero did not create, so the principal inherits whatever it already grants", status) + } + if !strings.Contains(err.Error(), windowsSandboxGroupName) { + t.Errorf("the refusal must name the group in the way, got %q", err) + } + } +} + +// Our own group is still adopted, or re-running setup would fail on the group it +// created a moment ago and provisioning would never converge. +func TestAdoptingOurOwnUsersGroupSucceeds(t *testing.T) { + prevAdd, prevOwned := addWindowsLocalGroupFn, windowsLocalGroupOwnedByZeroFn + t.Cleanup(func() { addWindowsLocalGroupFn, windowsLocalGroupOwnedByZeroFn = prevAdd, prevOwned }) + + addWindowsLocalGroupFn = func(string, string) (uintptr, error) { return nerrGroupExists, nil } + windowsLocalGroupOwnedByZeroFn = func(name, comment string) (bool, error) { + if name != windowsSandboxGroupName || comment != windowsSandboxGroupComment { + t.Errorf("ownership probed for %q/%q, want the users group", name, comment) + } + return true, nil + } + if err := ensureWindowsSandboxGroup(); err != nil { + t.Errorf("refused a group carrying Zero's own comment: %v", err) + } +} + +// A freshly created group is ours by construction, so no ownership probe runs. +func TestCreatingTheUsersGroupDoesNotProbeOwnership(t *testing.T) { + prevAdd, prevOwned := addWindowsLocalGroupFn, windowsLocalGroupOwnedByZeroFn + t.Cleanup(func() { addWindowsLocalGroupFn, windowsLocalGroupOwnedByZeroFn = prevAdd, prevOwned }) + + probed := false + addWindowsLocalGroupFn = func(string, string) (uintptr, error) { return nerrSuccess, nil } + windowsLocalGroupOwnedByZeroFn = func(string, string) (bool, error) { probed = true; return false, nil } + + if err := ensureWindowsSandboxGroup(); err != nil { + t.Fatalf("a successful create was rejected: %v", err) + } + if probed { + t.Error("ownership was probed for a group we had just created ourselves") + } +} + +// A failed ownership probe is not read as "ours" and not as "not ours" either. +func TestAnUnreadableUsersGroupIsNotGuessedEitherWay(t *testing.T) { + prevAdd, prevOwned := addWindowsLocalGroupFn, windowsLocalGroupOwnedByZeroFn + t.Cleanup(func() { addWindowsLocalGroupFn, windowsLocalGroupOwnedByZeroFn = prevAdd, prevOwned }) + + sentinel := errors.New("NetLocalGroupGetInfo: status 5") + addWindowsLocalGroupFn = func(string, string) (uintptr, error) { return nerrGroupExists, nil } + windowsLocalGroupOwnedByZeroFn = func(string, string) (bool, error) { return false, sentinel } + + if err := ensureWindowsSandboxGroup(); !errors.Is(err, sentinel) { + t.Fatalf("a failed ownership probe was swallowed, got %v", err) + } +} diff --git a/internal/sandbox/windows_identity_acl.go b/internal/sandbox/windows_identity_acl.go new file mode 100644 index 000000000..1d9c47162 --- /dev/null +++ b/internal/sandbox/windows_identity_acl.go @@ -0,0 +1,231 @@ +package sandbox + +// ACLs for a sandbox principal. +// +// The capability-SID model this sits beside starts from "the caller can already +// read everything" and narrows writes, because the sandboxed child runs as the +// caller. A principal inverts that: a separate local account has no access to +// the caller's profile at all, so the interesting direction is what to GRANT. +// +// That inversion is the point. Credential stores under the user's profile are +// unreachable because the principal is a different account, not because a deny +// rule enumerated them, which is what makes this able to close #662 and #675 on +// Windows where a deny-read ACE against the caller's own SID never could. Deny +// rules stay useful only for objects that are readable by everyone. +// +// Grants are explicit and narrow: the workspace and any extra write roots get +// read+write, declared read-only roots get read, and the protected metadata +// carve-outs the profile already defines stay denied so .git internals and +// .zero/.agents cannot be rewritten from inside the sandbox. + +import ( + "errors" + "fmt" + "path/filepath" +) + +// WindowsACLAllowRead grants read and execute without write. It exists for the +// principal model, where a read root must be granted rather than assumed. +const WindowsACLAllowRead WindowsACLAction = "allow-read" + +// windowsPrincipalACLInput is everything needed to describe a principal's +// access. It is deliberately a plain struct rather than the full command config +// so the plan can be built and tested without a live sandbox. +type windowsPrincipalACLInput struct { + // PrincipalSID is the string SID of the sandbox account every ACE names. + PrincipalSID string + // WriteRoots receive read+write+execute. The workspace lives here. + WriteRoots []WritableRoot + // ReadRoots receive read+execute only. + ReadRoots []string + // DenyRead covers objects a principal could otherwise reach because they are + // world-readable; per-user secrets need no entry. + DenyRead []string + // DenyWrite carries the policy's own deny-write paths. The capability plan + // has always emitted these; the principal plan denied write only on + // protected metadata and read-only subpaths inside write roots, so a policy + // deny sitting anywhere else was simply not enforced once the runner used a + // principal token, and a shell child could write where the restricted-token + // backend would have blocked it. + DenyWrite []string +} + +// buildWindowsPrincipalACLPlan turns a principal's access into ACL entries. +// +// Ordering matters at apply time: deny entries are emitted before allows so a +// carve-out inside a granted root survives, which mirrors how Windows evaluates +// an explicit DACL (deny ACEs first). +func buildWindowsPrincipalACLPlan(input windowsPrincipalACLInput) (WindowsACLPlan, error) { + if input.PrincipalSID == "" { + return WindowsACLPlan{}, errors.New("windows principal ACL plan requires a principal SID") + } + if len(input.WriteRoots) == 0 && len(input.ReadRoots) == 0 { + return WindowsACLPlan{}, errors.New("windows principal ACL plan requires at least one root") + } + + entries := make([]WindowsACLEntry, 0, len(input.WriteRoots)*2+len(input.ReadRoots)+len(input.DenyRead)) + + // Deny first. A deny ACE inside a write root (protected metadata, git + // internals) has to win over the grant that follows it. + // Materialized, matching the capability plan. applyWindowsACLPlan skips a + // target that does not exist, so without this a deny-read path created after + // setup ran never got an ACE at all and the principal could read it. The + // deny has to be in place before the object is. + for _, path := range normalizeProfilePaths(input.DenyRead) { + entries = append(entries, WindowsACLEntry{ + Action: WindowsACLDenyRead, + Path: path, + Capability: input.PrincipalSID, + Materialize: true, + }) + } + for _, path := range normalizeProfilePaths(input.DenyWrite) { + entries = append(entries, WindowsACLEntry{ + Action: WindowsACLDenyWrite, + Path: path, + Capability: input.PrincipalSID, + Materialize: true, + }) + } + for _, root := range input.WriteRoots { + // Normalized the same way as read and deny paths: a write root may arrive + // with "~" or as a relative path, and an ACE has to name the same absolute, + // symlink-resolved object the deny entries do or the two disagree. + cleaned := normalizeProfilePath(root.Root) + if cleaned == "" { + return WindowsACLPlan{}, fmt.Errorf("windows principal ACL plan: unusable write root %q", root.Root) + } + // Materialized, like the metadata and policy denies below and above. + // + // These are the git control-plane carveouts (.git/config, .git/hooks). On a + // workspace where git has not run yet they do not exist at setup time, and + // applyWindowsACLPlan skips a target that is absent, so the ACEs were never + // written. Once git created those paths the principal still held inherited + // write access to the workspace and could install a hook or rewrite + // credential.helper. The capability plan gets away without this because its + // child runs as the caller; a separate principal account does not. + // Which carveouts are files rather than directories comes from the same + // spec list the profile built ReadOnlySubpaths from, so a new carveout + // cannot be added without its shape coming along. + for _, subpath := range normalizeProfilePaths(root.ReadOnlySubpaths) { + entries = append(entries, WindowsACLEntry{ + Action: WindowsACLDenyWrite, + Path: subpath, + Capability: input.PrincipalSID, + Materialize: true, + MaterializeFile: gitMetadataCarveoutIsFile(subpath), + }) + } + for _, name := range root.ProtectedMetadataNames { + // These are documented as NAMES, and the join below is what makes that + // documentation load-bearing rather than descriptive: ".." or a + // separator would place this deny ACE, and the directory it + // materializes, outside the write root entirely. Today every caller + // passes a package constant, so this is unreachable; it is here so that + // stays true when a future caller sources these from config. + if err := validateWindowsACLComponent(name); err != nil { + return WindowsACLPlan{}, fmt.Errorf("windows principal ACL plan: protected metadata name: %w", err) + } + entries = append(entries, WindowsACLEntry{ + Action: WindowsACLDenyWrite, + Path: filepath.Join(cleaned, name), + Capability: input.PrincipalSID, + Materialize: true, + }) + } + // .git gets DELETE denied on the directory itself, because the carveouts + // that protect it are attached to .git/config and .git/hooks as OBJECTS. + // Rename .git aside and recreate it and those objects are gone, so the + // fresh config and hooks inherit the workspace allow with no deny of their + // own, handing back credential.helper and core.hooksPath. + // + // Not DenyWrite (git writes index, objects and refs), not materialized + // (git creates .git, and an empty one breaks git init), and not inherited, + // so everything underneath stays writable. + entries = append(entries, WindowsACLEntry{ + Action: WindowsACLDenyDelete, + Path: windowsRenameProtectedObject(cleaned), + Capability: input.PrincipalSID, + }) + } + + // Then the grants the principal cannot work without. + for _, root := range input.WriteRoots { + cleaned := normalizeProfilePath(root.Root) + if cleaned == "" { + continue + } + entries = append(entries, WindowsACLEntry{ + Action: WindowsACLAllowWrite, + Path: cleaned, + Capability: input.PrincipalSID, + }) + } + for _, path := range normalizeProfilePaths(input.ReadRoots) { + // NEVER grant at a volume root. + // + // permissionProfileReadRoots seeds its list with profileRootPath(), which + // is the separator alone, because the workspace-write posture is + // read-all/write-jail. That is harmless for the capability backend, whose + // child runs as the CALLER and therefore reads what the caller could read + // anyway. It is not harmless here: a principal is a separate local + // account, so this loop turns that synthetic entry into a real, + // persistent, inheritable allow-read ACE for that account at the root of + // the current drive, reaching every directory that does not block + // inheritance. + // + // Dropping it does not take away the reads the principal needs to run + // commands. NetUserAdd with USER_PRIV_USER puts the account in the + // built-in Users group (see usrPrivUser in windows_identity_windows.go), + // and the machine's own ACLs already grant Users read on the system and + // program directories. What the grant added on top was read access to the + // places Users are deliberately kept out of, which is the opposite of what + // a sandbox is for. + // + // Note this inherits rather than asserts: nothing here checks that those + // default ACLs are actually in place, so a hardened image that strips + // Users read would need an explicit bounded read set instead. + if isWindowsVolumeRoot(path) { + continue + } + entries = append(entries, WindowsACLEntry{ + Action: WindowsACLAllowRead, + Path: path, + Capability: input.PrincipalSID, + }) + } + + return WindowsACLPlan{Entries: dedupeWindowsACLEntries(entries)}, nil +} + +// windowsPrincipalRevokePlan returns the entries whose ACEs should be removed +// when a principal is retired. Revocation is by TRUSTEE rather than by path: +// every ACE naming this principal is dropped, so cleanup does not depend on +// remembering which paths were granted, and a grant added by an older version +// is still removed. +// +// This is the removal path the capability-SID model never had, where a synthetic +// SID left ACEs behind on the user's tree with nothing to match them against. +func windowsPrincipalRevokePlan(principalSID string, paths []string) (WindowsACLPlan, error) { + if principalSID == "" { + return WindowsACLPlan{}, errors.New("windows principal revoke plan requires a principal SID") + } + cleaned := normalizeProfilePaths(paths) + entries := make([]WindowsACLEntry, 0, len(cleaned)) + for _, path := range cleaned { + entries = append(entries, WindowsACLEntry{ + Action: windowsACLRevoke, + Path: path, + Capability: principalSID, + }) + } + return WindowsACLPlan{Entries: entries}, nil +} + +// windowsACLRevoke removes every ACE naming the trustee on a path, whatever +// access it granted or denied. +const windowsACLRevoke WindowsACLAction = "revoke" + +// windowsACLPlanPaths lives in windows_identity_runtime_windows.go, beside its +// only callers. It was here, in the portable file, which made it dead code on +// every non-Windows build and failed the static analysis gate. diff --git a/internal/sandbox/windows_identity_acl_test.go b/internal/sandbox/windows_identity_acl_test.go new file mode 100644 index 000000000..79b5ff85b --- /dev/null +++ b/internal/sandbox/windows_identity_acl_test.go @@ -0,0 +1,275 @@ +package sandbox + +import ( + "path/filepath" + "strings" + "testing" +) + +const testPrincipalSID = "S-1-5-21-1111111111-2222222222-3333333333-1005" + +func testPrincipalInput() windowsPrincipalACLInput { + return windowsPrincipalACLInput{ + PrincipalSID: testPrincipalSID, + WriteRoots: []WritableRoot{{ + Root: filepath.FromSlash("/ws/project"), + ReadOnlySubpaths: []string{filepath.FromSlash("/ws/project/.git/config")}, + ProtectedMetadataNames: []string{".zero", ".agents"}, + }}, + ReadRoots: []string{filepath.FromSlash("/usr/lib")}, + DenyRead: []string{filepath.FromSlash("/shared/secrets")}, + } +} + +// Windows evaluates an explicit DACL deny-before-allow, so a carve-out inside a +// granted root only survives if its deny ACE is written first. If the grant on +// the workspace landed before the deny on .zero, the protected metadata would be +// writable from inside the sandbox. +func TestPrincipalACLPlanEmitsDeniesBeforeAllows(t *testing.T) { + plan, err := buildWindowsPrincipalACLPlan(testPrincipalInput()) + if err != nil { + t.Fatalf("build: %v", err) + } + lastDeny, firstAllow := -1, -1 + for index, entry := range plan.Entries { + switch entry.Action { + case WindowsACLDenyRead, WindowsACLDenyWrite: + lastDeny = index + case WindowsACLAllowWrite, WindowsACLAllowRead: + if firstAllow == -1 { + firstAllow = index + } + } + } + if firstAllow == -1 || lastDeny == -1 { + t.Fatalf("plan is missing a deny or an allow: %+v", plan.Entries) + } + if lastDeny > firstAllow { + t.Fatalf("deny at %d comes after allow at %d; carve-outs would be overridden", lastDeny, firstAllow) + } +} + +// Every ACE must name the sandbox principal. An entry with any other trustee +// would change access for a real user. +func TestPrincipalACLPlanNamesOnlyThePrincipal(t *testing.T) { + plan, err := buildWindowsPrincipalACLPlan(testPrincipalInput()) + if err != nil { + t.Fatalf("build: %v", err) + } + for _, entry := range plan.Entries { + if entry.Capability != testPrincipalSID { + t.Fatalf("entry %+v names %q, want the principal SID", entry, entry.Capability) + } + } +} + +// A principal is a separate account with no inherent access, so a write root +// must be granted read+write and a read root granted read. Without the grant the +// sandbox cannot open its own workspace. +func TestPrincipalACLPlanGrantsRoots(t *testing.T) { + plan, err := buildWindowsPrincipalACLPlan(testPrincipalInput()) + if err != nil { + t.Fatalf("build: %v", err) + } + var grantedWrite, grantedRead bool + for _, entry := range plan.Entries { + if entry.Action == WindowsACLAllowWrite && entry.Path == normalizeProfilePath(filepath.FromSlash("/ws/project")) { + grantedWrite = true + } + if entry.Action == WindowsACLAllowRead && entry.Path == normalizeProfilePath(filepath.FromSlash("/usr/lib")) { + grantedRead = true + } + } + if !grantedWrite { + t.Fatal("write root was not granted; the sandbox could not write its workspace") + } + if !grantedRead { + t.Fatal("read root was not granted; the sandbox could not read it") + } +} + +// Protected metadata is denied write and marked Materialize so the ACE is +// created even when the directory does not exist yet, closing the window where +// a sandboxed command creates .zero before the deny lands. +func TestPrincipalACLPlanProtectsMetadata(t *testing.T) { + plan, err := buildWindowsPrincipalACLPlan(testPrincipalInput()) + if err != nil { + t.Fatalf("build: %v", err) + } + found := map[string]WindowsACLEntry{} + for _, entry := range plan.Entries { + found[entry.Path] = entry + } + for _, name := range []string{".zero", ".agents"} { + path := filepath.Join(normalizeProfilePath(filepath.FromSlash("/ws/project")), name) + entry, ok := found[path] + if !ok { + t.Fatalf("no entry protecting %s", path) + } + if entry.Action != WindowsACLDenyWrite { + t.Fatalf("%s has action %q, want deny-write", path, entry.Action) + } + if !entry.Materialize { + t.Fatalf("%s must be materialized so the deny exists before the directory does", path) + } + } +} + +// A missing principal SID must be a hard error: an empty trustee would either +// fail at apply time or, worse, be interpreted as some other account. +func TestPrincipalACLPlanRequiresSID(t *testing.T) { + input := testPrincipalInput() + input.PrincipalSID = "" + if _, err := buildWindowsPrincipalACLPlan(input); err == nil { + t.Fatal("an empty principal SID must be rejected") + } +} + +// A plan with no roots at all is a caller mistake rather than a valid empty +// grant, since the resulting sandbox could not run anything. +func TestPrincipalACLPlanRequiresRoots(t *testing.T) { + input := windowsPrincipalACLInput{PrincipalSID: testPrincipalSID} + if _, err := buildWindowsPrincipalACLPlan(input); err == nil { + t.Fatal("a plan with no roots must be rejected") + } +} + +// Revocation is keyed to the trustee, so retiring a principal removes every ACE +// naming it without having to remember what was granted. This is the cleanup +// path the capability-SID model lacks. +func TestPrincipalRevokePlanTargetsTrustee(t *testing.T) { + paths := []string{filepath.FromSlash("/ws/project"), filepath.FromSlash("/usr/lib")} + plan, err := windowsPrincipalRevokePlan(testPrincipalSID, paths) + if err != nil { + t.Fatalf("revoke plan: %v", err) + } + if len(plan.Entries) != len(paths) { + t.Fatalf("got %d entries, want %d", len(plan.Entries), len(paths)) + } + for _, entry := range plan.Entries { + if entry.Action != windowsACLRevoke { + t.Fatalf("entry %+v is not a revoke", entry) + } + if entry.Capability != testPrincipalSID { + t.Fatalf("revoke names %q, want the principal", entry.Capability) + } + } +} + +func TestPrincipalRevokePlanRequiresSID(t *testing.T) { + if _, err := windowsPrincipalRevokePlan("", []string{"/ws"}); err == nil { + t.Fatal("revoking without a principal SID must be rejected") + } +} + +// The action strings end up in a serialized plan consumed by the elevated +// helper, so they must stay stable and distinct from the existing actions. +func TestPrincipalACLActionsAreDistinct(t *testing.T) { + actions := []WindowsACLAction{ + WindowsACLAllowWrite, WindowsACLAllowRead, + WindowsACLDenyRead, WindowsACLDenyWrite, windowsACLRevoke, + } + seen := map[WindowsACLAction]bool{} + for _, action := range actions { + if strings.TrimSpace(string(action)) == "" { + t.Fatal("an action string is empty") + } + if seen[action] { + t.Fatalf("duplicate action %q", action) + } + seen[action] = true + } +} + +// ProtectedMetadataNames is joined onto the write root to place a deny ACE and to +// materialize the directory it names. A value that is not a single component +// therefore puts both OUTSIDE the workspace: ".." walks up out of it, and a +// separator reaches through whatever sits in between. Every caller passes a +// package constant today, so this is the guard that keeps it true if one ever +// sources these from config. +func TestPrincipalACLPlanRefusesProtectedNamesThatEscapeTheWriteRoot(t *testing.T) { + root := filepath.FromSlash("/ws/project") + for _, name := range []string{"..", ".", "", `..\..\Windows\System32`, "nested/child", `nested\child`, "C:", "stream:name"} { + t.Run(name, func(t *testing.T) { + input := testPrincipalInput() + input.WriteRoots = []WritableRoot{{ + Root: root, + ProtectedMetadataNames: []string{name}, + }} + plan, err := buildWindowsPrincipalACLPlan(input) + if err == nil { + t.Fatalf("accepted protected metadata name %q, which would place a deny ACE outside %s:\n%#v", name, root, plan.Entries) + } + if len(plan.Entries) != 0 { + t.Errorf("returned %d entries alongside the error, so a caller ignoring err would still apply them", len(plan.Entries)) + } + }) + } +} + +// A PRINCIPAL MUST NEVER BE GRANTED READ AT A VOLUME ROOT. +// +// permissionProfileReadRoots seeds its list with the bare separator, because +// the workspace-write posture is read-all with a write jail. That costs nothing +// for the capability backend, whose child runs as the caller and could read +// those paths anyway. For a principal it is a real, persistent, inheritable +// allow-read ACE for a separate local account at the root of the drive. +// +// Built from the production profile rather than a synthetic fixture, because +// the whole point is that the shipped configuration produced it. +func TestPrincipalACLPlanNeverGrantsReadAtAVolumeRoot(t *testing.T) { + workspace := filepath.FromSlash("/ws/project") + profile := DefaultPermissionProfile(workspace) + + // If the profile ever stops carrying a volume root, this test proves nothing + // and should be retired rather than left passing vacuously. + seeded := false + for _, root := range profile.FileSystem.ReadRoots { + if isWindowsVolumeRoot(normalizeProfilePath(root)) { + seeded = true + break + } + } + if !seeded { + t.Skipf("the production profile no longer contains a volume read root: %v", profile.FileSystem.ReadRoots) + } + + plan, err := buildWindowsPrincipalACLPlan(windowsPrincipalACLInput{ + PrincipalSID: testPrincipalSID, + ReadRoots: profile.FileSystem.ReadRoots, + WriteRoots: profile.FileSystem.WriteRoots, + }) + if err != nil { + t.Fatalf("buildWindowsPrincipalACLPlan: %v", err) + } + for _, entry := range plan.Entries { + if entry.Action == WindowsACLAllowRead && isWindowsVolumeRoot(entry.Path) { + t.Errorf("plan grants the principal read at the volume root %q, which inherits into every directory on the drive", entry.Path) + } + } + // And the workspace itself must still be reachable, or this traded a real + // grant for a broken sandbox. + wantWorkspace := normalizeProfilePath(workspace) + reachable := false + for _, entry := range plan.Entries { + if entry.Path == wantWorkspace && (entry.Action == WindowsACLAllowRead || entry.Action == WindowsACLAllowWrite) { + reachable = true + break + } + } + if !reachable { + t.Errorf("no grant for the workspace root %q, so the principal could not read its own workspace", wantWorkspace) + } +} + +// The ordinary names must keep working, or the guard above is just a break. +func TestPrincipalACLPlanStillAcceptsTheRealProtectedNames(t *testing.T) { + input := testPrincipalInput() + input.WriteRoots = []WritableRoot{{ + Root: filepath.FromSlash("/ws/project"), + ProtectedMetadataNames: sandboxFullyProtectedMetadataNames, + }} + if _, err := buildWindowsPrincipalACLPlan(input); err != nil { + t.Fatalf("the shipped protected names were rejected: %v", err) + } +} diff --git a/internal/sandbox/windows_identity_dpapi_windows.go b/internal/sandbox/windows_identity_dpapi_windows.go new file mode 100644 index 000000000..d9a0bbf1d --- /dev/null +++ b/internal/sandbox/windows_identity_dpapi_windows.go @@ -0,0 +1,76 @@ +//go:build windows + +package sandbox + +// DPAPI wrapping for the stored sandbox principal password. +// +// The file ACL is the primary control and remains the thing that keeps the +// sandbox principal itself from reading its own credential. This adds the layer +// the ACL cannot: an ACL is only meaningful while the filesystem is being asked +// to enforce it, so a backup, a mounted disk image, or a copy taken by anyone +// who can bypass the DACL yields the password in the clear. CryptProtectData +// binds the ciphertext to the invoking user's logon secret, so an offline copy +// is inert without that user's credentials. +// +// The principal name is passed as optional entropy, which makes a blob usable +// only for the account it was minted for; moving one secret file over another +// then fails to decrypt instead of silently authenticating the wrong principal. + +import ( + "errors" + "fmt" + "unsafe" + + "golang.org/x/sys/windows" +) + +// protectWindowsSecret encrypts a password to the current user. +// +// CRYPTPROTECT_UI_FORBIDDEN matters here: this runs inside a CLI and, on the +// setup path, potentially without an interactive desktop, so DPAPI must fail +// rather than try to prompt. +func protectWindowsSecret(plaintext string, entropy string) ([]byte, error) { + if plaintext == "" { + return nil, errors.New("windows sandbox secret: refusing to protect an empty password") + } + in := windows.DataBlob{ + Size: uint32(len(plaintext)), + Data: &[]byte(plaintext)[0], + } + entropyBytes := []byte(entropy) + var entropyBlob *windows.DataBlob + if len(entropyBytes) > 0 { + entropyBlob = &windows.DataBlob{Size: uint32(len(entropyBytes)), Data: &entropyBytes[0]} + } + var out windows.DataBlob + if err := windows.CryptProtectData(&in, nil, entropyBlob, 0, nil, windows.CRYPTPROTECT_UI_FORBIDDEN, &out); err != nil { + return nil, fmt.Errorf("protect sandbox secret: %w", err) + } + // DPAPI allocates the output with LocalAlloc; copy it out and hand it back. + defer windows.LocalFree(windows.Handle(unsafe.Pointer(out.Data))) + return append([]byte(nil), unsafe.Slice(out.Data, out.Size)...), nil +} + +// unprotectWindowsSecret reverses protectWindowsSecret. It fails for any user +// other than the one that wrote the blob, and for a blob minted with a different +// principal name as entropy. +func unprotectWindowsSecret(ciphertext []byte, entropy string) (string, error) { + if len(ciphertext) == 0 { + return "", errWindowsSandboxIdentityUnavailable + } + in := windows.DataBlob{ + Size: uint32(len(ciphertext)), + Data: &ciphertext[0], + } + entropyBytes := []byte(entropy) + var entropyBlob *windows.DataBlob + if len(entropyBytes) > 0 { + entropyBlob = &windows.DataBlob{Size: uint32(len(entropyBytes)), Data: &entropyBytes[0]} + } + var out windows.DataBlob + if err := windows.CryptUnprotectData(&in, nil, entropyBlob, 0, nil, windows.CRYPTPROTECT_UI_FORBIDDEN, &out); err != nil { + return "", fmt.Errorf("unprotect sandbox secret: %w", err) + } + defer windows.LocalFree(windows.Handle(unsafe.Pointer(out.Data))) + return string(unsafe.Slice(out.Data, out.Size)), nil +} diff --git a/internal/sandbox/windows_identity_logon_windows.go b/internal/sandbox/windows_identity_logon_windows.go new file mode 100644 index 000000000..45ebdaa5f --- /dev/null +++ b/internal/sandbox/windows_identity_logon_windows.go @@ -0,0 +1,270 @@ +//go:build windows + +package sandbox + +// Minting a token for a sandbox principal. +// +// A provisioned account is inert until something can log on as it. Windows +// gates that behind account rights held in the local security policy, so setup +// grants the principal exactly one: the right to be logged on as a batch job, +// which is what a non-interactive service-style logon needs. It is deliberately +// NOT granted interactive, network or remote-interactive logon, and those three +// are explicitly DENIED, so the account cannot be used to sign in at the +// console, over SMB, or through RDP even if its password leaked. The password +// exists only so LogonUser can mint a token; nobody is meant to type it. +// +// Rights are granted at setup (elevated) because LsaAddAccountRights requires +// administrator privileges. The per-command path only calls LogonUser, which +// needs no special privilege once the batch right is in place. + +import ( + "errors" + "fmt" + "runtime" + "unsafe" + + "golang.org/x/sys/windows" +) + +const ( + // Logon type/provider for a non-interactive token. Batch is the closest + // match for "run this command as a service-like principal": it produces a + // full token without a desktop or network-credential footprint. + logon32LogonBatch = 4 + logon32ProviderDefault = 0 + + // LSA policy access rights needed to add account rights. + policyCreateAccount = 0x00000010 + policyLookupNames = 0x00000800 + + // Account rights. The sandbox principal gets the batch right and is denied + // every interactive path. + seBatchLogonRight = "SeBatchLogonRight" + seDenyInteractiveLogonRight = "SeDenyInteractiveLogonRight" + seDenyNetworkLogonRight = "SeDenyNetworkLogonRight" + seDenyRemoteInteractiveRight = "SeDenyRemoteInteractiveLogonRight" + seDenyServiceLogonRightName = "SeDenyServiceLogonRight" + windowsIdentityLogonRightsNote = "granted by `zero sandbox setup`" +) + +var ( + procLogonUserW = windows.NewLazySystemDLL("advapi32.dll").NewProc("LogonUserW") + procLsaOpenPolicy = windows.NewLazySystemDLL("advapi32.dll").NewProc("LsaOpenPolicy") + procLsaClose = windows.NewLazySystemDLL("advapi32.dll").NewProc("LsaClose") + procLsaAddAccountRights = windows.NewLazySystemDLL("advapi32.dll").NewProc("LsaAddAccountRights") + // Retiring a principal has to drop its rights as well as its account, or the + // LSA policy database keeps an entry keyed to a SID that no longer resolves. + procLsaRemoveAccountRights = windows.NewLazySystemDLL("advapi32.dll").NewProc("LsaRemoveAccountRights") + procLsaNtStatusToWinErr = windows.NewLazySystemDLL("advapi32.dll").NewProc("LsaNtStatusToWinError") +) + +// lsaUnicodeString mirrors LSA_UNICODE_STRING. Length and MaximumLength are +// BYTE counts, not rune counts, which is the usual source of bugs here. +type lsaUnicodeString struct { + Length uint16 + MaximumLength uint16 + Buffer *uint16 +} + +// lsaObjectAttributes mirrors LSA_OBJECT_ATTRIBUTES. Every field except Length +// is unused for LsaOpenPolicy, but the struct must still be the right size. +type lsaObjectAttributes struct { + Length uint32 + RootDirectory windows.Handle + ObjectName *lsaUnicodeString + Attributes uint32 + SecurityDescriptor unsafe.Pointer + SecurityQualityOfService unsafe.Pointer +} + +// newLSAString builds an LSA_UNICODE_STRING over a UTF-16 buffer the caller +// keeps alive. The returned value borrows that buffer, so the buffer must +// outlive every use of the string. +func newLSAString(buffer []uint16) lsaUnicodeString { + if len(buffer) == 0 { + return lsaUnicodeString{} + } + // The buffer is NUL-terminated; the LSA length counts bytes WITHOUT the + // terminator, while MaximumLength counts bytes WITH it. + runes := len(buffer) - 1 + return lsaUnicodeString{ + Length: uint16(runes * 2), + MaximumLength: uint16(len(buffer) * 2), + Buffer: &buffer[0], + } +} + +// lsaStatusError converts an NTSTATUS from an Lsa* call into a Go error, going +// through LsaNtStatusToWinError so the message is the familiar Win32 one rather +// than a raw NTSTATUS. +func lsaStatusError(call string, status uintptr) error { + if status == 0 { + return nil + } + winErr, _, _ := procLsaNtStatusToWinErr.Call(status) + return fmt.Errorf("%s: %w", call, windows.Errno(winErr)) +} + +// grantWindowsSandboxLogonRights gives the principal the batch logon right and +// denies every interactive logon path. Idempotent: LsaAddAccountRights silently +// succeeds when the account already holds a right, so setup can re-run. +// +// Requires an elevated caller. +func grantWindowsSandboxLogonRights(sid *windows.SID) error { + if sid == nil { + return fmt.Errorf("grant sandbox logon rights: nil SID") + } + var attributes lsaObjectAttributes + attributes.Length = uint32(unsafe.Sizeof(attributes)) + var policy windows.Handle + status, _, _ := procLsaOpenPolicy.Call( + 0, // local system + uintptr(unsafe.Pointer(&attributes)), + uintptr(policyCreateAccount|policyLookupNames), + uintptr(unsafe.Pointer(&policy)), + ) + // LsaOpenPolicy borrows the attributes struct by address, so it has to stay + // reachable until the call has returned. + runtime.KeepAlive(attributes) + if err := lsaStatusError("LsaOpenPolicy", status); err != nil { + return err + } + defer procLsaClose.Call(uintptr(policy)) + + rights := []string{ + seBatchLogonRight, + seDenyInteractiveLogonRight, + seDenyNetworkLogonRight, + seDenyRemoteInteractiveRight, + seDenyServiceLogonRightName, + } + // Each right is added on its own call so one unsupported name on an odd SKU + // cannot silently drop the others. + for _, right := range rights { + buffer, err := windows.UTF16FromString(right) + if err != nil { + return err + } + entry := newLSAString(buffer) + status, _, _ := procLsaAddAccountRights.Call( + uintptr(policy), + uintptr(unsafe.Pointer(sid)), + uintptr(unsafe.Pointer(&entry)), + 1, + ) + // Both the descriptor and the buffer it points at are borrowed by the + // call. Kept alive before the error check, not after, so the failure path + // does not return with them already collectable. + runtime.KeepAlive(entry) + runtimeKeepAliveUint16(buffer) + if err := lsaStatusError("LsaAddAccountRights("+right+")", status); err != nil { + return err + } + } + return nil +} + +// revokeWindowsSandboxLogonRights drops every account right held by a principal +// and removes its entry from the LSA policy database. +// +// This is the logon-rights counterpart to revoking ACEs by trustee, and it has +// the same reason to exist: deleting the account on its own leaves the rights +// behind, keyed to a SID that no longer resolves, which is the orphaned residue +// this model is supposed to avoid. It must therefore run BEFORE the account is +// deleted, while the SID is still resolvable. +// +// Removing all rights rather than naming them is deliberate. The principal is +// being retired, so anything keyed to it should go, including rights a previous +// version of setup granted and this one no longer knows about. +// +// Requires an elevated caller. A principal that holds no rights is not an error: +// LsaRemoveAccountRights reports ERROR_FILE_NOT_FOUND for an account with no LSA +// entry, which is the state teardown is trying to reach anyway. +func revokeWindowsSandboxLogonRights(sid *windows.SID) error { + if sid == nil { + return fmt.Errorf("revoke sandbox logon rights: nil SID") + } + var attributes lsaObjectAttributes + attributes.Length = uint32(unsafe.Sizeof(attributes)) + var policy windows.Handle + status, _, _ := procLsaOpenPolicy.Call( + 0, // local system + uintptr(unsafe.Pointer(&attributes)), + uintptr(policyCreateAccount|policyLookupNames), + uintptr(unsafe.Pointer(&policy)), + ) + runtime.KeepAlive(attributes) + if err := lsaStatusError("LsaOpenPolicy", status); err != nil { + return err + } + defer procLsaClose.Call(uintptr(policy)) + + status, _, _ = procLsaRemoveAccountRights.Call( + uintptr(policy), + uintptr(unsafe.Pointer(sid)), + 1, // AllRights: drop everything and delete the LSA account object + 0, // UserRights ignored when AllRights is set + 0, // CountOfRights likewise + ) + runtime.KeepAlive(sid) + if err := lsaStatusError("LsaRemoveAccountRights", status); err != nil { + if errors.Is(err, windows.ERROR_FILE_NOT_FOUND) { + return nil + } + return err + } + return nil +} + +// logonWindowsSandboxPrincipal mints a primary token for the sandbox account. +// The caller owns the returned token and must Close it. +// +// This needs no elevation: the batch logon right granted at setup is what makes +// it work, which is why the per-command path can run unelevated once setup has +// been done once. +func logonWindowsSandboxPrincipal(username string, password string) (windows.Token, error) { + user, err := windows.UTF16PtrFromString(username) + if err != nil { + return 0, err + } + // "." is the local machine, so the lookup never leaves this host even if the + // machine is domain-joined and a same-named domain account exists. + domain, err := windows.UTF16PtrFromString(".") + if err != nil { + return 0, err + } + secret, err := windows.UTF16PtrFromString(password) + if err != nil { + return 0, err + } + var token windows.Token + result, _, callErr := procLogonUserW.Call( + uintptr(unsafe.Pointer(user)), + uintptr(unsafe.Pointer(domain)), + uintptr(unsafe.Pointer(secret)), + logon32LogonBatch, + logon32ProviderDefault, + uintptr(unsafe.Pointer(&token)), + ) + // The three strings are borrowed for the duration of the call. + runtime.KeepAlive(user) + runtime.KeepAlive(domain) + runtime.KeepAlive(secret) + if result == 0 { + if callErr != nil && callErr != windows.ERROR_SUCCESS { + return 0, fmt.Errorf("LogonUser(%s): %w", username, callErr) + } + return 0, fmt.Errorf("LogonUser(%s) failed", username) + } + return token, nil +} + +// runtimeKeepAliveUint16 keeps a UTF-16 buffer reachable across a syscall that +// borrows it. Declared rather than inlined so the intent is explicit at each +// call site; the compiler must not free the slice while LSA holds the pointer. +func runtimeKeepAliveUint16(buffer []uint16) { + if len(buffer) == 0 { + return + } + _ = buffer[0] +} diff --git a/internal/sandbox/windows_identity_logon_windows_test.go b/internal/sandbox/windows_identity_logon_windows_test.go new file mode 100644 index 000000000..503d12922 --- /dev/null +++ b/internal/sandbox/windows_identity_logon_windows_test.go @@ -0,0 +1,41 @@ +//go:build windows + +package sandbox + +import ( + "errors" + "testing" + + "golang.org/x/sys/windows" +) + +// revokeWindowsSandboxLogonRights treats "this account holds no rights" as +// success, because that is the state teardown is trying to reach anyway. That +// relies on the NTSTATUS for it surviving the trip through LsaNtStatusToWinError +// as an error errors.Is can still match, which is exactly the kind of Windows +// errno assumption that quietly turns out to be false. Assert it rather than +// trust it. +// +// Needs no privilege: LsaNtStatusToWinError is a pure status translation, so +// this runs everywhere rather than joining the gated set. +func TestLsaStatusErrorMapsObjectNameNotFound(t *testing.T) { + // STATUS_OBJECT_NAME_NOT_FOUND, what LsaRemoveAccountRights reports for an + // account that has no LSA entry. + const statusObjectNameNotFound = 0xC0000034 + + err := lsaStatusError("LsaRemoveAccountRights", statusObjectNameNotFound) + if err == nil { + t.Fatal("a nonzero NTSTATUS produced no error") + } + if !errors.Is(err, windows.ERROR_FILE_NOT_FOUND) { + t.Fatalf("error = %v, want one errors.Is matches against ERROR_FILE_NOT_FOUND; "+ + "without that, revoking a principal that simply holds no rights fails teardown", err) + } + + // The tolerance must be specific. If any failure matched it, revoke would + // swallow a real one and teardown would report success having done nothing. + const statusAccessDenied = 0xC0000022 + if other := lsaStatusError("LsaRemoveAccountRights", statusAccessDenied); errors.Is(other, windows.ERROR_FILE_NOT_FOUND) { + t.Fatalf("access denied matched the not-found tolerance: %v", other) + } +} diff --git a/internal/sandbox/windows_identity_policy_windows_test.go b/internal/sandbox/windows_identity_policy_windows_test.go new file mode 100644 index 000000000..f22ee57ff --- /dev/null +++ b/internal/sandbox/windows_identity_policy_windows_test.go @@ -0,0 +1,421 @@ +//go:build windows + +package sandbox + +import ( + "errors" + "os" + "path/filepath" + "strings" + "testing" + + "golang.org/x/sys/windows" +) + +// Policy deny-write has to reach the principal plan. +// +// The capability plan has always emitted these. The principal plan denied write +// only on protected metadata and read-only subpaths inside write roots, so once +// the runner used a principal token a policy deny sitting anywhere else was not +// enforced at the OS layer at all, and a shell child could write where the +// restricted-token backend would have stopped it. +func TestPrincipalACLPlanCarriesPolicyDenyWrite(t *testing.T) { + root := filepath.Join(t.TempDir(), "workspace") + denied := filepath.Join(root, "protected", "keep-out") + + plan, err := buildWindowsPrincipalACLPlan(windowsPrincipalACLInput{ + PrincipalSID: "S-1-5-21-0-0-0-1001", + WriteRoots: []WritableRoot{{Root: root}}, + DenyWrite: []string{denied}, + }) + if err != nil { + t.Fatalf("buildWindowsPrincipalACLPlan: %v", err) + } + + entry, ok := findPrincipalACLEntry(plan, WindowsACLDenyWrite, denied) + if !ok { + t.Fatalf("no deny-write ACE for the policy path; plan = %+v", plan.Entries) + } + // Materialized for the same reason the capability plan does it: the applier + // skips targets that do not exist, so a deny on a path created after setup + // would never be written. + if !entry.Materialize { + t.Error("policy deny-write ACE is not materialized, so it is skipped when the path does not exist yet") + } +} + +// Deny-read has to be materialized too, which it was not. +func TestPrincipalACLPlanMaterializesDenyRead(t *testing.T) { + root := filepath.Join(t.TempDir(), "workspace") + secret := filepath.Join(t.TempDir(), "elsewhere", "creds") + + plan, err := buildWindowsPrincipalACLPlan(windowsPrincipalACLInput{ + PrincipalSID: "S-1-5-21-0-0-0-1001", + WriteRoots: []WritableRoot{{Root: root}}, + DenyRead: []string{secret}, + }) + if err != nil { + t.Fatalf("buildWindowsPrincipalACLPlan: %v", err) + } + entry, ok := findPrincipalACLEntry(plan, WindowsACLDenyRead, secret) + if !ok { + t.Fatalf("no deny-read ACE emitted; plan = %+v", plan.Entries) + } + if !entry.Materialize { + t.Fatal("deny-read ACE is not materialized, so a path created after setup never gets one") + } +} + +// Deny entries must still precede the grants they carve out of, which is what +// makes them win under Windows DACL evaluation. Adding deny-write to the plan is +// only safe if it did not disturb that ordering. +func TestPrincipalACLPlanKeepsDeniesBeforeGrants(t *testing.T) { + root := filepath.Join(t.TempDir(), "workspace") + plan, err := buildWindowsPrincipalACLPlan(windowsPrincipalACLInput{ + PrincipalSID: "S-1-5-21-0-0-0-1001", + WriteRoots: []WritableRoot{{Root: root}}, + DenyWrite: []string{filepath.Join(root, "nope")}, + DenyRead: []string{filepath.Join(root, "secret")}, + }) + if err != nil { + t.Fatalf("buildWindowsPrincipalACLPlan: %v", err) + } + firstGrant := -1 + for i, entry := range plan.Entries { + switch entry.Action { + case WindowsACLAllowWrite, WindowsACLAllowRead: + if firstGrant == -1 { + firstGrant = i + } + case WindowsACLDenyRead, WindowsACLDenyWrite: + if firstGrant != -1 { + t.Fatalf("deny entry at %d follows a grant at %d; the grant would win", i, firstGrant) + } + } + } +} + +func findPrincipalACLEntry(plan WindowsACLPlan, action WindowsACLAction, path string) (WindowsACLEntry, bool) { + want := normalizeProfilePath(path) + for _, entry := range plan.Entries { + if entry.Action == action && entry.Path == want { + return entry, true + } + } + return WindowsACLEntry{}, false +} + +// An adopted account must not have its password rotated during provisioning. +// +// Rotating there left every later step running against an account whose password +// had already been replaced with nothing on disk holding it. Any failure in +// between stranded a working principal: the rollback correctly declined to +// delete an account it had not created, so what remained was a live account +// authenticated by a password no longer stored anywhere, and the command path +// read the absent secret as "not provisioned" and quietly fell back to the +// weaker backend. +func TestProvisionWindowsSandboxIdentityDefersPasswordRotation(t *testing.T) { + stubWindowsProvisioning(t, true, nil, nil, nil) + + rotated := false + previous := resetWindowsSandboxUserPasswordFn + t.Cleanup(func() { resetWindowsSandboxUserPasswordFn = previous }) + resetWindowsSandboxUserPasswordFn = func(string, string) error { + rotated = true + return nil + } + + if _, _, _, err := provisionWindowsSandboxIdentity("workspacekey", windowsSandboxRoleOffline); err != nil { + t.Fatalf("provisioning an adopted account: %v", err) + } + if rotated { + t.Fatal("provisioning rotated the password; the window this closes lasts until the secret is committed") + } +} + +// The ownership comment carries the full workspace key, so two workspaces whose +// digests collide in the 11 characters the account name can hold are refused +// rather than silently sharing one account, one secret and one ACL identity. +func TestWindowsSandboxUserCommentDistinguishesWorkspaces(t *testing.T) { + first := windowsSandboxUserCommentFor("aaaaaaaaaaaabbbbbbbb") + second := windowsSandboxUserCommentFor("aaaaaaaaaaaacccccccc") + if first == second { + t.Fatal("two workspaces produced the same ownership comment, so a name collision would be adopted") + } + if !strings.HasPrefix(first, windowsSandboxUserComment) { + t.Fatalf("comment %q lost the marker prefix that identifies it as ours", first) + } + // The names DO collide, which is the whole reason the comment has to carry + // the key. If this stops being true the test is no longer covering anything. + if windowsSandboxUserName("aaaaaaaaaaaabbbbbbbb", windowsSandboxRoleOffline) != windowsSandboxUserName("aaaaaaaaaaaacccccccc", windowsSandboxRoleOffline) { + t.Skip("account names no longer collide for these keys; revisit what this test is for") + } +} + +// Rollback must not strip an adopted principal's logon rights. +// +// revokeWindowsSandboxLogonRights passes AllRights, which drops every right the +// account holds and deletes its LSA object. On an account this run created that +// is a rollback; on one it adopted it destroys the SeBatchLogonRight and +// deny-logon rights an earlier setup established, which is the working +// principal this path exists to preserve. +func TestSetupRollbackRevokesRightsOnlyForCreatedPrincipals(t *testing.T) { + for name, testCase := range map[string]struct { + existed bool + wantRevoked bool + }{ + "adopted principal": {existed: true, wantRevoked: false}, + "created principal": {existed: false, wantRevoked: true}, + } { + t.Run(name, func(t *testing.T) { + stubWindowsProvisioning(t, testCase.existed, nil, nil, nil) + + revoked := false + prevGrant, prevRevoke := grantWindowsSandboxLogonRightsFn, revokeWindowsSandboxLogonRightsFn + t.Cleanup(func() { + grantWindowsSandboxLogonRightsFn, revokeWindowsSandboxLogonRightsFn = prevGrant, prevRevoke + }) + // Fail the grant so the undo path runs with rights already attempted, + // which is the state that used to revoke unconditionally. + grantWindowsSandboxLogonRightsFn = func(*windows.SID) error { + return errors.New("LSA grant refused by policy") + } + revokeWindowsSandboxLogonRightsFn = func(*windows.SID) error { + revoked = true + return nil + } + + config := WindowsSandboxCommandConfig{ + SandboxHome: t.TempDir(), + WorkspaceRoots: []string{`C:\ws`}, + } + if _, _, err := provisionWindowsSandboxPrincipalForSetup(config, windowsSandboxRoleOffline); err == nil { + t.Fatal("provisioning reported success despite an injected grant failure") + } + if revoked != testCase.wantRevoked { + if testCase.wantRevoked { + t.Fatal("rights were not revoked for an account this run created, leaving LSA entries keyed to a SID about to be deleted") + } + t.Fatal("rights were revoked for an adopted account; AllRights drops its pre-existing rights and deletes the LSA object") + } + // And the account itself: requested through the fake for the created + // case, never for an adopted one. This fixture fakes creation, so the + // only acceptable deletion is one the fake sees. + removals := len(stubbedWindowsIdentityRemovals) + if testCase.existed && removals != 0 { + t.Fatalf("rollback deleted an adopted account: %v", stubbedWindowsIdentityRemovals) + } + if !testCase.existed && removals != 1 { + t.Fatalf("rollback of a created account requested %d removals through the fake, want exactly one: %v", removals, stubbedWindowsIdentityRemovals) + } + }) + } +} + +// A secret the current user cannot read is unavailability, not breakage. +// +// The secret's DACL names whoever ran setup. An operator who elevated with a +// separate administrative account, via runas or an over-the-shoulder UAC +// prompt, leaves a secret their ordinary account cannot open. Treating that as a +// hard error made every sandboxed command fail on a machine that was merely set +// up by a different admin; it belongs in the same fail-soft path as a missing +// secret, so the warning fires and the restricted token takes over. +func TestReadWindowsSandboxSecretTreatsPermissionDeniedAsUnavailable(t *testing.T) { + previous := readWindowsSandboxSecretFile + t.Cleanup(func() { readWindowsSandboxSecretFile = previous }) + readWindowsSandboxSecretFile = func(string) ([]byte, error) { + return nil, &os.PathError{Op: "open", Path: "secret", Err: windows.ERROR_ACCESS_DENIED} + } + + if _, err := readWindowsSandboxSecret(`C:\anything.secret`); err != errWindowsSandboxIdentityUnavailable { + t.Fatalf("permission-denied read returned %v, want errWindowsSandboxIdentityUnavailable so the command falls back", err) + } +} + +// A workspace must not bind to another workspace's principal. +// +// The account name carries only 11 characters of the workspace digest, so two +// workspaces can derive the same name. Provisioning refuses that case by +// checking the full key in the account comment, but the command path resolved +// the name straight to a SID. The workspace that lost the race would have failed +// setup and then quietly run as the other one's principal, using its secret and +// its ACL identity. +func TestLookupWindowsSandboxIdentityRejectsForeignWorkspace(t *testing.T) { + previous := windowsSandboxUserIsManagedFn + t.Cleanup(func() { windowsSandboxUserIsManagedFn = previous }) + + prevSID := resolveWindowsSandboxSIDFn + t.Cleanup(func() { resolveWindowsSandboxSIDFn = prevSID }) + // The account resolves; whether it BELONGS to this workspace is the question. + resolveWindowsSandboxSIDFn = func(string) (*windows.SID, error) { + return windows.CreateWellKnownSid(windows.WinLocalSystemSid) + } + + var askedKey string + windowsSandboxUserIsManagedFn = func(_ string, workspaceKey string) (bool, error) { + askedKey = workspaceKey + return false, nil + } + _, err := lookupWindowsSandboxIdentity("workspacekey", windowsSandboxRoleOffline) + if err == nil { + t.Fatal("lookup accepted an account belonging to another workspace") + } + if errors.Is(err, errWindowsSandboxIdentityUnavailable) { + t.Fatal("a foreign account must not read as unprovisioned; that would silently fall back instead of reporting the conflict") + } + if askedKey != "" && askedKey != "workspacekey" { + t.Fatalf("ownership was checked against %q, want the caller's workspace key", askedKey) + } +} + +// An account that does not exist must stay the unavailable sentinel rather than +// becoming a collision error, since that is the ordinary not-set-up state. +func TestLookupWindowsSandboxIdentityAbsentAccountIsUnavailable(t *testing.T) { + previous := windowsSandboxUserIsManagedFn + t.Cleanup(func() { windowsSandboxUserIsManagedFn = previous }) + windowsSandboxUserIsManagedFn = func(string, string) (bool, error) { + t.Fatal("ownership must not be consulted for an account that does not resolve") + return false, nil + } + if _, err := lookupWindowsSandboxIdentity("nosuchworkspacekey", windowsSandboxRoleOffline); !errors.Is(err, errWindowsSandboxIdentityUnavailable) { + t.Fatalf("absent account returned %v, want errWindowsSandboxIdentityUnavailable", err) + } +} + +// The git control-plane carveouts must be materialized. +// +// gitMetadataWriteCarveouts supplies .git/config and .git/hooks as +// ReadOnlySubpaths of the workspace. On a workspace where git has not run yet +// they do not exist when setup applies the plan, and applyWindowsACLPlan skips +// an absent target, so the deny ACEs were never written. Once git created those +// paths the principal still held inherited write access and could install a +// hook or rewrite credential.helper. +func TestPrincipalACLPlanMaterializesReadOnlySubpaths(t *testing.T) { + root := filepath.Join(t.TempDir(), "workspace") + carveout := filepath.Join(root, ".git", "config") + + plan, err := buildWindowsPrincipalACLPlan(windowsPrincipalACLInput{ + PrincipalSID: "S-1-5-21-0-0-0-1001", + WriteRoots: []WritableRoot{{Root: root, ReadOnlySubpaths: []string{carveout}}}, + }) + if err != nil { + t.Fatalf("buildWindowsPrincipalACLPlan: %v", err) + } + entry, ok := findPrincipalACLEntry(plan, WindowsACLDenyWrite, carveout) + if !ok { + t.Fatalf("no deny-write ACE for the git carveout; plan = %+v", plan.Entries) + } + if !entry.Materialize { + t.Fatal("git carveout deny-write is not materialized, so it is skipped on a workspace where .git does not exist yet") + } +} + +// Setup must grant the principal the same runtime root that commands write to. +// +// permissionProfileWithRuntime appends this root to WriteRoots on every command +// and redirects HOME, GOCACHE and npm_config_cache into it, but it lives under +// the user cache rather than the workspace, so the profile setup sees never +// contains it. A principal is a separate account with no rights there, so +// without a grant every npm install or go build fails on a cache write. +// +// The assertion that matters is that the two derivations agree. If they drift, +// setup writes the ACE on one directory while commands use another, and the +// symptom is a bare ACCESS_DENIED with nothing pointing at the sandbox. +func TestSetupGrantsTheRuntimeRootCommandsActuallyUse(t *testing.T) { + isolateSandboxRuntimeRoots(t) + workspace := filepath.Join(t.TempDir(), "ws") + if err := os.MkdirAll(workspace, 0o700); err != nil { + t.Fatal(err) + } + + granted, err := setupWindowsSandboxRuntimeRoot(WindowsSandboxCommandConfig{ + WorkspaceRoots: []string{workspace}, + }) + if err != nil { + t.Fatalf("setupWindowsSandboxRuntimeRoot: %v", err) + } + if len(granted) == 0 { + t.Fatal("no runtime root resolved for a configured workspace") + } + // EVERY granted candidate has to exist, not just the one this process would + // pick: applyWindowsACLPlan fails the whole run on a target that is absent. + for _, root := range granted { + if info, err := os.Stat(root); err != nil || !info.IsDir() { + t.Fatalf("runtime root %q was not created; applyWindowsACLPlan skips absent targets so the grant would no-op (stat err %v)", root, err) + } + } + + // What a command would actually use. + runtimeState, release, err := prepareSandboxRuntime(workspace) + if err != nil { + t.Fatalf("prepareSandboxRuntime: %v", err) + } + if release != nil { + defer release() + } + // The command's choice has to be one setup provisioned. Setup grants the whole + // candidate set precisely because the choice is made per process, so the + // contract is membership rather than equality. + if !grantedRuntimeRootsCover(granted, runtimeState.Root) { + t.Fatalf("setup granted %q but commands write to %q", granted, runtimeState.Root) + } +} + +// grantedRuntimeRootsCover reports whether the root a command selected is one of +// the roots setup granted an ACE on. +func grantedRuntimeRootsCover(granted []string, selected string) bool { + for _, root := range granted { + if filepath.Clean(root) == filepath.Clean(selected) { + return true + } + } + return false +} + +// No workspace root means nothing to grant, which is not an error. +func TestSetupRuntimeRootWithoutWorkspaceIsNotAnError(t *testing.T) { + isolateSandboxRuntimeRoots(t) + granted, err := setupWindowsSandboxRuntimeRoot(WindowsSandboxCommandConfig{}) + if err != nil { + t.Fatalf("no workspace root should not error: %v", err) + } + if len(granted) != 0 { + t.Fatalf("granted %q with no workspace configured", granted) + } +} + +// An account in a privileged group must not be adopted, even when name and +// ownership comment say it is ours. +// +// Adoption resets the password and hands the account to the sandbox. If that +// account is also in Administrators, the sandbox gains the rights the sandbox +// exists to withhold: rewriting the ACLs confining it, reading the secret locked +// to the invoking user, and stopping Zero. +func TestProvisionWindowsSandboxIdentityRefusesPrivilegedAccount(t *testing.T) { + stubWindowsProvisioning(t, true, nil, nil, nil) + + previous := windowsSandboxUserIsPrivilegedFn + t.Cleanup(func() { windowsSandboxUserIsPrivilegedFn = previous }) + windowsSandboxUserIsPrivilegedFn = func(string) (bool, error) { return true, nil } + + _, _, created, err := provisionWindowsSandboxIdentity("workspacekey", windowsSandboxRoleOffline) + if !errors.Is(err, errWindowsSandboxPrivilegedAccount) { + t.Fatalf("provisioning adopted a privileged account, err = %v", err) + } + if created { + t.Fatal("created must stay false for an account this run refused to adopt") + } +} + +// The ordinary adopted account is unaffected. +func TestProvisionWindowsSandboxIdentityAdoptsUnprivilegedAccount(t *testing.T) { + stubWindowsProvisioning(t, true, nil, nil, nil) + + previous := windowsSandboxUserIsPrivilegedFn + t.Cleanup(func() { windowsSandboxUserIsPrivilegedFn = previous }) + windowsSandboxUserIsPrivilegedFn = func(string) (bool, error) { return false, nil } + + if _, _, _, err := provisionWindowsSandboxIdentity("workspacekey", windowsSandboxRoleOffline); err != nil { + t.Fatalf("an unprivileged managed account must still be adopted: %v", err) + } +} diff --git a/internal/sandbox/windows_identity_privilege_recheck_windows_test.go b/internal/sandbox/windows_identity_privilege_recheck_windows_test.go new file mode 100644 index 000000000..796c80831 --- /dev/null +++ b/internal/sandbox/windows_identity_privilege_recheck_windows_test.go @@ -0,0 +1,92 @@ +//go:build windows + +package sandbox + +import ( + "errors" + "testing" + + "golang.org/x/sys/windows" +) + +// Group membership is not frozen at setup. An account provisioned clean can be +// added to Administrators afterwards — by an operator, or by an attacker who +// already has that access and wants the sandbox to hand it back. Provisioning's +// refusal cannot see that; only the path that mints the token can. +func TestLookupPrincipalForCommandRefusesAnAccountThatBecamePrivileged(t *testing.T) { + sid, err := windows.CreateWellKnownSid(windows.WinLocalSystemSid) + if err != nil { + t.Fatalf("CreateWellKnownSid: %v", err) + } + restoreLookupSeams(t, sid) + + privilegedCalls := 0 + windowsSandboxUserIsPrivilegedFn = func(string) (bool, error) { + privilegedCalls++ + return true, nil + } + + _, err = lookupWindowsSandboxPrincipalForCommand("workspace-key", windowsSandboxRoleOnline) + if !errors.Is(err, errWindowsSandboxPrivilegedAccount) { + t.Fatalf("err = %v, want errWindowsSandboxPrivilegedAccount", err) + } + if privilegedCalls != 1 { + t.Errorf("privileged check ran %d times, want exactly 1", privilegedCalls) + } + // It must be a hard refusal, not the unavailable sentinel — that one is the + // quiet "not provisioned" fallback and would silently drop the sandbox back + // to the restricted token instead of telling the operator. + if errors.Is(err, errWindowsSandboxIdentityUnavailable) { + t.Error("privileged refusal must not read as the not-provisioned fallback") + } +} + +func TestLookupPrincipalForCommandAcceptsAnUnprivilegedAccount(t *testing.T) { + sid, err := windows.CreateWellKnownSid(windows.WinLocalSystemSid) + if err != nil { + t.Fatalf("CreateWellKnownSid: %v", err) + } + restoreLookupSeams(t, sid) + windowsSandboxUserIsPrivilegedFn = func(string) (bool, error) { return false, nil } + + identity, err := lookupWindowsSandboxPrincipalForCommand("workspace-key", windowsSandboxRoleOnline) + if err != nil { + t.Fatalf("lookupWindowsSandboxPrincipalForCommand: %v", err) + } + if identity.Username == "" { + t.Error("expected the resolved principal") + } +} + +// Teardown must stay able to clean up an account that has become privileged. +// If the refusal lived inside lookupWindowsSandboxIdentity, the account this +// guard exists to catch would become undeletable by Zero. +func TestLookupIdentityItselfDoesNotConsultPrivilege(t *testing.T) { + sid, err := windows.CreateWellKnownSid(windows.WinLocalSystemSid) + if err != nil { + t.Fatalf("CreateWellKnownSid: %v", err) + } + restoreLookupSeams(t, sid) + windowsSandboxUserIsPrivilegedFn = func(string) (bool, error) { + t.Error("teardown's lookup must not be gated on privilege") + return true, nil + } + + if _, err := lookupWindowsSandboxIdentity("workspace-key", windowsSandboxRoleOnline); err != nil { + t.Fatalf("lookupWindowsSandboxIdentity: %v", err) + } +} + +func restoreLookupSeams(t *testing.T, sid *windows.SID) { + t.Helper() + prevResolve := resolveWindowsSandboxSIDFn + prevManaged := windowsSandboxUserIsManagedFn + prevPrivileged := windowsSandboxUserIsPrivilegedFn + t.Cleanup(func() { + resolveWindowsSandboxSIDFn = prevResolve + windowsSandboxUserIsManagedFn = prevManaged + windowsSandboxUserIsPrivilegedFn = prevPrivileged + }) + resolveWindowsSandboxSIDFn = func(string) (*windows.SID, error) { return sid, nil } + windowsSandboxUserIsManagedFn = func(string, string) (bool, error) { return true, nil } +} diff --git a/internal/sandbox/windows_identity_rollback_windows_test.go b/internal/sandbox/windows_identity_rollback_windows_test.go new file mode 100644 index 000000000..09d3d47f1 --- /dev/null +++ b/internal/sandbox/windows_identity_rollback_windows_test.go @@ -0,0 +1,203 @@ +//go:build windows + +package sandbox + +import ( + "errors" + "testing" + + "golang.org/x/sys/windows" +) + +// stubWindowsProvisioning replaces the four provisioning syscalls so the +// function can run on an ordinary machine. Every one of them needs an elevated +// caller and a real local account, so without this the test would stop at the +// first call and never reach the behaviour it is named for. +func stubWindowsProvisioning(t *testing.T, existed bool, groupErr error, offlineErr error, sidErr error) { + t.Helper() + prevGroup, prevUser := ensureWindowsSandboxGroupFn, ensureWindowsSandboxUserFn + prevOfflineGroup := ensureWindowsSandboxOfflineGroupFn + prevAdd, prevSID := addWindowsSandboxUserToGroupFn, resolveWindowsSandboxSIDFn + prevManaged := windowsSandboxUserIsManagedFn + prevOffline := addWindowsSandboxUserToOfflineGroupFn + prevReset := resetWindowsSandboxUserPasswordFn + prevRemove, prevUpgrade := removeWindowsSandboxIdentityFn, upgradeWindowsSandboxUserCommentFn + stubbedWindowsIdentityRemovals = nil + t.Cleanup(func() { + ensureWindowsSandboxGroupFn, ensureWindowsSandboxUserFn = prevGroup, prevUser + ensureWindowsSandboxOfflineGroupFn = prevOfflineGroup + addWindowsSandboxUserToGroupFn, resolveWindowsSandboxSIDFn = prevAdd, prevSID + windowsSandboxUserIsManagedFn = prevManaged + addWindowsSandboxUserToOfflineGroupFn = prevOffline + resetWindowsSandboxUserPasswordFn = prevReset + removeWindowsSandboxIdentityFn, upgradeWindowsSandboxUserCommentFn = prevRemove, prevUpgrade + }) + + // THE INVERSE IS FAKED TOO. Creation is faked and ownership is faked, so a + // rollback branch that trusts both would otherwise call the real NetUserDel + // on the current user's deterministic principal for C:ws, an account this + // fixture never created. Removals are recorded for the tests to assert on. + removeWindowsSandboxIdentityFn = func(username, workspaceKey string) error { + stubbedWindowsIdentityRemovals = append(stubbedWindowsIdentityRemovals, username+"|"+workspaceKey) + return nil + } + upgradeWindowsSandboxUserCommentFn = func(string, string) error { return nil } + ensureWindowsSandboxGroupFn = func() error { return nil } + ensureWindowsSandboxOfflineGroupFn = func() error { return nil } + // Adopted accounts are ours in these tests; the ownership check is a real + // syscall and would otherwise refuse before the code under test runs. + windowsSandboxUserIsManagedFn = func(string, string) (bool, error) { return true, nil } + // Stubbed even though provisioning no longer rotates: rotation moved to the + // caller, so nothing under test reaches this today. Left in so that if it + // ever moves back, these tests fail rather than resetting the password of a + // real managed account that happens to exist on the machine running them. + resetWindowsSandboxUserPasswordFn = func(string, string) error { + t.Fatal("provisioning must not rotate an account password; rotation belongs to the caller, immediately before the secret is written") + return nil + } + ensureWindowsSandboxUserFn = func(string, string, string) (bool, error) { return existed, nil } + addWindowsSandboxUserToGroupFn = func(string) error { return groupErr } + addWindowsSandboxUserToOfflineGroupFn = func(string) error { return offlineErr } + resolveWindowsSandboxSIDFn = func(username string) (*windows.SID, error) { + if sidErr != nil { + return nil, sidErr + } + return windows.CreateWellKnownSid(windows.WinLocalSystemSid) + } +} + +// A failure after the account has been created must still hand back the name. +// +// The caller's rollback deletes by identity.Username, so returning a zero +// identity alongside created=true asked it to delete "" and quietly left the +// account this run had just made. Group attachment is the case that matters +// most: it is the enforcement boundary and it can fail under local policy. +func TestProvisionWindowsSandboxIdentityReturnsNameForRollback(t *testing.T) { + groupFailure := errors.New("group attachment refused by policy") + offlineFailure := errors.New("offline group attachment refused by policy") + sidFailure := errors.New("sid lookup failed") + + for name, testCase := range map[string]struct { + groupErr error + offlineErr error + sidErr error + }{ + "group attachment fails": {groupErr: groupFailure}, + "offline group attachment fails": {offlineErr: offlineFailure}, + "sid resolution fails": {sidErr: sidFailure}, + } { + t.Run(name, func(t *testing.T) { + stubWindowsProvisioning(t, false, testCase.groupErr, testCase.offlineErr, testCase.sidErr) + + identity, _, created, err := provisionWindowsSandboxIdentity("workspacekey", windowsSandboxRoleOffline) + if err == nil { + t.Fatal("provisioning reported success despite an injected failure") + } + if !created { + t.Fatal("created = false, so the rollback would skip an account this run made") + } + // The whole point: without a name there is nothing to delete. + if identity.Username == "" { + t.Fatal("identity carries no username, so the rollback deletes \"\" and strands the account") + } + if want := windowsSandboxUserName("workspacekey", windowsSandboxRoleOffline); identity.Username != want { + t.Fatalf("username = %q, want %q", identity.Username, want) + } + }) + } +} + +// An account that already existed must not be deleted because a later step +// failed. created=false is what stops the rollback turning a partial failure +// into the loss of a working principal from an earlier setup. +func TestProvisionWindowsSandboxIdentityDoesNotClaimPreexistingAccount(t *testing.T) { + stubWindowsProvisioning(t, true, errors.New("group attachment refused"), nil, nil) + + _, _, created, err := provisionWindowsSandboxIdentity("workspacekey", windowsSandboxRoleOffline) + if err == nil { + t.Fatal("provisioning reported success despite an injected failure") + } + // created is the whole assertion. It is what stops the rollback deleting an + // account it did not make, and it must stay false however provisioning + // fails. The identity is deliberately not checked: an adopted account exits + // early at the ownership check, which is a real syscall and not stubbed + // here, so asserting on the name would be testing the stub rather than the + // contract. + if created { + t.Fatal("created = true for an account this run did not create; rollback would delete a working principal") + } +} + +// The write grant has to include delete. +// +// FILE_GENERIC_WRITE covers creating and modifying but not removing or +// renaming, and a rename needs delete on the source. The old same-user token hid +// this because the caller already held inherited rights on its own tree; a +// principal is a separate account with none, so without these it can write files +// it can never remove, which fails ordinary editing and most git operations +// rather than an edge case. +func TestWindowsACLAllowWriteGrantsDelete(t *testing.T) { + mode, mask, err := windowsACLAccess(WindowsACLAllowWrite) + if err != nil { + t.Fatalf("windowsACLAccess: %v", err) + } + if mode != windows.GRANT_ACCESS { + t.Fatalf("mode = %v, want GRANT_ACCESS", mode) + } + // Atomic bits only. FILE_GENERIC_READ and FILE_GENERIC_WRITE both carry + // READ_CONTROL and SYNCHRONIZE, so testing a composite constant with & is + // satisfied by any grant at all and proves nothing. + for label, bit := range map[string]windows.ACCESS_MASK{ + "DELETE": windows.DELETE, + "FILE_WRITE_DATA": windows.FILE_WRITE_DATA, + } { + if mask&bit == 0 { + t.Errorf("write grant is missing %s", label) + } + } + // Granting these would let the principal rewrite the very restrictions + // placed on it. They are in the deny mask for that reason and must not + // appear here. + // + // FILE_DELETE_CHILD belongs in this set and was originally in the one + // above, on the reasoning that the grant should mirror the deny mask. On a + // parent it authorises deleting a child whatever the child's own DACL says, + // so on a write root it hands back the write-denied carve-outs underneath: + // delete .git/config, recreate it, and the replacement inherits the grant + // with no deny of its own. Mirroring the deny mask is the wrong instinct — + // denying a capability is not a reason to grant it. + for label, bit := range map[string]windows.ACCESS_MASK{ + "WRITE_DAC": windows.WRITE_DAC, + "WRITE_OWNER": windows.WRITE_OWNER, + "FILE_DELETE_CHILD": windowsFileDeleteChild, + } { + if mask&bit != 0 { + t.Errorf("write grant unexpectedly includes %s", label) + } + } +} + +// The read grant must stay read-only. Widening the write mask above is only +// safe if this one did not move with it. +func TestWindowsACLAllowReadGrantsNoDelete(t *testing.T) { + _, mask, err := windowsACLAccess(WindowsACLAllowRead) + if err != nil { + t.Fatalf("windowsACLAccess: %v", err) + } + // Atomic bits, for the same reason as above: the read and write composites + // overlap on the standard rights, so a composite check here would report a + // failure that is not real. + for label, bit := range map[string]windows.ACCESS_MASK{ + "DELETE": windows.DELETE, + "FILE_DELETE_CHILD": windowsFileDeleteChild, + "FILE_WRITE_DATA": windows.FILE_WRITE_DATA, + } { + if mask&bit != 0 { + t.Errorf("read grant unexpectedly includes %s", label) + } + } +} + +// stubbedWindowsIdentityRemovals records every account deletion the mocked +// provisioning fixture was asked for, newest last, as "username|workspaceKey". +var stubbedWindowsIdentityRemovals []string diff --git a/internal/sandbox/windows_identity_runtime_windows.go b/internal/sandbox/windows_identity_runtime_windows.go new file mode 100644 index 000000000..5f4631d50 --- /dev/null +++ b/internal/sandbox/windows_identity_runtime_windows.go @@ -0,0 +1,1024 @@ +//go:build windows + +package sandbox + +// Using a sandbox principal at command time. +// +// This is the seam between the identity model and the existing runner. It is +// deliberately fail-soft: when no principal is provisioned, when the secret is +// missing, or when the opt-in is off, it reports "not available" and the caller +// keeps using today's restricted-token backend. Only an outright failure to log +// on with a principal that IS provisioned surfaces as an error, because that +// means setup ran but the identity is broken, and silently downgrading the +// sandbox in that case would be the wrong kind of quiet. + +import ( + "crypto/sha256" + "encoding/hex" + "errors" + "fmt" + "os" + "strings" + "sync" + + "golang.org/x/sys/windows" +) + +// The opt-in itself (windowsSandboxIdentityEnv and windowsSandboxIdentityEnabled) +// lives in windows_setup.go: it is part of the setup protocol, which the elevated +// half and the command half both have to read the same way, so it cannot be +// Windows-only. + +// windowsSandboxWorkspaceKey derives the per-workspace key a principal is named +// after. It hashes the workspace root the same way the sandbox runtime keys its +// own state, so the account name leaks no path and one workspace always maps to +// one principal. +func windowsSandboxWorkspaceKey(workspaceRoots []string) string { + root := "" + for _, candidate := range workspaceRoots { + if trimmed := strings.TrimSpace(candidate); trimmed != "" { + root = normalizeProfilePath(trimmed) + break + } + } + if root == "" { + root = "default" + } + digest := sha256.Sum256([]byte(strings.ToLower(root))) + return hex.EncodeToString(digest[:]) +} + +// windowsSandboxPrincipalEligible reports whether the principal backend may be +// used for this command at all, before any account or secret is consulted. +// +// Kept separate from the lookup so the decision is observable on its own: on a +// machine with nothing provisioned the lookup declines anyway, which would let a +// missing guard here pass unnoticed. +func windowsSandboxPrincipalEligible(config WindowsSandboxCommandConfig) bool { + return windowsSandboxIdentityEnabled(config.Env) +} + +// windowsSandboxRoleForNetwork maps a command's network mode onto the principal +// that enforces it. +// +// The two roles differ only in membership of the offline group, which the block +// filters are keyed to. That indirection exists because network denial cannot be +// expressed on a logon token the way it is on a restricted token: the restricted +// token carries the offline-marker SID as a restricting SID, and LogonUser has +// no equivalent, since it builds a token from the account's real memberships. +// +// Defaulting anything that is not an explicit allow to the offline principal is +// deliberate. A mode this does not recognise should lose the network, not keep +// it. +func windowsSandboxRoleForNetwork(mode NetworkMode) windowsSandboxRole { + // Delegated so this and `zero doctor` answer from one rule. Doctor reporting a + // different account than the runtime actually uses is the drift that made the + // old inactive-reason helper wrong. + if WindowsSandboxPrincipalRoleForNetwork(mode) == string(windowsSandboxRoleOnline) { + return windowsSandboxRoleOnline + } + return windowsSandboxRoleOffline +} + +// windowsSandboxPrincipalToken returns a token for this workspace's sandbox +// principal. +// +// ok is false, with a nil error, whenever the principal backend simply is not in +// play: the opt-in is off, setup has not provisioned an account, or no secret is +// stored. The caller falls back to the restricted token in those cases. An error +// means the identity exists but could not be used, which is worth surfacing +// rather than downgrading around. +func windowsSandboxPrincipalToken(config WindowsSandboxCommandConfig) (windows.Token, bool, error) { + if !windowsSandboxPrincipalEligible(config) { + // Deliberately silent, and this is a change of mind worth recording. + // + // Announcing it looks right: deny is the DEFAULT network mode, so an + // operator who opted in never gets a principal for ordinary commands, and + // that is worth knowing. But the warning cannot be delivered here. This + // runner is re-exec'd per command as `zero __windows-command-runner`, so the + // sync.Once below is once per COMMAND, not once per session — the notice + // would land on the stderr of essentially every sandboxed tool call. Noise + // that repeats gets filtered by the reader rather than acted on, which is + // the exact failure the helper's own comment warns about. + // + // It is also not actionable per command: windowsSandboxPrincipalEligible + // prefers network enforcement over read confinement on purpose, so there is + // nothing to do differently. A standing configuration fact belongs on a + // surface read once — `zero doctor`, which carries the opt-in now. + return 0, false, nil + } + key := windowsSandboxPrincipalKey(config) + // The network mode picks the principal. Both are provisioned by setup; the + // offline one is in the group the block filters match, so choosing here is + // what enforces the mode. + role := windowsSandboxRoleForNetwork(config.PermissionProfile.Network.Mode) + identity, err := lookupWindowsSandboxPrincipalForCommandFn(key, role) + if err != nil { + if errors.Is(err, errWindowsSandboxIdentityUnavailable) { + // Not provisioned. On the restricted-token tier the marker check has + // already refused a command whose opt-in disagrees with setup, so reaching + // here means the unelevated tier, which validates no marker at all and + // cannot provision an account (that needs Administrator). Falling back is + // right — refusing would break every machine-wide opt-in that relies on + // the unelevated tier — but it must not be silent. + warnWindowsSandboxPrincipalNotUsed("no sandbox principal is provisioned for this workspace; `zero sandbox setup` from an elevated (Administrator) terminal provisions one") + return 0, false, nil + } + // The name resolves to something that is not a usable principal, most + // likely squatted by a local group or alias. That is a conflict an + // operator has to see, not a reason to pretend setup never ran. + return 0, false, err + } + // Selecting the offline account is only half of what denies it the network. + // The WFP filters match the OFFLINE GROUP'S SID, so an account that has + // drifted out of that group — local policy, an administrator, a re-setup that + // could not re-add it — still logs on from a stored secret and its token no + // longer satisfies the filter condition. It would get full egress under a + // profile that asked for none. Re-check the membership the mode depends on + // rather than trusting the marker written when setup last succeeded. + if role == windowsSandboxRoleOffline { + member, err := windowsSandboxUserInLocalGroupFn(identity.Username, windowsSandboxOfflineGroupName) + if err != nil { + return 0, false, err + } + if !member { + // Fail closed toward the weaker-but-still-denied path: the restricted + // token carries the offline marker the same filters match, so egress + // stays blocked. Handing back this principal token would not. + warnWindowsSandboxOfflineMembershipMissing(identity.Username) + return 0, false, nil + } + } + secretPath, err := windowsSandboxSecretPath(config.SandboxHome, identity.Username) + if err != nil { + return 0, false, err + } + password, err := readWindowsSandboxSecretFn(secretPath) + if err != nil { + if errors.Is(err, errWindowsSandboxIdentityUnavailable) { + // The account exists but its password does not. Setup was interrupted + // or the secret was removed; fall back rather than fail the command. + // + // Falling back is right, staying quiet about it was not. The opt-in is + // set and an account IS provisioned, so the operator asked for + // principal isolation and is silently getting the weaker same-user + // restricted token instead. That is the one fail-soft case worth + // announcing: the others mean the backend was never set up, while this + // one means it was and has broken since. + warnWindowsSandboxPrincipalUnavailable(identity.Username) + return 0, false, nil + } + return 0, false, err + } + token, err := logonWindowsSandboxPrincipal(identity.Username, password) + if err != nil { + // Provisioned but unusable. Surface it: a wrong password or a revoked + // batch-logon right is a broken sandbox, not an absent one. + return 0, false, err + } + return token, true, nil +} + +// warnWindowsSandboxPrincipalUnavailable tells the operator once per process +// that the backend they opted into is not the one running. +// +// Once, because this sits on the command path: a warning per command would be +// noise on every tool call for the whole session, and noise that repeats gets +// filtered out by the reader rather than acted on. Indirected through a var so a +// test can observe it without capturing stderr. +var warnWindowsSandboxPrincipalUnavailable = func(username string) { + windowsSandboxPrincipalWarnOnce.Do(func() { + fmt.Fprintf(os.Stderr, + "[zero] %s is set and sandbox principal %q is provisioned, but its stored password is missing or unreadable. "+ + "Falling back to the restricted-token sandbox, which does not confine reads. "+ + "Re-run `zero sandbox setup` from an elevated terminal to restore it.\n", + windowsSandboxIdentityEnv, username) + }) +} + +var windowsSandboxPrincipalWarnOnce sync.Once + +// warnWindowsSandboxPrincipalNotUsed covers the other ways an opted-in command +// ends up on the restricted token: the principal is ineligible for this +// command's policy, or none is provisioned on a tier that validates no marker. +// Neither is an error — both are correct fallbacks — but both leave the operator +// believing an account boundary is isolating them when it is not, which is the +// one thing this backend must never do quietly. +// +// Once per process and behind its own sync.Once, so it neither silences nor is +// silenced by the provisioned-but-secretless warning above. +var warnWindowsSandboxPrincipalNotUsed = func(reason string) { + windowsSandboxPrincipalNotUsedWarnOnce.Do(func() { + fmt.Fprintf(os.Stderr, + "[zero] %s is set, but this command is not running as a sandbox principal: %s. "+ + "Falling back to the restricted-token sandbox, which does not confine reads.\n", + windowsSandboxIdentityEnv, reason) + }) +} + +var windowsSandboxPrincipalNotUsedWarnOnce sync.Once + +// Seamed so the dual-role setup rollback can be exercised without an elevated +// machine. The contract under test is which principals the outer rollback is +// willing to delete, which is a data-loss decision and the one thing here worth +// proving rather than reasoning about. +var ( + provisionWindowsSandboxPrincipalForSetupFn = provisionWindowsSandboxPrincipalForSetup + removeWindowsSandboxPrincipalForSetupFn = removeWindowsSandboxPrincipalForSetup + // The both-roles retirement setup runs when the opt-in is off. It belongs + // beside its single-role sibling for the reason recorded below: one seam + // declared in two places lets a test stub one while production uses the + // other. + removeWindowsSandboxPrincipalsForSetupFn = removeWindowsSandboxPrincipalsForSetup + // applyWindowsACLPlanFn lives with the other identity seams in + // windows_identity_windows.go — the ACL-ordering guarantee it exists for + // predates the dual-role split, and two declarations of the same seam would + // let a test stub one while production used the other. +) + +// provisionWindowsSandboxPrincipalForSetup does the elevated half: create the +// account, grant it the batch logon right, and store its password locked to the +// invoking user. Called from `zero sandbox setup`. +// +// The password is written BEFORE the caller applies any ACL plan, so a setup +// that fails partway leaves a principal that can at least be logged on and +// therefore cleaned up, rather than an account nothing holds the secret for. +func provisionWindowsSandboxPrincipalForSetup(config WindowsSandboxCommandConfig, role windowsSandboxRole) (windowsSandboxIdentity, bool, error) { + key := windowsSandboxPrincipalKey(config) + identity, password, created, err := provisionWindowsSandboxIdentityFn(key, role) + + // Undo whatever this run actually did, in reverse, on any failure after the + // account exists. Without it a failure between creating the account and + // storing its secret left the account behind with no caller able to remove + // it: the rollback the setup path installs is only built once this function + // has returned successfully. + // + // Scoped to what THIS run created on purpose. An account that already existed + // and belongs to Zero is a working principal from an earlier setup, and + // deleting it because a later run failed would turn a partial failure into a + // total one. + rightsAttempted := false + rotated := false + // Resolved from the account name rather than the identity, so it is known + // before anything can fail. + secretPath, secretPathErr := windowsSandboxSecretPath(config.SandboxHome, windowsSandboxUserName(key, role)) + undo := func() error { + // Only when this run invalidated it. The secret is removed if this run + // created the account, or if it rotated an existing account's password, + // because in both cases what is on disk cannot authenticate and absent + // beats stale: the command path treats a missing secret as "not + // provisioned" and falls back, while a stale one fails the logon and + // reports a broken sandbox. + // + // Removing it unconditionally, as this used to, destroyed a WORKING + // secret whenever setup failed before rotation on a machine that was + // already provisioned. The account kept its old password, the only copy + // of it was deleted, and the sandbox silently degraded. + // + // This one failure is reported rather than swallowed. The others below + // leave residue; this one leaves a CREDENTIAL for a password that no + // longer works, which is the stale-secret state the invariant above + // exists to prevent. Staying quiet would claim the invariant was restored + // when it was not, and the operator would instead meet a + // provisioned-but-unusable principal on the next command. + var cleanupErr error + if secretPath != "" && (created || rotated) { + if err := removeWindowsSandboxSecretFn(secretPath); err != nil { + cleanupErr = fmt.Errorf( + "sandbox secret %s is stale and could not be removed, so the next command will "+ + "fail to log the principal on rather than falling back; delete it and re-run "+ + "`zero sandbox setup`: %w", secretPath, err) + } + } + // Only for an account this run created, and attempted rather than + // completed. + // + // Attempted, because grantWindowsSandboxLogonRights adds rights one at a + // time and returns on the first failure, so a partial grant is possible + // and gating on success left those entries behind, keyed to a SID that + // deleting the account then made unresolvable. + // + // Created, because revokeWindowsSandboxLogonRights passes AllRights, which + // drops every right the account holds and deletes its LSA object outright. + // On an adopted principal that is not a rollback, it is destruction: a + // transient failure anywhere below would strip the SeBatchLogonRight and + // deny-logon rights a previous setup established, leaving exactly the + // broken-but-present principal this whole function exists to avoid. The + // rights this run granted are the ones the account is supposed to have, so + // leaving them in place on an adopted account is the safe direction. + if identity.SID != nil && rightsAttempted && created { + _ = revokeWindowsSandboxLogonRightsFn(identity.SID) + } + if created { + _ = removeWindowsSandboxIdentityFn(identity.Username, key) + } + return cleanupErr + } + + if err != nil { + // provisionWindowsSandboxIdentity can fail after creating the account, so + // this path needs the same cleanup even though nothing below ran. + return windowsSandboxIdentity{}, false, errors.Join(err, undo()) + } + rightsAttempted = true + if err := grantWindowsSandboxLogonRightsFn(identity.SID); err != nil { + return windowsSandboxIdentity{}, false, errors.Join(err, undo()) + } + if secretPathErr != nil { + return windowsSandboxIdentity{}, false, errors.Join(secretPathErr, undo()) + } + // Rotation happens HERE, immediately before the secret is committed, rather + // than inside provisioning where it used to. + // + // A new account already has this password from NetUserAdd, so only an + // adopted one needs setting. Doing it at the top of provisioning meant every + // step in between ran with the account's password already replaced and no + // copy of it stored, so any failure there stranded a working principal. The + // two operations are now adjacent, which is the smallest window this can + // have without a way to restore the previous password, which Windows does + // not offer. + if !created { + if err := resetWindowsSandboxUserPasswordFn(identity.Username, password); err != nil { + return windowsSandboxIdentity{}, false, errors.Join(err, undo()) + } + rotated = true + } + if err := writeWindowsSandboxSecretFn(secretPath, password); err != nil { + return windowsSandboxIdentity{}, false, errors.Join(err, undo()) + } + return identity, created, nil +} + +// setupWindowsSandboxPrincipal provisions this workspace's principal and grants +// it the filesystem access its permission profile describes. It returns a +// rollback that undoes everything it created, so a later setup step failing does +// not leave a half-provisioned account behind. +// +// Rollback order is the inverse of creation and matters: ACEs are revoked BEFORE +// the account is deleted, because removing the account first would leave ACEs +// naming a SID that no longer resolves, which is the orphaned-entry residue this +// model exists to avoid. +func setupWindowsSandboxPrincipal(config WindowsSandboxCommandConfig) (func() error, error) { + key := windowsSandboxPrincipalKey(config) + // Both roles are provisioned regardless of the profile this setup ran with. + // Setup needs elevation and a command does not, so a command whose network + // mode differs from the one setup happened to see must still find a principal + // waiting; provisioning lazily would mean an unelevated command discovering it + // needs an account it cannot create. + roles := windowsSandboxLiveRoles + + var undo []func() error + rollback := func() error { + // Unwind in reverse, and keep going after a failure so one broken step + // cannot strand everything behind it. The first error is reported. + var firstErr error + for i := len(undo) - 1; i >= 0; i-- { + if err := undo[i](); err != nil && firstErr == nil { + firstErr = err + } + } + return firstErr + } + + filesystem := config.PermissionProfile.FileSystem + // Resolved once, before the loop: both principals need write access to the + // same runtime tree. + // + // permissionProfileWithRuntime appends this root to WriteRoots on every + // COMMAND and redirects HOME, GOCACHE, npm_config_cache and friends into it, + // but it lives under the user cache rather than the workspace, so the profile + // setup sees never contains it. On the restricted-token path that costs + // nothing, since the child still runs as the caller. A principal is a separate + // local account with none of those rights, so without this every npm install, + // go build or pip install fails on a cache write with a bare ACCESS_DENIED and + // nothing pointing at the sandbox as the cause. + runtimeRoots, err := setupWindowsSandboxRuntimeRoot(config) + if err != nil { + return nil, err + } + + for _, role := range roles { + username := windowsSandboxUserName(key, role) + // Retire a principal whose grants were never recorded, BEFORE provisioning + // adopts it. + // + // This is the one case where the prior grant set is not empty but unknowable: + // an account from an earlier setup exists, and nothing on Windows can + // enumerate the paths whose DACL names its SID. Carrying on would revoke only + // what the new plan happens to name and leave the rest — the fail-open the + // record exists to close, reached on the single path where it cannot be ruled + // out. + // + // Retiring is a real fix rather than a gesture because Windows never reuses a + // deleted local account's RID: whatever ACEs survive name a principal that no + // longer exists and grant access to nobody, and the SID minted below is one + // no DACL on this machine can already carry. It also needs no new operator + // action, which matters — there is no `zero sandbox teardown` to send anyone + // to, so refusing here would strand the workspace instead of fixing it. + // + // Per role, like everything else in this loop: the two principals are + // separate accounts with separate records, and one having lost its record + // says nothing about the other. + if _, recorded := readWindowsPrincipalACLLedger(config.SandboxHome, username); !recorded { + if err := retireUnrecordedWindowsSandboxPrincipal(config, role); err != nil { + _ = rollback() + return nil, err + } + } + identity, created, err := provisionWindowsSandboxPrincipalForSetupFn(config, role) + if err != nil { + _ = rollback() + return nil, err + } + // Only for a principal this run created. + // + // Appending unconditionally meant that if the offline role adopted an + // account that already existed and the online role, or any later step, + // then failed, the outer rollback deleted a principal that was working + // before setup started. Dual-role provisioning makes that likely rather + // than unlucky: the first role usually succeeds, so there is almost + // always something for a later failure to destroy. This mirrors the + // contract provisioning already applies to its own inner rollback. + if created { + undo = append(undo, func() error { return removeWindowsSandboxPrincipalForSetupFn(config, role) }) + } + + // Both principals get the same filesystem access. They differ only in + // network reach, so granting the offline one less would make an approved + // network command see a different filesystem than an ordinary one. + writeRoots := filesystem.WriteRoots + for _, root := range runtimeRoots { + writeRoots = append(append([]WritableRoot{}, writeRoots...), WritableRoot{Root: root}) + } + // Revokes this trustee's existing ACEs on the paths the plan touches AND + // on the paths this role's record names, before applying it, so a re-run + // after narrowing a root does not leave the wider grant behind. Per role: + // the two principals are separate trustees with separate records, and + // revoking one must not disturb the other. + revertACL, err := applyWindowsPrincipalACLs(config.SandboxHome, username, identity.SID.String(), filesystem, writeRoots) + if err != nil { + _ = rollback() + return nil, err + } + // ACEs are reverted before the account they name is deleted, because + // removing the account first leaves ACEs naming a SID that no longer + // resolves, which is the orphaned residue this model exists to avoid. + // Appending after the removal closure puts it earlier in the reverse + // unwind, which is what gets that ordering. + undo = append(undo, revertACL) + } + return rollback, nil +} + +// retireUnrecordedWindowsSandboxPrincipal removes this role's principal when one +// exists, and does nothing when one does not. +// +// The absent case is the ordinary one and is not a problem: with no account +// there is nothing that could be holding an ACE, so a missing record is simply +// a machine where setup has not run yet. +// +// Scoped to one role. The other role is a separate account holding its own ACEs +// under its own record, so retiring both because one lost its record would +// destroy a principal there is nothing wrong with. +func retireUnrecordedWindowsSandboxPrincipal(config WindowsSandboxCommandConfig, role windowsSandboxRole) error { + _, err := lookupWindowsSandboxIdentityFn(windowsSandboxPrincipalKey(config), role) + if err != nil { + if errors.Is(err, errWindowsSandboxIdentityUnavailable) { + return nil + } + return err + } + return removeWindowsSandboxPrincipalForSetupFn(config, role) +} + +// removeWindowsSandboxPrincipalForSetup retires one role's principal in the +// order that leaves nothing behind: secret, then ACEs, then LSA logon rights, +// then the account itself. Everything keyed to the SID has to go while the SID +// still resolves. +func removeWindowsSandboxPrincipalForSetup(config WindowsSandboxCommandConfig, role windowsSandboxRole) error { + key := windowsSandboxPrincipalKey(config) + username := windowsSandboxUserName(key, role) + secretPath, err := windowsSandboxSecretPath(config.SandboxHome, username) + if err != nil { + return err + } + // A secret owned by another administrator's setup is not a reason to abandon + // the rest of teardown. Aborting here left the account, its logon rights, its + // ACEs and its ledger all installed because one file could not be unlinked, + // which is the worst of both outcomes: the operator asked for removal, got a + // failure, and kept a working principal. + // + // Remembered and reported at the end rather than swallowed, so the leftover + // secret is visible to whoever has to clean it up. + var secretErr error + if err := removeWindowsSandboxSecret(secretPath); err != nil { + if !errors.Is(err, errWindowsSandboxSecretNotOurs) { + return err + } + secretErr = err + } + // Set when ACE revocation could not complete. Teardown continues regardless, + // but the ledger is kept and the error surfaced, so the residue stays + // findable instead of being silently orphaned. + var revokeErr error + // Drop the LSA account rights before the account itself. Deleting the account + // first would leave its rights behind keyed to a SID that no longer resolves, + // which is the same orphaned residue the trustee-keyed ACE revocation exists + // to avoid. A principal that was never provisioned has no SID to resolve and + // nothing to revoke, so that case is not an error. + if identity, err := lookupWindowsSandboxIdentity(key, role); err == nil { + // ACEs first, for the same reason: once the account is gone its SID stops + // resolving and every ACE naming it becomes an orphaned raw-SID entry on + // the user's own tree, which is precisely the residue the capability-SID + // model left behind and this one exists to avoid. Revocation is by + // trustee, so it clears grants written by older versions too, and it is + // scoped to THIS role's principal — the other role is a separate trustee + // whose ACEs must survive one role being retired. + // + // Failing to revoke is not fatal. A path the user has since deleted or + // renamed cannot be cleaned, and refusing to remove the account over it + // would strand the principal and its logon rights permanently — a worse + // outcome than a leftover ACE on a path that may not exist any more. + // + // The rollback is discarded here on purpose, unlike at setup: this is + // teardown, the account is about to be deleted, and putting its ACEs back + // is the opposite of what the caller asked for. + // + // The ERROR is not discarded, though it used to be. Teardown carried on + // and reported success, which meant a failed revocation left ACEs naming + // this SID on the user's tree while the ledger recording which paths they + // sat on was deleted moments later: residue nothing could find again. + // Remembered below rather than returned here, so removing the account + // still happens and the principal is not stranded. + // + // The revocation is per ROLE: each workspace has two principals and they + // are separate trustees, so the ledger has to be asked for this one's + // paths rather than the workspace's. + paths, pathsErr := windowsPrincipalRevocationPaths(config, identity.SID.String(), role) + if pathsErr != nil { + revokeErr = fmt.Errorf("resolve the paths holding ACEs for sandbox principal %s: %w", username, pathsErr) + } else if _, err := revokeWindowsPrincipalACEs(identity.SID.String(), paths); err != nil { + revokeErr = fmt.Errorf("revoke ACEs for sandbox principal %s: %w", username, err) + } + if err := revokeWindowsSandboxLogonRights(identity.SID); err != nil { + return err + } + } else if !errors.Is(err, errWindowsSandboxIdentityUnavailable) { + return err + } + // A foreign account under our derived name is left in place deliberately, and + // that is the goal state for teardown: no principal of ours exists under it. + // Surfacing it as a teardown failure would make setup rollback report an + // error for having correctly declined to delete somebody else's account. + if err := removeWindowsSandboxIdentity(username, key); err != nil { + if !errors.Is(err, errWindowsSandboxForeignAccountRetained) { + return err + } + // The record is KEPT in that case. It names paths this role's principal + // was granted, and a foreign account holding the name is not evidence + // those grants are gone — dropping the record would leave them + // unrevokable, which is the whole failure it exists to prevent. A record + // with no principal is revoked harmlessly by the next setup. + return nil + } + // Otherwise last, and only once the account is actually gone, so a failure + // anywhere above leaves the record describing a principal that still exists. + // + // It describes grants for a SID that no longer resolves, and leaving it would + // have the next setup revoke those paths on behalf of a freshly minted SID + // that never held them. That is a harmless no-op rather than a hole — the + // deleted account's RID is never reused, but a record that outlives its + // principal is a lie the next reader has no way to detect. + // + // Unless revocation failed. Then the ledger is the ONLY surviving record of + // which paths still carry ACEs for this SID, and deleting it turns a + // reportable leftover into permanent unfindable residue. Keeping a record + // that outlives its principal is the lesser problem, and the error says so. + if revokeErr != nil { + return errors.Join(secretErr, fmt.Errorf("%w; the principal ACL ledger has been kept so the remaining ACEs can still be found", revokeErr)) + } + // The account and its ACEs are gone either way. A secret left behind is + // residue worth naming, not a reason to keep the ledger: there are no ACEs + // left for it to describe. + return errors.Join(secretErr, removeWindowsPrincipalACLLedger(config.SandboxHome, username)) +} + +// setupWindowsSandboxRuntimeRoot resolves this workspace's runtime root and +// makes sure it exists, so the principal ACL plan can name it. +// +// It is created here rather than left to the first command because +// applyWindowsACLPlan skips a target that does not exist: granting write on a +// directory that setup never made would silently no-op, and the failure would +// only show up later as a denied cache write. Creating it under the elevated +// setup process is safe, since it lives under the invoking user's own cache +// directory and prepareSandboxRuntime would create it on the same path anyway. +// +// An empty return means there is no runtime root to grant (no workspace root +// configured), which is not an error: the caller simply grants nothing extra. +func windowsSandboxRuntimeRootPath(config WindowsSandboxCommandConfig) ([]string, error) { + // EVERY candidate, not just the cache-derived one. + // + // This used to take the deterministic cache-derived root alone and return + // nothing when the cache sat inside the workspace, on the reasoning that the + // other branch minted a random per-process directory through os.MkdirTemp and + // so had no name both sides could agree on. That reasoning is now stale: + // fallbackSandboxRuntimeRoot derives its path by hashing the workspace and + // creates nothing, so every process reaches the same answer. + // + // Leaving it stale had a cost. Commands still SELECT the fallback in that + // layout, and prepareSandboxRuntime redirects TMP, GOCACHE and the package + // caches into it, while setup granted neither principal an ACE on it. A + // principal command in a supported layout then failed ordinary cache writes + // with a bare ACCESS_DENIED and nothing naming the sandbox as the cause. + // + // So this now answers with the same candidate set the capability plan already + // grants, and for the same reason: setup covers both so a command lands on a + // provisioned tree whichever one it picks. + return windowsSandboxRuntimeCandidates(config.WorkspaceRoots), nil +} + +// setupWindowsSandboxRuntimeRoot resolves the runtime root AND creates it. +// Teardown wants the name without the side effect, so the derivation lives in +// windowsSandboxRuntimeRootPath above and this only adds the mkdir. +func setupWindowsSandboxRuntimeRoot(config WindowsSandboxCommandConfig) ([]string, error) { + roots, err := windowsSandboxRuntimeRootPath(config) + if err != nil { + return nil, err + } + // Every candidate is created, because setup grants an ACE on every candidate + // and applyWindowsACLPlan fails the whole run on a target that does not exist. + for _, root := range roots { + // Built relative to handles rather than by pathname, for the reason the + // secret directory is: this runs elevated over a tree the invoking user + // owns, so a junction planted mid-walk turns an os.MkdirAll into an + // elevated create wherever that junction points. + created, parent, err := makeWindowsACLDirChainNoFollow(root) + if parent != 0 { + _ = windows.CloseHandle(parent) + } + if err != nil { + if _, unwindErr := rollbackWindowsACLMaterialization(created); unwindErr != nil { + err = errors.Join(err, unwindErr) + } + return nil, fmt.Errorf("create sandbox runtime root: %w", err) + } + } + return roots, nil +} + +// revokeWindowsPrincipalACEs drops every ACE naming principalSID on paths and +// returns a rollback that puts them back. +// +// A path that does not exist is skipped rather than failing: revocation is +// cleanup, and there is nothing to clean on a path that was never created. +func revokeWindowsPrincipalACEs(principalSID string, paths []string) (func() error, error) { + if len(paths) == 0 { + return func() error { return nil }, nil + } + plan, err := windowsPrincipalRevokePlan(principalSID, paths) + if err != nil { + return nil, err + } + return applyWindowsACLPlanFn(plan) +} + +// applyWindowsPrincipalACLs writes the principal's ACEs for one policy: it +// revokes whatever this trustee already had on the paths the new plan touches +// AND on the paths an earlier setup recorded, then applies the plan. +// +// The order is the whole point. applyWindowsACLPlan MERGES into the existing +// DACL, so without the revocation first a re-run after narrowing a write root +// or shortening a deny list leaves the previous, wider ACEs beside the new ones +// and the principal keeps access the current policy no longer grants — the +// sandbox silently widens as a result of tightening it. Setup does get the +// chance to notice: marker validation refuses commands with "permission roots +// or deny lists changed" until setup runs again. +// +// The recorded paths are what makes that revocation complete. Revoking only the +// new plan's paths could never reach a root that had LEFT the policy, which is +// precisely the root whose ACE has to go: absent from the new plan, it was +// skipped, so the re-setup that was supposed to resolve the widening preserved +// it instead. +// +// Revocation is by TRUSTEE, so it drops every ACE naming this principal on +// these paths whatever an older version of Zero granted, and a recorded path +// that no longer exists or never held an ACE costs nothing. +func applyWindowsPrincipalACLs(sandboxHome string, username string, principalSID string, filesystem FileSystemPolicy, writeRoots []WritableRoot) (func() error, error) { + plan, err := buildWindowsPrincipalACLPlan(windowsPrincipalACLInput{ + PrincipalSID: principalSID, + WriteRoots: writeRoots, + ReadRoots: filesystem.ReadRoots, + DenyRead: filesystem.DenyRead, + DenyWrite: filesystem.DenyWrite, + }) + if err != nil { + return nil, err + } + granted := windowsACLPlanPaths(plan) + // An unreadable record arrives here as an empty prior set, which taken alone + // would be the fail-open. It cannot be reached: setupWindowsSandboxPrincipal + // retires any principal whose record is missing BEFORE provisioning, so by + // this point either the record is trustworthy or principalSID is one that no + // DACL on this machine has ever been able to name. + recorded, _ := readWindowsPrincipalACLLedger(sandboxHome, username) + stale := unionWindowsPrincipalACLPaths(recorded, granted) + + // Recorded BEFORE a single DACL changes, and as the union rather than the new + // set. A crash between the grant below and the narrowing write would otherwise + // leave a record that omits paths this run granted, and the next policy change + // would strand them in exactly the way this fix exists to prevent. A superset + // is the safe direction to be wrong in: revoking a path that holds no ACE for + // this trustee is a no-op. + if err := writeWindowsPrincipalACLLedger(sandboxHome, username, stale); err != nil { + return nil, err + } + + // The revocation's own rollback matters, and discarding it was a real bug. + // + // It was discarded on the reasoning that the only failure path from here + // removes the principal outright, so restoring stale ACEs for an account + // about to be deleted would be pointless. That holds for a principal this + // run CREATED. It is false for one this run ADOPTED: setup keeps a + // pre-existing account on failure rather than destroying someone else's + // working principal, so discarding the snapshot left that account alive with + // its previous ACEs stripped and the new ones rolled back — logged on and + // unable to reach its own workspace. + // + // Restoring the pre-revocation DACL first, then the grant, unwinds in the + // reverse order they were applied. + restoreRevoked, err := revokeWindowsPrincipalACEs(principalSID, stale) + if err != nil { + return nil, err + } + revertGrant, err := applyWindowsACLPlanFn(plan) + if err != nil { + if restoreRevoked != nil { + _ = restoreRevoked() + } + return nil, err + } + // Narrow the record to what is granted now, so it tracks the policy instead + // of accumulating every root the workspace has ever had. + // + // A failure here fails the setup rather than being shrugged off. The same + // file was written successfully moments ago, so failing now means the sandbox + // home has stopped being writable, and reporting a provisioned sandbox on a + // sandbox home that cannot hold its own state is the kind of quiet this + // backend must not have. Unwinding leaves the union recorded, which is the + // safe direction. + if err := writeWindowsPrincipalACLLedger(sandboxHome, username, granted); err != nil { + if revertErr := revertGrant(); revertErr != nil { + err = errors.Join(err, revertErr) + } + if restoreRevoked != nil { + if restoreErr := restoreRevoked(); restoreErr != nil { + err = errors.Join(err, restoreErr) + } + } + return nil, err + } + return func() error { + grantErr := revertGrant() + // Restore the pre-revocation ACEs even when reverting the grant failed: + // leaving the principal with neither set is the state this exists to + // avoid. Report the grant error, since that is the one leaving residue. + var restoreErr error + if restoreRevoked != nil { + restoreErr = restoreRevoked() + } + // THE LEDGER HAS TO COME BACK TOO. + // + // The narrowed set was written above only once everything had succeeded. + // This closure puts the old ACEs back, so leaving that narrowed ledger in + // place describes a principal holding ACEs the ledger does not name, and + // a later setup or teardown reads the ledger to decide what to revoke. + // The restored paths would never be revisited: access outside the current + // policy, held by an account nothing knows to clean up. + // + // The union is what was recorded before any DACL changed, so writing it + // back returns the ledger to the state the restored ACEs belong to. + // Over-recording is the safe direction: a path named but no longer held + // costs one redundant revoke, while a path held but unnamed is residue + // nothing can find. + if ledgerErr := writeWindowsPrincipalACLLedger(sandboxHome, username, stale); ledgerErr != nil { + restoreErr = errors.Join(restoreErr, ledgerErr) + } + if grantErr != nil { + return grantErr + } + return restoreErr + }, nil +} + +// windowsPrincipalTeardownPaths names every path this principal could hold an +// ACE on: the policy's roots plus the per-workspace runtime tree. +// +// The runtime root is derived through the same resolver setup uses, so teardown +// revokes ACEs on the tree commands actually wrote to. That was not possible +// while the fallback minted a random per-process directory: nothing here could +// name it, so an opted-out machine kept principal ACEs on whatever tree the +// fallback had produced. Both branches are creation-free now, so teardown can +// ask for the real answer without materializing anything on its way out. +func windowsPrincipalTeardownPaths(config WindowsSandboxCommandConfig, principalSID string) ([]string, error) { + filesystem := config.PermissionProfile.FileSystem + writeRoots := filesystem.WriteRoots + runtimeRoots, err := windowsSandboxRuntimeRootPath(config) + if err != nil { + return nil, err + } + // Every candidate, matching what setup granted: teardown that revoked only + // one would leave principal ACEs on the tree commands actually used. + for _, root := range runtimeRoots { + writeRoots = append(append([]WritableRoot{}, writeRoots...), WritableRoot{Root: root}) + } + plan, err := buildWindowsPrincipalACLPlan(windowsPrincipalACLInput{ + PrincipalSID: principalSID, + WriteRoots: writeRoots, + ReadRoots: filesystem.ReadRoots, + DenyRead: filesystem.DenyRead, + DenyWrite: filesystem.DenyWrite, + }) + if err != nil { + return nil, err + } + return windowsACLPlanPaths(plan), nil +} + +// windowsPrincipalRevocationPaths is what teardown actually has to revoke: the +// paths the CURRENT policy describes, plus every path an earlier setup recorded. +// +// Teardown used the current policy alone, which reproduced the setup-side bug on +// the way out. A root the user removed from their policy is missing from today's +// plan, so retiring the principal revoked every ACE except the one that was +// widening the sandbox — and then deleted the account, leaving that ACE naming a +// SID nothing could resolve to clean it up later. +func windowsPrincipalRevocationPaths(config WindowsSandboxCommandConfig, principalSID string, role windowsSandboxRole) ([]string, error) { + current, err := windowsPrincipalTeardownPaths(config, principalSID) + if err != nil { + return nil, err + } + // This role's record, not the workspace's: the other role is a separate + // trustee whose recorded paths are none of this revocation's business. + recorded, _ := readWindowsPrincipalACLLedger( + config.SandboxHome, windowsSandboxUserName(windowsSandboxPrincipalKey(config), role)) + return unionWindowsPrincipalACLPaths(recorded, current), nil +} + +// Seams for the two elevated calls the provisioning rollback depends on, so the +// stale-secret recovery path is reachable in tests without an elevated machine. +var ( + provisionWindowsSandboxIdentityFn = provisionWindowsSandboxIdentity + removeWindowsSandboxSecretFn = removeWindowsSandboxSecret + writeWindowsSandboxSecretFn = writeWindowsSandboxSecret +) + +// warnWindowsSandboxOfflineMembershipMissing reports the one drift that would +// otherwise silently hand a no-network profile full egress. +// +// Separate from the missing-secret warning because the remedy differs and +// because this one is a containment failure rather than a setup gap: the +// account is fine, its group membership is not. Once per process, for the same +// reason as its sibling — this sits on the command path. +var warnWindowsSandboxOfflineMembershipMissing = func(username string) { + windowsSandboxOfflineMembershipWarnOnce.Do(func() { + fmt.Fprintf(os.Stderr, + "[zero] sandbox principal %q is no longer a member of %q, which is the group the network "+ + "block filters match. Falling back to the restricted-token sandbox, which still denies "+ + "the network but does not confine reads. "+ + "Re-run `zero sandbox setup` from an elevated terminal to restore it.\n", + username, windowsSandboxOfflineGroupName) + }) +} + +var windowsSandboxOfflineMembershipWarnOnce sync.Once + +// Seam for the principal lookup, so the command path's mode-enforcement checks +// are reachable in tests without a provisioned machine. +var lookupWindowsSandboxPrincipalForCommandFn = lookupWindowsSandboxPrincipalForCommand + +// Seam for the secret read on the command path, so the mode-enforcement gates +// ahead of it can be exercised without a provisioned secret on disk. +var readWindowsSandboxSecretFn = readWindowsSandboxSecret + +// removeWindowsSandboxIdentityFn is the rollback's account deletion. A seam so a +// fixture that fakes creation can also fake its inverse: a mocked provisioning +// test that reports "created" and then injects a failure must not reach the +// real NetUserDel on an account it never made. +var removeWindowsSandboxIdentityFn = removeWindowsSandboxIdentity + +// Seam for the lookup the unrecorded-principal retirement decides on, so that +// decision is observable in a test without a provisioned machine — on which the +// lookup declines for its own reasons and would report success whether or not +// the guard existed. +// +// The retirement's other elevated call, removeWindowsSandboxPrincipalForSetup, +// is seamed with the dual-role rollback seams above rather than here, for the +// reason recorded there: two declarations of one seam let a test stub one while +// production uses the other. +var lookupWindowsSandboxIdentityFn = lookupWindowsSandboxIdentity + +// windowsACLPlanPaths returns each distinct path a plan touches, in plan order. +// +// Windows-tagged deliberately: every caller is, so defining it in the portable +// file made it unused on Linux and macOS builds and failed static analysis. +func windowsACLPlanPaths(plan WindowsACLPlan) []string { + seen := make(map[string]struct{}, len(plan.Entries)) + paths := make([]string, 0, len(plan.Entries)) + for _, entry := range plan.Entries { + if _, ok := seen[entry.Path]; ok { + continue + } + seen[entry.Path] = struct{}{} + paths = append(paths, entry.Path) + } + return paths +} + +// removeWindowsSandboxPrincipalsForSetup retires BOTH of a workspace's +// principals. +// +// Each workspace has an offline and an online account, and they are separate +// trustees with their own secret, logon rights, ACEs and ledger entries. +// Retiring one and calling it done is what the opt-out path is there to +// prevent: the marker would flip to opted-out while the other account and +// everything it owns stayed on the machine, invisible, because nothing +// afterwards looks for a principal it believes was never provisioned. +// +// Both are attempted even if the first fails, and the errors are joined, so one +// stubborn account cannot hide residue left by the other. +// windowsSandboxRoleLegacy is retired alongside the two live roles. A machine +// set up before the roles were split holds one untagged "zero-sbx-" +// account, and neither role name resolves to it, so every other path on this +// branch reports it as never provisioned. Opting out or re-running setup would +// otherwise leave that account, its secret, its logon rights, its ACEs and its +// ledger installed and unreferenced. Retirement is idempotent and a principal +// that was never provisioned is not an error, so this costs a lookup on the +// machines that never had one. +func removeWindowsSandboxPrincipalsForSetup(config WindowsSandboxCommandConfig) error { + var errs []error + for _, role := range windowsSandboxRetirableRoles { + // Through the seam, like retireUnrecordedWindowsSandboxPrincipal, so the + // set of roles this retires is assertable without provisioning real + // accounts on the machine running the tests. + if err := removeWindowsSandboxPrincipalForSetupFn(config, role); err != nil { + errs = append(errs, fmt.Errorf("retire the %s sandbox principal: %w", role, err)) + } + } + return errors.Join(errs...) +} + +// windowsCurrentUserSID returns the SID of the user this process runs as, or +// empty when the token cannot be read. Elevation does not change it: a UAC +// consent prompt splits the caller's token but keeps the user, so this differs +// from the caller's SID only when a DIFFERENT account supplied the credentials +// (over-the-shoulder elevation, or `runas /user:`). +func windowsCurrentUserSID() string { + user, err := windows.GetCurrentProcessToken().GetTokenUser() + if err != nil || user == nil || user.User.Sid == nil { + return "" + } + return user.User.Sid.String() +} + +// windowsSandboxPrincipalKey derives the key a principal ACCOUNT is named after, +// scoped to the invoking user as well as the workspace. +// +// The workspace alone is not enough. The account name is machine-wide, but its +// DPAPI secret and its ACL ledger both live under the invoking user's sandbox +// home, so two Windows users sharing one workspace path derive the same account +// while keeping private bookkeeping for it. The second user finds an account +// with no ledger of their own and either retires the first user's working +// principal or adopts it and rotates its password while storing only their own +// secret. Either way the first user's marker still validates and their next +// command cannot log on. +// +// Deliberately NOT used for the setup lock, which stays keyed to the workspace +// alone. Two users setting up one shared workspace still write DACLs on the same +// paths, so they must continue to serialize against each other even though they +// now provision separate accounts. +// +// The invoking user is config.CallerSID when the caller supplied one, and only +// then the current process token. Those are the same identity for an ordinary +// command, which runs as the caller. They are NOT the same inside elevated +// setup: over-the-shoulder UAC runs the helper as whichever administrator typed +// the credentials, so sampling its own token here named the account after the +// administrator while every later command derived the ordinary user's name, +// found nothing, and fell back to the restricted token — leaving an admin-keyed +// account, secret and ledger that nothing afterwards would ever reclaim. +// +// Falls back to the workspace-only key when neither is available. That is the +// pre-existing behaviour rather than a new failure mode, and refusing to derive +// a name at all would break setup on a machine whose token query fails for +// unrelated reasons. +func windowsSandboxPrincipalKey(config WindowsSandboxCommandConfig) string { + workspace := windowsSandboxWorkspaceKey(config.WorkspaceRoots) + sid := strings.TrimSpace(config.CallerSID) + if sid == "" { + sid = windowsCurrentUserSID() + } + if sid == "" { + return workspace + } + digest := sha256.Sum256([]byte(workspace + "\x00" + sid)) + return hex.EncodeToString(digest[:]) +} diff --git a/internal/sandbox/windows_identity_runtime_windows_test.go b/internal/sandbox/windows_identity_runtime_windows_test.go new file mode 100644 index 000000000..7d5fff53b --- /dev/null +++ b/internal/sandbox/windows_identity_runtime_windows_test.go @@ -0,0 +1,292 @@ +//go:build windows + +package sandbox + +import ( + "os" + "strings" + "sync" + "testing" +) + +// An opted-in command that ends up on the restricted token anyway must say so. +// Both cases below are correct fallbacks, not errors — but silence leaves the +// operator believing an account boundary is isolating them when it is not, which +// is the same failure the setup-protocol opt-in check exists to prevent, reached +// from the other side. The deny case matters most: deny is the DEFAULT network +// mode, so a fully provisioned, fully agreeing setup still never uses the +// principal for an ordinary command. +func TestWindowsSandboxPrincipalFallbackIsAnnounced(t *testing.T) { + testCases := []struct { + name string + mode NetworkMode + reason string + }{ + // The network-deny case is deliberately absent. It used to warn here, but + // this runner is re-exec'd per command, so the sync.Once guarding the notice + // is once per COMMAND — and deny is the default mode, so the warning landed + // on nearly every tool call. That fact belongs to `zero doctor` now, which is + // read once. The deny-mode BEHAVIOUR is still pinned, by + // TestPrincipalBackendDefersToRestrictedTokenWhenNetworkDenied below. + {name: "no principal provisioned on this machine", mode: NetworkAllow, reason: "no sandbox principal is provisioned"}, + } + for _, testCase := range testCases { + t.Run(testCase.name, func(t *testing.T) { + var warned []string + originalWarn := warnWindowsSandboxPrincipalNotUsed + warnWindowsSandboxPrincipalNotUsed = func(reason string) { warned = append(warned, reason) } + windowsSandboxPrincipalNotUsedWarnOnce = sync.Once{} + t.Cleanup(func() { + warnWindowsSandboxPrincipalNotUsed = originalWarn + windowsSandboxPrincipalNotUsedWarnOnce = sync.Once{} + }) + + config := WindowsSandboxCommandConfig{ + SandboxHome: t.TempDir(), + CommandCWD: `C:\workspace`, + WorkspaceRoots: []string{`C:\workspace`}, + PermissionProfile: PermissionProfile{ + FileSystem: FileSystemPolicy{Kind: FileSystemRestricted, WriteRoots: []WritableRoot{{Root: `C:\workspace`}}}, + Network: NetworkPolicy{Mode: testCase.mode}, + }, + Env: map[string]string{windowsSandboxIdentityEnv: "1"}, + SandboxLevel: WindowsSandboxLevelRestrictedToken, + Command: []string{"cmd.exe", "/c", "echo"}, + } + // Assert the precondition rather than assume it: this must be the quiet + // fallback path, not a token this host actually minted and not an error. + token, ok, err := windowsSandboxPrincipalToken(config) + if ok { + token.Close() + t.Fatalf("host unexpectedly provisioned a principal; this test cannot measure the fallback") + } + if err != nil { + t.Fatalf("windowsSandboxPrincipalToken error = %v, want the quiet fallback", err) + } + if len(warned) != 1 { + t.Fatalf("opted-in fallback warnings = %v, want exactly one naming %q", warned, testCase.reason) + } + if !strings.Contains(warned[0], testCase.reason) { + t.Fatalf("warning = %q, want it to name %q", warned[0], testCase.reason) + } + }) + } +} + +// The opt-out must stay silent, or the warning becomes noise every user learns +// to ignore. +func TestWindowsSandboxPrincipalFallbackIsSilentWhenOptedOut(t *testing.T) { + var warned []string + originalWarn := warnWindowsSandboxPrincipalNotUsed + warnWindowsSandboxPrincipalNotUsed = func(reason string) { warned = append(warned, reason) } + windowsSandboxPrincipalNotUsedWarnOnce = sync.Once{} + t.Cleanup(func() { + warnWindowsSandboxPrincipalNotUsed = originalWarn + windowsSandboxPrincipalNotUsedWarnOnce = sync.Once{} + }) + + config := WindowsSandboxCommandConfig{ + SandboxHome: t.TempDir(), + CommandCWD: `C:\workspace`, + WorkspaceRoots: []string{`C:\workspace`}, + PermissionProfile: PermissionProfile{ + FileSystem: FileSystemPolicy{Kind: FileSystemRestricted, WriteRoots: []WritableRoot{{Root: `C:\workspace`}}}, + Network: NetworkPolicy{Mode: NetworkDeny}, + }, + Env: map[string]string{windowsSandboxIdentityEnv: "0"}, + SandboxLevel: WindowsSandboxLevelRestrictedToken, + Command: []string{"cmd.exe", "/c", "echo"}, + } + if _, ok, err := windowsSandboxPrincipalToken(config); ok || err != nil { + t.Fatalf("windowsSandboxPrincipalToken ok=%v err=%v, want the quiet opted-out fallback", ok, err) + } + if len(warned) != 0 { + t.Fatalf("opted-out command warned %v, want silence", warned) + } +} + +// Setup must stay inert unless the principal backend is explicitly opted into. +// This is the property that makes the branch safe to merge while the privileged +// paths are still being validated: without the opt-in, `zero sandbox setup` +// creates no local account and the capability-SID backend is the whole of setup. +func TestWindowsSandboxIdentityGating(t *testing.T) { + for name, testCase := range map[string]struct { + env map[string]string + want bool + }{ + // An explicit map entry is authoritative; these cases never reach the + // process environment. + "empty": {env: map[string]string{windowsSandboxIdentityEnv: ""}, want: false}, + "zero": {env: map[string]string{windowsSandboxIdentityEnv: "0"}, want: false}, + "true not one": {env: map[string]string{windowsSandboxIdentityEnv: "true"}, want: false}, + "one": {env: map[string]string{windowsSandboxIdentityEnv: "1"}, want: true}, + "one with space": {env: map[string]string{windowsSandboxIdentityEnv: " 1 "}, want: true}, + } { + t.Run(name, func(t *testing.T) { + // Pin the process variable too. Every case here supplies an explicit + // map entry so none of them should consult it, and pinning proves + // that rather than assuming it: without this a developer who exports + // the opt-in would see different results from CI. + t.Setenv(windowsSandboxIdentityEnv, "1") + if got := windowsSandboxIdentityEnabled(testCase.env); got != testCase.want { + t.Fatalf("enabled = %v, want %v for %q", got, testCase.want, testCase.env[windowsSandboxIdentityEnv]) + } + }) + } +} + +// With no map entry the process environment decides. That fallback is what the +// elevated setup path actually runs on — it passes no Env — so it needs its own +// coverage rather than riding on a case that also has a map entry. +func TestWindowsSandboxIdentityGatingFallsBackToTheProcessEnvironment(t *testing.T) { + for name, testCase := range map[string]struct { + value string + set bool + want bool + }{ + "unset": {set: false, want: false}, + "empty": {value: "", set: true, want: false}, + "zero": {value: "0", set: true, want: false}, + "one": {value: "1", set: true, want: true}, + "one with space": {value: " 1 ", set: true, want: true}, + } { + t.Run(name, func(t *testing.T) { + // t.Setenv registers the restore even when the variable is then + // cleared, which is the only way to test a genuinely absent variable + // without leaking that state into the rest of the package. + t.Setenv(windowsSandboxIdentityEnv, testCase.value) + if !testCase.set { + if err := os.Unsetenv(windowsSandboxIdentityEnv); err != nil { + t.Fatalf("unset %s: %v", windowsSandboxIdentityEnv, err) + } + } + if got := windowsSandboxIdentityEnabled(nil); got != testCase.want { + t.Fatalf("enabled = %v, want %v (set=%v value=%q)", got, testCase.want, testCase.set, testCase.value) + } + }) + } +} + +// The command environment wins over the process environment, so a run can opt in +// or out without depending on how the parent shell was launched. +func TestWindowsSandboxIdentityEnvOverridesProcess(t *testing.T) { + t.Setenv(windowsSandboxIdentityEnv, "1") + if windowsSandboxIdentityEnabled(map[string]string{windowsSandboxIdentityEnv: "0"}) { + t.Fatal("command env set to 0 must override a process env of 1") + } + if !windowsSandboxIdentityEnabled(map[string]string{}) { + t.Fatal("with no command-env entry the process env should apply") + } +} + +// One workspace maps to one principal, and different workspaces must not share +// an account, or two projects would run under the same identity and could reach +// each other's granted roots. +func TestWindowsSandboxWorkspaceKeyIsStableAndDistinct(t *testing.T) { + first := windowsSandboxWorkspaceKey([]string{`C:\ws\alpha`}) + again := windowsSandboxWorkspaceKey([]string{`C:\ws\alpha`}) + other := windowsSandboxWorkspaceKey([]string{`C:\ws\beta`}) + if first != again { + t.Fatalf("key is not stable: %q vs %q", first, again) + } + if first == other { + t.Fatal("two different workspaces produced the same principal key") + } + if first == "" { + t.Fatal("empty key") + } + // An empty root list still has to yield a usable key rather than a blank one. + if windowsSandboxWorkspaceKey(nil) == "" { + t.Fatal("no workspace roots produced an empty key") + } +} + +// Network denial is enforced by WFP filters keyed to the offline-marker SID, +// which only the restricted token carries. A principal token would leave those +// filters matching nothing, so the principal backend must stand down whenever +// the network is denied rather than silently trading network enforcement for +// read confinement. +// The backend now runs under both network modes, so the opt-in is the only thing +// eligibility turns on. The mode selects which principal is used, not whether one +// is used at all. +func TestPrincipalBackendEligibilityTurnsOnlyOnTheOptIn(t *testing.T) { + eligible := func(mode NetworkMode, optIn string) bool { + config := WindowsSandboxCommandConfig{ + SandboxHome: t.TempDir(), + WorkspaceRoots: []string{t.TempDir()}, + Env: map[string]string{windowsSandboxIdentityEnv: optIn}, + } + config.PermissionProfile.Network.Mode = mode + return windowsSandboxPrincipalEligible(config) + } + + for _, mode := range []NetworkMode{NetworkDeny, NetworkAllow} { + if !eligible(mode, "1") { + t.Fatalf("principal backend refused with network %q; the mode should choose a principal, not disable the backend", mode) + } + if eligible(mode, "0") { + t.Fatalf("principal backend eligible without the opt-in for network %q", mode) + } + } +} + +// The network mode is enforced by WHICH principal runs, because the offline one +// is a member of the group the block filters match. Picking the wrong role is +// therefore a silent loss of network enforcement, so the mapping is asserted +// directly rather than inferred from a token that only exists on a provisioned +// machine. +func TestPrincipalRoleFollowsNetworkMode(t *testing.T) { + if got := windowsSandboxRoleForNetwork(NetworkDeny); got != windowsSandboxRoleOffline { + t.Fatalf("deny selected %q, want the offline principal; the online one is not in the blocked group and would have the network", got) + } + if got := windowsSandboxRoleForNetwork(NetworkAllow); got != windowsSandboxRoleOnline { + t.Fatalf("allow selected %q, want the online principal", got) + } + // Fail closed. An unrecognised mode must lose the network rather than keep it, + // so anything that is not an explicit allow maps to the offline principal. + for _, mode := range []NetworkMode{NetworkMode(""), NetworkMode("bogus"), NetworkMode("ALLOW")} { + if got := windowsSandboxRoleForNetwork(mode); got != windowsSandboxRoleOffline { + t.Fatalf("unrecognised mode %q selected %q, want the offline principal so an unknown mode fails closed", mode, got) + } + } +} + +// The account key has to separate USERS, not only workspaces. +// +// The name is machine-wide while the DPAPI secret and the ACL ledger both live +// under the invoking user's sandbox home. Keyed on the workspace alone, two +// Windows users sharing one path derive the same account with private +// bookkeeping: the second finds no ledger of its own and either retires the +// first user's working principal or adopts it and rotates the password while +// storing only its own secret. The first user's marker still validates and its +// next command cannot log on. +// +// Asserting the principal key DIFFERS from the workspace key is what pins that. +// It is the cheapest observable difference; a test that only checked the key was +// stable would pass against the broken version. +func TestPrincipalKeyIsScopedToTheUserNotOnlyTheWorkspace(t *testing.T) { + roots := []string{`C:\ws\shared`} + workspace := windowsSandboxWorkspaceKey(roots) + config := WindowsSandboxCommandConfig{WorkspaceRoots: roots} + principal := windowsSandboxPrincipalKey(config) + + if principal == workspace { + t.Fatal("the principal key equals the workspace key, so two users sharing this workspace would derive one account with separate secrets and ledgers") + } + if principal != windowsSandboxPrincipalKey(config) { + t.Error("the principal key is not stable for one user, so a second setup would not find its own account") + } + if other := windowsSandboxPrincipalKey(WindowsSandboxCommandConfig{WorkspaceRoots: []string{`C:\ws\other`}}); other == principal { + t.Error("two workspaces derived the same principal key, so they would share one account") + } +} + +// The setup lock must stay keyed to the WORKSPACE alone. Two users setting up +// one shared workspace now provision separate accounts, but they still write +// DACLs on the same paths, so they have to keep serializing against each other. +func TestSetupLockStaysWorkspaceScoped(t *testing.T) { + roots := []string{`C:\ws\shared`} + if windowsSandboxWorkspaceKey(roots) == windowsSandboxPrincipalKey(WindowsSandboxCommandConfig{WorkspaceRoots: roots}) { + t.Fatal("the lock key and the principal key are the same, so making the account per-user also made the lock per-user") + } +} diff --git a/internal/sandbox/windows_identity_secret_chain_windows_test.go b/internal/sandbox/windows_identity_secret_chain_windows_test.go new file mode 100644 index 000000000..431329875 --- /dev/null +++ b/internal/sandbox/windows_identity_secret_chain_windows_test.go @@ -0,0 +1,140 @@ +//go:build windows + +package sandbox + +import ( + "os" + "path/filepath" + "testing" +) + +// AND THE ANCESTOR IS PART OF THE PATH TOO. +// +// The sibling test puts the junction at the directory the secret is addressed +// into, which the no-follow open of that directory refuses. This one puts it +// ABOVE a component that does not exist yet, which is the case the refusal never +// saw: os.MkdirAll walked the pathname, followed the junction, and created the +// missing component on the other side of it as Administrator. The no-follow open +// that came next then found a perfectly ordinary directory at the requested +// pathname, because the caller-controlled redirection had already happened one +// level up, and the elevated write landed outside the sandbox home. +// +// A junction rather than a symlink for the same reason as the sibling test: it +// needs no privilege, so it is reachable by exactly the unprivileged user this +// guards against. +func TestWriteWindowsSandboxSecretRefusesAJunctionedAncestor(t *testing.T) { + base := t.TempDir() + outside := filepath.Join(base, "outside") + if err := os.MkdirAll(outside, 0o700); err != nil { + t.Fatal(err) + } + // home is the junction. The secret is addressed two levels below it, and the + // level in between does not exist, so it has to be created. + home := filepath.Join(base, "home") + makeJunction(t, home, outside) + + path := filepath.Join(home, "identity", "zero-sbx-test.secret") + err := writeWindowsSandboxSecret(path, "Zs1!EXAMPLEPASSWORDVALUE") + if err == nil { + t.Error("wrote the principal secret below a junctioned ancestor, so an elevated setup placed it in a tree the caller controls") + } + + // The refusal is only half of it: nothing may have been created on the other + // side of the junction, because an elevated create there is a write primitive + // at a path the caller chose even when no secret follows it. + entries, readErr := os.ReadDir(outside) + if readErr != nil { + t.Fatal(readErr) + } + if len(entries) != 0 { + names := []string{} + for _, entry := range entries { + names = append(names, entry.Name()) + } + t.Errorf("the junction target gained %v, so an elevated create followed the ancestor", names) + } +} + +// The same shape for the runtime root, which is the other elevated pathname +// create in this setup path. Its candidates sit under the invoking user's own +// cache directory, which is to say under a tree the party this sandbox contains +// can rearrange before elevated setup runs. +func TestSetupWindowsSandboxRuntimeRootRefusesAJunctionedAncestor(t *testing.T) { + isolateSandboxRuntimeRoots(t) + cache := t.TempDir() + t.Setenv("LOCALAPPDATA", cache) + t.Setenv("XDG_CACHE_HOME", cache) + workspace := filepath.Join(t.TempDir(), "ws") + if err := os.MkdirAll(workspace, 0o700); err != nil { + t.Fatal(err) + } + config := WindowsSandboxCommandConfig{WorkspaceRoots: []string{workspace}} + + candidates := windowsSandboxRuntimeCandidates(config.WorkspaceRoots) + if len(candidates) == 0 { + t.Fatal("SETUP INVALID: no runtime root candidate for a configured workspace") + } + // PICKED BY EXCLUDING THE OTHER ONE, NOT BY PATH PREFIX. The second candidate + // is the shared fallback under the system temp directory, which this test + // must not junction. Selecting the cache-derived one by prefix looks + // equivalent and is not: a CI runner spells one directory two ways + // (RUNNER~1 against runneradmin), so os.UserCacheDir and t.TempDir disagree + // on a path they both mean and the prefix matches nothing. + fallback, _ := fallbackSandboxRuntimeRoot(workspace) + cacheCandidate := "" + for _, candidate := range candidates { + if candidate != fallback { + cacheCandidate = candidate + break + } + } + if cacheCandidate == "" { + t.Fatalf("SETUP INVALID: no cache-derived candidate among %v (fallback %q)", candidates, fallback) + } + + // The swap: the candidate's parent is a junction out of the cache tree, and + // the candidate itself does not exist, so it has to be created. + outside := filepath.Join(t.TempDir(), "outside") + if err := os.MkdirAll(outside, 0o700); err != nil { + t.Fatal(err) + } + parent := filepath.Dir(cacheCandidate) + if err := os.MkdirAll(filepath.Dir(parent), 0o700); err != nil { + t.Fatal(err) + } + makeJunction(t, parent, outside) + + if _, err := setupWindowsSandboxRuntimeRoot(config); err == nil { + t.Error("created the runtime root below a junctioned ancestor, as an elevated process, in a tree the caller controls") + } + + entries, readErr := os.ReadDir(outside) + if readErr != nil { + t.Fatal(readErr) + } + if len(entries) != 0 { + names := []string{} + for _, entry := range entries { + names = append(names, entry.Name()) + } + t.Errorf("the junction target gained %v, so an elevated create followed the ancestor", names) + } +} + +// And an ordinary nested directory is still created, or the refusals above +// prove only that these functions fail. +func TestWriteWindowsSandboxSecretCreatesAMissingChain(t *testing.T) { + base := t.TempDir() + path := filepath.Join(base, "home", "identity", "nested", "zero-sbx-test.secret") + const password = "Zs1!EXAMPLEPASSWORDVALUE" + if err := writeWindowsSandboxSecret(path, password); err != nil { + t.Fatalf("write: %v", err) + } + got, err := readWindowsSandboxSecret(path) + if err != nil { + t.Fatalf("read: %v", err) + } + if got != password { + t.Fatalf("read %q, want the stored password", got) + } +} diff --git a/internal/sandbox/windows_identity_secret_cleanup_windows_test.go b/internal/sandbox/windows_identity_secret_cleanup_windows_test.go new file mode 100644 index 000000000..c66eed952 --- /dev/null +++ b/internal/sandbox/windows_identity_secret_cleanup_windows_test.go @@ -0,0 +1,107 @@ +//go:build windows + +package sandbox + +import ( + "errors" + "os" + "path/filepath" + "strings" + "testing" + + "golang.org/x/sys/windows" +) + +// CLEANUP HAS TO UNDO WHAT CREATE DID, ON THE SAME OBJECT. +// +// Everything between creating the secret file and writing it is handle-relative +// and no-follow. The failure path was not: it called os.Remove on the pathname, +// which is the one resolution the rest of the function exists to avoid, and it +// did not even work. The leaf is created with FILE_SHARE_READ alone, so a delete +// by name while that handle is open is a sharing violation; the error was +// discarded, the directory unwind did not cover the file, and a failed +// provisioning left an empty secret behind that a later run would read as a +// stale password. +func TestWriteWindowsSandboxSecretRemovesTheLeafWhenLockingFails(t *testing.T) { + restore := failWindowsSecretLock(t, errors.New("SetSecurityInfo refused")) + defer restore() + + path := filepath.Join(t.TempDir(), "cfg", "zero-sbx-test.secret") + err := writeWindowsSandboxSecret(path, "Zs1!EXAMPLEPASSWORDVALUE") + + if err == nil { + t.Fatal("SETUP INVALID: the injected lock failure was not reported") + } + if !strings.Contains(err.Error(), "SetSecurityInfo refused") { + t.Fatalf("err = %v, want the injected cause", err) + } + if _, statErr := os.Stat(path); !errors.Is(statErr, os.ErrNotExist) { + t.Fatalf("the secret file survived a failed lock (stat: %v); a later run would read it as a stale password", statErr) + } + // The directory this attempt created goes too, which is the half that + // already worked and must keep working. + if _, statErr := os.Stat(filepath.Dir(path)); !errors.Is(statErr, os.ErrNotExist) { + t.Errorf("the directory chain this attempt created was left behind (stat: %v)", statErr) + } +} + +// AND A CLEANUP THAT CANNOT FINISH SAYS SO. Leftover state after a failed +// provisioning is the caller's business, and an empty secret file is exactly +// what a later run would mistake for a stored password. +// +// The delete is injected rather than blocked for real. Making a Windows delete +// fail means arranging a sharing conflict, and my first attempt at that held a +// second handle on the leaf and did not block anything: the cleanup succeeded +// and the test asserted its way to a false negative. A seam says what the code +// does with a failure, which is the thing under test here. +func TestWriteWindowsSandboxSecretReportsAFailedCleanup(t *testing.T) { + restoreLock := failWindowsSecretLock(t, errors.New("SetSecurityInfo refused")) + defer restoreLock() + + previousDelete := deleteWindowsSecretLeafFn + deleteWindowsSecretLeafFn = func(windows.Handle, string) error { + return errors.New("STATUS_CANNOT_DELETE") + } + defer func() { deleteWindowsSecretLeafFn = previousDelete }() + + path := filepath.Join(t.TempDir(), "cfg", "zero-sbx-test.secret") + err := writeWindowsSandboxSecret(path, "Zs1!EXAMPLEPASSWORDVALUE") + + if err == nil { + t.Fatal("SETUP INVALID: the injected lock failure was not reported") + } + if !strings.Contains(err.Error(), "SetSecurityInfo refused") { + t.Errorf("err = %v, want the original cause preserved", err) + } + if !strings.Contains(err.Error(), "remove partially written secret") { + t.Errorf("err = %v, want the failed cleanup reported rather than discarded", err) + } + if !strings.Contains(err.Error(), "STATUS_CANNOT_DELETE") { + t.Errorf("err = %v, want the delete failure carried too", err) + } +} + +// The ordinary path is unchanged, or the two refusals above would prove only +// that this function fails. +func TestWriteWindowsSandboxSecretStillStoresAndReadsBack(t *testing.T) { + path := filepath.Join(t.TempDir(), "cfg", "zero-sbx-test.secret") + const password = "Zs1!EXAMPLEPASSWORDVALUE" + if err := writeWindowsSandboxSecret(path, password); err != nil { + t.Fatalf("write: %v", err) + } + got, err := readWindowsSandboxSecret(path) + if err != nil { + t.Fatalf("read: %v", err) + } + if got != password { + t.Fatalf("read %q, want the stored password", got) + } +} + +// failWindowsSecretLock makes the DACL step fail for one test. +func failWindowsSecretLock(t *testing.T, cause error) func() { + t.Helper() + previous := lockWindowsSecretHandleToOwnerFn + lockWindowsSecretHandleToOwnerFn = func(windows.Handle, *windows.SID) error { return cause } + return func() { lockWindowsSecretHandleToOwnerFn = previous } +} diff --git a/internal/sandbox/windows_identity_secret_handle_windows.go b/internal/sandbox/windows_identity_secret_handle_windows.go new file mode 100644 index 000000000..4d79ed7c9 --- /dev/null +++ b/internal/sandbox/windows_identity_secret_handle_windows.go @@ -0,0 +1,112 @@ +//go:build windows + +package sandbox + +import ( + "fmt" + "unsafe" + + "golang.org/x/sys/windows" +) + +// createWindowsSecretFileNoFollow creates (or replaces) the secret file as a +// CHILD of a pinned parent handle and returns a handle to it. +// +// The pathname version this replaces resolved the path four separate times in an +// elevated process: MkdirAll, an O_TRUNC open, SetNamedSecurityInfo by name, then +// WriteFile by name. The sandbox home belongs to the invoking user, who is the +// party this sandbox exists to contain, so each of those resolutions was a place +// to swap a component. Worst case an attacker-chosen file gets truncated and then +// has its DACL rewritten by an Administrator; the milder junction-only case still +// plants the deterministic secret somewhere the caller then owns. +// +// One handle, one resolution. The parent chain is built no-follow by the caller +// and its deepest handle is passed in here, the leaf is created relative to that +// handle so its name can never be resolved again, and both the DACL and the +// bytes are applied to the handle rather than to a path. +// +// FILE_OVERWRITE_IF, not FILE_CREATE: an existing secret must be replaced, since +// a stale password makes LogonUser fail in a way that reads as a sandbox bug. +// FILE_OPEN_REPARSE_POINT means a leaf that IS a reparse point comes back as +// itself rather than being followed, and the attribute check below then refuses +// it instead of writing through it. +func createWindowsSecretFileAt(parent windows.Handle, name string) (windows.Handle, error) { + if err := validateWindowsACLComponent(name); err != nil { + return 0, fmt.Errorf("secret file name: %w", err) + } + + objectName, err := windows.NewNTUnicodeString(name) + if err != nil { + return 0, fmt.Errorf("encode secret file name %s: %w", name, err) + } + attributes := windows.OBJECT_ATTRIBUTES{ + RootDirectory: parent, + ObjectName: objectName, + Attributes: windows.OBJ_CASE_INSENSITIVE, + } + attributes.Length = uint32(unsafe.Sizeof(attributes)) + + var handle windows.Handle + var status windows.IO_STATUS_BLOCK + if err := windows.NtCreateFile( + &handle, + // WRITE_DAC to lock it down and FILE_WRITE_DATA to fill it, both through + // this one handle. READ_CONTROL because SetSecurityInfo reads the existing + // descriptor before replacing it, and FILE_READ_ATTRIBUTES because the + // reparse-point check below asks the handle what it landed on. + windows.FILE_WRITE_DATA|windows.FILE_READ_ATTRIBUTES|windows.WRITE_DAC|windows.READ_CONTROL|windows.SYNCHRONIZE, + &attributes, + &status, + nil, + windows.FILE_ATTRIBUTE_NORMAL, + windows.FILE_SHARE_READ, + windows.FILE_OVERWRITE_IF, + windows.FILE_NON_DIRECTORY_FILE|windows.FILE_OPEN_REPARSE_POINT|windows.FILE_SYNCHRONOUS_IO_NONALERT, + 0, + 0, + ); err != nil { + return 0, fmt.Errorf("create secret file %s: %w", name, err) + } + if err := rejectWindowsACLReparseHandle(handle, name); err != nil { + _ = windows.CloseHandle(handle) + return 0, err + } + return handle, nil +} + +// lockWindowsSecretHandleToOwner applies the owner-and-SYSTEM-only DACL to an +// open handle instead of to a pathname, so the object being locked is provably +// the object that was just created rather than whatever the name resolves to now. +func lockWindowsSecretHandleToOwner(handle windows.Handle, owner *windows.SID) error { + acl, err := windowsSecretOwnerACL(owner) + if err != nil { + return err + } + if err := windows.SetSecurityInfo( + handle, + windows.SE_FILE_OBJECT, + windows.DACL_SECURITY_INFORMATION|windows.PROTECTED_DACL_SECURITY_INFORMATION, + nil, + nil, + acl, + nil, + ); err != nil { + return fmt.Errorf("lock secret to owner: %w", err) + } + return nil +} + +// writeWindowsSecretHandle writes the whole payload through the handle. +func writeWindowsSecretHandle(handle windows.Handle, payload []byte) error { + for written := 0; written < len(payload); { + var n uint32 + if err := windows.WriteFile(handle, payload[written:], &n, nil); err != nil { + return fmt.Errorf("write secret: %w", err) + } + if n == 0 { + return fmt.Errorf("write secret: wrote no bytes with %d remaining", len(payload)-written) + } + written += int(n) + } + return nil +} diff --git a/internal/sandbox/windows_identity_secret_junction_windows_test.go b/internal/sandbox/windows_identity_secret_junction_windows_test.go new file mode 100644 index 000000000..ef918b183 --- /dev/null +++ b/internal/sandbox/windows_identity_secret_junction_windows_test.go @@ -0,0 +1,71 @@ +//go:build windows + +package sandbox + +import ( + "os" + "path/filepath" + "strings" + "testing" +) + +// THE SECRET MUST NOT BE WRITTEN THROUGH A REPARSE POINT. +// +// The sandbox home belongs to the invoking user, who is the party this sandbox +// exists to contain, so they can put a junction where the secret directory is +// expected before elevated setup runs. The pathname version resolved the path +// four times over (MkdirAll, an O_TRUNC open, SetNamedSecurityInfo by name, then +// WriteFile by name) and every one of them followed it, so an Administrator +// process created the principal's password in a directory of the caller's +// choosing and then granted them control of it. +// +// A junction rather than a symlink on purpose: it needs no privilege, so it is +// reachable by exactly the unprivileged user this guards against. os.Symlink +// would need SeCreateSymbolicLinkPrivilege and would skip on an ordinary +// developer or CI account, which is where this most needs to run. +func TestWriteWindowsSandboxSecretRefusesAJunctionedDirectory(t *testing.T) { + base := t.TempDir() + outside := filepath.Join(base, "outside") + if err := os.MkdirAll(outside, 0o700); err != nil { + t.Fatalf("MkdirAll: %v", err) + } + // The swap: the directory the secret is addressed into is a junction that + // leads out of the sandbox home. The pathname never changes. + cfg := filepath.Join(base, "cfg") + makeJunction(t, cfg, outside) + + path := filepath.Join(cfg, "zero-sbx-test.secret") + err := writeWindowsSandboxSecret(path, "Zs1!EXAMPLEPASSWORDVALUE") + if err == nil { + t.Fatal("wrote the principal secret through a junction, so an elevated setup would have placed it in a directory the caller controls") + } + if !strings.Contains(err.Error(), "reparse") { + t.Fatalf("refused for the wrong reason: %v", err) + } + // Refusing is only half of it. Nothing may survive on the other side, or the + // caller still ends up holding a file the sandbox created for them. + entries, readErr := os.ReadDir(outside) + if readErr != nil { + t.Fatalf("ReadDir: %v", readErr) + } + if len(entries) != 0 { + t.Fatalf("junction target holds %d entries after the refusal, want none", len(entries)) + } +} + +// The ordinary path must still work, or the refusal above proves only that the +// function fails. +func TestWriteWindowsSandboxSecretStillWritesAnOrdinaryDirectory(t *testing.T) { + path := filepath.Join(t.TempDir(), "cfg", "zero-sbx-test.secret") + const password = "Zs1!EXAMPLEPASSWORDVALUE" + if err := writeWindowsSandboxSecret(path, password); err != nil { + t.Fatalf("write: %v", err) + } + got, err := readWindowsSandboxSecret(path) + if err != nil { + t.Fatalf("read: %v", err) + } + if got != password { + t.Fatalf("read %q, want the stored password", got) + } +} diff --git a/internal/sandbox/windows_identity_secret_windows.go b/internal/sandbox/windows_identity_secret_windows.go new file mode 100644 index 000000000..04065a441 --- /dev/null +++ b/internal/sandbox/windows_identity_secret_windows.go @@ -0,0 +1,313 @@ +//go:build windows + +package sandbox + +// Storing a sandbox principal's password. +// +// The elevated setup path provisions the account, but the per-command path runs +// UNELEVATED and needs the password to call LogonUser. So the secret has to +// cross that boundary on disk, and the only thing standing between it and the +// sandboxed child is the file's ACL. +// +// The file is locked to the invoking user: an explicit, INHERITANCE-PROTECTED +// DACL granting that user and SYSTEM, and nobody else. The sandbox principal is +// deliberately absent from it, which is the property that matters, because a +// principal that could read this file could mint its own token and the whole +// identity boundary would be decorative. Administrators are not added either; +// an admin can already take ownership, so naming them buys nothing and widens +// the visible grant. +// +// Ordering is load-bearing: the ACL is applied to an EMPTY file before the +// password is written, so the bytes never exist under the directory's inherited +// permissions even briefly. + +import ( + "errors" + "fmt" + "os" + "path/filepath" + "strings" + + "golang.org/x/sys/windows" +) + +// windowsSandboxSecretDirName holds per-principal secrets under the Zero config +// directory. Kept in its own directory so the whole set can be removed when the +// sandbox is torn down. +const windowsSandboxSecretDirName = "windows-sandbox" + +// windowsSandboxSecretPath returns where a principal's password lives. The +// account name is already sanitised to [a-z0-9-] by windowsSandboxUserName, so +// it cannot escape the directory. +func windowsSandboxSecretPath(configDir string, username string) (string, error) { + if strings.TrimSpace(configDir) == "" { + return "", errors.New("windows sandbox secret: empty config directory") + } + if strings.TrimSpace(username) == "" { + return "", errors.New("windows sandbox secret: empty principal name") + } + // Defence in depth against a caller passing something windowsSandboxUserName + // did not produce: refuse anything with a separator or a parent reference. + if strings.ContainsAny(username, `\/:`) || strings.Contains(username, "..") { + return "", fmt.Errorf("windows sandbox secret: unsafe principal name %q", username) + } + return filepath.Join(configDir, windowsSandboxSecretDirName, username+".secret"), nil +} + +// currentTokenUserSID returns the SID of the user this process runs as. Under +// UAC the elevated token keeps the same user SID as the desktop session, so +// setup and the later unelevated command path agree on the owner, which is what +// makes an owner-scoped ACL usable across the elevation boundary. +func currentTokenUserSID() (*windows.SID, error) { + var token windows.Token + if err := windows.OpenProcessToken(windows.CurrentProcess(), windows.TOKEN_QUERY, &token); err != nil { + return nil, fmt.Errorf("open process token: %w", err) + } + defer token.Close() + user, err := token.GetTokenUser() + if err != nil { + return nil, fmt.Errorf("get token user: %w", err) + } + // The SID points into a buffer owned by the Tokenuser, so copy it out before + // that buffer goes away. + copied, err := user.User.Sid.Copy() + if err != nil { + return nil, fmt.Errorf("copy token user SID: %w", err) + } + return copied, nil +} + +// windowsSecretOwnerACL builds the explicit, inheritance-protected DACL granting +// only owner and SYSTEM. PROTECTED is what drops any ACE inherited from the +// config directory; without it a permissive parent would still grant access to +// whoever it names. +// +// Returns the ACL rather than applying it, because the caller applies it to an +// open HANDLE. Applying by pathname meant the object being locked was whatever +// the name resolved to at that instant, which is not necessarily the object that +// was just created. +func windowsSecretOwnerACL(owner *windows.SID) (*windows.ACL, error) { + if owner == nil { + return nil, errors.New("windows sandbox secret: nil owner SID") + } + system, err := windows.CreateWellKnownSid(windows.WinLocalSystemSid) + if err != nil { + return nil, fmt.Errorf("resolve SYSTEM SID: %w", err) + } + entries := []windows.EXPLICIT_ACCESS{ + { + AccessPermissions: windows.GENERIC_ALL, + AccessMode: windows.GRANT_ACCESS, + Inheritance: windows.NO_INHERITANCE, + Trustee: windows.TRUSTEE{ + TrusteeForm: windows.TRUSTEE_IS_SID, + TrusteeType: windows.TRUSTEE_IS_USER, + TrusteeValue: windows.TrusteeValueFromSID(owner), + }, + }, + { + AccessPermissions: windows.GENERIC_ALL, + AccessMode: windows.GRANT_ACCESS, + Inheritance: windows.NO_INHERITANCE, + Trustee: windows.TRUSTEE{ + TrusteeForm: windows.TRUSTEE_IS_SID, + TrusteeType: windows.TRUSTEE_IS_USER, + TrusteeValue: windows.TrusteeValueFromSID(system), + }, + }, + } + acl, err := windows.ACLFromEntries(entries, nil) + if err != nil { + return nil, fmt.Errorf("build secret ACL: %w", err) + } + return acl, nil +} + +// writeWindowsSandboxSecret stores a principal's password readable only by the +// invoking user. +// +// The file is created empty, locked down, and only then written, so the secret +// is never on disk under the directory's inherited ACL. An existing file is +// replaced rather than appended, since a stale password would make LogonUser +// fail in a way that looks like a sandbox bug. +func writeWindowsSandboxSecret(path string, password string) error { + owner, err := currentTokenUserSID() + if err != nil { + return err + } + // THE PARENT CHAIN IS PART OF THE RESOLUTION COUNT. os.MkdirAll walks a + // pathname, so a junction planted on the way down pointed this elevated + // create at a directory of the caller's choosing, and the no-follow open + // below then refused a tree this process had already made outside the + // sandbox home. The chain is built the same way the ACL targets are: from + // the deepest existing ancestor, one component at a time relative to the + // handle above it, with a record of what was created so a later failure + // unwinds exactly that and nothing else. + created, parent, err := makeWindowsACLDirChainNoFollow(filepath.Dir(path)) + if err != nil { + if _, unwindErr := rollbackWindowsACLMaterialization(created); unwindErr != nil { + err = errors.Join(err, unwindErr) + } + return fmt.Errorf("create secret directory: %w", err) + } + defer func() { _ = windows.CloseHandle(parent) }() + unwindChain := func(cause error) error { + if _, unwindErr := rollbackWindowsACLMaterialization(created); unwindErr != nil { + return errors.Join(cause, unwindErr) + } + return cause + } + // Sealed BEFORE the file exists, so a failure here leaves nothing on disk at + // all rather than an empty file for someone else to win a race on. + // + // Encrypted to the invoking user on top of the ACL, so a copy taken outside + // the filesystem's enforcement (backup, disk image) is inert. The principal + // name is the entropy, which keeps one principal's blob from authenticating + // another. + sealed, err := protectWindowsSecret(password, windowsSandboxSecretEntropy(path)) + if err != nil { + return unwindChain(err) + } + // ONE resolution of the path, not four. The parent chain above was built + // relative to handles, the leaf is created relative to the deepest of them, + // and the DACL and the bytes are both applied to the leaf handle. Resolving + // the name again between any two of those steps is what let a junction or + // symlink swap point an elevated write at a file of the caller's choosing. + leaf := filepath.Base(path) + handle, err := createWindowsSecretFileAt(parent, leaf) + if err != nil { + return unwindChain(err) + } + leafClosed := false + closeLeaf := func() { + if leafClosed { + return + } + leafClosed = true + _ = windows.CloseHandle(handle) + } + defer closeLeaf() + // CLEANUP ACTS ON THE OBJECT THE CREATE MADE, NOT ON THE NAME AGAIN. + // + // This used to call os.Remove(path), which is the one pathname resolution + // the rest of this function was written to remove, and it did not even + // work: the leaf is created with FILE_SHARE_READ alone, so a delete by + // name while this handle is open comes back as a sharing violation. The + // error was discarded, unwindChain rolled back only the directories, and a + // failed provisioning left an empty secret file it believed it had removed. + // + // The handle is closed first and the leaf is then deleted relative to the + // pinned parent, the same way rollbackWindowsACLMaterialization removes + // what it created. Closing first rather than widening the share mask keeps + // the secret file unshareable for delete while it exists, which is the + // narrower of the two ways to make this work. + removeLeaf := func(cause error) error { + closeLeaf() + if deleteErr := deleteWindowsSecretLeafFn(parent, leaf); deleteErr != nil { + // Reported, not discarded: a secret file left behind after a failed + // provisioning is state the caller has to know about, and the empty + // file is the thing a later run would read as a stale password. + cause = errors.Join(cause, fmt.Errorf("remove partially written secret %s: %w", path, deleteErr)) + } + return unwindChain(cause) + } + if err := lockWindowsSecretHandleToOwnerFn(handle, owner); err != nil { + return removeLeaf(err) + } + if err := writeWindowsSecretHandle(handle, sealed); err != nil { + return removeLeaf(err) + } + return nil +} + +// lockWindowsSecretHandleToOwnerFn is a seam for the one step between creating +// the secret file and writing it, so the cleanup that runs when it fails is +// reachable without an elevated machine or a way to make SetSecurityInfo fail. +var lockWindowsSecretHandleToOwnerFn = lockWindowsSecretHandleToOwner + +// deleteWindowsSecretLeafFn is the matching seam for the cleanup itself, so the +// case where cleanup cannot finish is reachable without arranging a Windows +// sharing conflict, which is not something a test can stage reliably. +var deleteWindowsSecretLeafFn = deleteWindowsACLChildFile + +// windowsSandboxSecretEntropy derives the DPAPI entropy from the secret's own +// filename, which is the principal name. Deriving it rather than threading the +// name through keeps read and write agreeing by construction. +func windowsSandboxSecretEntropy(path string) string { + return strings.TrimSuffix(filepath.Base(path), ".secret") +} + +// Seamed so the permission-denied mapping in readWindowsSandboxSecret is +// testable. Producing a real ERROR_ACCESS_DENIED needs DACL surgery on Windows, +// since a 0000 file is still readable and reading a directory reports +// "Incorrect function", so a test built that way would exercise the platform +// rather than the mapping. +var readWindowsSandboxSecretFile = os.ReadFile + +// readWindowsSandboxSecret loads a principal's password. A missing file means +// setup has not run for this workspace, which the caller turns into a fallback +// rather than a hard failure. +func readWindowsSandboxSecret(path string) (string, error) { + data, err := readWindowsSandboxSecretFile(path) + if err != nil { + if os.IsNotExist(err) { + return "", errWindowsSandboxIdentityUnavailable + } + // Permission denied is unavailability, not breakage. The secret's DACL + // names whoever ran setup, so an operator who elevated with a separate + // administrative account, through runas or an over-the-shoulder UAC + // prompt, ends up with a secret their ordinary account cannot open. That + // is the documented fail-soft case: fall back to the restricted token and + // let the warning say so. Treating it as a hard error instead made every + // sandboxed command fail on a machine that was merely set up by a + // different admin, which is a common way to run an elevated setup. + if os.IsPermission(err) { + return "", errWindowsSandboxIdentityUnavailable + } + return "", fmt.Errorf("read sandbox secret: %w", err) + } + if len(data) == 0 { + return "", errWindowsSandboxIdentityUnavailable + } + secret, err := unprotectWindowsSecret(data, windowsSandboxSecretEntropy(path)) + if err != nil { + // A blob written by another user, for another principal, or by an older + // build that stored the password in the clear. Report it as unavailable so + // the caller falls back to the restricted token; the next elevated setup + // rewrites the secret in the current format. + return "", errWindowsSandboxIdentityUnavailable + } + if strings.TrimSpace(secret) == "" { + return "", errWindowsSandboxIdentityUnavailable + } + return secret, nil +} + +// removeWindowsSandboxSecret deletes a stored password. Called before the +// account itself is removed so a secret never outlives the principal it +// authenticates. +// errWindowsSandboxSecretNotOurs reports a secret this account cannot delete +// because another administrator's setup owns its DACL. +// +// Distinguishable so teardown can carry on. The alternative is what it used to +// do: abort before removing the account, its logon rights, its ACEs and its +// ledger, leaving a fully provisioned principal on the machine because one file +// could not be unlinked. +var errWindowsSandboxSecretNotOurs = errors.New("the sandbox secret belongs to a different administrator's setup") + +func removeWindowsSandboxSecret(path string) error { + err := os.Remove(path) + if err == nil || os.IsNotExist(err) { + return nil + } + // The read path already treats permission-denied as unavailability rather + // than breakage, for the reason documented on readWindowsSandboxSecret: an + // operator who elevated with a separate administrative account gets a secret + // their ordinary account cannot open. Removal has to agree. It did not, so on + // exactly those machines teardown stopped at the first file and stranded + // everything after it. + if os.IsPermission(err) { + return fmt.Errorf("%w: %s", errWindowsSandboxSecretNotOurs, path) + } + return fmt.Errorf("remove sandbox secret: %w", err) +} diff --git a/internal/sandbox/windows_identity_secret_windows_test.go b/internal/sandbox/windows_identity_secret_windows_test.go new file mode 100644 index 000000000..0293ec573 --- /dev/null +++ b/internal/sandbox/windows_identity_secret_windows_test.go @@ -0,0 +1,300 @@ +//go:build windows + +package sandbox + +import ( + "bytes" + "errors" + "fmt" + "os" + "path/filepath" + "strings" + "testing" + "unsafe" + + "golang.org/x/sys/windows" +) + +func TestWindowsSandboxSecretRoundTrip(t *testing.T) { + path := filepath.Join(t.TempDir(), "cfg", "zero-sbx-test.secret") + const password = "Zs1!EXAMPLEPASSWORDVALUE" + + if err := writeWindowsSandboxSecret(path, password); err != nil { + t.Fatalf("write: %v", err) + } + got, err := readWindowsSandboxSecret(path) + if err != nil { + t.Fatalf("read: %v", err) + } + if got != password { + t.Fatalf("read %q, want the stored password", got) + } +} + +// THE security property: the stored password must be readable only by the user +// who owns it. If any other trustee appears in the DACL, and in particular the +// sandbox principal, that account could mint its own token and the identity +// boundary would be worthless. +func TestWindowsSandboxSecretIsLockedToOwner(t *testing.T) { + path := filepath.Join(t.TempDir(), "locked.secret") + if err := writeWindowsSandboxSecret(path, "Zs1!SECRET"); err != nil { + t.Fatalf("write: %v", err) + } + + descriptor, err := windows.GetNamedSecurityInfo( + path, + windows.SE_FILE_OBJECT, + windows.DACL_SECURITY_INFORMATION, + ) + if err != nil { + t.Fatalf("read back security info: %v", err) + } + dacl, _, err := descriptor.DACL() + if err != nil { + t.Fatalf("read DACL: %v", err) + } + if dacl == nil { + t.Fatal("secret has a nil DACL, which grants everyone access") + } + + owner, err := currentTokenUserSID() + if err != nil { + t.Fatalf("owner SID: %v", err) + } + system, err := windows.CreateWellKnownSid(windows.WinLocalSystemSid) + if err != nil { + t.Fatalf("SYSTEM SID: %v", err) + } + + entries, err := windowsSecretACEList(dacl) + if err != nil { + t.Fatalf("enumerate ACEs: %v", err) + } + if len(entries) == 0 { + t.Fatal("secret DACL has no ACEs") + } + for _, sid := range entries { + if sid.Equals(owner) || sid.Equals(system) { + continue + } + t.Fatalf("secret DACL grants an unexpected trustee %s; only the owner and SYSTEM may appear", sid) + } +} + +// The DACL must be inheritance-protected, otherwise a permissive ACE on the +// config directory would still reach the secret. +func TestWindowsSandboxSecretDaclIsProtected(t *testing.T) { + path := filepath.Join(t.TempDir(), "protected.secret") + if err := writeWindowsSandboxSecret(path, "Zs1!SECRET"); err != nil { + t.Fatalf("write: %v", err) + } + descriptor, err := windows.GetNamedSecurityInfo(path, windows.SE_FILE_OBJECT, windows.DACL_SECURITY_INFORMATION) + if err != nil { + t.Fatalf("read back security info: %v", err) + } + control, _, err := descriptor.Control() + if err != nil { + t.Fatalf("read control bits: %v", err) + } + if control&windows.SE_DACL_PROTECTED == 0 { + t.Fatal("secret DACL is not protected, so inherited ACEs still apply") + } +} + +// Rewriting must replace the previous secret rather than append to it, or +// LogonUser would be handed two concatenated passwords. +func TestWindowsSandboxSecretOverwrites(t *testing.T) { + path := filepath.Join(t.TempDir(), "rewrite.secret") + if err := writeWindowsSandboxSecret(path, "Zs1!FIRST"); err != nil { + t.Fatalf("first write: %v", err) + } + if err := writeWindowsSandboxSecret(path, "Zs1!SECOND"); err != nil { + t.Fatalf("second write: %v", err) + } + got, err := readWindowsSandboxSecret(path) + if err != nil { + t.Fatalf("read: %v", err) + } + if got != "Zs1!SECOND" { + t.Fatalf("read %q, want only the newest password", got) + } +} + +// A workspace whose setup has not run must report the actionable sentinel so the +// command path falls back to the restricted-token backend instead of failing. +func TestWindowsSandboxSecretMissingIsSentinel(t *testing.T) { + path := filepath.Join(t.TempDir(), "absent.secret") + if _, err := readWindowsSandboxSecret(path); err != errWindowsSandboxIdentityUnavailable { + t.Fatalf("missing secret returned %v, want errWindowsSandboxIdentityUnavailable", err) + } +} + +// An empty file is a half-written secret, not a valid empty password. +// +// Both seeds matter and only one of them tests what the name says. A +// zero-length file is the truncated-write case, and it is the only one that +// reaches the length check. Whitespace is several bytes, so it travels on to +// DPAPI and fails to unprotect instead, which is the path +// TestWindowsSandboxSecretRejectsLegacyPlaintext already covers. Seeding only +// the whitespace, as this did, left the branch in the test's own name +// unexercised. +func TestWindowsSandboxSecretEmptyIsSentinel(t *testing.T) { + for name, seed := range map[string][]byte{ + "truncated write": {}, + "whitespace only": []byte(" \r\n"), + } { + t.Run(name, func(t *testing.T) { + path := filepath.Join(t.TempDir(), "empty.secret") + if err := os.WriteFile(path, seed, 0o600); err != nil { + t.Fatalf("seed: %v", err) + } + if _, err := readWindowsSandboxSecret(path); err != errWindowsSandboxIdentityUnavailable { + t.Fatalf("secret of %d bytes returned %v, want the unavailable sentinel", len(seed), err) + } + }) + } +} + +// The principal name lands in a filename, so anything that could escape the +// directory has to be refused even though windowsSandboxUserName already +// sanitises its output. +func TestWindowsSandboxSecretPathRejectsTraversal(t *testing.T) { + for _, name := range []string{`..\evil`, "sub/dir", `C:\abs`, "..", ""} { + if _, err := windowsSandboxSecretPath(`C:\cfg`, name); err == nil { + t.Fatalf("principal name %q was accepted", name) + } + } + if _, err := windowsSandboxSecretPath("", "zero-sbx-a"); err == nil { + t.Fatal("empty config directory was accepted") + } + path, err := windowsSandboxSecretPath(`C:\cfg`, "zero-sbx-abc") + if err != nil { + t.Fatalf("valid name rejected: %v", err) + } + if !strings.HasSuffix(path, `zero-sbx-abc.secret`) { + t.Fatalf("unexpected secret path %q", path) + } +} + +// Removal must be idempotent so teardown converges the same way provisioning +// does, and must actually delete the secret. +func TestWindowsSandboxSecretRemoveIsIdempotent(t *testing.T) { + path := filepath.Join(t.TempDir(), "gone.secret") + if err := writeWindowsSandboxSecret(path, "Zs1!SECRET"); err != nil { + t.Fatalf("write: %v", err) + } + if err := removeWindowsSandboxSecret(path); err != nil { + t.Fatalf("first remove: %v", err) + } + if _, err := os.Stat(path); !os.IsNotExist(err) { + t.Fatalf("secret still present after removal (stat err %v)", err) + } + if err := removeWindowsSandboxSecret(path); err != nil { + t.Fatalf("removing an absent secret must succeed, got %v", err) + } +} + +// windowsSecretACEList returns the trustee SID of every ACE in a DACL so a test +// can assert exactly who is named. +func windowsSecretACEList(dacl *windows.ACL) ([]*windows.SID, error) { + var out []*windows.SID + for index := uint32(0); index < uint32(dacl.AceCount); index++ { + var ace *windows.ACCESS_ALLOWED_ACE + if err := windows.GetAce(dacl, index, &ace); err != nil { + return nil, err + } + // GetAce hands back a generic ACE_HEADER and we reinterpret it. That is + // only sound for the fixed-layout types: an object ACE carries Flags and + // two GUIDs ahead of the trustee, so SidStart would land mid-structure + // and Copy would read whatever bytes happen to be there. The caller's + // "unexpected trustee" assertion would then print a nonsense SID instead + // of naming the ACE that does not belong, which is the opposite of what + // a failing test should do. Nothing under test builds anything but + // allowed ACEs today, so this exists to keep the failure legible if that + // ever changes. + if ace.Header.AceType != windows.ACCESS_ALLOWED_ACE_TYPE { + return nil, fmt.Errorf("ACE %d has type %d, want ACCESS_ALLOWED_ACE_TYPE (%d); refusing to decode its trustee", index, ace.Header.AceType, windows.ACCESS_ALLOWED_ACE_TYPE) + } + sid := (*windows.SID)(unsafe.Pointer(&ace.SidStart)) + copied, err := sid.Copy() + if err != nil { + return nil, err + } + out = append(out, copied) + } + return out, nil +} + +// DPAPI round-trip through the real store, which needs no privilege and so is +// genuine coverage rather than a gated stub. +func TestWindowsSandboxSecretRoundTripsThroughDPAPI(t *testing.T) { + path, err := windowsSandboxSecretPath(t.TempDir(), "zero-sbx-roundtrip") + if err != nil { + t.Fatalf("secret path: %v", err) + } + const password = "S0me-Sandbox-P@ssw0rd-value" + if err := writeWindowsSandboxSecret(path, password); err != nil { + t.Fatalf("write secret: %v", err) + } + got, err := readWindowsSandboxSecret(path) + if err != nil { + t.Fatalf("read secret: %v", err) + } + if got != password { + t.Fatalf("round-trip returned %q, want %q", got, password) + } + // The point of the exercise: the password must not be recoverable by reading + // the file, or the encryption layer is decorative. + raw, err := os.ReadFile(path) + if err != nil { + t.Fatalf("read raw secret file: %v", err) + } + if bytes.Contains(raw, []byte(password)) { + t.Fatal("the password appears verbatim in the stored file; it was not encrypted") + } +} + +// Entropy is the principal name, so a blob moved onto another principal's secret +// path must fail to decrypt rather than authenticate the wrong account. +func TestWindowsSandboxSecretDoesNotTransferBetweenPrincipals(t *testing.T) { + home := t.TempDir() + minePath, err := windowsSandboxSecretPath(home, "zero-sbx-mine") + if err != nil { + t.Fatalf("secret path: %v", err) + } + theirsPath, err := windowsSandboxSecretPath(home, "zero-sbx-theirs") + if err != nil { + t.Fatalf("secret path: %v", err) + } + if err := writeWindowsSandboxSecret(minePath, "a-password-for-mine"); err != nil { + t.Fatalf("write secret: %v", err) + } + blob, err := os.ReadFile(minePath) + if err != nil { + t.Fatalf("read blob: %v", err) + } + if err := os.WriteFile(theirsPath, blob, 0o600); err != nil { + t.Fatalf("plant blob: %v", err) + } + if _, err := readWindowsSandboxSecret(theirsPath); !errors.Is(err, errWindowsSandboxIdentityUnavailable) { + t.Fatalf("a blob planted at another principal's path decrypted; got err = %v", err) + } +} + +// An older plaintext secret must degrade to a fallback rather than being handed +// to LogonUser as if it were a password. +func TestWindowsSandboxSecretRejectsLegacyPlaintext(t *testing.T) { + path, err := windowsSandboxSecretPath(t.TempDir(), "zero-sbx-legacy") + if err != nil { + t.Fatalf("secret path: %v", err) + } + if err := os.MkdirAll(filepath.Dir(path), 0o700); err != nil { + t.Fatalf("mkdir: %v", err) + } + if err := os.WriteFile(path, []byte("plaintext-password"), 0o600); err != nil { + t.Fatalf("write legacy secret: %v", err) + } + if _, err := readWindowsSandboxSecret(path); !errors.Is(err, errWindowsSandboxIdentityUnavailable) { + t.Fatalf("legacy plaintext secret was accepted; got err = %v", err) + } +} diff --git a/internal/sandbox/windows_identity_windows.go b/internal/sandbox/windows_identity_windows.go new file mode 100644 index 000000000..b26885d2f --- /dev/null +++ b/internal/sandbox/windows_identity_windows.go @@ -0,0 +1,1131 @@ +//go:build windows + +package sandbox + +// Windows sandbox principals. +// +// Every other Windows backend here derives its token from the CALLING user via +// CreateRestrictedToken, which is why the sandbox can constrain writes but not +// reads: a deny ACE that would stop the sandboxed child reading a credential +// store names the same account Zero itself runs as, so it would lock Zero out +// too. Reads therefore stay on the caller's identity and +// credentialDenyReadPaths is a no-op on Windows (#662, #675). +// +// This file provisions a SEPARATE local account per workspace, held in one +// managed local group, so the sandbox has an identity of its own. A deny-read +// ACE naming that principal denies the sandboxed child and nothing else, and +// the same SID is what a firewall rule or a write grant can be keyed to. The +// accounts are created by the elevated `zero sandbox setup` path because +// NetUserAdd requires administrator rights; nothing here runs unelevated. +// +// Provisioning is idempotent: the "already exists" status from each API is a +// success, so setup can be re-run safely and a partially provisioned machine +// converges. + +import ( + "crypto/rand" + "encoding/base32" + "errors" + "fmt" + "os" + "runtime" + "strings" + "unsafe" + + "golang.org/x/sys/windows" +) + +const ( + // windowsSandboxGroupName holds every sandbox principal. Grouping them means + // an ACE can name the group once instead of enumerating accounts, and it + // gives setup a single place to find what it previously created. + windowsSandboxGroupName = "ZeroSandboxUsers" + windowsSandboxGroupComment = "Zero sandbox principals (managed by zero sandbox setup)" + + // windowsSandboxUserPrefix keeps the accounts recognisable in `net user` and + // lets cleanup identify what belongs to Zero. Windows caps a local account + // name at 20 characters, which windowsSandboxUserName respects. + windowsSandboxUserPrefix = "zero-sbx-" + // The comment doubles as the ownership marker AND records which workspace + // the account belongs to. The account NAME can only carry 11 characters of + // the workspace digest because of the 20-character local-account limit, so + // two workspaces whose digests share that prefix derive the same name. The + // full key here turns that from a silent share of one account, one secret + // and one ACL identity into a refusal. + windowsSandboxUserComment = "Zero sandbox principal (managed)" + windowsSandboxUserCommentKey = windowsSandboxUserComment + " key=" + windowsSandboxUserNameMax = 20 + + // windowsSandboxOfflineGroupName is what the network block filters are keyed + // to. Network denial cannot be expressed on a principal's own token the way + // it is on a restricted token: the block filters match the offline-marker + // SID, which is a synthetic capability SID, and LogonUser mints a token from + // an account's real group memberships rather than from arbitrary SIDs. A + // principal is therefore denied the network by being a MEMBER of this group, + // whose SID the filters also match. + // + // A group rather than the offline principal's own SID because principals are + // per workspace: one filter set covers every offline principal on the machine + // instead of needing a filter per workspace. + windowsSandboxOfflineGroupName = "ZeroSandboxOffline" + windowsSandboxOfflineGroupComment = "Zero sandbox principals denied network access (managed)" +) + +// windowsSandboxRole distinguishes the two principals a workspace gets. They are +// separate accounts rather than one account reconfigured per command, because +// the network decision is baked into group membership at setup time and setup +// needs elevation; flipping it per command would need an elevated hop on every +// command. +type windowsSandboxRole string + +const ( + // windowsSandboxRoleOffline is a member of the offline group, so the block + // filters match its token and it has no network. + windowsSandboxRoleOffline windowsSandboxRole = "offline" + // windowsSandboxRoleOnline is not, so an approved network command reaches the + // network while still being read-confined by having its own identity. + windowsSandboxRoleOnline windowsSandboxRole = "online" + // windowsSandboxRoleLegacy names the SINGLE UNTAGGED account this branch's + // predecessor provisioned as "zero-sbx-", before the roles were split. + // + // It is never provisioned and must never be passed to a provisioning path: + // it exists so the ordered retirement below can still derive that name and + // remove the account, its secret, its logon rights, its ACEs and its ledger. + // Without it an upgraded machine keeps a fully privileged principal that + // nothing afterwards looks for, because both roles report themselves absent. + windowsSandboxRoleLegacy windowsSandboxRole = "legacy" +) + +// TWO INVENTORIES, because they are two different questions and answering one +// with the other is a bug in both directions. +// +// windowsSandboxLiveRoles are the roles setup PROVISIONS. Legacy must never +// appear here: it names the single untagged pre-split account, and provisioning +// it would create that account fresh on every setup of an already-upgraded +// machine. +// +// windowsSandboxRetirableRoles are every role that may EXIST on a machine, which +// is the live pair plus legacy. Teardown and the is-it-still-installed check +// both belong to this one. Asking the live list instead let an opted-out marker +// report a workspace clean while the legacy account was still on the box, which +// is precisely the lie that check exists to prevent. +var ( + windowsSandboxLiveRoles = []windowsSandboxRole{windowsSandboxRoleOffline, windowsSandboxRoleOnline} + windowsSandboxRetirableRoles = append(append([]windowsSandboxRole{}, windowsSandboxLiveRoles...), windowsSandboxRoleLegacy) +) + +// roleTag is the single character that distinguishes the two accounts for a +// workspace. One character because the 20-character account-name limit is +// already tight and every character spent here is a bit of workspace hash lost. +func (role windowsSandboxRole) roleTag() string { + switch role { + case windowsSandboxRoleOnline: + return "n" + case windowsSandboxRoleLegacy: + // Deliberately empty: this reproduces the pre-split name exactly, which + // is the only way retirement can find what the predecessor installed. + return "" + default: + return "d" + } +} + +// Win32 status codes that mean "already there". Treated as success so +// provisioning converges instead of failing on a second run. +const ( + nerrSuccess = 0 + nerrGroupExists = 2223 + // NERR_GroupNotFound, the group half of nerrUserNotFound below. + nerrGroupNotFound = 2220 + nerrUserExists = 2224 + errorAliasExists = 1379 + errorMemberInAlias = 1378 + errorAccessDenied32 = 5 + nerrUserNotFound = 2221 +) + +// USER_INFO_1 privilege and flag values. +const ( + usrPrivUser = 1 + ufScript = 0x0001 + ufNormalAccount = 0x0200 + ufDontExpirePasswd = 0x10000 + windowsPasswordLength = 24 +) + +var ( + netapi32 = windows.NewLazySystemDLL("netapi32.dll") + procNetUserAdd = netapi32.NewProc("NetUserAdd") + procNetLocalGroupAdd = netapi32.NewProc("NetLocalGroupAdd") + procNetLocalGroupGetInfo = netapi32.NewProc("NetLocalGroupGetInfo") + procNetLocalGroupAddMembers = netapi32.NewProc("NetLocalGroupAddMembers") + procNetUserDel = netapi32.NewProc("NetUserDel") + procNetUserSetInfo = netapi32.NewProc("NetUserSetInfo") + procNetUserGetInfo = netapi32.NewProc("NetUserGetInfo") + procNetApiBufferFree = netapi32.NewProc("NetApiBufferFree") + procNetUserGetLocalGroups = netapi32.NewProc("NetUserGetLocalGroups") +) + +// userInfo1 mirrors USER_INFO_1. Field order and widths must match the Win32 +// struct exactly; it is passed to NetUserAdd as a raw buffer. +type userInfo1 struct { + Name *uint16 + Password *uint16 + PasswordAge uint32 + Priv uint32 + HomeDir *uint16 + Comment *uint16 + Flags uint32 + ScriptPath *uint16 +} + +// userInfo1003 mirrors USER_INFO_1003, the password-only form NetUserSetInfo +// takes when nothing else about the account should change. +type userInfo1003 struct { + Password *uint16 +} + +// userInfo1007 mirrors USER_INFO_1007, the comment-only form. Used to finish +// the legacy-comment upgrade windowsSandboxUserIsManaged promises. +type userInfo1007 struct { + Comment *uint16 +} + +// localGroupInfo1 mirrors LOCALGROUP_INFO_1. +type localGroupInfo1 struct { + Name *uint16 + Comment *uint16 +} + +// localGroupMembersInfo3 mirrors LOCALGROUP_MEMBERS_INFO_3, which identifies a +// member by name rather than SID. +type localGroupMembersInfo3 struct { + DomainAndName *uint16 +} + +// windowsSandboxIdentity is a provisioned sandbox principal: the account name +// and the SID that ACEs, tokens and firewall rules are keyed to. +type windowsSandboxIdentity struct { + Username string + SID *windows.SID +} + +// String renders the identity for logs without exposing the password, which is +// never stored on this struct. +func (identity windowsSandboxIdentity) String() string { + if identity.SID == nil { + return identity.Username + } + return identity.Username + " (" + identity.SID.String() + ")" +} + +// windowsSandboxUserName derives a stable account name for a workspace key and +// role. The key is hashed by the caller (see windowsSandboxWorkspaceKey) so the +// name reveals no path, and it is truncated to the 20-character local-account +// limit. The same workspace and role always map to the same account, so +// re-running setup reuses the principals instead of accumulating accounts. +// +// The role tag sits before the hash rather than after it, so truncation eats +// hash characters and never the tag. A name that lost its tag would collide the +// two roles onto one account, which would silently put an online principal in +// the offline group or the reverse. +func windowsSandboxUserName(workspaceKey string, role windowsSandboxRole) string { + cleaned := strings.Map(func(r rune) rune { + switch { + case r >= 'a' && r <= 'z', r >= '0' && r <= '9': + return r + case r >= 'A' && r <= 'Z': + return r + ('a' - 'A') + default: + return -1 + } + }, workspaceKey) + if cleaned == "" { + cleaned = "default" + } + name := windowsSandboxUserPrefix + role.roleTag() + cleaned + if len(name) > windowsSandboxUserNameMax { + name = name[:windowsSandboxUserNameMax] + } + return name +} + +// windowsSandboxUserCommentFor returns the ownership comment for a workspace, +// carrying the full key the account name could only hold 11 characters of. +func windowsSandboxUserCommentFor(workspaceKey string) string { + return windowsSandboxUserCommentKey + workspaceKey +} + +// newWindowsSandboxPassword returns a random password for a sandbox principal. +// The account is never signed into interactively: the password exists only so +// LogonUser can mint a token for it, so it is generated per provisioning run, +// handed straight to the caller, and never persisted by this file. Base32 of +// crypto/rand bytes keeps it alphanumeric, which satisfies complexity policies +// that reject unusual punctuation, and a fixed suffix guarantees the mixed-case +// and digit classes even if the random draw happens to omit one. +func newWindowsSandboxPassword() (string, error) { + raw := make([]byte, windowsPasswordLength) + if _, err := rand.Read(raw); err != nil { + return "", fmt.Errorf("generate sandbox password: %w", err) + } + encoded := base32.StdEncoding.WithPadding(base32.NoPadding).EncodeToString(raw) + if len(encoded) > windowsPasswordLength { + encoded = encoded[:windowsPasswordLength] + } + return "Zs1!" + encoded, nil +} + +// netAPIStatus converts a netapi32 return value into an error, treating the +// supplied status codes as success so callers can spell out which "already +// exists" results are expected. +func netAPIStatus(call string, status uintptr, okStatuses ...uintptr) error { + if status == nerrSuccess { + return nil + } + for _, ok := range okStatuses { + if status == ok { + return nil + } + } + if status == errorAccessDenied32 { + return fmt.Errorf("%s: access denied (run `zero sandbox setup` from an elevated terminal)", call) + } + return fmt.Errorf("%s: status %d", call, status) +} + +// ensureWindowsSandboxGroup creates the managed local group, or adopts it when +// it already exists AND carries our ownership comment. +func ensureWindowsSandboxGroup() error { + return ensureWindowsLocalGroup(windowsSandboxGroupName, windowsSandboxGroupComment) +} + +// ensureWindowsSandboxOfflineGroup creates the group the network block filters +// are keyed to. Membership of it is what denies a principal the network, so it +// has to exist before any offline principal is created. +func ensureWindowsSandboxOfflineGroup() error { + return ensureWindowsLocalGroup(windowsSandboxOfflineGroupName, windowsSandboxOfflineGroupComment) +} + +// Wires the portable network planner to the Windows group lookup. Done here so +// windows_network.go can stay free of Win32 calls while still folding the group +// into the filter identity set. +func init() { + resolveWindowsSandboxOfflineGroupSIDHook = resolveWindowsSandboxOfflineGroupSID +} + +// resolveWindowsSandboxOfflineGroupSID returns the SID the block filters key to. +// +// Reports ("", nil) when the group does not exist, which is the state before +// setup has ever run. Callers fold that into "no extra identity", so the plan +// they compute matches what an un-provisioned machine would produce rather than +// failing to build at all. +func resolveWindowsSandboxOfflineGroupSID() (string, error) { + sid, _, accountType, err := windows.LookupSID("", windowsSandboxOfflineGroupName) + if err != nil { + if errors.Is(err, windows.ERROR_NONE_MAPPED) { + return "", nil + } + return "", fmt.Errorf("look up %s: %w", windowsSandboxOfflineGroupName, err) + } + // A local group is an alias. Anything else means the name has been taken by + // something that is not ours, and keying network filters to it would be both + // wrong and a way to affect unrelated accounts. + if accountType != windows.SidTypeAlias && accountType != windows.SidTypeGroup { + return "", fmt.Errorf("%s resolves to a non-group account (type %d)", windowsSandboxOfflineGroupName, accountType) + } + // Being a group is not enough: it has to be OUR group. + // + // The ownership check used to live only on the principal-provisioning path, + // so an opt-out setup reached here and adopted whatever carried the name. + // applyWindowsNetworkPlan turns every SID in the plan into an allowed-to-match + // WFP descriptor, so adopting a foreign group installs global deny filters + // against every one of ITS members: someone who happens to have a local group + // by this name loses the network for those accounts because we ran setup. + // + // Refused rather than skipped. Returning "" would build a plan whose filters + // cover no principal at all while reporting a successful setup, and a network + // deny that silently covers nothing is the failure this backend exists to + // prevent. A squatted name is an operator problem and has to say so. + owned, err := windowsLocalGroupOwnedByZero(windowsSandboxOfflineGroupName, windowsSandboxOfflineGroupComment) + if err != nil { + return "", err + } + if !owned { + return "", fmt.Errorf("%s exists but is not the group this setup manages; rename or remove it, then re-run `zero sandbox setup`", windowsSandboxOfflineGroupName) + } + return sid.String(), nil +} + +// ensureWindowsLocalGroup creates a local group, or leaves it alone when it +// already exists. +func ensureWindowsLocalGroup(groupName string, groupComment string) error { + status, err := addWindowsLocalGroupFn(groupName, groupComment) + if err != nil { + return err + } + if status == nerrGroupExists || status == errorAliasExists { + // A group with this name already exists, which is the normal re-run case + // — but only if it is OURS. The offline group's SID is installed on the + // persistent WFP deny filters and the sandbox principal is made a member + // of it, so silently adopting a same-named group created by some other + // tool or by policy would cut off every existing member's outbound + // traffic and hand our principal that group's permissions. + owned, err := windowsLocalGroupOwnedByZeroFn(groupName, groupComment) + if err != nil { + return err + } + if !owned { + return fmt.Errorf("local group %q already exists and is not managed by zero; "+ + "rename or remove it before running sandbox setup", groupName) + } + return nil + } + return netAPIStatus("NetLocalGroupAdd", status) +} + +// Seams for the two Win32 calls behind group creation, so the already-exists +// branch is reachable in tests without an elevated machine. +var ( + addWindowsLocalGroupFn = addWindowsLocalGroup + windowsLocalGroupOwnedByZeroFn = windowsLocalGroupOwnedByZero +) + +// addWindowsLocalGroup issues NetLocalGroupAdd and hands back its raw status so +// the caller can distinguish "already exists" from a real failure. +func addWindowsLocalGroup(groupName string, groupComment string) (uintptr, error) { + name, err := windows.UTF16PtrFromString(groupName) + if err != nil { + return 0, err + } + comment, err := windows.UTF16PtrFromString(groupComment) + if err != nil { + return 0, err + } + info := localGroupInfo1{Name: name, Comment: comment} + status, _, _ := procNetLocalGroupAdd.Call( + 0, // local machine + 1, // level: LOCALGROUP_INFO_1 + uintptr(unsafe.Pointer(&info)), + 0, + ) + // The struct holds pointers into Go memory that the syscall dereferences, so + // it has to stay reachable until the call has returned. + runtime.KeepAlive(info) + runtime.KeepAlive(name) + runtime.KeepAlive(comment) + return status, nil +} + +// windowsLocalGroupOwnedByZero reports whether an existing local group carries +// the managed-group marker this setup writes, so a foreign group that merely +// shares the name is never adopted. +func windowsLocalGroupOwnedByZero(groupName string, wantComment string) (bool, error) { + name, err := windows.UTF16PtrFromString(groupName) + if err != nil { + return false, err + } + var buffer *byte + status, _, _ := procNetLocalGroupGetInfo.Call( + 0, // local machine + uintptr(unsafe.Pointer(name)), + 1, // level: LOCALGROUP_INFO_1 + uintptr(unsafe.Pointer(&buffer)), + ) + runtime.KeepAlive(name) + if status == nerrGroupNotFound { + // It existed a moment ago and does not now. Treat that as not-ours + // rather than guessing; the next setup run recreates it cleanly. + return false, nil + } + if err := netAPIStatus("NetLocalGroupGetInfo", status); err != nil { + return false, err + } + if buffer == nil { + return false, nil + } + defer procNetApiBufferFree.Call(uintptr(unsafe.Pointer(buffer))) + info := (*localGroupInfo1)(unsafe.Pointer(buffer)) + if info.Comment == nil { + return false, nil + } + return windows.UTF16PtrToString(info.Comment) == wantComment, nil +} + +// ensureWindowsSandboxUser creates a sandbox account with the supplied password. +// The account is a plain local user with no home directory or logon script, +// flagged so its password never expires (nobody is there to rotate it) and so it +// is a normal, enabled account LogonUser can authenticate. +// +// It reports whether the account already existed, because NetUserAdd leaves such +// an account completely untouched, password included. The caller has to reset it +// or the secret it goes on to store would not be the account's password at all. +func ensureWindowsSandboxUser(username string, password string, workspaceKey string) (bool, error) { + name, err := windows.UTF16PtrFromString(username) + if err != nil { + return false, err + } + secret, err := windows.UTF16PtrFromString(password) + if err != nil { + return false, err + } + comment, err := windows.UTF16PtrFromString(windowsSandboxUserCommentFor(workspaceKey)) + if err != nil { + return false, err + } + info := userInfo1{ + Name: name, + Password: secret, + Priv: usrPrivUser, + Comment: comment, + Flags: ufScript | ufNormalAccount | ufDontExpirePasswd, + } + status, _, _ := procNetUserAdd.Call( + 0, // local machine + 1, // level: USER_INFO_1 + uintptr(unsafe.Pointer(&info)), + 0, + ) + // The struct holds pointers into Go memory that the call dereferences, so + // everything it borrows has to outlive the call. + runtime.KeepAlive(info) + runtime.KeepAlive(name) + runtime.KeepAlive(secret) + runtime.KeepAlive(comment) + if status == nerrUserExists { + return true, nil + } + return false, netAPIStatus("NetUserAdd", status) +} + +// resetWindowsSandboxUserPassword sets the password on an account that already +// existed, so the secret the caller stores is actually the account's password. +// +// Without this, re-running setup produced a fresh random password, wrote it to +// disk, and left the account authenticating with the old one, so every later +// command failed to log on with a principal that looked correctly provisioned. +func resetWindowsSandboxUserPassword(username string, password string) error { + name, err := windows.UTF16PtrFromString(username) + if err != nil { + return err + } + secret, err := windows.UTF16PtrFromString(password) + if err != nil { + return err + } + // USER_INFO_1003 is a password-only update, so nothing else about the + // account is disturbed. + info := userInfo1003{Password: secret} + status, _, _ := procNetUserSetInfo.Call( + 0, // local machine + uintptr(unsafe.Pointer(name)), + 1003, // level: USER_INFO_1003 + uintptr(unsafe.Pointer(&info)), + 0, + ) + runtime.KeepAlive(info) + runtime.KeepAlive(name) + runtime.KeepAlive(secret) + return netAPIStatus("NetUserSetInfo", status) +} + +// addWindowsSandboxUserToGroup puts a principal in the managed group, ignoring +// the status that means it is already a member. +func addWindowsSandboxUserToGroup(username string) error { + return addWindowsUserToLocalGroup(windowsSandboxGroupName, username) +} + +// addWindowsSandboxUserToOfflineGroup is what actually denies a principal the +// network: the block filters match this group's SID, and LogonUser puts the +// group into the token it mints for a member. +func addWindowsSandboxUserToOfflineGroup(username string) error { + return addWindowsUserToLocalGroup(windowsSandboxOfflineGroupName, username) +} + +// addWindowsUserToLocalGroup puts an account in a local group, ignoring the +// status that means it is already a member. +func addWindowsUserToLocalGroup(groupName string, username string) error { + group, err := windows.UTF16PtrFromString(groupName) + if err != nil { + return err + } + member, err := windows.UTF16PtrFromString(username) + if err != nil { + return err + } + entry := localGroupMembersInfo3{DomainAndName: member} + status, _, _ := procNetLocalGroupAddMembers.Call( + 0, // local machine + uintptr(unsafe.Pointer(group)), + 3, // level: LOCALGROUP_MEMBERS_INFO_3 + uintptr(unsafe.Pointer(&entry)), + 1, // one member + ) + runtime.KeepAlive(entry) + runtime.KeepAlive(group) + runtime.KeepAlive(member) + return netAPIStatus("NetLocalGroupAddMembers", status, errorMemberInAlias) +} + +// errWindowsSandboxNameCollision reports that the derived account name is taken +// by a local account Zero did not create. Setup refuses rather than adopting it. +var errWindowsSandboxPrivilegedAccount = errors.New("the local account matching Zero's derived sandbox name belongs to a privileged group (Administrators, Power Users or Backup Operators); refusing to adopt it as a sandbox principal") + +var errWindowsSandboxNameCollision = errors.New("a local account with Zero's derived sandbox name already exists and was not created by Zero") + +// windowsSandboxUserIsManaged reports whether a local account is one Zero +// created, by reading back the comment provisioning stamps on it. +// +// This is the gate on adopting an existing account. The name is derived, not +// discovered, so an account can be sitting on it for reasons that have nothing +// to do with Zero, and taking it over means resetting a stranger's password. +// +// A missing account is not managed rather than an error, so callers can use this +// as a plain question without special-casing absence. +func windowsSandboxUserIsManaged(username string, workspaceKey string) (bool, error) { + name, err := windows.UTF16PtrFromString(username) + if err != nil { + return false, err + } + var buffer *byte + status, _, _ := procNetUserGetInfo.Call( + 0, // local machine + uintptr(unsafe.Pointer(name)), + 1, // level: USER_INFO_1 + uintptr(unsafe.Pointer(&buffer)), + ) + runtime.KeepAlive(name) + if status == nerrUserNotFound { + return false, nil + } + if err := netAPIStatus("NetUserGetInfo", status); err != nil { + return false, err + } + if buffer == nil { + return false, nil + } + defer procNetApiBufferFree.Call(uintptr(unsafe.Pointer(buffer))) + info := (*userInfo1)(unsafe.Pointer(buffer)) + if info.Comment == nil { + return false, nil + } + comment := windows.UTF16PtrToString(info.Comment) + // An account provisioned before the key was recorded is still ours; it + // predates this check and cannot be attributed to a workspace, so it is + // adopted and its comment rewritten on the way through. The rewrite is + // upgradeWindowsSandboxUserComment, called from the adoption path. + if comment == windowsSandboxUserComment { + return true, nil + } + return comment == windowsSandboxUserCommentFor(workspaceKey), nil +} + +// windowsSandboxUserHasLegacyComment reports an account carrying the OLD bare +// ownership comment, the one with no workspace key. +// +// Separate from windowsSandboxUserIsManaged rather than folded into it so that +// function keeps its (bool, error) shape and the test seam around it does not +// have to change. Both read the same field; this one runs on the adoption path +// only, at setup time, so the second lookup costs nothing that matters. +func windowsSandboxUserHasLegacyComment(username string) (bool, error) { + name, err := windows.UTF16PtrFromString(username) + if err != nil { + return false, err + } + var buffer *byte + status, _, _ := procNetUserGetInfo.Call( + 0, // local machine + uintptr(unsafe.Pointer(name)), + 1, // level: USER_INFO_1 + uintptr(unsafe.Pointer(&buffer)), + ) + runtime.KeepAlive(name) + if status == nerrUserNotFound { + return false, nil + } + if err := netAPIStatus("NetUserGetInfo", status); err != nil { + return false, err + } + if buffer == nil { + return false, nil + } + defer procNetApiBufferFree.Call(uintptr(unsafe.Pointer(buffer))) + info := (*userInfo1)(unsafe.Pointer(buffer)) + if info.Comment == nil { + return false, nil + } + return windows.UTF16PtrToString(info.Comment) == windowsSandboxUserComment, nil +} + +// upgradeWindowsSandboxUserComment stamps the workspace-keyed ownership comment +// onto an account still carrying the legacy bare one. +// +// windowsSandboxUserIsManaged has always said the comment is "rewritten on the +// way through", and it never was: provisioning only ever set the password. So a +// pre-key account stayed permanently unattributable, and every later run had to +// keep accepting the bare comment to avoid orphaning it. Writing the key ends +// that, and lets the bare form eventually be retired. +// +// Not fatal on failure. The account is adopted and usable either way; refusing +// to provision because a cosmetic stamp did not take would strand a working +// sandbox over bookkeeping. +func upgradeWindowsSandboxUserComment(username string, workspaceKey string) error { + name, err := windows.UTF16PtrFromString(username) + if err != nil { + return err + } + comment, err := windows.UTF16PtrFromString(windowsSandboxUserCommentFor(workspaceKey)) + if err != nil { + return err + } + info := userInfo1007{Comment: comment} + status, _, _ := procNetUserSetInfo.Call( + 0, // local machine + uintptr(unsafe.Pointer(name)), + 1007, // level: USER_INFO_1007, comment only + uintptr(unsafe.Pointer(&info)), + 0, // no parameter-error index + ) + runtime.KeepAlive(name) + runtime.KeepAlive(comment) + return netAPIStatus("NetUserSetInfo", status) +} + +// localGroupUsersInfo0 mirrors LOCALGROUP_USERS_INFO_0: one group name pointer. +type localGroupUsersInfo0 struct { + Name *uint16 +} + +// windowsSandboxUserIsPrivileged reports whether an account belongs to a local +// group that would make it a poor sandbox principal. +// +// Adoption is the reason this exists. Provisioning will take over an account +// whose name and ownership comment match, and an account that is also in +// Administrators would hand the sandbox exactly the rights the sandbox is meant +// to withhold: it could rewrite the ACLs confining it, read the secret locked to +// the invoking user, and terminate Zero. The name is derived rather than +// discovered, so an account can end up matching without anyone intending it. +// +// Membership is resolved by SID rather than by name so a localised install, where +// the group is called Administrateurs or Administratoren, is still recognised. +func windowsSandboxUserIsPrivileged(username string) (bool, error) { + name, err := windows.UTF16PtrFromString(username) + if err != nil { + return false, err + } + var ( + buffer *byte + entries uint32 + total uint32 + ) + status, _, _ := procNetUserGetLocalGroups.Call( + 0, // local machine + uintptr(unsafe.Pointer(name)), + 0, // level: LOCALGROUP_USERS_INFO_0 + 0, // flags: direct membership only + uintptr(unsafe.Pointer(&buffer)), + uintptr(^uint32(0)), // MAX_PREFERRED_LENGTH + uintptr(unsafe.Pointer(&entries)), + uintptr(unsafe.Pointer(&total)), + ) + runtime.KeepAlive(name) + if status == nerrUserNotFound { + return false, nil + } + if err := netAPIStatus("NetUserGetLocalGroups", status); err != nil { + return false, err + } + if buffer == nil || entries == 0 { + return false, nil + } + defer procNetApiBufferFree.Call(uintptr(unsafe.Pointer(buffer))) + + privileged, err := privilegedLocalGroupNames() + if err != nil { + return false, err + } + groups := unsafe.Slice((*localGroupUsersInfo0)(unsafe.Pointer(buffer)), entries) + for _, group := range groups { + if group.Name == nil { + continue + } + if privileged[strings.ToLower(windows.UTF16PtrToString(group.Name))] { + return true, nil + } + } + return false, nil +} + +// privilegedLocalGroupNames resolves the local names of the groups a sandbox +// principal must not belong to. Resolved from well-known SIDs so the comparison +// survives a localised Windows install. +func privilegedLocalGroupNames() (map[string]bool, error) { + out := map[string]bool{} + for _, wellKnown := range []windows.WELL_KNOWN_SID_TYPE{ + windows.WinBuiltinAdministratorsSid, + windows.WinBuiltinPowerUsersSid, + windows.WinBuiltinBackupOperatorsSid, + } { + sid, err := windows.CreateWellKnownSid(wellKnown) + if err != nil { + // A group this build of Windows does not define is not a membership + // anyone can hold, so it cannot make an account privileged. + continue + } + account, _, _, err := sid.LookupAccount("") + if err != nil { + continue + } + out[strings.ToLower(account)] = true + } + if len(out) == 0 { + return nil, errors.New("could not resolve any privileged local group name") + } + return out, nil +} + +// resolveWindowsSandboxSID looks up the SID for a provisioned principal. The SID +// is the durable handle: account names can collide with a pre-existing local +// user, so every ACE and firewall rule is keyed to the SID rather than the name. +func resolveWindowsSandboxSID(username string) (*windows.SID, error) { + sid, _, accountType, err := windows.LookupSID("", username) + if err != nil { + return nil, fmt.Errorf("look up sandbox principal %q: %w", username, err) + } + if accountType != windows.SidTypeUser { + return nil, fmt.Errorf("sandbox principal %q resolves to a non-user account (type %d)", username, accountType) + } + return sid, nil +} + +// Indirected so a test can drive provisioning end to end and inject a failure +// at the two points that occur AFTER the account exists. Those are the paths +// whose return value the caller's rollback depends on. +// +// All four are seamed rather than just the last two: every step here needs an +// elevated caller and a real local account, so a test that only replaced the +// post-creation pair would never get past ensureWindowsSandboxGroup on an +// ordinary machine and would pass without reaching the code it names. +var ( + ensureWindowsSandboxGroupFn = ensureWindowsSandboxGroup + ensureWindowsSandboxOfflineGroupFn = ensureWindowsSandboxOfflineGroup + ensureWindowsSandboxUserFn = ensureWindowsSandboxUser + addWindowsSandboxUserToGroupFn = addWindowsSandboxUserToGroup + addWindowsSandboxUserToOfflineGroupFn = addWindowsSandboxUserToOfflineGroup + resolveWindowsSandboxSIDFn = resolveWindowsSandboxSID + resetWindowsSandboxUserPasswordFn = resetWindowsSandboxUserPassword + windowsSandboxUserIsManagedFn = windowsSandboxUserIsManaged + windowsSandboxUserHasLegacyCommentFn = windowsSandboxUserHasLegacyComment + // Seamed so the upgrade CALL SITE can be observed. The legacy-comment probe + // was already seamed and the upgrade itself was not, so nothing could tell + // whether it ran: deleting the call left the whole suite green, which is + // exactly how a fix gets silently reverted by a later change. + upgradeWindowsSandboxUserCommentFn = upgradeWindowsSandboxUserComment + windowsSandboxUserIsPrivilegedFn = windowsSandboxUserIsPrivileged + grantWindowsSandboxLogonRightsFn = grantWindowsSandboxLogonRights + revokeWindowsSandboxLogonRightsFn = revokeWindowsSandboxLogonRights + // applyWindowsACLPlanFn is a seam so a test can pin the ORDER of setup's ACL + // work. The revocation only prevents a stale grant if it runs before the plan + // that re-adds the current one; a test that exercised the revoke helper on its + // own would pass just as happily with the call site deleted. + applyWindowsACLPlanFn = applyWindowsACLPlan +) + +// provisionWindowsSandboxIdentity ensures the managed group and one sandbox +// principal for workspaceKey exist, and returns the identity plus the password +// the caller needs to mint a token with LogonUser. It is idempotent, so setup +// can run repeatedly. +// +// The password is returned rather than stored, so no credential is written to +// disk here; that happens a layer up where the secret has somewhere safe to +// live. The returned value is always the account's actual password, including +// when the account already existed, because that case is reset explicitly +// below. +func provisionWindowsSandboxIdentity(workspaceKey string, role windowsSandboxRole) (windowsSandboxIdentity, string, bool, error) { + if err := ensureWindowsSandboxGroupFn(); err != nil { + return windowsSandboxIdentity{}, "", false, err + } + // The offline group has to exist before an offline principal joins it, and it + // is created unconditionally so the network filters can be keyed to its SID + // whether or not an offline principal has been provisioned yet. + if err := ensureWindowsSandboxOfflineGroupFn(); err != nil { + return windowsSandboxIdentity{}, "", false, err + } + username := windowsSandboxUserName(workspaceKey, role) + password, err := newWindowsSandboxPassword() + if err != nil { + return windowsSandboxIdentity{}, "", false, err + } + existed, err := ensureWindowsSandboxUserFn(username, password, workspaceKey) + if err != nil { + return windowsSandboxIdentity{}, "", false, err + } + if existed { + // Prove the account is ours before touching it. The name is derived from a + // workspace hash rather than discovered, so it can be occupied by an + // account that has nothing to do with Zero, whether by coincidence or + // because somebody created it deliberately. Adopting one means resetting + // its password, which is not something to do on the strength of a name + // matching a pattern we generate ourselves. + managed, err := windowsSandboxUserIsManagedFn(username, workspaceKey) + if err != nil { + return windowsSandboxIdentity{}, "", false, err + } + if !managed { + return windowsSandboxIdentity{}, "", false, fmt.Errorf("%w: %q", errWindowsSandboxNameCollision, username) + } + // Ours by name and comment is not enough to adopt it. An account that also + // sits in Administrators (or Power Users, or Backup Operators) would give + // the sandbox the rights the sandbox exists to withhold: it could rewrite + // the ACLs confining it, read the secret locked to the invoking user, and + // stop Zero. Refuse rather than quietly take it over, and say which account + // so an operator can look at it. + privileged, err := windowsSandboxUserIsPrivilegedFn(username) + if err != nil { + return windowsSandboxIdentity{}, "", false, err + } + if privileged { + return windowsSandboxIdentity{}, "", false, fmt.Errorf("%w: %q", errWindowsSandboxPrivilegedAccount, username) + } + // Finish the upgrade windowsSandboxUserIsManaged has always promised. An + // account carrying the legacy bare comment gets the workspace key stamped + // on now, so it stops being unattributable and the bare form can + // eventually be retired rather than accepted forever. + // + // Best effort on purpose: the account is adopted and working either way, + // and failing provisioning over a comment would strand a usable sandbox + // for bookkeeping. Reported rather than swallowed so a run that keeps + // re-adopting the same legacy account is visible. + if legacy, err := windowsSandboxUserHasLegacyCommentFn(username); err == nil && legacy { + if err := upgradeWindowsSandboxUserCommentFn(username, workspaceKey); err != nil { + fmt.Fprintf(os.Stderr, "%s: could not stamp the workspace key onto sandbox principal %s: %v\n", + WindowsSandboxSetupName, username, err) + } + } + // Deliberately NOT resetting the password here. + // + // NetUserAdd left an existing account untouched, so the password above is + // not yet its password and something has to set it. Doing that here, at + // the top of provisioning, opened a window that lasted until the secret + // was written several steps later: a failure anywhere in between left a + // live account whose password nothing on disk knew, and because the + // account already existed the rollback correctly declined to delete it. + // The command path then read the absent secret as "not provisioned" and + // quietly fell back to the weaker backend, so the sandbox was downgraded + // for good with nothing to show for it. + // + // The caller rotates instead, immediately before committing the secret, + // which narrows that window to a single operation. Until it does, the + // account keeps its old password and the old secret on disk still + // authenticates, so a failure before that point costs nothing. + } + // Both failures below can happen AFTER NetUserAdd created the account, so the + // name has to come back with them. The caller's rollback deletes by + // identity.Username, and returning a zero identity alongside created=true + // asked it to delete "", which silently stranded the account this run had + // just made. Group attachment in particular is not a formality: it can fail + // under local policy, and it is the enforcement boundary, so a half-created + // principal is exactly the state worth not leaving behind. The SID is absent + // here, which the rollback already tolerates, since nothing has been granted + // to it yet. + if err := addWindowsSandboxUserToGroupFn(username); err != nil { + return windowsSandboxIdentity{Username: username}, "", !existed, err + } + // Membership of the offline group IS the network denial, so it is applied + // here rather than left to a later step: a principal that reached the runner + // without it would look correctly provisioned and quietly have the network. + // + // This is the failure most worth surviving cleanly. It happens after the + // account exists, and local policy can refuse a group join, so the name has + // to come back with the error or the rollback deletes "" and strands a + // principal that has network access and no offline membership. + if role == windowsSandboxRoleOffline { + if err := addWindowsSandboxUserToOfflineGroupFn(username); err != nil { + return windowsSandboxIdentity{Username: username}, "", !existed, err + } + } + sid, err := resolveWindowsSandboxSIDFn(username) + if err != nil { + return windowsSandboxIdentity{Username: username}, "", !existed, err + } + return windowsSandboxIdentity{Username: username, SID: sid}, password, !existed, nil +} + +// removeWindowsSandboxIdentity deletes a provisioned principal. Callers must +// revoke the principal's ACEs FIRST (see windowsPrincipalRevokePlan): deleting +// the account leaves any surviving ACE naming an unresolvable SID, which is what +// shows up in Explorer as an orphaned entry and is exactly the residue this +// model is meant to avoid. +// +// A missing account is success, so teardown converges the same way provisioning +// does. Requires an elevated caller. +func removeWindowsSandboxIdentity(username string, workspaceKey string) error { + // Account names here are derived, not discovered, so this could be pointed at + // a name that happens to belong to somebody else's local account. Deleting a + // user is not a recoverable mistake, so ownership is proven before deleting + // rather than inferred from the name matching a pattern we generate. + // Keyed to the workspace as well as to the marker: on a name collision the + // account belongs to a DIFFERENT workspace, and deleting it would be the + // same unrecoverable mistake as deleting a stranger's account. + managed, err := windowsSandboxUserIsManagedFn(username, workspaceKey) + if err != nil { + return err + } + if !managed { + // Distinguishable rather than a bare nil. Leaving the account alone is + // right, and callers that only want the goal state ("no principal of ours + // under this name") can treat it as success by matching this sentinel. But + // reporting a plain success would tell an operator that cleanup completed + // when a name they may care about was deliberately left in place, which is + // the one detail worth surfacing here. + return fmt.Errorf("%w: %q", errWindowsSandboxForeignAccountRetained, username) + } + name, err := windows.UTF16PtrFromString(username) + if err != nil { + return err + } + status, _, _ := procNetUserDel.Call(0, uintptr(unsafe.Pointer(name))) + runtime.KeepAlive(name) + return netAPIStatus("NetUserDel", status, nerrUserNotFound) +} + +// errWindowsSandboxForeignAccountRetained reports that removal left an account +// in place because it is not one Zero created. It is not a failure to clean up, +// it is a refusal to delete somebody else's account, and teardown paths treat it +// as success. +var errWindowsSandboxForeignAccountRetained = errors.New("left a local account in place because Zero did not create it") + +// errWindowsSandboxIdentityUnavailable reports that no sandbox principal has +// been provisioned yet, so callers can fall back to the restricted-token +// backend instead of failing the command. +var errWindowsSandboxIdentityUnavailable = errors.New("no Zero sandbox principal is provisioned; run `zero sandbox setup` from an elevated terminal") + +// lookupWindowsSandboxIdentity resolves an already-provisioned principal without +// creating anything, so the unelevated command path can discover whether an +// identity exists. It returns errWindowsSandboxIdentityUnavailable when setup +// has not run. +func lookupWindowsSandboxIdentity(workspaceKey string, role windowsSandboxRole) (windowsSandboxIdentity, error) { + username := windowsSandboxUserName(workspaceKey, role) + // SID resolution runs FIRST so "no such account" stays the unavailable + // sentinel. windowsSandboxUserIsManaged answers false for both an absent + // account and one belonging to someone else, so checking it before this would + // report an unprovisioned workspace as a name collision and turn the ordinary + // not-set-up case into an error the operator has to interpret. + sid, err := resolveWindowsSandboxSIDFn(username) + if err != nil { + return windowsSandboxIdentity{}, classifyWindowsSandboxLookupError(err) + } + managed, err := windowsSandboxUserIsManagedFn(username, workspaceKey) + if err != nil { + return windowsSandboxIdentity{}, err + } + if !managed { + return windowsSandboxIdentity{}, fmt.Errorf("%w: %q", errWindowsSandboxNameCollision, username) + } + return windowsSandboxIdentity{Username: username, SID: sid}, nil +} + +// lookupWindowsSandboxPrincipalForCommand resolves the principal a command will +// actually run as, and refuses one that has since joined a privileged group. +// +// Provisioning already refuses a privileged account, but group membership is not +// frozen at setup: the account can be added to Administrators, Power Users or +// Backup Operators afterwards. Minting a token for it would hand the sandboxed +// command exactly the privileges the sandbox exists to withhold, so the check has +// to run again on the path that mints the token, not only on the path that +// created the account. +// +// Deliberately NOT folded into lookupWindowsSandboxIdentity: teardown resolves +// the same identity to revoke its logon rights before deleting it, and it must +// stay able to clean up an account that has become privileged rather than +// refusing to touch it. Refusing there would leave the very account this guards +// against permanently undeletable by Zero. +func lookupWindowsSandboxPrincipalForCommand(workspaceKey string, role windowsSandboxRole) (windowsSandboxIdentity, error) { + identity, err := lookupWindowsSandboxIdentity(workspaceKey, role) + if err != nil { + return windowsSandboxIdentity{}, err + } + privileged, err := windowsSandboxUserIsPrivilegedFn(identity.Username) + if err != nil { + return windowsSandboxIdentity{}, err + } + if privileged { + return windowsSandboxIdentity{}, fmt.Errorf("%w: %q", errWindowsSandboxPrivilegedAccount, identity.Username) + } + return identity, nil +} + +// classifyWindowsSandboxLookupError decides whether a failed SID resolution +// means "setup has not run" or "this principal exists but is unusable". +// +// Only "no such account" is the former. Every other failure is a principal the +// caller must not paper over, including the deliberate refusal in +// resolveWindowsSandboxSID of a name squatted by a group or alias. Collapsing +// those into the unavailable sentinel would turn a real conflict into a silent +// fall back to the restricted token, which is exactly the case that should +// reach the operator rather than be absorbed. +// +// Split out from the lookup so the decision can be asserted on its own: the +// lookup derives its account name from a workspace key, so a test cannot hand +// it a name that resolves to a group. +func classifyWindowsSandboxLookupError(err error) error { + if err == nil { + return nil + } + if errors.Is(err, windows.ERROR_NONE_MAPPED) { + return errWindowsSandboxIdentityUnavailable + } + return err +} + +// windowsSandboxUserInLocalGroup reports direct membership of one named local +// group. Network denial is keyed to the offline group's SID on the WFP filters, +// so the command path uses this to confirm the account it is about to log on +// still carries the membership that makes those filters apply to it. +// +// Indirected through a var so the command path can be tested without an +// elevated machine. +var windowsSandboxUserInLocalGroupFn = windowsSandboxUserInLocalGroup + +func windowsSandboxUserInLocalGroup(username string, groupName string) (bool, error) { + name, err := windows.UTF16PtrFromString(username) + if err != nil { + return false, err + } + var ( + buffer *byte + entries uint32 + total uint32 + ) + status, _, _ := procNetUserGetLocalGroups.Call( + 0, // local machine + uintptr(unsafe.Pointer(name)), + 0, // level: LOCALGROUP_USERS_INFO_0 + 0, // flags: direct membership only + uintptr(unsafe.Pointer(&buffer)), + uintptr(^uint32(0)), // MAX_PREFERRED_LENGTH + uintptr(unsafe.Pointer(&entries)), + uintptr(unsafe.Pointer(&total)), + ) + runtime.KeepAlive(name) + if status == nerrUserNotFound { + return false, nil + } + if err := netAPIStatus("NetUserGetLocalGroups", status); err != nil { + return false, err + } + if buffer == nil || entries == 0 { + return false, nil + } + defer procNetApiBufferFree.Call(uintptr(unsafe.Pointer(buffer))) + want := strings.ToLower(groupName) + groups := unsafe.Slice((*localGroupUsersInfo0)(unsafe.Pointer(buffer)), entries) + for _, group := range groups { + if group.Name == nil { + continue + } + if strings.ToLower(windows.UTF16PtrToString(group.Name)) == want { + return true, nil + } + } + return false, nil +} diff --git a/internal/sandbox/windows_identity_windows_test.go b/internal/sandbox/windows_identity_windows_test.go new file mode 100644 index 000000000..34ca83d9f --- /dev/null +++ b/internal/sandbox/windows_identity_windows_test.go @@ -0,0 +1,513 @@ +//go:build windows + +package sandbox + +import ( + "errors" + "fmt" + "os" + "path/filepath" + "strings" + "testing" + "unsafe" + + "golang.org/x/sys/windows" +) + +// A local Windows account name is capped at 20 characters, so the derived name +// must truncate rather than produce a name NetUserAdd rejects. +func TestWindowsSandboxUserNameRespectsLengthLimit(t *testing.T) { + name := windowsSandboxUserName(strings.Repeat("a", 64), windowsSandboxRoleOffline) + if len(name) > windowsSandboxUserNameMax { + t.Fatalf("name %q is %d chars, want at most %d", name, len(name), windowsSandboxUserNameMax) + } + if !strings.HasPrefix(name, windowsSandboxUserPrefix) { + t.Fatalf("name %q lost the managed prefix", name) + } +} + +// The same workspace must map to the same principal, otherwise re-running setup +// would accumulate a new local account every time. +func TestWindowsSandboxUserNameIsStable(t *testing.T) { + first := windowsSandboxUserName("abc123", windowsSandboxRoleOffline) + second := windowsSandboxUserName("abc123", windowsSandboxRoleOffline) + if first != second { + t.Fatalf("name is not stable: %q vs %q", first, second) + } + if other := windowsSandboxUserName("def456", windowsSandboxRoleOffline); other == first { + t.Fatalf("different workspaces produced the same principal %q", first) + } +} + +// The key is sanitised to characters a local account name accepts, so a hash or +// path fragment cannot smuggle a separator or a space into the name. +func TestWindowsSandboxUserNameRejectsUnsafeCharacters(t *testing.T) { + name := windowsSandboxUserName(`C:\Users\me\proj ect`, windowsSandboxRoleOffline) + for _, r := range strings.TrimPrefix(name, windowsSandboxUserPrefix) { + isLower := r >= 'a' && r <= 'z' + isDigit := r >= '0' && r <= '9' + if !isLower && !isDigit { + t.Fatalf("name %q contains unsafe rune %q", name, r) + } + } + if name == windowsSandboxUserPrefix { + t.Fatal("sanitising removed every character, leaving a bare prefix") + } +} + +// An empty or fully-sanitised-away key must still yield a usable name rather +// than the bare prefix. +func TestWindowsSandboxUserNameHandlesEmptyKey(t *testing.T) { + for _, key := range []string{"", "///", " "} { + if got := windowsSandboxUserName(key, windowsSandboxRoleOffline); got == windowsSandboxUserPrefix { + t.Fatalf("key %q produced a bare prefix", key) + } + } +} + +// The password must be fresh per call and carry the character classes a default +// Windows complexity policy demands, or NetUserAdd fails with ERROR_PASSWORD_RESTRICTION. +func TestNewWindowsSandboxPasswordIsRandomAndComplex(t *testing.T) { + first, err := newWindowsSandboxPassword() + if err != nil { + t.Fatalf("generate: %v", err) + } + second, err := newWindowsSandboxPassword() + if err != nil { + t.Fatalf("generate: %v", err) + } + if first == second { + t.Fatal("two generated passwords are identical, so they are not random") + } + if len(first) < 12 { + t.Fatalf("password is only %d chars", len(first)) + } + var hasUpper, hasLower, hasDigit bool + for _, r := range first { + switch { + case r >= 'A' && r <= 'Z': + hasUpper = true + case r >= 'a' && r <= 'z': + hasLower = true + case r >= '0' && r <= '9': + hasDigit = true + } + } + if !hasUpper || !hasLower || !hasDigit { + t.Fatalf("password %q lacks a required character class", first) + } +} + +// "Already exists" is the normal result of re-running setup and must not surface +// as an error, while a genuine failure must. +func TestNetAPIStatusTreatsExistingAsSuccess(t *testing.T) { + if err := netAPIStatus("NetUserAdd", nerrSuccess); err != nil { + t.Fatalf("success status returned %v", err) + } + if err := netAPIStatus("NetUserAdd", nerrUserExists, nerrUserExists); err != nil { + t.Fatalf("existing user must be success, got %v", err) + } + if err := netAPIStatus("NetLocalGroupAdd", nerrGroupExists, nerrGroupExists, errorAliasExists); err != nil { + t.Fatalf("existing group must be success, got %v", err) + } + if err := netAPIStatus("NetUserAdd", 2245); err == nil { + t.Fatal("an unexpected status must surface as an error") + } +} + +// Access-denied is the status an unelevated run gets, and it must say so rather +// than reporting a bare number the user cannot act on. +func TestNetAPIStatusExplainsAccessDenied(t *testing.T) { + err := netAPIStatus("NetUserAdd", errorAccessDenied32) + if err == nil { + t.Fatal("access denied must be an error") + } + if !strings.Contains(err.Error(), "elevated") { + t.Fatalf("error %q should point at elevation", err) + } +} + +// The Win32 structs are passed to netapi32 as raw buffers, so their layout must +// match what the API expects. A wrong size means silent memory corruption. +func TestWindowsIdentityStructLayouts(t *testing.T) { + ptr := unsafe.Sizeof(uintptr(0)) + if got, want := unsafe.Sizeof(localGroupMembersInfo3{}), ptr; got != want { + t.Fatalf("LOCALGROUP_MEMBERS_INFO_3 size = %d, want %d", got, want) + } + if got, want := unsafe.Sizeof(localGroupInfo1{}), 2*ptr; got != want { + t.Fatalf("LOCALGROUP_INFO_1 size = %d, want %d", got, want) + } + // USER_INFO_1 is four pointers plus three DWORDs, with the compiler padding + // each DWORD pair up to pointer alignment on amd64. + if got := unsafe.Sizeof(userInfo1{}); got < 4*ptr { + t.Fatalf("USER_INFO_1 size = %d, smaller than its four pointer fields", got) + } + if unsafe.Offsetof(userInfo1{}.Password) != ptr { + t.Fatal("USER_INFO_1.Password must directly follow Name") + } +} + +// LSA_UNICODE_STRING counts BYTES, not runes, and excludes the NUL terminator +// from Length while including it in MaximumLength. Getting either wrong makes +// LsaAddAccountRights read past the buffer or silently match no right, so pin it. +func TestNewLSAStringUsesByteLengths(t *testing.T) { + buffer, err := windows.UTF16FromString("SeBatchLogonRight") + if err != nil { + t.Fatalf("encode: %v", err) + } + entry := newLSAString(buffer) + const runes uint16 = uint16(len("SeBatchLogonRight")) + if entry.Length != runes*2 { + t.Fatalf("Length = %d, want %d (bytes, excluding NUL)", entry.Length, runes*2) + } + if entry.MaximumLength != (runes+1)*2 { + t.Fatalf("MaximumLength = %d, want %d (bytes, including NUL)", entry.MaximumLength, (runes+1)*2) + } + if entry.Buffer == nil { + t.Fatal("Buffer must point at the encoded string") + } +} + +// An empty buffer must not produce a struct pointing at nothing with a nonzero +// length, which would hand LSA a wild pointer. +func TestNewLSAStringHandlesEmptyBuffer(t *testing.T) { + entry := newLSAString(nil) + if entry.Buffer != nil || entry.Length != 0 || entry.MaximumLength != 0 { + t.Fatalf("empty buffer produced %+v, want a zero value", entry) + } +} + +// The LSA structs are passed to advapi32 as raw buffers, so their sizes must +// match the Win32 definitions. +func TestLSAStructLayouts(t *testing.T) { + ptr := unsafe.Sizeof(uintptr(0)) + // LSA_UNICODE_STRING: two uint16 then a pointer, padded to pointer alignment. + if got, want := unsafe.Sizeof(lsaUnicodeString{}), 2*ptr; got != want { + t.Fatalf("LSA_UNICODE_STRING size = %d, want %d", got, want) + } + if unsafe.Offsetof(lsaUnicodeString{}.Buffer) != ptr { + t.Fatal("LSA_UNICODE_STRING.Buffer must sit at the second pointer slot") + } + var attributes lsaObjectAttributes + if unsafe.Sizeof(attributes) < 6*ptr-ptr { + t.Fatalf("LSA_OBJECT_ATTRIBUTES size = %d, smaller than its fields", unsafe.Sizeof(attributes)) + } + if unsafe.Offsetof(attributes.ObjectName) == 0 { + t.Fatal("LSA_OBJECT_ATTRIBUTES.ObjectName must not alias Length") + } +} + +// Provisioning creates real local accounts, so it only runs when explicitly +// opted into on an elevated machine. Everything above covers the logic that can +// be exercised without touching the account database. +func TestProvisionWindowsSandboxIdentityRoundTrip(t *testing.T) { + if os.Getenv("ZERO_WINDOWS_IDENTITY_PROVISION_TEST") != "1" { + // Spelled out per shell because `set VAR=1` is cmd syntax and silently + // sets a shell variable rather than an environment variable in + // PowerShell, which makes this skip look like the elevation check failing. + t.Skip("provisioning test not enabled: PowerShell `$env:ZERO_WINDOWS_IDENTITY_PROVISION_TEST = \"1\"`, cmd `set ZERO_WINDOWS_IDENTITY_PROVISION_TEST=1`, bash `export ZERO_WINDOWS_IDENTITY_PROVISION_TEST=1` (also needs an elevated terminal)") + } + if !windowsProcessIsElevated() { + t.Skip("provisioning requires an elevated process") + } + // Starting clean keeps a failure here from being explained by residue from a + // previous run. Provisioning no longer resets an adopted account's password, + // so a leftover account would otherwise be adopted with a password this test + // never learns. + _ = removeWindowsSandboxIdentity(windowsSandboxUserName("ziptest01", windowsSandboxRoleOffline), "ziptest01") + + identity, password, _, err := provisionWindowsSandboxIdentity("ziptest01", windowsSandboxRoleOffline) + if err != nil { + t.Fatalf("provision: %v", err) + } + // Registered immediately after provisioning so every failure path below is + // covered. This test grants a real batch-logon right to a real local account; + // leaving either behind on a developer machine is not acceptable residue, and + // rights are revoked before the account so nothing is left keyed to a SID that + // no longer resolves. + t.Cleanup(func() { + if err := revokeWindowsSandboxLogonRights(identity.SID); err != nil { + t.Errorf("cleanup: revoke logon rights: %v", err) + } + if err := removeWindowsSandboxIdentity(identity.Username, "ziptest01"); err != nil && !errors.Is(err, errWindowsSandboxForeignAccountRetained) { + t.Errorf("cleanup: remove principal: %v", err) + } + }) + if identity.SID == nil { + t.Fatal("provisioned identity has no SID") + } + if password == "" { + t.Fatal("provisioning returned an empty password") + } + // Re-running must converge on the same principal rather than failing or + // creating a second account. + again, secondPassword, _, err := provisionWindowsSandboxIdentity("ziptest01", windowsSandboxRoleOffline) + if err != nil { + t.Fatalf("second provision: %v", err) + } + if again.Username != identity.Username || !again.SID.Equals(identity.SID) { + t.Fatalf("provisioning is not idempotent: %s then %s", identity, again) + } + if secondPassword == "" { + t.Fatal("second provision returned an empty password") + } + // Deliberately NOT logging on with secondPassword. Provisioning does not + // rotate an adopted account any more, so that value is a fresh random string + // the account does not hold; rotation happens in + // provisionWindowsSandboxPrincipalForSetup, immediately before the secret is + // written, to keep the window where no stored password authenticates as small + // as possible. + // + // The guarantee worth asserting is therefore the one the setup path makes: + // after it returns, the stored secret logs the principal on. That covers + // rotation, the secret write and the logon right in one assertion, and it is + // the thing a broken re-setup would actually break. + // A SECOND ACCOUNT THIS FIXTURE OWNS, CLEANED UP BY ITS OWN KEY. The key is + // derived from the workspace, so a fixture-owned workspace gives a key no + // other test or developer state can collide with, and cleanup uses that same + // derived key rather than a literal the ownership guard would refuse. It is + // registered the moment ownership exists, before the secret read and the + // logon below can fail, and it only deletes what this run created. + config := WindowsSandboxCommandConfig{ + SandboxHome: t.TempDir(), + WorkspaceRoots: []string{filepath.Join(t.TempDir(), "ws")}, + } + secondKey := windowsSandboxPrincipalKey(config) + setupIdentity, secondCreated, err := provisionWindowsSandboxPrincipalForSetup(config, windowsSandboxRoleOffline) + if err != nil { + t.Fatalf("setup provision: %v", err) + } + t.Cleanup(func() { + if !secondCreated { + return + } + if err := revokeWindowsSandboxLogonRights(setupIdentity.SID); err != nil { + t.Errorf("revoke logon rights for the second account: %v", err) + } + if err := removeWindowsSandboxIdentity(setupIdentity.Username, secondKey); err != nil { + t.Errorf("remove the second account %s: %v", setupIdentity.Username, err) + } + }) + secretPath, err := windowsSandboxSecretPath(config.SandboxHome, setupIdentity.Username) + if err != nil { + t.Fatalf("secret path: %v", err) + } + storedPassword, err := readWindowsSandboxSecret(secretPath) + if err != nil { + t.Fatalf("read stored secret: %v", err) + } + token, err := logonWindowsSandboxPrincipal(setupIdentity.Username, storedPassword) + if err != nil { + t.Fatalf("logon with the secret the setup path stored: %v", err) + } + _ = token.Close() + // Lookup must find what provisioning created. + found, err := lookupWindowsSandboxIdentity("ziptest01", windowsSandboxRoleOffline) + if err != nil { + t.Fatalf("lookup after provision: %v", err) + } + if !found.SID.Equals(identity.SID) { + t.Fatalf("lookup returned %s, want %s", found, identity) + } +} + +// The other half of the privileged chain: granting logon rights and actually +// minting a token. Provisioning proves the account exists; this proves it is +// USABLE, which is the part the runner depends on. +// +// The batch logon doubles as the assertion that LsaAddAccountRights worked. A +// LOGON32_LOGON_BATCH logon fails with ERROR_LOGON_TYPE_NOT_GRANTED (1385) +// unless SeBatchLogonRight is actually held, so a token coming back is proof the +// grant landed rather than merely that the call returned success. +// +// Creates a real local account and removes it again, so it is gated the same way +// as the provisioning round-trip. +func TestGrantLogonRightsAndMintPrincipalToken(t *testing.T) { + if os.Getenv("ZERO_WINDOWS_IDENTITY_PROVISION_TEST") != "1" { + t.Skip("provisioning test not enabled: PowerShell `$env:ZERO_WINDOWS_IDENTITY_PROVISION_TEST = \"1\"`, cmd `set ZERO_WINDOWS_IDENTITY_PROVISION_TEST=1`, bash `export ZERO_WINDOWS_IDENTITY_PROVISION_TEST=1` (also needs an elevated terminal)") + } + if !windowsProcessIsElevated() { + t.Skip("granting logon rights requires an elevated process") + } + + const key = "ziplogon01" + // A leftover account from an interrupted run would keep its old password, + // which the freshly generated one will not match, so start from a clean slate. + _ = removeWindowsSandboxIdentity(windowsSandboxUserName(key, windowsSandboxRoleOffline), key) + + identity, password, _, err := provisionWindowsSandboxIdentity(key, windowsSandboxRoleOffline) + if err != nil { + t.Fatalf("provision: %v", err) + } + t.Cleanup(func() { + // Rights first, then the account: the reverse order strands an LSA entry + // keyed to a SID that no longer resolves. + if err := revokeWindowsSandboxLogonRights(identity.SID); err != nil { + t.Errorf("cleanup: revoke logon rights: %v", err) + } + if err := removeWindowsSandboxIdentity(identity.Username, key); err != nil && !errors.Is(err, errWindowsSandboxForeignAccountRetained) { + t.Errorf("cleanup: remove principal: %v", err) + } + }) + + // Exercises LsaAddAccountRights, including the LSA_UNICODE_STRING byte-length + // handling that nothing else has run. + if err := grantWindowsSandboxLogonRights(identity.SID); err != nil { + t.Fatalf("grant logon rights: %v", err) + } + // Idempotent: setup re-runs must not fail on rights the account already holds. + if err := grantWindowsSandboxLogonRights(identity.SID); err != nil { + t.Fatalf("granting logon rights twice must succeed: %v", err) + } + + token, err := logonWindowsSandboxPrincipal(identity.Username, password) + if err != nil { + t.Fatalf("logon as principal: %v", err) + } + defer token.Close() + + // The token must BE the principal. If this came back as the caller, the whole + // identity boundary would be an illusion and reads would still run as the user. + user, err := token.GetTokenUser() + if err != nil { + t.Fatalf("token user: %v", err) + } + if !user.User.Sid.Equals(identity.SID) { + t.Fatalf("token belongs to %s, want the principal %s", user.User.Sid, identity.SID) + } + t.Logf("minted a token for %s", identity) +} + +// A workspace with no provisioned principal must report the actionable +// "run setup" error rather than a raw lookup failure, so the command path can +// fall back instead of surfacing a Win32 code. +func TestLookupWindowsSandboxIdentityUnprovisioned(t *testing.T) { + _, err := lookupWindowsSandboxIdentity("nosuchworkspacekey9z", windowsSandboxRoleOffline) + if err == nil { + t.Skip("a principal for this key unexpectedly exists on this machine") + } + if err != errWindowsSandboxIdentityUnavailable { + t.Fatalf("error = %v, want errWindowsSandboxIdentityUnavailable", err) + } +} + +// A name that resolves to something other than a user account is a conflict, +// not an absent principal, and must not be reported as "setup has not run": the +// command path treats that sentinel as permission to fall back silently, so +// collapsing the two would hide a squatted account behind a quiet downgrade to +// the restricted token. +// +// Every machine already has well-known non-user names to test against, so this +// needs no privilege and no provisioning. +func TestLookupWindowsSandboxIdentityRejectsNonUserAccount(t *testing.T) { + // Groups that exist on any Windows install. Whichever resolves first is + // enough; localized machines may not carry the English name. + for _, group := range []string{"Administrators", "Users", "Guests"} { + sid, _, accountType, err := windows.LookupSID("", group) + if err != nil || sid == nil { + continue + } + if accountType == windows.SidTypeUser { + continue + } + resolveErr := func() error { + _, err := resolveWindowsSandboxSID(group) + return err + }() + if resolveErr == nil { + t.Fatalf("resolveWindowsSandboxSID(%q) accepted a non-user account (type %d)", group, accountType) + } + // The classification is the part that matters: the sentinel is what + // licenses the command path to fall back silently, so this refusal must + // survive it rather than be folded into it. + if classified := classifyWindowsSandboxLookupError(resolveErr); errors.Is(classified, errWindowsSandboxIdentityUnavailable) { + t.Fatalf("non-user account %q classified as unprovisioned, which would silently downgrade to the restricted token: %v", group, classified) + } + return + } + t.Skip("no well-known non-user account resolved on this machine") +} + +// The account name is derived from a workspace hash, not discovered, so it can +// be occupied by a local account that has nothing to do with Zero. Adopting one +// means resetting a stranger's password, so provisioning has to prove ownership +// first and refuse otherwise. +// +// Driven against real accounts every Windows install carries, which are +// definitively not ours. Unprivileged: it only has to establish that they are +// not classified as managed, so nothing is ever created or modified. +func TestWindowsSandboxUserIsManagedRefusesForeignAccounts(t *testing.T) { + checked := 0 + for _, name := range []string{"Administrator", "Guest", "DefaultAccount"} { + managed, err := windowsSandboxUserIsManaged(name, "workspacekey") + if err != nil { + // Localized or disabled installs may not carry every one of these. + continue + } + checked++ + if managed { + t.Fatalf("%q classified as a Zero sandbox principal; provisioning would reset its password", name) + } + } + if checked == 0 { + t.Skip("no well-known local account could be queried on this machine") + } + // An absent account must answer false rather than error, since provisioning + // asks this question about names that usually do not exist yet. + managed, err := windowsSandboxUserIsManaged("zero-sbx-nosuchacct", "workspacekey") + if err != nil { + t.Fatalf("querying a missing account: %v", err) + } + if managed { + t.Fatal("a missing account was classified as managed") + } +} + +// The refusal has to be a typed, recognisable collision rather than a generic +// failure, so setup can say what is wrong instead of reporting a Win32 status. +func TestWindowsSandboxNameCollisionIsTyped(t *testing.T) { + wrapped := fmt.Errorf("%w: %q", errWindowsSandboxNameCollision, "zero-sbx-dexample") + if !errors.Is(wrapped, errWindowsSandboxNameCollision) { + t.Fatal("collision error does not match its sentinel") + } + if !strings.Contains(wrapped.Error(), "not created by Zero") { + t.Fatalf("collision message = %q, want it to say the account is not ours", wrapped.Error()) + } +} + +// The two roles must never collide onto one account. If they did, the online +// principal would be whatever the offline one is, which means either an approved +// network command silently has no network, or an ordinary command silently has +// one. The second is the dangerous direction. +func TestWindowsSandboxUserNameSeparatesRoles(t *testing.T) { + const key = "abc123" + offline := windowsSandboxUserName(key, windowsSandboxRoleOffline) + online := windowsSandboxUserName(key, windowsSandboxRoleOnline) + if offline == online { + t.Fatalf("both roles produced %q", offline) + } + // Still stable per role, or setup would create a new account every run. + if again := windowsSandboxUserName(key, windowsSandboxRoleOnline); again != online { + t.Fatalf("online name is not stable: %q then %q", online, again) + } + + // The role tag has to survive truncation. A very long key is exactly when the + // name is cut to the 20-character limit, and losing the tag there is what + // would collide the roles on precisely the workspaces most likely to hit it. + longKey := strings.Repeat("f", 200) + longOffline := windowsSandboxUserName(longKey, windowsSandboxRoleOffline) + longOnline := windowsSandboxUserName(longKey, windowsSandboxRoleOnline) + if longOffline == longOnline { + t.Fatalf("truncation collided the roles onto %q", longOffline) + } + for _, name := range []string{longOffline, longOnline} { + if len(name) > windowsSandboxUserNameMax { + t.Fatalf("name %q is %d chars, want at most %d", name, len(name), windowsSandboxUserNameMax) + } + } + + // Different workspaces still get different accounts within a role. + if other := windowsSandboxUserName("def456", windowsSandboxRoleOffline); other == offline { + t.Fatalf("two workspaces produced the same offline account %q", other) + } +} diff --git a/internal/sandbox/windows_legacy_comment_upgrade_windows_test.go b/internal/sandbox/windows_legacy_comment_upgrade_windows_test.go new file mode 100644 index 000000000..37d7d93c7 --- /dev/null +++ b/internal/sandbox/windows_legacy_comment_upgrade_windows_test.go @@ -0,0 +1,90 @@ +//go:build windows + +package sandbox + +import ( + "testing" + + "golang.org/x/sys/windows" +) + +// THE UPGRADE HAS TO BE OBSERVABLE, or it is not really there. +// +// windowsSandboxUserIsManaged accepts the old bare account comment and promises +// the workspace key gets stamped on so the bare form can eventually be retired. +// The probe for the legacy comment was seamed and the upgrade itself was not, so +// nothing in the suite could tell whether the call ran: deleting it left +// everything green. This repository has already had a fix silently reverted by a +// later change, which is the failure this pins. +func TestAdoptingALegacyAccountStampsTheWorkspaceKey(t *testing.T) { + stub := func(t *testing.T, legacy bool) *string { + t.Helper() + var upgraded string + + prevGroup := ensureWindowsSandboxGroupFn + prevOffline := ensureWindowsSandboxOfflineGroupFn + prevEnsure := ensureWindowsSandboxUserFn + prevManaged := windowsSandboxUserIsManagedFn + prevLegacy := windowsSandboxUserHasLegacyCommentFn + prevPrivileged := windowsSandboxUserIsPrivilegedFn + prevUpgrade := upgradeWindowsSandboxUserCommentFn + prevReset := resetWindowsSandboxUserPasswordFn + prevGroupAdd := addWindowsSandboxUserToGroupFn + prevOfflineAdd := addWindowsSandboxUserToOfflineGroupFn + prevSID := resolveWindowsSandboxSIDFn + t.Cleanup(func() { + ensureWindowsSandboxGroupFn = prevGroup + ensureWindowsSandboxOfflineGroupFn = prevOffline + ensureWindowsSandboxUserFn = prevEnsure + windowsSandboxUserIsManagedFn = prevManaged + windowsSandboxUserHasLegacyCommentFn = prevLegacy + windowsSandboxUserIsPrivilegedFn = prevPrivileged + upgradeWindowsSandboxUserCommentFn = prevUpgrade + resetWindowsSandboxUserPasswordFn = prevReset + addWindowsSandboxUserToGroupFn = prevGroupAdd + addWindowsSandboxUserToOfflineGroupFn = prevOfflineAdd + resolveWindowsSandboxSIDFn = prevSID + }) + + ensureWindowsSandboxGroupFn = func() error { return nil } + ensureWindowsSandboxOfflineGroupFn = func() error { return nil } + // existed = true: the adoption path, which is the only one that upgrades. + ensureWindowsSandboxUserFn = func(string, string, string) (bool, error) { return true, nil } + windowsSandboxUserIsManagedFn = func(string, string) (bool, error) { return true, nil } + windowsSandboxUserHasLegacyCommentFn = func(string) (bool, error) { return legacy, nil } + windowsSandboxUserIsPrivilegedFn = func(string) (bool, error) { return false, nil } + upgradeWindowsSandboxUserCommentFn = func(_ string, workspaceKey string) error { + upgraded = workspaceKey + return nil + } + resetWindowsSandboxUserPasswordFn = func(string, string) error { return nil } + addWindowsSandboxUserToGroupFn = func(string) error { return nil } + addWindowsSandboxUserToOfflineGroupFn = func(string) error { return nil } + resolveWindowsSandboxSIDFn = func(string) (*windows.SID, error) { + return windows.StringToSid("S-1-5-32-546") + } + return &upgraded + } + + t.Run("a legacy comment is upgraded", func(t *testing.T) { + upgraded := stub(t, true) + if _, _, _, err := provisionWindowsSandboxIdentity("workspacekey", windowsSandboxRoleOffline); err != nil { + t.Fatalf("provisionWindowsSandboxIdentity: %v", err) + } + if *upgraded != "workspacekey" { + t.Errorf("the workspace key was never stamped onto the adopted legacy account (got %q); windowsSandboxUserIsManaged goes on accepting the bare comment forever", *upgraded) + } + }) + + // And an account that already carries the key is left alone, or the + // assertion above would be satisfied by an unconditional rewrite. + t.Run("a current comment is left alone", func(t *testing.T) { + upgraded := stub(t, false) + if _, _, _, err := provisionWindowsSandboxIdentity("workspacekey", windowsSandboxRoleOffline); err != nil { + t.Fatalf("provisionWindowsSandboxIdentity: %v", err) + } + if *upgraded != "" { + t.Errorf("an account that already carried the workspace key was rewritten anyway: %q", *upgraded) + } + }) +} diff --git a/internal/sandbox/windows_legacy_principal_windows_test.go b/internal/sandbox/windows_legacy_principal_windows_test.go new file mode 100644 index 000000000..3beb1c3dc --- /dev/null +++ b/internal/sandbox/windows_legacy_principal_windows_test.go @@ -0,0 +1,108 @@ +package sandbox + +import ( + "errors" + "sort" + "strings" + "testing" +) + +// The predecessor of this branch provisioned ONE account per workspace, named +// "zero-sbx-" with no role tag. Splitting the roles changed both names, so +// on an upgraded machine neither role resolves to the account that is actually +// installed. Every other path on this branch then reports the principal as never +// provisioned, and the untagged account keeps its secret, its logon rights, its +// ACEs and its ledger with nothing left that looks for them. +// +// windowsSandboxRoleLegacy exists to reproduce that name exactly so retirement +// can still find it. If its tag ever stops being empty, this stops matching what +// the predecessor wrote and the residue becomes unreachable again. +func TestLegacyRoleReproducesTheUntaggedPredecessorName(t *testing.T) { + // Short enough that the 20-character account-name limit does not truncate, + // so the comparison is against the spelling itself rather than the cap. + const key = "abc123" + + legacy := windowsSandboxUserName(key, windowsSandboxRoleLegacy) + if want := windowsSandboxUserPrefix + key; legacy != want { + t.Fatalf("legacy account name = %q, want the pre-split spelling %q", legacy, want) + } + // And the cap still applies to it exactly as it did before the split, since + // the predecessor truncated the same way. + long := windowsSandboxUserName("averylongworkspacekeyindeed", windowsSandboxRoleLegacy) + if len(long) > windowsSandboxUserNameMax { + t.Errorf("legacy name %q is %d characters, over the %d-character limit", long, len(long), windowsSandboxUserNameMax) + } + if tag := windowsSandboxRoleLegacy.roleTag(); tag != "" { + t.Errorf("legacy role tag = %q, want empty; a tag here would stop matching the installed account", tag) + } + + // It must not collide with either live role, or retiring the predecessor + // would delete an account this branch is still using. + offline := windowsSandboxUserName(key, windowsSandboxRoleOffline) + online := windowsSandboxUserName(key, windowsSandboxRoleOnline) + for _, live := range []string{offline, online} { + if strings.EqualFold(legacy, live) { + t.Errorf("legacy name %q collides with a live role name %q", legacy, live) + } + } + if strings.EqualFold(offline, online) { + t.Fatalf("the two live roles derived the same account name %q", offline) + } +} + +// THE LEGACY ROLE IS RETIRED, NEVER PROVISIONED. It derives the untagged name +// on purpose, so provisioning it would create the very account this branch +// replaced, on every setup, forever. Adding it to the retirement list and the +// provisioning list are one character apart in the source and the difference is +// invisible at a glance, so it is asserted rather than left to review. +func TestSetupNeverProvisionsTheLegacyPrincipal(t *testing.T) { + previous := provisionWindowsSandboxPrincipalForSetupFn + t.Cleanup(func() { provisionWindowsSandboxPrincipalForSetupFn = previous }) + + var provisioned []windowsSandboxRole + provisionWindowsSandboxPrincipalForSetupFn = func(_ WindowsSandboxCommandConfig, role windowsSandboxRole) (windowsSandboxIdentity, bool, error) { + provisioned = append(provisioned, role) + return windowsSandboxIdentity{}, false, errStopProvisioningForTest + } + + // The error stops setup at the first role; what matters is which roles it + // was willing to ask for, and legacy must never be among them. + _, _ = setupWindowsSandboxPrincipal(WindowsSandboxCommandConfig{SandboxHome: t.TempDir()}) + + for _, role := range provisioned { + if role == windowsSandboxRoleLegacy { + t.Fatalf("setup tried to provision the legacy role, which would recreate the pre-split account: %v", provisioned) + } + } +} + +var errStopProvisioningForTest = errors.New("stop provisioning for test") + +// Opting out or re-running setup has to retire the predecessor as well as both +// live roles. Retiring only the two this branch knows about is what leaves a +// fully provisioned account on a machine whose marker says it has none. +func TestSetupRetirementCoversTheLegacyPrincipal(t *testing.T) { + previous := removeWindowsSandboxPrincipalForSetupFn + t.Cleanup(func() { removeWindowsSandboxPrincipalForSetupFn = previous }) + + var retired []string + removeWindowsSandboxPrincipalForSetupFn = func(_ WindowsSandboxCommandConfig, role windowsSandboxRole) error { + retired = append(retired, string(role)) + return nil + } + + if err := removeWindowsSandboxPrincipalsForSetup(WindowsSandboxCommandConfig{SandboxHome: t.TempDir()}); err != nil { + t.Fatalf("removeWindowsSandboxPrincipalsForSetup: %v", err) + } + + sort.Strings(retired) + want := []string{"legacy", "offline", "online"} + if len(retired) != len(want) { + t.Fatalf("retired %v, want all of %v; a role left out keeps its account and everything it owns", retired, want) + } + for index := range want { + if retired[index] != want[index] { + t.Fatalf("retired %v, want %v", retired, want) + } + } +} diff --git a/internal/sandbox/windows_network.go b/internal/sandbox/windows_network.go index 0d614a0fc..3530603d4 100644 --- a/internal/sandbox/windows_network.go +++ b/internal/sandbox/windows_network.go @@ -62,15 +62,190 @@ func BuildWindowsNetworkInfraPlan(config WindowsSandboxCommandConfig) (WindowsNe if err != nil { return WindowsNetworkPlan{}, err } + // The offline-marker SID stays first: the setup marker records + // IdentitySIDs[0] as the offline filter identity. + identitySIDs := []string{offlineSID} + // A sandbox principal cannot carry the offline-marker SID, because LogonUser + // builds a token from an account's real group memberships and the marker is a + // synthetic capability SID. So the same filters additionally match a real + // local group that network-denied principals belong to. + // + // Gated on THIS home's opt-in, not merely on the group existing. + // + // The group is machine-global, so keying off its existence meant the first + // workspace to opt in changed the computed plan for every OTHER sandbox home + // on the machine. The plan is hashed into the setup marker and compared on + // every command, so those homes started failing every command with "setup is + // out of date" until each was re-run from an elevated terminal, having never + // opted into anything. Existing markers have to keep validating. + // + // Reading the opt-in instead keeps an opted-out home computing exactly the + // plan it computed before any of this existed. An opted-in home records the + // group in its own marker, and setup and the command path both derive the + // flag from the same environment, so they agree. Opting in AFTER setup does + // invalidate that home's marker, which is correct: it has no principals yet + // and setup is genuinely required. + if windowsSandboxIdentityEnabled(config.Env) && resolveWindowsSandboxOfflineGroupSIDHook != nil { + groupSID, err := resolveWindowsSandboxOfflineGroupSIDHook() + if err != nil { + return WindowsNetworkPlan{}, err + } + if trimmed := strings.TrimSpace(groupSID); trimmed != "" { + identitySIDs = append(identitySIDs, trimmed) + } + } return WindowsNetworkPlan{ Mode: NetworkDeny, ProviderKey: windowsWFPProviderKey, SubLayerKey: windowsWFPSubLayerKey, - IdentitySIDs: []string{offlineSID}, + IdentitySIDs: identitySIDs, Filters: windowsDenyWFPFilterSpecs(), }, nil } +// WindowsNetworkPlanForApply returns the plan to INSTALL, which is not always +// the plan a home fingerprints. +// +// The filters are machine-global and every setup installs them by deleting and +// recreating the fixed set. So an opted-OUT setup for workspace B was replacing +// filters that workspace A's opted-in setup had installed, dropping the offline +// group SID as it went. A's offline principal still passed the runtime +// membership check and A's marker still validated, but no filter matched its +// token any more, so a NetworkDeny command in A quietly gained egress. Nothing +// on either side reported anything: B did exactly what it was asked to. +// +// The gate in BuildWindowsNetworkInfraPlan stays as it is, and this is +// deliberately a SEPARATE function rather than a change to it. That gate exists +// because the plan is hashed into each home's setup marker: keying it on the +// group's existence made the first workspace to opt in invalidate every other +// home's marker on the machine, and those homes then failed every command with +// "setup is out of date" having opted into nothing. That has to keep working. +// +// The two questions are simply different. What a home RECORDS is about that +// home's own configuration; what setup INSTALLS is about the machine, where the +// group either exists or does not. Answering both from one plan is what forced a +// choice between stale markers and a silent hole. +// +// The group not existing is the ordinary opted-out machine, and it adds nothing. +// +// A resolution FAILURE is a different answer and is fatal, because the filters +// are replaced wholesale. The earlier reasoning here was that refusing to +// install would trade a partial denial for no denial at all, and that is wrong: +// the alternative to installing is leaving the filters that are already there +// alone. Returning a marker-only plan does not decline to add coverage, it +// actively REMOVES coverage another workspace is relying on. An opted-out home +// skips assertWindowsNetworkPlanCoversOfflineGroup entirely (it is gated on +// provisioned), so this is the only thing between a transient lookup error and +// workspace A silently regaining egress under NetworkDeny. +// +// "Absent" and "could not be read" have to be different answers, which is the +// same distinction the marker/apply split above is built on. +func WindowsNetworkPlanForApply(plan WindowsNetworkPlan, resolveOfflineGroupSID func() (string, error)) (WindowsNetworkPlan, error) { + if resolveOfflineGroupSID == nil { + return plan, nil + } + groupSID, err := resolveOfflineGroupSID() + if err != nil { + return WindowsNetworkPlan{}, fmt.Errorf("resolve the sandbox offline group before replacing the machine network filters: %w", err) + } + trimmed := strings.TrimSpace(groupSID) + if trimmed == "" { + return plan, nil + } + for _, existing := range plan.IdentitySIDs { + if strings.EqualFold(strings.TrimSpace(existing), trimmed) { + return plan, nil + } + } + // Copied rather than appended in place: the caller's plan is what gets + // fingerprinted, and growing its backing array would risk changing the + // recorded plan as a side effect of installing one. + augmented := plan + augmented.IdentitySIDs = append(append([]string{}, plan.IdentitySIDs...), trimmed) + return augmented, nil +} + +// WindowsNetworkPlanCoversPrincipals reports whether a plan's block filters name +// the offline group, which is the only thing that makes them apply to a sandbox +// principal. +// +// This exists to be asserted, not consulted. The plan must be built AFTER +// provisioning, because provisioning is what creates the group it resolves; a +// plan built first names only the offline marker and leaves every offline +// principal with an open network while setup reports success. That is a control +// that enforces nothing while claiming to work, and the ordering which prevents +// it is invisible at the call site, so a later refactor can undo it silently. +func WindowsNetworkPlanCoversPrincipals(plan WindowsNetworkPlan, offlineGroupSID string) bool { + offlineGroupSID = strings.TrimSpace(offlineGroupSID) + if offlineGroupSID == "" { + // No group provisioned on this host, so there is no principal for the + // filters to miss and nothing to assert. + return true + } + for _, sid := range plan.IdentitySIDs { + if strings.EqualFold(strings.TrimSpace(sid), offlineGroupSID) { + return true + } + } + return false +} + +// assertWindowsNetworkPlanCoversOfflineGroup is the post-provisioning check that +// setup runs before it installs filters and reports success. +// +// It fails closed on every answer that is not a definite yes, because the thing +// it guards is invisible when it goes wrong: filters that do not name the +// offline group leave every offline principal with an open network on a machine +// whose setup marker says it is protected. Nobody sees an error, and the sandbox +// looks correctly installed. +// +// Two answers other than "the plan omits the group" also mean the check did not +// happen and must not be treated as a pass: +// +// - The lookup failed. Whether the group is covered is then unknown, and the +// Win32 reason is worth surfacing: an operator seeing "access is denied" +// knows to re-run elevated. +// - The lookup succeeded and found nothing. ("", nil) means the group does not +// exist, which is the ordinary state BEFORE provisioning and an impossible +// one after it, so here it means the group vanished or was never created. +// WindowsNetworkPlanCoversPrincipals answers true for an empty SID, correctly +// for the pre-provisioning callers that ask it, so the emptiness has to be +// rejected here rather than delegated to it. +// +// provisioned says whether principals were actually provisioned on this run. +// When they were not there is no offline group, no principal carrying it, and +// nothing for the filters to miss, so there is nothing to assert. Running the +// check anyway turns the ordinary opt-out setup into a hard failure, because a +// machine with no group resolves to ("", nil) and the empty-SID rejection below +// is correct only after provisioning. +func assertWindowsNetworkPlanCoversOfflineGroup(plan WindowsNetworkPlan, resolve func() (string, error), provisioned bool) error { + if !provisioned { + return nil + } + if resolve == nil { + return errors.New("sandbox offline group resolver is not wired up, so filter coverage cannot be verified") + } + groupSID, err := resolve() + if err != nil { + return fmt.Errorf("resolve the sandbox offline group after provisioning, so filter coverage cannot be verified: %w", err) + } + if strings.TrimSpace(groupSID) == "" { + return errors.New("the sandbox offline group does not exist after provisioning, so the block filters would not apply to any sandbox principal") + } + if !WindowsNetworkPlanCoversPrincipals(plan, groupSID) { + return errors.New("network block filters do not name the sandbox offline group, so they would not apply to any sandbox principal; the network plan must be built after principals are provisioned") + } + return nil +} + +// resolveWindowsSandboxOfflineGroupSIDHook resolves the local group that +// network-denied principals belong to. It is wired up on Windows only, so this +// file stays free of Win32 calls and the plan on other platforms is unchanged. +// +// Returning ("", nil) means the group does not exist yet, which is the state +// before principals have ever been provisioned. +var resolveWindowsSandboxOfflineGroupSIDHook func() (string, error) + // WindowsNetworkInfraHash fingerprints the provisioned (mode-independent) network // infrastructure so the setup marker validates against the same setup for BOTH // command modes. It never folds in the per-command network mode. diff --git a/internal/sandbox/windows_network_coverage_assert_test.go b/internal/sandbox/windows_network_coverage_assert_test.go new file mode 100644 index 000000000..7b07176b9 --- /dev/null +++ b/internal/sandbox/windows_network_coverage_assert_test.go @@ -0,0 +1,129 @@ +package sandbox + +import ( + "errors" + "strings" + "testing" +) + +// The post-provisioning coverage assertion is the last thing standing between a +// mis-ordered plan and a machine that reports a successful setup while every +// offline principal has an open network. It has to fail closed on every way of +// not knowing the answer, not just on a definite negative. +func TestAssertWindowsNetworkPlanCoversOfflineGroupFailsClosed(t *testing.T) { + const groupSID = "S-1-5-32-9999" + covering := WindowsNetworkPlan{IdentitySIDs: []string{"S-1-5-21-marker", groupSID}} + notCovering := WindowsNetworkPlan{IdentitySIDs: []string{"S-1-5-21-marker"}} + lookupFailed := errors.New("look up zero-sandbox-offline: access is denied") + + tests := []struct { + name string + plan WindowsNetworkPlan + resolve func() (string, error) + wantErr bool + wantText string + }{ + { + name: "plan names the group", + plan: covering, + resolve: func() (string, error) { return groupSID, nil }, + }, + { + // The finding anandh8x raised: a failed lookup skipped the assertion + // entirely and setup carried on to write a success marker. + name: "lookup fails", + plan: covering, + resolve: func() (string, error) { return "", lookupFailed }, + wantErr: true, + wantText: "access is denied", + }, + { + // Empty-and-no-error means "the group does not exist", which is a + // legitimate answer BEFORE provisioning and an impossible one after it. + // WindowsNetworkPlanCoversPrincipals answers true for an empty SID, so + // without an explicit check the assertion passes vacuously. + name: "group missing after provisioning", + plan: covering, + resolve: func() (string, error) { return "", nil }, + wantErr: true, + wantText: "does not exist", + }, + { + name: "whitespace-only SID", + plan: covering, + resolve: func() (string, error) { return " ", nil }, + wantErr: true, + wantText: "does not exist", + }, + { + name: "plan omits the group", + plan: notCovering, + resolve: func() (string, error) { return groupSID, nil }, + wantErr: true, + wantText: "do not name the sandbox offline group", + }, + } + + for _, test := range tests { + t.Run(test.name, func(t *testing.T) { + err := assertWindowsNetworkPlanCoversOfflineGroup(test.plan, test.resolve, true) + if test.wantErr && err == nil { + t.Fatalf("assertWindowsNetworkPlanCoversOfflineGroup() = nil, want an error so setup rolls back") + } + if !test.wantErr { + if err != nil { + t.Fatalf("assertWindowsNetworkPlanCoversOfflineGroup() = %v, want nil", err) + } + return + } + if !strings.Contains(err.Error(), test.wantText) { + t.Fatalf("error = %q, want it to mention %q so the operator can tell the cases apart", err, test.wantText) + } + }) + } + + // The lookup failure has to stay unwrapped-comparable: setup logs it and an + // operator needs the underlying Win32 reason, not a flattened string. + err := assertWindowsNetworkPlanCoversOfflineGroup(covering, func() (string, error) { return "", lookupFailed }, true) + if !errors.Is(err, lookupFailed) { + t.Fatalf("errors.Is(err, lookupFailed) = false; the cause was dropped: %v", err) + } +} + +// The assert only means something AFTER principals were provisioned. The offline +// group is created inside provisionWindowsSandboxIdentity, which runs only under +// the ZERO_WINDOWS_SANDBOX_IDENTITY opt-in, so on a default machine there is no +// group, no principal, and nothing for the filters to miss. +// +// Rejecting an empty SID is right once provisioning ran (that is the vacuous +// pass anandh8x found) and wrong before it: it turned the ordinary opt-out setup +// into a hard failure, which is the regression jatmn caught on #812. +func TestAssertWindowsNetworkPlanCoversOfflineGroupSkipsWhenNothingWasProvisioned(t *testing.T) { + plan := WindowsNetworkPlan{IdentitySIDs: []string{"S-1-5-21-marker"}} + + // The default machine: no group exists, so the resolver reports ("", nil). + if err := assertWindowsNetworkPlanCoversOfflineGroup(plan, func() (string, error) { return "", nil }, false); err != nil { + t.Fatalf("opt-out setup was refused: %v", err) + } + // Same state WITH provisioning claimed: now the missing group is a real fault. + if err := assertWindowsNetworkPlanCoversOfflineGroup(plan, func() (string, error) { return "", nil }, true); err == nil { + t.Fatal("a missing group after provisioning must still fail closed") + } + // Not provisioned must not become a licence to skip a lookup failure either: + // the resolver is never consulted, so a broken lookup cannot matter here. + if err := assertWindowsNetworkPlanCoversOfflineGroup(plan, func() (string, error) { + t.Fatal("the resolver must not be consulted when nothing was provisioned") + return "", nil + }, false); err != nil { + t.Fatalf("opt-out setup was refused: %v", err) + } +} + +// A nil resolver is a wiring mistake, and guessing that coverage is fine is the +// one thing it must not do. +func TestAssertWindowsNetworkPlanCoversOfflineGroupRejectsNilResolver(t *testing.T) { + plan := WindowsNetworkPlan{IdentitySIDs: []string{"S-1-5-32-9999"}} + if err := assertWindowsNetworkPlanCoversOfflineGroup(plan, nil, true); err == nil { + t.Fatal("assertWindowsNetworkPlanCoversOfflineGroup(nil resolver) = nil, want an error") + } +} diff --git a/internal/sandbox/windows_network_mixed_optin_test.go b/internal/sandbox/windows_network_mixed_optin_test.go new file mode 100644 index 000000000..4a96b0000 --- /dev/null +++ b/internal/sandbox/windows_network_mixed_optin_test.go @@ -0,0 +1,163 @@ +package sandbox + +import ( + "errors" + "strings" + "testing" +) + +// THE BLOCK FILTERS ARE MACHINE-GLOBAL, SO ONE WORKSPACE'S SETUP DECIDES ANOTHER +// WORKSPACE'S NETWORK. +// +// Every setup installs them by deleting and recreating one fixed set. The plan +// each home builds names the offline group only when THAT home opted in, which +// is right for the marker it fingerprints and wrong for the filters it installs: +// an ordinary opted-out setup for a second workspace replaced the filters +// without the group SID, and the first workspace's offline principal — still in +// the group, still passing the runtime membership check, still holding a valid +// marker — was no longer matched by any filter. A NetworkDeny command there +// gained egress, silently, because the second setup did exactly what it was +// asked to do. +// +// The plan a home installs and the plan it records are therefore different +// questions, and these pin both. + +const testOfflineGroupSID = "S-1-5-32-9990" + +func offlineGroupResolver(sid string) func() (string, error) { + return func() (string, error) { return sid, nil } +} + +// The regression itself: an opted-OUT home must still install filters covering +// the managed offline group, because an opted-IN home's principals depend on it. +func TestOptedOutSetupKeepsOfflineGroupCoverageForOtherWorkspaces(t *testing.T) { + optedOut, err := BuildWindowsNetworkInfraPlan(WindowsSandboxCommandConfig{ + SandboxHome: t.TempDir(), + // No opt-in in the environment: this is the ordinary workspace. + Env: map[string]string{}, + }) + if err != nil { + t.Fatalf("BuildWindowsNetworkInfraPlan: %v", err) + } + + // What this home RECORDS must not change, or every opted-out marker on the + // machine goes stale the moment another workspace opts in. + if WindowsNetworkPlanCoversPrincipals(optedOut, testOfflineGroupSID) { + t.Fatal("the recorded plan named the offline group; that is what invalidated other homes' markers") + } + + // What this home INSTALLS must cover the group, since the group exists. + applied, err := WindowsNetworkPlanForApply(optedOut, offlineGroupResolver(testOfflineGroupSID)) + if err != nil { + t.Fatalf("WindowsNetworkPlanForApply: %v", err) + } + if !WindowsNetworkPlanCoversPrincipals(applied, testOfflineGroupSID) { + t.Errorf("an opted-out setup would install filters that do not match the offline group, so another workspace's NetworkDeny commands gain egress: %v", applied.IdentitySIDs) + } + // The offline-marker SID has to survive alongside it: the restricted-token + // tier is matched by that one, and it is the whole denial on machines with no + // principals at all. + if len(applied.IdentitySIDs) == 0 || strings.TrimSpace(applied.IdentitySIDs[0]) == "" { + t.Fatalf("the offline-marker SID was lost from the applied plan: %v", applied.IdentitySIDs) + } + if applied.IdentitySIDs[0] != optedOut.IdentitySIDs[0] { + t.Errorf("the marker SID must stay first, since the setup marker records IdentitySIDs[0]: %v", applied.IdentitySIDs) + } +} + +// Augmenting must not mutate the plan the caller fingerprints. Appending in +// place would let installing a plan change the plan that was recorded. +func TestPlanForApplyDoesNotMutateTheRecordedPlan(t *testing.T) { + recorded, err := BuildWindowsNetworkInfraPlan(WindowsSandboxCommandConfig{ + SandboxHome: t.TempDir(), + Env: map[string]string{}, + }) + if err != nil { + t.Fatalf("BuildWindowsNetworkInfraPlan: %v", err) + } + before := append([]string{}, recorded.IdentitySIDs...) + + if _, err := WindowsNetworkPlanForApply(recorded, offlineGroupResolver(testOfflineGroupSID)); err != nil { + t.Fatalf("WindowsNetworkPlanForApply: %v", err) + } + + if len(recorded.IdentitySIDs) != len(before) { + t.Fatalf("the recorded plan grew from %v to %v", before, recorded.IdentitySIDs) + } + for index := range before { + if recorded.IdentitySIDs[index] != before[index] { + t.Errorf("the recorded plan changed at %d: %v vs %v", index, before, recorded.IdentitySIDs) + } + } +} + +// An opted-IN home already names the group, and must not name it twice: a +// duplicate identity would change the installed filter set for no reason. +func TestPlanForApplyDoesNotDuplicateAnAlreadyCoveredGroup(t *testing.T) { + plan := WindowsNetworkPlan{IdentitySIDs: []string{"S-1-5-21-marker", testOfflineGroupSID}} + applied, err := WindowsNetworkPlanForApply(plan, offlineGroupResolver(strings.ToLower(testOfflineGroupSID))) + if err != nil { + t.Fatalf("WindowsNetworkPlanForApply: %v", err) + } + if len(applied.IdentitySIDs) != 2 { + t.Errorf("the group was added again despite already being covered: %v", applied.IdentitySIDs) + } +} + +// On an ordinary machine with no principals the group does not exist, and the +// plan must be left exactly as it was. +func TestPlanForApplyLeavesAMachineWithoutTheGroupAlone(t *testing.T) { + plan := WindowsNetworkPlan{IdentitySIDs: []string{"S-1-5-21-marker"}} + + for _, testCase := range []struct { + name string + resolve func() (string, error) + }{ + {name: "no resolver", resolve: nil}, + {name: "group absent", resolve: func() (string, error) { return "", nil }}, + {name: "blank SID", resolve: func() (string, error) { return " ", nil }}, + } { + t.Run(testCase.name, func(t *testing.T) { + applied, err := WindowsNetworkPlanForApply(plan, testCase.resolve) + if err != nil { + t.Fatalf("a machine with no offline group is not a failure: %v", err) + } + if len(applied.IdentitySIDs) != 1 || applied.IdentitySIDs[0] != "S-1-5-21-marker" { + t.Errorf("plan changed on a machine with no offline group: %v", applied.IdentitySIDs) + } + }) + } +} + +// A LOOKUP FAILURE IS NOT AN ABSENT GROUP. The filters are machine-global and +// are replaced wholesale, so returning a marker-only plan here does not decline +// to add coverage, it removes coverage another workspace is relying on. +// +// This is reachable only from an opted-OUT home, because +// assertWindowsNetworkPlanCoversOfflineGroup is gated on provisioned and never +// runs for one. Workspace A opts in, the group SID lands in the global filters; +// workspace B is opted out and its lookup fails transiently; B installs a plan +// without the group and A's NetworkDeny commands silently regain egress while +// A's marker and its direct membership check both still pass. +func TestPlanForApplyRefusesToReplaceFiltersWhenTheGroupCannotBeRead(t *testing.T) { + // Opted-out home B: the coverage assert is skipped for it, so nothing else + // stands between the failure and the wholesale replace. + if err := assertWindowsNetworkPlanCoversOfflineGroup( + WindowsNetworkPlan{IdentitySIDs: []string{"S-1-5-21-B-marker"}}, + func() (string, error) { return "", errors.New("group lookup failed") }, + false, + ); err != nil { + t.Fatalf("the coverage assert is meant to be skipped for an opted-out home: %v", err) + } + + plan := WindowsNetworkPlan{IdentitySIDs: []string{"S-1-5-21-B-marker"}} + applied, err := WindowsNetworkPlanForApply(plan, func() (string, error) { + return "", errors.New("group lookup failed") + }) + if err == nil { + t.Fatalf("a failed group lookup installed %v, stripping the offline group from the machine's filters", applied.IdentitySIDs) + } + if !strings.Contains(err.Error(), "group lookup failed") { + t.Errorf("the underlying failure must be reported so the operator can act on it: %v", err) + } +} diff --git a/internal/sandbox/windows_network_test.go b/internal/sandbox/windows_network_test.go index 4253916f8..3a31388af 100644 --- a/internal/sandbox/windows_network_test.go +++ b/internal/sandbox/windows_network_test.go @@ -118,3 +118,154 @@ func assertWindowsWFPCommonFilter(t *testing.T, specs map[string]WindowsWFPFilte // Coverage for the network infra plan + hash and the per-mode token-SID // composition lives in windows_online_offline_test.go. + +// The block filters have to name the offline group as well as the offline-marker +// SID, or a sandbox principal is not covered by them at all: LogonUser builds a +// token from real group memberships and cannot carry the synthetic marker. This +// is the single property that makes network denial work for the principal +// backend, so assert it on the plan rather than trusting the wiring. +func TestNetworkInfraPlanIncludesOfflineGroupIdentity(t *testing.T) { + previous := resolveWindowsSandboxOfflineGroupSIDHook + t.Cleanup(func() { resolveWindowsSandboxOfflineGroupSIDHook = previous }) + + config := WindowsSandboxCommandConfig{SandboxHome: t.TempDir(), Env: map[string]string{windowsSandboxIdentityEnv: "1"}} + + // Before principals have ever been provisioned the group does not exist, and + // the plan must be exactly what it was before this feature. The plan is hashed + // into the setup marker and re-derived on every command, so an identity set + // that appeared out of nowhere would fail every command with "setup is out of + // date". + resolveWindowsSandboxOfflineGroupSIDHook = func() (string, error) { return "", nil } + base, err := BuildWindowsNetworkInfraPlan(config) + if err != nil { + t.Fatalf("build plan without the group: %v", err) + } + if len(base.IdentitySIDs) != 1 { + t.Fatalf("identity SIDs = %v, want only the offline marker when the group is absent", base.IdentitySIDs) + } + + const groupSID = "S-1-5-32-9999" + resolveWindowsSandboxOfflineGroupSIDHook = func() (string, error) { return groupSID, nil } + withGroup, err := BuildWindowsNetworkInfraPlan(config) + if err != nil { + t.Fatalf("build plan with the group: %v", err) + } + if len(withGroup.IdentitySIDs) != 2 || withGroup.IdentitySIDs[1] != groupSID { + t.Fatalf("identity SIDs = %v, want the offline marker plus %s", withGroup.IdentitySIDs, groupSID) + } + // The marker records IdentitySIDs[0] as the offline filter identity, so the + // marker SID has to stay first. + if withGroup.IdentitySIDs[0] != base.IdentitySIDs[0] { + t.Fatalf("offline marker moved from position 0: %v", withGroup.IdentitySIDs) + } + // Adding the group must change the fingerprint, or setup and the command path + // could disagree about the installed filters without anything noticing. + baseHash, err := WindowsNetworkInfraHash(base) + if err != nil { + t.Fatalf("hash base: %v", err) + } + groupHash, err := WindowsNetworkInfraHash(withGroup) + if err != nil { + t.Fatalf("hash with group: %v", err) + } + if baseHash == groupHash { + t.Fatal("the infra hash ignored the offline group identity") + } +} + +// A lookup failure must not be swallowed into "no group", because that silently +// produces a plan whose filters do not cover sandbox principals. +func TestNetworkInfraPlanPropagatesOfflineGroupLookupFailure(t *testing.T) { + previous := resolveWindowsSandboxOfflineGroupSIDHook + t.Cleanup(func() { resolveWindowsSandboxOfflineGroupSIDHook = previous }) + + resolveWindowsSandboxOfflineGroupSIDHook = func() (string, error) { + return "", errors.New("boom") + } + if _, err := BuildWindowsNetworkInfraPlan(WindowsSandboxCommandConfig{SandboxHome: t.TempDir(), Env: map[string]string{windowsSandboxIdentityEnv: "1"}}); err == nil { + t.Fatal("a failed offline-group lookup produced a plan; the filters would not cover any principal") + } +} + +// The ordering that makes the block filters apply to sandbox principals is +// invisible at the call site: the plan must be built AFTER provisioning, +// because provisioning creates the group the filters name. A plan built first +// names only the offline marker, and a machine would then report a successful +// setup while every offline principal had an open network. +// +// Asserted on the predicate setup checks before installing anything, so moving +// the plan build back ahead of provisioning fails loudly instead of silently +// producing a control that enforces nothing. +func TestNetworkPlanCoverageDetectsPrincipalsLeftUncovered(t *testing.T) { + const groupSID = "S-1-5-32-4242" + + // What a plan built BEFORE provisioning looks like: marker only. + tooEarly := WindowsNetworkPlan{IdentitySIDs: []string{"S-1-15-3-1111"}} + if WindowsNetworkPlanCoversPrincipals(tooEarly, groupSID) { + t.Fatal("a plan naming only the offline marker was reported as covering principals; the ordering guard would not fire") + } + + // And after: marker plus the group. + correct := WindowsNetworkPlan{IdentitySIDs: []string{"S-1-15-3-1111", groupSID}} + if !WindowsNetworkPlanCoversPrincipals(correct, groupSID) { + t.Fatal("a plan naming the offline group was reported as not covering principals; setup would refuse a correct plan") + } + // SID comparison is case-insensitive, so a differently-cased resolve does not + // read as a missing group and block a valid setup. + mixed := WindowsNetworkPlan{IdentitySIDs: []string{strings.ToLower(groupSID)}} + if !WindowsNetworkPlanCoversPrincipals(mixed, groupSID) { + t.Fatal("case difference read as a missing group") + } + + // A host with no group provisioned has no principal to miss, so there is + // nothing to assert and setup must not refuse. + if !WindowsNetworkPlanCoversPrincipals(tooEarly, "") { + t.Fatal("an unprovisioned host was treated as a coverage failure") + } +} + +// AN OPTED-OUT HOME MUST KEEP ITS EXISTING MARKER VALID. +// +// The offline group is machine-global. Keyed on its existence, the first +// workspace to opt in changed the computed plan for every OTHER sandbox home on +// the machine, so their stored NetworkInfraHash stopped matching and every one +// of their commands failed with "setup is out of date" until each was re-run +// from an elevated terminal. They had opted into nothing. +// +// The hashes are compared rather than the SID lists because the hash is what the +// marker actually stores and what validation compares. +func TestOfflineGroupDoesNotChangeAnOptedOutHomesInfraHash(t *testing.T) { + previous := resolveWindowsSandboxOfflineGroupSIDHook + t.Cleanup(func() { resolveWindowsSandboxOfflineGroupSIDHook = previous }) + + optedOut := WindowsSandboxCommandConfig{SandboxHome: t.TempDir()} + + // No group on the machine yet: the state before anyone opted in. + resolveWindowsSandboxOfflineGroupSIDHook = nil + before, err := BuildWindowsNetworkInfraPlan(optedOut) + if err != nil { + t.Fatalf("build plan before any opt-in: %v", err) + } + beforeHash, err := WindowsNetworkInfraHash(before) + if err != nil { + t.Fatalf("hash before: %v", err) + } + + // Another workspace has now opted in and created the machine-global group. + resolveWindowsSandboxOfflineGroupSIDHook = func() (string, error) { + return "S-1-5-21-1111111111-1111111111-1111111111-4002", nil + } + after, err := BuildWindowsNetworkInfraPlan(optedOut) + if err != nil { + t.Fatalf("build plan after another home opted in: %v", err) + } + afterHash, err := WindowsNetworkInfraHash(after) + if err != nil { + t.Fatalf("hash after: %v", err) + } + + if beforeHash != afterHash { + t.Fatalf("another workspace opting in changed this home's infra hash (%s -> %s), so its stored marker stops validating and every command fails until it is re-set-up", + beforeHash, afterHash) + } +} diff --git a/internal/sandbox/windows_offline_group_ownership_windows_test.go b/internal/sandbox/windows_offline_group_ownership_windows_test.go new file mode 100644 index 000000000..4ca357ba7 --- /dev/null +++ b/internal/sandbox/windows_offline_group_ownership_windows_test.go @@ -0,0 +1,96 @@ +//go:build windows + +package sandbox + +import ( + "errors" + "strings" + "testing" +) + +// The offline group's SID goes onto the persistent WFP deny filters and the +// sandbox principal is made a member of it. Adopting a same-named group that +// something else owns therefore cuts off every existing member's outbound +// traffic and hands our principal that group's permissions, so an unmarked +// group has to be refused rather than reused. +func TestEnsureWindowsLocalGroupRefusesForeignSameNameGroup(t *testing.T) { + for name, testCase := range map[string]struct { + status uintptr + owned bool + ownedErr error + wantError string + }{ + "foreign group with our name": { + status: nerrGroupExists, owned: false, + wantError: "not managed by zero", + }, + "foreign group reported via ERROR_ALIAS_EXISTS": { + status: errorAliasExists, owned: false, + wantError: "not managed by zero", + }, + "our own group on a re-run": { + status: nerrGroupExists, owned: true, + }, + "ownership lookup fails": { + status: nerrGroupExists, ownedErr: errors.New("lookup refused"), + wantError: "lookup refused", + }, + "group did not exist": { + status: nerrSuccess, + }, + } { + t.Run(name, func(t *testing.T) { + prevAdd, prevOwned := addWindowsLocalGroupFn, windowsLocalGroupOwnedByZeroFn + t.Cleanup(func() { + addWindowsLocalGroupFn = prevAdd + windowsLocalGroupOwnedByZeroFn = prevOwned + }) + addWindowsLocalGroupFn = func(string, string) (uintptr, error) { return testCase.status, nil } + lookups := 0 + windowsLocalGroupOwnedByZeroFn = func(string, string) (bool, error) { + lookups++ + return testCase.owned, testCase.ownedErr + } + + err := ensureWindowsSandboxOfflineGroup() + if testCase.wantError == "" { + if err != nil { + t.Fatalf("ensureWindowsSandboxOfflineGroup: %v", err) + } + return + } + if err == nil { + t.Fatalf("adopted a group it should have refused (lookups=%d)", lookups) + } + if !strings.Contains(err.Error(), testCase.wantError) { + t.Fatalf("error = %q, want it to mention %q", err, testCase.wantError) + } + }) + } +} + +// The marker compared against is the offline group's own comment, not the +// principals group's: two managed groups with different markers must not be +// mistaken for each other. +func TestEnsureWindowsLocalGroupChecksTheGroupsOwnMarker(t *testing.T) { + prevAdd, prevOwned := addWindowsLocalGroupFn, windowsLocalGroupOwnedByZeroFn + t.Cleanup(func() { + addWindowsLocalGroupFn = prevAdd + windowsLocalGroupOwnedByZeroFn = prevOwned + }) + addWindowsLocalGroupFn = func(string, string) (uintptr, error) { return nerrGroupExists, nil } + var gotName, gotComment string + windowsLocalGroupOwnedByZeroFn = func(name, comment string) (bool, error) { + gotName, gotComment = name, comment + return true, nil + } + if err := ensureWindowsSandboxOfflineGroup(); err != nil { + t.Fatalf("ensureWindowsSandboxOfflineGroup: %v", err) + } + if gotName != windowsSandboxOfflineGroupName { + t.Fatalf("checked group %q, want %q", gotName, windowsSandboxOfflineGroupName) + } + if gotComment != windowsSandboxOfflineGroupComment { + t.Fatalf("compared marker %q, want %q", gotComment, windowsSandboxOfflineGroupComment) + } +} diff --git a/internal/sandbox/windows_offline_membership_windows_test.go b/internal/sandbox/windows_offline_membership_windows_test.go new file mode 100644 index 000000000..5c6f946a3 --- /dev/null +++ b/internal/sandbox/windows_offline_membership_windows_test.go @@ -0,0 +1,102 @@ +//go:build windows + +package sandbox + +import ( + "errors" + "testing" + + "golang.org/x/sys/windows" +) + +// Choosing the offline account is only half of what denies it the network: the +// WFP filters match the offline GROUP'S SID. An account that has drifted out of +// that group still logs on from its stored secret, and its token no longer +// satisfies the filter condition — so a no-network profile would get full +// egress. The command path has to re-check the membership, not trust the marker +// setup wrote when it last succeeded. +func TestPrincipalTokenRechecksOfflineGroupMembership(t *testing.T) { + for name, testCase := range map[string]struct { + mode NetworkMode + member bool + memberErr error + wantChecked bool + wantReached bool + wantErr bool + }{ + "offline principal drifted out of the group": { + mode: NetworkDeny, member: false, + wantChecked: true, wantReached: false, + }, + "offline principal still a member": { + mode: NetworkDeny, member: true, + wantChecked: true, wantReached: true, + }, + "membership lookup fails": { + mode: NetworkDeny, memberErr: errors.New("group lookup refused"), + wantChecked: true, wantReached: false, wantErr: true, + }, + // An allow-network command uses the online principal, which is not in the + // offline group by design. Checking it there would refuse every approved + // network command. + "online principal is not checked": { + mode: NetworkAllow, member: false, + wantChecked: false, wantReached: true, + }, + } { + t.Run(name, func(t *testing.T) { + prevLookup := lookupWindowsSandboxPrincipalForCommandFn + prevMember := windowsSandboxUserInLocalGroupFn + prevSecret := readWindowsSandboxSecretFn + prevWarn := warnWindowsSandboxOfflineMembershipMissing + t.Cleanup(func() { + lookupWindowsSandboxPrincipalForCommandFn = prevLookup + windowsSandboxUserInLocalGroupFn = prevMember + readWindowsSandboxSecretFn = prevSecret + warnWindowsSandboxOfflineMembershipMissing = prevWarn + }) + + sid, err := windows.CreateWellKnownSid(windows.WinLocalSystemSid) + if err != nil { + t.Fatal(err) + } + lookupWindowsSandboxPrincipalForCommandFn = func(string, windowsSandboxRole) (windowsSandboxIdentity, error) { + return windowsSandboxIdentity{Username: "zero-sbx-test", SID: sid}, nil + } + checkedGroup := "" + windowsSandboxUserInLocalGroupFn = func(_ string, group string) (bool, error) { + checkedGroup = group + return testCase.member, testCase.memberErr + } + secretRead := false + readWindowsSandboxSecretFn = func(string) (string, error) { + secretRead = true + return "", errWindowsSandboxIdentityUnavailable + } + warnWindowsSandboxOfflineMembershipMissing = func(string) {} + + config := windowsSandboxTestConfig(t) + config.Env = map[string]string{windowsSandboxIdentityEnv: "1"} + config.PermissionProfile.Network.Mode = testCase.mode + + _, ok, err := windowsSandboxPrincipalToken(config) + if testCase.wantErr { + if err == nil { + t.Fatal("a failed membership lookup was swallowed") + } + } else if err != nil { + t.Fatalf("windowsSandboxPrincipalToken: %v", err) + } + _ = ok + if secretRead != testCase.wantReached { + t.Fatalf("reached the secret read = %v, want %v (gate must short-circuit before it)", secretRead, testCase.wantReached) + } + if checked := checkedGroup != ""; checked != testCase.wantChecked { + t.Fatalf("membership checked = %v, want %v", checked, testCase.wantChecked) + } + if testCase.wantChecked && checkedGroup != windowsSandboxOfflineGroupName { + t.Fatalf("checked group %q, want %q", checkedGroup, windowsSandboxOfflineGroupName) + } + }) + } +} diff --git a/internal/sandbox/windows_online_offline_test.go b/internal/sandbox/windows_online_offline_test.go index 370abee73..759113cd4 100644 --- a/internal/sandbox/windows_online_offline_test.go +++ b/internal/sandbox/windows_online_offline_test.go @@ -37,9 +37,25 @@ func TestWindowsRuntimeTokenSIDs(t *testing.T) { // setup serves both modes (and its fingerprint is stable across modes). func TestBuildWindowsNetworkInfraPlanIsModeIndependent(t *testing.T) { home := t.TempDir() + // Pinned, because the count below depends on whether this machine happens to + // have the offline group already. BuildWindowsNetworkInfraPlan folds that + // group's SID in when it resolves, so on a Windows host where an earlier + // elevated setup created ZeroSandboxOffline the plan legitimately carries two + // identity SIDs and this test failed on a correct plan. CI never saw it: the + // Linux and macOS jobs leave the hook nil, and a fresh Windows runner has no + // group yet. Stubbing it makes the assertion about the plan rather than about + // the machine it runs on. + previousHook := resolveWindowsSandboxOfflineGroupSIDHook + t.Cleanup(func() { resolveWindowsSandboxOfflineGroupSIDHook = previousHook }) + resolveWindowsSandboxOfflineGroupSIDHook = nil + mk := func(mode NetworkMode) WindowsSandboxCommandConfig { return WindowsSandboxCommandConfig{ - SandboxHome: home, + SandboxHome: home, + // Opted in, because the offline group only enters the plan for a home + // that asked for principals. Without this the plan is the pre-principal + // one and the group assertions below have nothing to find. + Env: map[string]string{windowsSandboxIdentityEnv: "1"}, CommandCWD: `C:\ws`, WorkspaceRoots: []string{`C:\ws`}, PermissionProfile: PermissionProfile{ @@ -72,6 +88,40 @@ func TestBuildWindowsNetworkInfraPlanIsModeIndependent(t *testing.T) { if denyPlan.IdentitySIDs[0] != offline { t.Errorf("infra plan SID = %q, want offline-marker %q", denyPlan.IdentitySIDs[0], offline) } + + // Mode independence has to hold on a machine that already has the offline + // group too, which is where the pinned counts above would otherwise be + // asserting something about the host rather than the plan. The group SID is + // folded in for both modes, so the hashes must still agree; only the count + // changes. + resolveWindowsSandboxOfflineGroupSIDHook = func() (string, error) { return "S-1-5-32-9999", nil } + withGroupDeny, err := BuildWindowsNetworkInfraPlan(mk(NetworkDeny)) + if err != nil { + t.Fatalf("deny infra plan with the group present: %v", err) + } + withGroupAllow, err := BuildWindowsNetworkInfraPlan(mk(NetworkAllow)) + if err != nil { + t.Fatalf("allow infra plan with the group present: %v", err) + } + // Assert the SIDs themselves, not the count. A duplicate offline marker or an + // unrelated SID would satisfy a length check while meaning something quite + // different. + if len(withGroupDeny.IdentitySIDs) != 2 || + withGroupDeny.IdentitySIDs[0] != offline || + withGroupDeny.IdentitySIDs[1] != "S-1-5-32-9999" { + t.Fatalf("group present should add its SID after the offline marker, got %v (offline marker %q)", withGroupDeny.IdentitySIDs, offline) + } + groupDenyHash, _ := WindowsNetworkInfraHash(withGroupDeny) + groupAllowHash, _ := WindowsNetworkInfraHash(withGroupAllow) + if groupDenyHash != groupAllowHash || groupDenyHash == "" { + t.Fatalf("infra hash must stay mode-independent with the group present: deny=%q allow=%q", groupDenyHash, groupAllowHash) + } + // And it must differ from the no-group hash, which is the cross-workspace + // coupling called out for a maintainer decision: one workspace creating the + // group changes every other sandbox home's expected fingerprint. + if groupDenyHash == denyHash { + t.Error("group presence did not change the infra hash; the marker staleness this causes is a deliberate property and should be visible here") + } } // A pre-existing schema-1 capability file (no Offline SID) is upgraded in place: diff --git a/internal/sandbox/windows_principal_env_windows.go b/internal/sandbox/windows_principal_env_windows.go new file mode 100644 index 000000000..9de375dd3 --- /dev/null +++ b/internal/sandbox/windows_principal_env_windows.go @@ -0,0 +1,67 @@ +//go:build windows + +package sandbox + +import ( + "os" + "path/filepath" + "strings" +) + +// windowsPrincipalIdentityEnvironment rewrites the variables that name WHO the +// command is running as. +// +// The child environment starts life as the invoking user's, and the deliberate +// sandbox redirects only replace HOME, the temp variables and the per-tool cache +// dirs. Everything identifying the account was left alone, so a command running +// as the principal still read USERPROFILE, APPDATA, LOCALAPPDATA, HOMEDRIVE, +// HOMEPATH and USERNAME describing the CALLER, whose profile the principal +// deliberately cannot open. Native tools resolve config and per-user state +// through exactly those, so they either fail during startup or, worse, quietly +// look somewhere they have no business reading. +// +// The values point into the sandbox runtime tree rather than at the principal's +// real Windows profile. That tree is already granted to the principal and +// already holds its caches, so the paths are writable by construction. Naming +// the real profile would need LoadUserProfile to have run, and a variable +// pointing at a directory that does not exist yet is a worse answer than one +// pointing somewhere usable. +// +// Layered UNDER the deliberate redirects: HOME, TMPDIR, TMP and TEMP are not +// touched here, so sandboxRuntimeEnvironment stays the single owner of those. +func windowsPrincipalIdentityEnvironment(env map[string]string, username string, runtime *SandboxRuntime) map[string]string { + username = strings.TrimSpace(username) + if env == nil || username == "" || runtime == nil { + return env + } + base := strings.TrimSpace(runtime.Data) + if base == "" { + base = strings.TrimSpace(runtime.Root) + } + if base == "" { + // No runtime tree means no writable place to point at, and inventing one + // outside the grants would only move the failure. Leaving the caller's + // values would be worse, so name the account and stop there. + env["USERNAME"] = username + return env + } + + profile := filepath.Join(base, "profile") + env["USERPROFILE"] = profile + env["APPDATA"] = filepath.Join(profile, "AppData", "Roaming") + env["LOCALAPPDATA"] = filepath.Join(profile, "AppData", "Local") + env["USERNAME"] = username + if host, err := os.Hostname(); err == nil && strings.TrimSpace(host) != "" { + // A local account's domain is the machine. Left as the caller's it would + // name a domain the principal is not a member of. + env["USERDOMAIN"] = host + } + // HOMEDRIVE and HOMEPATH are a split of the same location, and tools join + // them back together, so they have to be split from the value above rather + // than carried over from the caller. + if volume := filepath.VolumeName(profile); volume != "" { + env["HOMEDRIVE"] = volume + env["HOMEPATH"] = strings.TrimPrefix(profile, volume) + } + return env +} diff --git a/internal/sandbox/windows_principal_env_windows_test.go b/internal/sandbox/windows_principal_env_windows_test.go new file mode 100644 index 000000000..6c580611a --- /dev/null +++ b/internal/sandbox/windows_principal_env_windows_test.go @@ -0,0 +1,90 @@ +//go:build windows + +package sandbox + +import ( + "strings" + "testing" +) + +// callerEnvironment is the shape the child starts from: the invoking user's +// values, with the deliberate sandbox redirects already layered on top. +func callerEnvironment() map[string]string { + return map[string]string{ + "USERPROFILE": `C:\Users\caller`, + "APPDATA": `C:\Users\caller\AppData\Roaming`, + "LOCALAPPDATA": `C:\Users\caller\AppData\Local`, + "HOMEDRIVE": `C:`, + "HOMEPATH": `\Users\caller`, + "USERNAME": "caller", + "USERDOMAIN": "CALLERDOMAIN", + // Owned by sandboxRuntimeEnvironment, and this must not touch them. + "HOME": `C:\runtime\home`, + "TEMP": `C:\runtime\temp`, + "TMP": `C:\runtime\temp`, + } +} + +// NOTHING MAY STILL NAME THE CALLER. +// +// A principal is a separate account that deliberately cannot open the invoking +// user's profile, so every variable a native tool resolves per-user state +// through has to describe the principal instead. Left alone they pointed at the +// caller, and tools either failed during startup or looked somewhere they had no +// business reading. +func TestPrincipalEnvironmentStopsNamingTheCaller(t *testing.T) { + runtime := &SandboxRuntime{ + Root: `C:\runtime`, + Data: `C:\runtime\data`, + } + env := windowsPrincipalIdentityEnvironment(callerEnvironment(), "zero-sbx-abc123", runtime) + + for _, key := range []string{"USERPROFILE", "APPDATA", "LOCALAPPDATA"} { + if strings.Contains(strings.ToLower(env[key]), "caller") { + t.Errorf("%s still points into the caller's profile: %s", key, env[key]) + } + if !strings.HasPrefix(strings.ToLower(env[key]), strings.ToLower(runtime.Data)) { + t.Errorf("%s is not inside the runtime tree the principal can write: %s", key, env[key]) + } + } + if strings.Contains(strings.ToLower(env["HOMEPATH"]), "caller") { + t.Errorf("HOMEPATH still points into the caller's profile: %s", env["HOMEPATH"]) + } + if env["USERNAME"] != "zero-sbx-abc123" { + t.Errorf("USERNAME = %q, want the principal account", env["USERNAME"]) + } + if env["USERDOMAIN"] == "CALLERDOMAIN" { + t.Error("USERDOMAIN still names the caller's domain, which the principal is not a member of") + } + // HOMEDRIVE and HOMEPATH are rejoined by tools, so they have to describe the + // same location as USERPROFILE rather than a mix of both identities. + if rejoined := env["HOMEDRIVE"] + env["HOMEPATH"]; rejoined != env["USERPROFILE"] { + t.Errorf("HOMEDRIVE+HOMEPATH = %s, want the same location as USERPROFILE %s", rejoined, env["USERPROFILE"]) + } +} + +// The deliberate redirects stay owned by sandboxRuntimeEnvironment. Clobbering +// them here would silently move the sandbox's temp and cache handling into a +// second place, which is how the two drift apart. +func TestPrincipalEnvironmentLeavesTheSandboxRedirectsAlone(t *testing.T) { + before := callerEnvironment() + env := windowsPrincipalIdentityEnvironment(callerEnvironment(), "zero-sbx-abc123", &SandboxRuntime{ + Root: `C:\runtime`, + Data: `C:\runtime\data`, + }) + for _, key := range []string{"HOME", "TEMP", "TMP"} { + if env[key] != before[key] { + t.Errorf("%s = %q, want it left at the sandbox redirect %q", key, env[key], before[key]) + } + } +} + +// Without a runtime tree there is nowhere writable to point at, and inventing a +// path outside the grants would only move the failure. Naming the account is +// still strictly better than advertising the caller's. +func TestPrincipalEnvironmentWithoutARuntimeTreeStillRenamesTheAccount(t *testing.T) { + env := windowsPrincipalIdentityEnvironment(callerEnvironment(), "zero-sbx-abc123", &SandboxRuntime{}) + if env["USERNAME"] != "zero-sbx-abc123" { + t.Errorf("USERNAME = %q, want the principal account even with no runtime tree", env["USERNAME"]) + } +} diff --git a/internal/sandbox/windows_principal_fingerprint_test.go b/internal/sandbox/windows_principal_fingerprint_test.go new file mode 100644 index 000000000..5f5d82fb2 --- /dev/null +++ b/internal/sandbox/windows_principal_fingerprint_test.go @@ -0,0 +1,47 @@ +package sandbox + +import "testing" + +// THE FINGERPRINT MUST DESCRIBE THE PLAN THAT ACTUALLY GETS APPLIED. +// +// windowsPrincipalPlanFingerprint builds a principal plan to hash into the setup +// marker, and it was the one of three buildWindowsPrincipalACLPlan call sites +// that did not pass DenyWrite. Apply and teardown both did. So a change to the +// policy's deny-write paths moved the plan those two build and left the marker's +// hash unchanged, which is the marker failing at the one job it has. +// +// It was not a live hole, because the capability ACLPlanHash covers the same +// paths and moves the marker anyway. That is exactly what would have kept it +// invisible until somebody changed the capability plan's shape. +func TestThePrincipalFingerprintCoversDenyWrite(t *testing.T) { + workspace := t.TempDir() + base := WindowsSandboxSetupConfig{ + SandboxHome: t.TempDir(), + CommandCWD: workspace, + WorkspaceRoots: []string{workspace}, + PrincipalOptIn: true, + PermissionProfile: PermissionProfile{ + FileSystem: FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{Root: workspace}}, + }, + Network: NetworkPolicy{Mode: NetworkDeny}, + }, + } + + without, err := windowsPrincipalPlanFingerprint(base) + if err != nil { + t.Fatalf("windowsPrincipalPlanFingerprint: %v", err) + } + + withDeny := base + withDeny.PermissionProfile.FileSystem.DenyWrite = []string{workspace + string('/') + "protected"} + changed, err := windowsPrincipalPlanFingerprint(withDeny) + if err != nil { + t.Fatalf("windowsPrincipalPlanFingerprint: %v", err) + } + + if without == changed { + t.Error("adding a deny-write path did not move the principal fingerprint, so the marker cannot tell that the applied plan changed") + } +} diff --git a/internal/sandbox/windows_principal_jail_sids_windows_test.go b/internal/sandbox/windows_principal_jail_sids_windows_test.go new file mode 100644 index 000000000..9d6bbd6c9 --- /dev/null +++ b/internal/sandbox/windows_principal_jail_sids_windows_test.go @@ -0,0 +1,87 @@ +//go:build windows + +package sandbox + +import ( + "strings" + "testing" +) + +// THE ACCOUNT SID MUST NOT BE A RESTRICTING SID. +// +// WRITE_RESTRICTED allows a write only when the normal token AND the +// restricting-SID list both allow it. The principal's account SID is already +// enabled in its normal token, so listing it as a restricting SID makes the +// second check a formality for anything granted to that account: every path +// carrying a direct principal ACE passes both halves and is writable wherever it +// sits, including the profile directory Windows creates on first logon outside +// every configured write root. +// +// Same shape as the World SID in the restricted list (#865). A SID already +// carried by the normal token cannot also restrict it. +// +// The pre-existing jail test could not catch this: it grants a GROUP +// (WinBuiltinGuestsSid), so it exercises a configuration the product never +// ships. This asserts the composition that actually ships. +func TestPrincipalJailExcludesTheAccountsOwnSID(t *testing.T) { + accountSID := "S-1-5-21-9-9-9-1500" + // The account SID is planted IN the capability list on purpose. Passing a + // list that never contained it would make this test unfalsifiable: it would + // pass against any implementation, including one that appends the account SID + // straight back. This is the route the World SID took in #865. + capabilities := []string{"S-1-5-21-1-2-3-1001", accountSID, "S-1-5-21-1-2-3-1002"} + + jail := windowsPrincipalJailSIDs(capabilities, accountSID) + + for _, sid := range jail { + if strings.EqualFold(sid, accountSID) { + t.Fatalf("the principal's own account SID is a restricting SID, so any path with a direct principal ACE escapes the write jail: %v", jail) + } + } + // ...and the capability SIDs are still all present, or the principal would be + // jailed out of the workspace it owns and the fix would be a denial of service + // rather than a fix. + for _, want := range []string{"S-1-5-21-1-2-3-1001", "S-1-5-21-1-2-3-1002"} { + if !containsString(jail, want) { + t.Errorf("capability SID %q missing from the jail, so the principal loses write access to its own workspace: %v", want, jail) + } + } +} + +// The jail must not alias the caller's slice. Appending to a shared backing +// array would let a later append mutate the restricting set of a token already +// being built. +// +// Both assertions read through to the BACKING ARRAY, which is the only place +// aliasing is observable. An earlier version of this test appended a sentinel and +// then ranged over the caller's slice looking for it, which can never fail: +// append returns a new header, so the caller's length never grows and the range +// never reaches the slot the write landed in. It passed against every +// implementation, including one that returns the caller's slice verbatim. +func TestPrincipalJailDoesNotAliasTheCapabilitySlice(t *testing.T) { + capabilities := make([]string, 2, 8) + capabilities[0] = "S-1-5-21-1-2-3-1001" + capabilities[1] = "S-1-5-21-1-2-3-1002" + + jail := windowsPrincipalJailSIDs(capabilities, "S-1-5-21-9-9-9-1500") + if len(jail) == 0 { + t.Fatal("jail is empty, so neither assertion below proves anything") + } + // Shared element storage: a mutation through jail[0] would rewrite a SID in + // the restricting set of a token already being built. + if &jail[0] == &capabilities[0] { + t.Fatal("jail shares element storage with the caller's capability slice") + } + // Spare capacity: appending to an aliased jail writes into the caller's + // backing array past its length, which is invisible through the caller's own + // header but very visible to anything else holding a longer view of it. + grown := append(jail, "S-1-1-0") + shared := capabilities[:cap(capabilities)] + if shared[len(capabilities)] == "S-1-1-0" { + t.Fatalf("appending to the jail wrote through into the caller's backing array: jail=%v shared=%v", + grown, shared[:len(capabilities)+1]) + } + if len(capabilities) != 2 { + t.Errorf("caller's slice grew to %d", len(capabilities)) + } +} diff --git a/internal/sandbox/windows_principal_jail_windows_test.go b/internal/sandbox/windows_principal_jail_windows_test.go new file mode 100644 index 000000000..873a710c3 --- /dev/null +++ b/internal/sandbox/windows_principal_jail_windows_test.go @@ -0,0 +1,77 @@ +//go:build windows + +package sandbox + +import ( + "os" + "path/filepath" + "testing" + + "golang.org/x/sys/windows" +) + +// The principal path must apply the write jail, not just hand over the account's +// own token. A LogonUser token carries every write the account's ambient +// memberships grant, so without this the profile's write roots are advisory. +// +// Driven with the process token as base because minting a real principal token +// needs an elevated, provisioned machine; the restriction machinery under test +// is identical either way. +func TestRestrictWindowsTokenJailsWritesOutsideCapabilitySIDs(t *testing.T) { + var base windows.Token + desired := uint32(windows.TOKEN_DUPLICATE | windows.TOKEN_QUERY | windows.TOKEN_ASSIGN_PRIMARY | + windows.TOKEN_ADJUST_DEFAULT | windows.TOKEN_ADJUST_SESSIONID | windows.TOKEN_ADJUST_PRIVILEGES) + if err := windows.OpenProcessToken(windows.CurrentProcess(), desired, &base); err != nil { + t.Skipf("cannot open the process token here: %v", err) + } + defer base.Close() + + // A capability SID granted nowhere near the probe directory. + capSID, err := windows.CreateWellKnownSid(windows.WinBuiltinGuestsSid) + if err != nil { + t.Fatal(err) + } + jailed, err := restrictWindowsTokenForCapabilitySIDs(base, []string{capSID.String()}, true) + if err != nil { + t.Skipf("cannot build a restricted token here: %v", err) + } + defer jailed.Close() + + // Setup assertion: the unrestricted process can write here, so a denial below + // is the jail and not a broken fixture. + dir := t.TempDir() + target := filepath.Join(dir, "written.txt") + if err := os.WriteFile(target, []byte("probe"), 0o600); err != nil { + t.Fatalf("SETUP INVALID: the test process itself cannot write %s: %v", target, err) + } + if err := os.Remove(target); err != nil { + t.Fatal(err) + } + + config := WindowsSandboxCommandConfig{ + CommandCWD: dir, + WorkspaceRoots: []string{dir}, + Command: []string{"cmd", "/c", "echo probe> " + target}, + } + if _, err := runWindowsCommandAsUser(jailed, config); err != nil { + t.Fatalf("run under the jailed token: %v", err) + } + if _, err := os.Stat(target); err == nil { + t.Error("the jailed token wrote a path no capability SID covers; the write jail is not applied") + } +} + +// parseWindowsCapabilitySIDs must reject an empty list rather than build an +// unrestricted token, and must not leak the SIDs it already parsed on failure. +func TestParseWindowsCapabilitySIDsRejectsEmptyAndBadInput(t *testing.T) { + if _, err := parseWindowsCapabilitySIDs(nil); err == nil { + t.Error("empty SID list accepted; that would build a token with no restriction") + } + valid, err := windows.CreateWellKnownSid(windows.WinBuiltinGuestsSid) + if err != nil { + t.Fatal(err) + } + if _, err := parseWindowsCapabilitySIDs([]string{valid.String(), "not-a-sid"}); err == nil { + t.Error("an unparseable SID was accepted") + } +} diff --git a/internal/sandbox/windows_principal_launch_windows.go b/internal/sandbox/windows_principal_launch_windows.go new file mode 100644 index 000000000..406a36302 --- /dev/null +++ b/internal/sandbox/windows_principal_launch_windows.go @@ -0,0 +1,147 @@ +//go:build windows + +package sandbox + +import ( + "errors" + "fmt" + "strings" + "unsafe" + + "golang.org/x/sys/windows" +) + +// Privileges CreateProcessAsUser requires when the token belongs to a DIFFERENT +// account than the caller. +// +// The documented exemption from SE_ASSIGNPRIMARYTOKEN_NAME is for a restricted +// version of the caller's OWN primary token, which is exactly what the ordinary +// restricted-token path passes and exactly why that path works without holding +// anything special. A principal token comes from LogonUser against a separate +// local account, so the exemption does not apply and both privileges are back in +// play. +const ( + seAssignPrimaryTokenPrivilege = "SeAssignPrimaryTokenPrivilege" + seIncreaseQuotaPrivilege = "SeIncreaseQuotaPrivilege" +) + +// enableWindowsPrincipalLaunchPrivileges enables what launching another account +// needs, and reports precisely what is missing when it cannot. +// +// Enabling matters on its own: a privilege present in a token but DISABLED still +// fails the access check, and an elevated administrator typically holds +// SeIncreaseQuotaPrivilege in exactly that state. So this is not only a +// diagnostic, it is the step that makes the call work wherever the rights are +// actually held. +// +// Where they are not held, which is the ordinary unelevated case, this turns a +// bare "Access is denied" from deep inside process creation into a statement of +// what the sandbox cannot do and why. That failure is otherwise indistinguishable +// from the command itself being rejected, and it happens before the command's +// executable is ever opened. +func enableWindowsPrincipalLaunchPrivileges() error { + var token windows.Token + if err := windows.OpenProcessToken( + windows.CurrentProcess(), + windows.TOKEN_ADJUST_PRIVILEGES|windows.TOKEN_QUERY, + &token, + ); err != nil { + return fmt.Errorf("open process token for principal launch: %w", err) + } + defer token.Close() + + // ENUMERATE the token, do not infer from AdjustTokenPrivileges. + // + // AdjustTokenPrivileges reports a privilege the token does not hold by + // returning success and setting ERROR_NOT_ALL_ASSIGNED, and this binding does + // not surface that: it returns nil for SeTcbPrivilege on an ordinary user + // process, which holds nothing of the sort. A check built on its error would + // pass on every machine, which is worse than no check, because it would report + // the sandbox ready in exactly the case it cannot run. + held, err := windowsTokenPrivilegeLUIDs(token) + if err != nil { + return err + } + var missing []string + for _, name := range []string{seAssignPrimaryTokenPrivilege, seIncreaseQuotaPrivilege} { + luid, err := windowsPrivilegeLUID(name) + if err != nil { + return err + } + if _, ok := held[luid]; !ok { + missing = append(missing, name) + continue + } + // Held but possibly disabled, which still fails the access check. This is + // the case an elevated administrator lands in. + if err := enableWindowsTokenPrivilege(token, name); err != nil { + return fmt.Errorf("enable %s: %w", name, err) + } + } + return principalLaunchPrivilegeError(missing) +} + +// principalLaunchPrivilegeError renders the refusal for a set of missing +// privileges, and is separate from the token work so it can be driven for every +// combination. +// +// It has to be, because WHICH privilege is missing depends on the account +// running the test. An ordinary unelevated user typically holds neither, but not +// always: this developer box holds SeAssignPrimaryTokenPrivilege and not +// SeIncreaseQuotaPrivilege, so a test that named one of them and ran the real +// preflight asserted a machine's configuration rather than the code. Splitting +// the rendering out makes each case deterministic and leaves the token path with +// only the question that genuinely varies. +func principalLaunchPrivilegeError(missing []string) error { + if len(missing) == 0 { + return nil + } + return fmt.Errorf( + "launching a command as the sandbox principal needs %s, which this process does not hold. "+ + "CreateProcessAsUser exempts only a restricted version of the caller's own token, and a principal "+ + "is a separate account, so an ordinary unelevated process cannot start one. "+ + "Unset %s to use the restricted-token sandbox, which needs no such privilege", + strings.Join(missing, " and "), windowsSandboxIdentityEnv) +} + +// windowsPrivilegeLUID resolves a privilege name to its locally unique id, which +// is how a token actually names the privileges it holds. +func windowsPrivilegeLUID(name string) (windows.LUID, error) { + namePtr, err := windows.UTF16PtrFromString(name) + if err != nil { + return windows.LUID{}, fmt.Errorf("encode privilege name %s: %w", name, err) + } + var luid windows.LUID + if err := windows.LookupPrivilegeValue(nil, namePtr, &luid); err != nil { + return windows.LUID{}, fmt.Errorf("look up privilege %s: %w", name, err) + } + return luid, nil +} + +// windowsTokenPrivilegeLUIDs returns every privilege the token holds, enabled or +// not. Presence and enabled-ness are separate questions: a privilege absent from +// this set can never be enabled, while one present but disabled only needs +// AdjustTokenPrivileges. +func windowsTokenPrivilegeLUIDs(token windows.Token) (map[windows.LUID]struct{}, error) { + var size uint32 + // First call sizes the buffer and is expected to fail with + // ERROR_INSUFFICIENT_BUFFER, so its error is deliberately not checked. + _ = windows.GetTokenInformation(token, windows.TokenPrivileges, nil, 0, &size) + if size == 0 { + return nil, errors.New("query token privileges: zero-length result") + } + buffer := make([]byte, size) + if err := windows.GetTokenInformation(token, windows.TokenPrivileges, &buffer[0], size, &size); err != nil { + return nil, fmt.Errorf("query token privileges: %w", err) + } + privileges := (*windows.Tokenprivileges)(unsafe.Pointer(&buffer[0])) + if privileges.PrivilegeCount == 0 { + return map[windows.LUID]struct{}{}, nil + } + entries := unsafe.Slice(&privileges.Privileges[0], privileges.PrivilegeCount) + held := make(map[windows.LUID]struct{}, len(entries)) + for _, entry := range entries { + held[entry.Luid] = struct{}{} + } + return held, nil +} diff --git a/internal/sandbox/windows_principal_launch_windows_test.go b/internal/sandbox/windows_principal_launch_windows_test.go new file mode 100644 index 000000000..338ffb66c --- /dev/null +++ b/internal/sandbox/windows_principal_launch_windows_test.go @@ -0,0 +1,112 @@ +//go:build windows + +package sandbox + +import ( + "strings" + "testing" +) + +// The launch preflight must name what is missing and what to do about it. +// +// Without it the failure surfaces as a bare "Access is denied" from inside +// CreateProcessAsUser, which is indistinguishable from the sandboxed command +// itself being rejected. It happens before the command's executable is opened, +// so there is nothing else in the output to tell the two apart. +// +// WHICH privilege is missing is a property of the account running the test, not +// of the code. An unelevated user often holds neither, but not always: the box +// this was written on holds SeAssignPrimaryTokenPrivilege and not +// SeIncreaseQuotaPrivilege. The previous version of this test ran the real +// preflight and then asserted SeAssignPrimaryTokenPrivilege appeared in the +// message, so it passed or failed on the tester's token rather than on +// anything in the source, and it failed here for a reason that was never a +// defect. Rendering is tested directly instead, once per combination. +func TestPrincipalLaunchPrivilegeErrorNamesWhatIsMissing(t *testing.T) { + cases := []struct { + name string + missing []string + }{ + {"neither held", []string{seAssignPrimaryTokenPrivilege, seIncreaseQuotaPrivilege}}, + {"only assign-primary-token missing", []string{seAssignPrimaryTokenPrivilege}}, + {"only increase-quota missing", []string{seIncreaseQuotaPrivilege}}, + } + for _, tc := range cases { + t.Run(tc.name, func(t *testing.T) { + err := principalLaunchPrivilegeError(tc.missing) + if err == nil { + t.Fatal("a missing privilege produced no error") + } + for _, want := range tc.missing { + if !strings.Contains(err.Error(), want) { + t.Errorf("the message does not name %s, so an operator cannot act on it: %v", want, err) + } + } + // And it must not name one that IS held, or the operator goes looking + // for a privilege that was never the problem. + for _, other := range []string{seAssignPrimaryTokenPrivilege, seIncreaseQuotaPrivilege} { + if contains(tc.missing, other) { + continue + } + if strings.Contains(err.Error(), other) { + t.Errorf("the message names %s, which this case holds: %v", other, err) + } + } + if !strings.Contains(err.Error(), windowsSandboxIdentityEnv) { + t.Errorf("the message does not name the opt-out: %v", err) + } + // The point of the message is the way out. Without it the operator is + // told only that something is denied. + if !strings.Contains(err.Error(), "restricted-token sandbox") { + t.Errorf("the message does not offer the fallback, so it reads as a dead end: %v", err) + } + }) + } +} + +// Holding both is not an error, or the sandbox would refuse on exactly the +// machines it is meant to run on. +func TestPrincipalLaunchPrivilegeErrorIsNilWhenNothingIsMissing(t *testing.T) { + if err := principalLaunchPrivilegeError(nil); err != nil { + t.Errorf("no missing privileges produced an error: %v", err) + } + if err := principalLaunchPrivilegeError([]string{}); err != nil { + t.Errorf("an empty missing set produced an error: %v", err) + } +} + +// The real preflight against this machine's token. The verdict legitimately +// differs by account, so the only thing asserted is that a refusal is +// actionable: it names at least one of the two privileges rather than failing +// with something the operator cannot use. +func TestPrincipalLaunchPreflightRefusalIsActionable(t *testing.T) { + err := enableWindowsPrincipalLaunchPrivileges() + if err == nil { + t.Log("this process holds the principal launch privileges; nothing to explain") + return + } + if !strings.Contains(err.Error(), seAssignPrimaryTokenPrivilege) && + !strings.Contains(err.Error(), seIncreaseQuotaPrivilege) { + t.Errorf("the preflight refused without naming either privilege: %v", err) + } +} + +// Repeated calls must agree. This runs once per sandboxed command, and a +// preflight that answered differently on the second call would let a command +// through that the first call refused. +func TestPrincipalLaunchPreflightIsStable(t *testing.T) { + first := enableWindowsPrincipalLaunchPrivileges() + second := enableWindowsPrincipalLaunchPrivileges() + if (first == nil) != (second == nil) { + t.Fatalf("preflight verdict changed between calls: first=%v second=%v", first, second) + } +} + +func contains(values []string, want string) bool { + for _, value := range values { + if value == want { + return true + } + } + return false +} diff --git a/internal/sandbox/windows_principal_ledger.go b/internal/sandbox/windows_principal_ledger.go new file mode 100644 index 000000000..13705c74c --- /dev/null +++ b/internal/sandbox/windows_principal_ledger.go @@ -0,0 +1,160 @@ +package sandbox + +import ( + "encoding/json" + "errors" + "fmt" + "os" + "path/filepath" + "strings" +) + +// A record of the paths a sandbox principal was last granted ACEs on. +// +// applyWindowsPrincipalACLs revokes this trustee before it re-applies, but the +// only paths it could name were the ones in the plan it was about to apply. A +// root that LEFT the policy is absent from that plan, so its ACE survived the +// re-setup marker validation forces and the principal kept write access the +// current policy no longer grants — the sandbox widened as a result of being +// tightened. Teardown repeated the same current-plan-only calculation, so it did +// not clean the leftover either. +// +// Nothing on Windows can answer "which paths hold an ACE for this SID" without +// walking every volume, so the grants have to be written down as they are made. +// +// Keyed by principal, beside the secret and for the same reason: one sandbox +// home serves every workspace on the machine, so a single shared file would let +// one workspace's setup overwrite another's record — reproducing exactly the +// stale-ACE bug this exists to close, one level up. + +const windowsPrincipalACLLedgerSchemaVersion = 1 + +const windowsPrincipalACLLedgerDirName = "windows-principal-acl" + +type windowsPrincipalACLLedger struct { + SchemaVersion int `json:"schemaVersion"` + Paths []string `json:"paths"` +} + +func windowsPrincipalACLLedgerPath(sandboxHome string, username string) (string, error) { + if strings.TrimSpace(sandboxHome) == "" { + return "", errors.New("windows principal ACL ledger: empty sandbox home") + } + if strings.TrimSpace(username) == "" { + return "", errors.New("windows principal ACL ledger: empty principal name") + } + // The same guard the secret path applies, against a caller passing something + // windowsSandboxUserName did not produce. + if strings.ContainsAny(username, `\/:`) || strings.Contains(username, "..") { + return "", fmt.Errorf("windows principal ACL ledger: unsafe principal name %q", username) + } + return filepath.Join(sandboxHome, windowsPrincipalACLLedgerDirName, username+".json"), nil +} + +// readWindowsPrincipalACLLedger returns the paths an earlier setup recorded for +// this principal. +// +// recorded is false for every reason the record cannot be trusted — absent, +// unreadable, malformed, or written to a schema this build does not know — and +// not merely for "absent". Collapsing them is deliberate: there is exactly one +// safe response to any of them, and it is the same one. The prior grant set is +// unknown, and a principal whose grants are unknown cannot be reused. Returning +// an error instead would invite a caller to report it and carry on with an empty +// prior set, which is the fail-open this record exists to close — the one case +// where the previous paths are not empty but unenumerable. +func readWindowsPrincipalACLLedger(sandboxHome string, username string) ([]string, bool) { + path, err := windowsPrincipalACLLedgerPath(sandboxHome, username) + if err != nil { + return nil, false + } + contents, err := os.ReadFile(path) + if err != nil { + return nil, false + } + var ledger windowsPrincipalACLLedger + if err := json.Unmarshal(contents, &ledger); err != nil { + return nil, false + } + if ledger.SchemaVersion != windowsPrincipalACLLedgerSchemaVersion { + return nil, false + } + return trimNonEmptyStrings(ledger.Paths), true +} + +// writeWindowsPrincipalACLLedger records paths for this principal, replacing any +// previous record atomically so an interrupted write cannot leave a truncated +// file — which the reader would then treat as "no principal was ever granted +// anything", the very state it must never guess. +func writeWindowsPrincipalACLLedger(sandboxHome string, username string, paths []string) error { + path, err := windowsPrincipalACLLedgerPath(sandboxHome, username) + if err != nil { + return err + } + if err := os.MkdirAll(filepath.Dir(path), 0o700); err != nil { + return fmt.Errorf("create windows principal ACL ledger dir: %w", err) + } + contents, err := json.MarshalIndent(windowsPrincipalACLLedger{ + SchemaVersion: windowsPrincipalACLLedgerSchemaVersion, + Paths: trimNonEmptyStrings(paths), + }, "", " ") + if err != nil { + return fmt.Errorf("marshal windows principal ACL ledger: %w", err) + } + tmp, err := os.CreateTemp(filepath.Dir(path), ".windows-principal-acl-*.tmp") + if err != nil { + return fmt.Errorf("create windows principal ACL ledger temp file: %w", err) + } + tmpPath := tmp.Name() + if _, err := tmp.Write(contents); err != nil { + _ = tmp.Close() + _ = os.Remove(tmpPath) + return fmt.Errorf("write windows principal ACL ledger temp file: %w", err) + } + if err := tmp.Close(); err != nil { + _ = os.Remove(tmpPath) + return fmt.Errorf("close windows principal ACL ledger temp file: %w", err) + } + if err := os.Rename(tmpPath, path); err != nil { + _ = os.Remove(tmpPath) + return fmt.Errorf("replace windows principal ACL ledger: %w", err) + } + return nil +} + +// removeWindowsPrincipalACLLedger drops the record. An absent one is not an +// error: this runs on the teardown path, where being gone is the goal. +func removeWindowsPrincipalACLLedger(sandboxHome string, username string) error { + path, err := windowsPrincipalACLLedgerPath(sandboxHome, username) + if err != nil { + return err + } + if err := os.Remove(path); err != nil && !errors.Is(err, os.ErrNotExist) { + return fmt.Errorf("remove windows principal ACL ledger: %w", err) + } + return nil +} + +// unionWindowsPrincipalACLPaths merges path sets for revocation, keeping the +// first spelling of each path. +// +// Deduplication uses the same case-insensitive key the ACL plans use, so a root +// recorded as C:\Ws by one setup and re-granted as c:\ws by the next is one path +// to revoke rather than two. +func unionWindowsPrincipalACLPaths(sets ...[]string) []string { + seen := make(map[string]struct{}) + union := make([]string, 0) + for _, set := range sets { + for _, path := range set { + key := windowsCapabilityPathKey(path) + if key == "" { + continue + } + if _, ok := seen[key]; ok { + continue + } + seen[key] = struct{}{} + union = append(union, path) + } + } + return union +} diff --git a/internal/sandbox/windows_principal_ledger_test.go b/internal/sandbox/windows_principal_ledger_test.go new file mode 100644 index 000000000..26fd36065 --- /dev/null +++ b/internal/sandbox/windows_principal_ledger_test.go @@ -0,0 +1,174 @@ +package sandbox + +import ( + "encoding/json" + "os" + "path/filepath" + "strings" + "testing" +) + +// The point of the record is that a LATER setup can name a root the CURRENT +// policy no longer mentions, so the round trip has to survive the process that +// wrote it having no memory of the paths. +func TestPrincipalACLLedgerRoundTripsRecordedPaths(t *testing.T) { + home := t.TempDir() + paths := []string{`C:\ws\alpha`, `C:\ws\beta`, `C:\cache\runtime`} + if err := writeWindowsPrincipalACLLedger(home, "zerosbx01", paths); err != nil { + t.Fatalf("write: %v", err) + } + got, recorded := readWindowsPrincipalACLLedger(home, "zerosbx01") + if !recorded { + t.Fatal("a record this process just wrote read back as untrusted") + } + if strings.Join(got, "|") != strings.Join(paths, "|") { + t.Errorf("read back %v, want %v", got, paths) + } +} + +// One sandbox home serves every workspace on the machine. If the record were a +// single shared file, workspace B's setup would overwrite workspace A's, and A's +// dropped roots would then be unnameable at the next re-setup — the same stale +// ACE this record exists to revoke, produced by the record itself. +func TestPrincipalACLLedgerIsPerPrincipal(t *testing.T) { + home := t.TempDir() + if err := writeWindowsPrincipalACLLedger(home, "zerosbx01", []string{`C:\ws\alpha`}); err != nil { + t.Fatalf("write alpha: %v", err) + } + if err := writeWindowsPrincipalACLLedger(home, "zerosbx02", []string{`C:\ws\beta`}); err != nil { + t.Fatalf("write beta: %v", err) + } + alpha, ok := readWindowsPrincipalACLLedger(home, "zerosbx01") + if !ok || len(alpha) != 1 || alpha[0] != `C:\ws\alpha` { + t.Errorf("first principal's record = %v (ok=%v); a second workspace's setup overwrote it", alpha, ok) + } +} + +// Every untrustworthy record has to read as untrustworthy, not as "nothing was +// ever granted". The caller retires the principal on false; treating a corrupt +// file as an empty prior set is the fail-open. +func TestPrincipalACLLedgerRefusesRecordsItCannotTrust(t *testing.T) { + for name, contents := range map[string]string{ + "truncated mid-write": `{"schemaVersion": 1, "pat`, + "not json at all": "\x00\x01garbage", + "a schema from later": `{"schemaVersion": 99, "paths": ["C:\\ws"]}`, + "a schema from before": `{"paths": ["C:\\ws"]}`, + } { + t.Run(name, func(t *testing.T) { + home := t.TempDir() + path, err := windowsPrincipalACLLedgerPath(home, "zerosbx01") + if err != nil { + t.Fatalf("path: %v", err) + } + if err := os.MkdirAll(filepath.Dir(path), 0o700); err != nil { + t.Fatalf("mkdir: %v", err) + } + if err := os.WriteFile(path, []byte(contents), 0o600); err != nil { + t.Fatalf("write: %v", err) + } + if paths, recorded := readWindowsPrincipalACLLedger(home, "zerosbx01"); recorded { + t.Errorf("read a record it cannot interpret as trustworthy (%v); the caller would then revoke nothing", paths) + } + }) + } + // And an absent one, which is the ordinary first-setup case. + if _, recorded := readWindowsPrincipalACLLedger(t.TempDir(), "zerosbx01"); recorded { + t.Error("a missing record read as trusted") + } +} + +// A partial write must not be readable at all, which is why the file is renamed +// into place rather than written in situ: a reader that saw half a record would +// treat the missing half as never granted. +func TestPrincipalACLLedgerWriteIsAtomic(t *testing.T) { + home := t.TempDir() + if err := writeWindowsPrincipalACLLedger(home, "zerosbx01", []string{`C:\ws\alpha`, `C:\ws\beta`}); err != nil { + t.Fatalf("first write: %v", err) + } + if err := writeWindowsPrincipalACLLedger(home, "zerosbx01", []string{`C:\ws\alpha`}); err != nil { + t.Fatalf("second write: %v", err) + } + path, err := windowsPrincipalACLLedgerPath(home, "zerosbx01") + if err != nil { + t.Fatalf("path: %v", err) + } + contents, err := os.ReadFile(path) + if err != nil { + t.Fatalf("read: %v", err) + } + var ledger windowsPrincipalACLLedger + if err := json.Unmarshal(contents, &ledger); err != nil { + t.Fatalf("the replaced record did not parse: %v", err) + } + if len(ledger.Paths) != 1 { + t.Errorf("record = %v, want the second write to have replaced the first outright", ledger.Paths) + } + // No temp files left behind to be mistaken for a record later. + entries, err := os.ReadDir(filepath.Dir(path)) + if err != nil { + t.Fatalf("read dir: %v", err) + } + if len(entries) != 1 { + t.Errorf("ledger directory holds %d entries, want just the record", len(entries)) + } +} + +// The name comes from windowsSandboxUserName, but the path builder is the last +// thing between a caller and the filesystem, so it refuses anything that could +// escape the directory. +func TestPrincipalACLLedgerPathRefusesUnsafeNames(t *testing.T) { + for _, username := range []string{"", " ", `..\..\evil`, "a/b", `a\b`, "c:evil", "..", "zerosbx..01"} { + if _, err := windowsPrincipalACLLedgerPath(`C:\home`, username); err == nil { + t.Errorf("accepted principal name %q", username) + } + } + if _, err := windowsPrincipalACLLedgerPath("", "zerosbx01"); err == nil { + t.Error("accepted an empty sandbox home") + } + if _, err := windowsPrincipalACLLedgerPath(`C:\home`, "zerosbx01"); err != nil { + t.Errorf("rejected a name windowsSandboxUserName would produce: %v", err) + } +} + +// Removal is idempotent because teardown only cares that the record is gone, +// and a setup that failed before writing one must not make teardown fail too. +func TestPrincipalACLLedgerRemovalToleratesAnAbsentRecord(t *testing.T) { + home := t.TempDir() + if err := removeWindowsPrincipalACLLedger(home, "zerosbx01"); err != nil { + t.Fatalf("remove an absent record: %v", err) + } + if err := writeWindowsPrincipalACLLedger(home, "zerosbx01", []string{`C:\ws`}); err != nil { + t.Fatalf("write: %v", err) + } + if err := removeWindowsPrincipalACLLedger(home, "zerosbx01"); err != nil { + t.Fatalf("remove: %v", err) + } + if _, recorded := readWindowsPrincipalACLLedger(home, "zerosbx01"); recorded { + t.Error("the record survived removal") + } +} + +// The union is what setup revokes over. A path in both sets must be revoked +// once, and a root respelled between setups — Windows opens a path whatever its +// casing — is one path, not two. +func TestUnionPrincipalACLPathsDedupesTheWayTheACLPlansDo(t *testing.T) { + union := unionWindowsPrincipalACLPaths( + []string{`C:\Ws\Alpha`, `C:\ws\beta`, " "}, + []string{`c:\ws\alpha`, `C:\ws\gamma`, `C:/ws/beta`}, + ) + if len(union) != 3 { + t.Fatalf("union = %v, want three distinct paths", union) + } + // The first spelling wins: revocation needs a real path, and the recorded one + // is the spelling that was actually granted. + if union[0] != `C:\Ws\Alpha` { + t.Errorf("union[0] = %q, want the recorded spelling kept", union[0]) + } + // The recorded set comes first so a dropped root cannot be crowded out. + if union[1] != `C:\ws\beta` || union[2] != `C:\ws\gamma` { + t.Errorf("union = %v, want the recorded paths before the newly granted ones", union) + } + if got := unionWindowsPrincipalACLPaths(nil, nil); len(got) != 0 { + t.Errorf("union of nothing = %v, want empty", got) + } +} diff --git a/internal/sandbox/windows_principal_ledger_windows_test.go b/internal/sandbox/windows_principal_ledger_windows_test.go new file mode 100644 index 000000000..ff3640a1f --- /dev/null +++ b/internal/sandbox/windows_principal_ledger_windows_test.go @@ -0,0 +1,354 @@ +//go:build windows + +package sandbox + +import ( + "errors" + "os" + "path/filepath" + "testing" + + "golang.org/x/sys/windows" +) + +// The finding this record exists for, end to end and against real DACLs: setup +// grants two roots, the user removes one from their policy, setup runs again, +// and the removed root must not still name the principal. +// +// Both halves go through the PRODUCTION applyWindowsPrincipalACLs rather than +// the revoke helper, because the mechanism already worked — what did not was the +// call site's idea of which paths to revoke. It could only name the paths of the +// plan it was about to apply, and the dropped root is by definition absent from +// that plan, so the re-setup marker validation forces preserved the very ACE it +// was supposed to clear. +func TestReSetupRevokesARootTheNarrowedPolicyDropped(t *testing.T) { + home := t.TempDir() + username := "zerosbxregression" + root := t.TempDir() + kept := filepath.Join(root, "kept") + dropped := filepath.Join(root, "dropped") + for _, dir := range []string{kept, dropped} { + if err := os.MkdirAll(dir, 0o700); err != nil { + t.Fatalf("mkdir %s: %v", dir, err) + } + } + // Guests: a real, resolvable trustee this process is not a member of, so the + // ACEs below are observable without affecting the test process. + principal := "S-1-5-32-546" + + wide := FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{Root: kept}, {Root: dropped}}, + } + if _, err := applyWindowsPrincipalACLs(home, username, principal, wide, wide.WriteRoots); err != nil { + t.Fatalf("first setup: %v", err) + } + if !hasACEForTrustee(t, dropped, principal) { + t.Fatal("precondition: the first setup should have granted the root that is about to leave the policy") + } + + // The user narrows their policy and re-runs setup. Nothing in this call + // mentions the dropped root; only the record does. + narrow := FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{Root: kept}}, + } + if _, err := applyWindowsPrincipalACLs(home, username, principal, narrow, narrow.WriteRoots); err != nil { + t.Fatalf("re-setup: %v", err) + } + + if hasACEForTrustee(t, dropped, principal) { + t.Error("the principal kept its grant on a root the narrowed policy removed") + } + if !hasACEForTrustee(t, kept, principal) { + t.Error("revoking over the recorded paths also dropped the grant the narrowed policy still wants") + } + // And the record narrows with the policy, or it would accumulate every root + // the workspace has ever had. + recorded, ok := readWindowsPrincipalACLLedger(home, username) + if !ok { + t.Fatal("no record survived the re-setup") + } + if containsPathFold(recorded, dropped) { + t.Errorf("the record still names %q after the policy dropped it", dropped) + } +} + +// The record is written BEFORE any DACL changes and as the union, not after and +// as the new set. A crash between the grant and a post-hoc write would otherwise +// leave a record missing paths this run granted, stranding them at the next +// policy change — the same bug, one interruption away. +func TestPrincipalACLRecordCoversTheGrantBeforeItIsMade(t *testing.T) { + prevApply := applyWindowsACLPlanFn + t.Cleanup(func() { applyWindowsACLPlanFn = prevApply }) + + home := t.TempDir() + username := "zerosbxtwophase" + workspace := t.TempDir() + stale := filepath.Join(t.TempDir(), "left-the-policy") + if err := writeWindowsPrincipalACLLedger(home, username, []string{stale}); err != nil { + t.Fatalf("seed the record: %v", err) + } + + var atGrant []string + applyWindowsACLPlanFn = func(plan WindowsACLPlan) (func() error, error) { + if len(plan.Entries) > 0 && plan.Entries[0].Action != windowsACLRevoke { + atGrant, _ = readWindowsPrincipalACLLedger(home, username) + } + return func() error { return nil }, nil + } + + filesystem := FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{Root: workspace}}, + } + if _, err := applyWindowsPrincipalACLs(home, username, "S-1-5-32-546", filesystem, filesystem.WriteRoots); err != nil { + t.Fatalf("applyWindowsPrincipalACLs: %v", err) + } + + if !containsPathFold(atGrant, stale) || !containsPathFold(atGrant, workspace) { + t.Errorf("record at grant time = %v, want the union of the recorded and newly granted paths", atGrant) + } + after, ok := readWindowsPrincipalACLLedger(home, username) + if !ok { + t.Fatal("no record after a successful setup") + } + if containsPathFold(after, stale) { + t.Errorf("record after = %v, want it narrowed to what is granted now", after) + } + if !containsPathFold(after, workspace) { + t.Errorf("record after = %v, want the granted workspace root", after) + } +} + +// A principal from an earlier setup whose grants were never recorded is the one +// case where the prior set is not empty but unenumerable, and carrying on with +// it is the fail-open: revocation would then cover only what the new plan +// happens to name. Retiring the account instead makes every ACE that cannot be +// found name a SID Windows never reuses. +// +// The record is per role because the principals are: seeding one role's record +// must retire the other and only the other. Retiring both would destroy an +// account there is nothing wrong with, and retiring neither is the fail-open. +func TestSetupRetiresAPrincipalWithNoRecordOfItsGrants(t *testing.T) { + for name, testCase := range map[string]struct { + seedRoles []windowsSandboxRole + identityFound bool + wantRetired []windowsSandboxRole + }{ + "no record and principals from an earlier setup": { + identityFound: true, + wantRetired: []windowsSandboxRole{windowsSandboxRoleOffline, windowsSandboxRoleOnline}, + }, + "no record and nothing provisioned": {identityFound: false}, + "only the offline role has a record": { + seedRoles: []windowsSandboxRole{windowsSandboxRoleOffline}, + identityFound: true, + wantRetired: []windowsSandboxRole{windowsSandboxRoleOnline}, + }, + "both roles have a record to reconcile against": { + seedRoles: []windowsSandboxRole{windowsSandboxRoleOffline, windowsSandboxRoleOnline}, + identityFound: true, + }, + } { + t.Run(name, func(t *testing.T) { + config := stubWindowsPrincipalSetup(t) + key := windowsSandboxPrincipalKey(config) + for _, role := range testCase.seedRoles { + username := windowsSandboxUserName(key, role) + if err := writeWindowsPrincipalACLLedger(config.SandboxHome, username, []string{`C:\ws\recorded`}); err != nil { + t.Fatalf("seed the %s record: %v", role, err) + } + } + + lookupWindowsSandboxIdentityFn = func(_ string, role windowsSandboxRole) (windowsSandboxIdentity, error) { + if testCase.identityFound { + return windowsSandboxIdentity{Username: windowsSandboxUserName(key, role), SID: guestsSID(t)}, nil + } + return windowsSandboxIdentity{}, errWindowsSandboxIdentityUnavailable + } + var retired []windowsSandboxRole + removeWindowsSandboxPrincipalForSetupFn = func(_ WindowsSandboxCommandConfig, role windowsSandboxRole) error { + retired = append(retired, role) + return nil + } + + if _, err := setupWindowsSandboxPrincipal(config); err != nil { + t.Fatalf("setupWindowsSandboxPrincipal: %v", err) + } + if !sameRoles(retired, testCase.wantRetired) { + t.Errorf("retired %v, want %v", retired, testCase.wantRetired) + } + }) + } +} + +func sameRoles(got []windowsSandboxRole, want []windowsSandboxRole) bool { + if len(got) != len(want) { + return false + } + for index := range got { + if got[index] != want[index] { + return false + } + } + return true +} + +// A failure to retire has to fail the setup. Reporting success would leave the +// operator believing the sandbox is provisioned while a principal whose grants +// nobody can enumerate is still holding them. +func TestSetupFailsWhenAnUnrecordedPrincipalCannotBeRetired(t *testing.T) { + config := stubWindowsPrincipalSetup(t) + lookupWindowsSandboxIdentityFn = func(string, windowsSandboxRole) (windowsSandboxIdentity, error) { + return windowsSandboxIdentity{Username: "zerosbx", SID: guestsSID(t)}, nil + } + removeWindowsSandboxPrincipalForSetupFn = func(WindowsSandboxCommandConfig, windowsSandboxRole) error { + return errors.New("account is in use") + } + if _, err := setupWindowsSandboxPrincipal(config); err == nil { + t.Fatal("setup reported success after failing to retire a principal it cannot reconcile") + } +} + +// Teardown had the same blind spot as setup: it computed the paths to revoke +// from the CURRENT policy, so retiring a principal cleared every ACE except the +// one on the root that had left the policy — and then deleted the account, which +// left that ACE naming a SID nothing could resolve to clean up later. +func TestTeardownRevokesRecordedPathsTheCurrentPolicyNoLongerNames(t *testing.T) { + home := t.TempDir() + workspace := t.TempDir() + dropped := filepath.Join(t.TempDir(), "left-the-policy") + config := WindowsSandboxCommandConfig{ + SandboxHome: home, + CommandCWD: workspace, + WorkspaceRoots: []string{workspace}, + PermissionProfile: PermissionProfile{ + FileSystem: FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{Root: workspace}}, + }, + }, + } + role := windowsSandboxRoleOffline + username := windowsSandboxUserName(windowsSandboxPrincipalKey(config), role) + if err := writeWindowsPrincipalACLLedger(home, username, []string{dropped}); err != nil { + t.Fatalf("seed the record: %v", err) + } + + paths, err := windowsPrincipalRevocationPaths(config, "S-1-5-32-546", role) + if err != nil { + t.Fatalf("windowsPrincipalRevocationPaths: %v", err) + } + + // The other role's revocation must not pick this up: separate accounts, + // separate records, and retiring one must leave the other's ACEs alone. + other, err := windowsPrincipalRevocationPaths(config, "S-1-5-32-546", windowsSandboxRoleOnline) + if err != nil { + t.Fatalf("windowsPrincipalRevocationPaths(online): %v", err) + } + if containsPathFold(other, dropped) { + t.Errorf("the online role would revoke %q, which only the offline role's record names", dropped) + } + if !containsPathFold(paths, dropped) { + t.Errorf("teardown would revoke %v, missing the recorded root %q the policy no longer names", paths, dropped) + } + if !containsPathFold(paths, workspace) { + t.Errorf("teardown would revoke %v, missing the workspace the current policy grants", paths) + } +} + +// stubWindowsPrincipalSetup replaces every elevated call +// setupWindowsSandboxPrincipal makes, so the decision under test is reachable on +// a machine with nothing provisioned. +func stubWindowsPrincipalSetup(t *testing.T) WindowsSandboxCommandConfig { + t.Helper() + home := t.TempDir() + workspace := t.TempDir() + + prevLookup := lookupWindowsSandboxIdentityFn + prevRemove := removeWindowsSandboxPrincipalForSetupFn + prevProvision := provisionWindowsSandboxIdentityFn + prevGrant := grantWindowsSandboxLogonRightsFn + prevReset := resetWindowsSandboxUserPasswordFn + prevSecret := writeWindowsSandboxSecretFn + prevApply := applyWindowsACLPlanFn + prevCache := sandboxUserCacheDir + t.Cleanup(func() { + lookupWindowsSandboxIdentityFn = prevLookup + removeWindowsSandboxPrincipalForSetupFn = prevRemove + provisionWindowsSandboxIdentityFn = prevProvision + grantWindowsSandboxLogonRightsFn = prevGrant + resetWindowsSandboxUserPasswordFn = prevReset + writeWindowsSandboxSecretFn = prevSecret + applyWindowsACLPlanFn = prevApply + sandboxUserCacheDir = prevCache + }) + + provisionWindowsSandboxIdentityFn = func(key string, role windowsSandboxRole) (windowsSandboxIdentity, string, bool, error) { + return windowsSandboxIdentity{Username: windowsSandboxUserName(key, role), SID: guestsSID(t)}, "pw", true, nil + } + grantWindowsSandboxLogonRightsFn = func(*windows.SID) error { return nil } + resetWindowsSandboxUserPasswordFn = func(string, string) error { return nil } + writeWindowsSandboxSecretFn = func(string, string) error { return nil } + applyWindowsACLPlanFn = func(WindowsACLPlan) (func() error, error) { + return func() error { return nil }, nil + } + cache := t.TempDir() + sandboxUserCacheDir = func() (string, error) { return cache, nil } + // And the temp-derived fallback, or setup and preparation reach for the real + // one whenever the cache candidate is not selected. + fixtureTemp := filepath.Join(t.TempDir(), "temp") + if err := os.MkdirAll(fixtureTemp, 0o700); err != nil { + t.Fatal(err) + } + t.Setenv("TMP", fixtureTemp) + t.Setenv("TEMP", fixtureTemp) + + return WindowsSandboxCommandConfig{ + SandboxHome: home, + CommandCWD: workspace, + WorkspaceRoots: []string{workspace}, + PermissionProfile: PermissionProfile{ + FileSystem: FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{Root: workspace}}, + }, + }, + } +} + +func guestsSID(t *testing.T) *windows.SID { + t.Helper() + sid, err := windows.StringToSid("S-1-5-32-546") + if err != nil { + t.Fatalf("StringToSid: %v", err) + } + return sid +} + +// containsPathFold compares the way the ACL plans do, which is the only +// comparison that means anything here. +// +// Comparing the raw spellings passed CI on nothing and failed on Windows: the +// plan builder runs every root through normalizeProfilePath, whose EvalSymlinks +// expands the 8.3 short name GitHub's runners hand out for TEMP, so the record +// holds C:\Users\runneradmin\... while t.TempDir() returned C:\Users\RUNNER~1\... +// and EqualFold called two spellings of one directory different paths. A +// developer whose TEMP has no short name never sees it. +// +// Production is self-consistent — both the recorded and the newly planned paths +// go through the same normalization — so this was only ever the test being +// naive about what "same path" means on Windows. +func containsPathFold(paths []string, want string) bool { + wanted := windowsCapabilityPathKey(normalizeProfilePath(want)) + if wanted == "" { + return false + } + for _, path := range paths { + if windowsCapabilityPathKey(normalizeProfilePath(path)) == wanted { + return true + } + } + return false +} diff --git a/internal/sandbox/windows_principal_preflight_test.go b/internal/sandbox/windows_principal_preflight_test.go new file mode 100644 index 000000000..dc199ff38 --- /dev/null +++ b/internal/sandbox/windows_principal_preflight_test.go @@ -0,0 +1,103 @@ +package sandbox + +import ( + "errors" + "strings" + "testing" +) + +// PROVISIONING STAYS CLOSED WHILE NO LAUNCH MECHANISM EXISTS. +// +// Launching a command as a separate account needs SeAssignPrimaryTokenPrivilege +// and SeIncreaseQuotaPrivilege. The process that needs them is the LATER, +// ordinary command process, not setup, and nothing carries launch authority +// across that boundary. +// +// The previous gate asked whether the CALLING process held the privileges, +// which is the wrong lifetime and got the answer backwards where it mattered: +// it refused unelevated setup, which was never dangerous, and it PASSED +// elevated setup, which is precisely the case that lands a local account, its +// password, its logon rights, workspace ACEs, the recovery ledger and network +// filter state for a backend no later command can use. The gate existed to +// prevent durable state serving something that cannot run, and it produced it. +// +// So the refusal must not be satisfiable by holding the privileges right now. +func TestSetupRefusesToProvisionAPrincipalNothingCanLaunch(t *testing.T) { + optIn := true + options := WindowsSandboxSetupArgsOptions{ + CommandCWD: t.TempDir(), + SandboxHome: t.TempDir(), + PrincipalOptIn: &optIn, + } + + args, err := BuildWindowsSandboxSetupArgs(options) + if err == nil { + t.Fatalf("setup args were built for a principal nothing can launch: %v", args) + } + if !errors.Is(err, errWindowsPrincipalLaunchUnavailable) { + t.Errorf("the refusal is not the launch-unavailable one: %v", err) + } + for _, want := range []string{"SeAssignPrimaryTokenPrivilege", "SeIncreaseQuotaPrivilege", "provision"} { + if !strings.Contains(err.Error(), want) { + t.Errorf("the refusal does not mention %q: %v", want, err) + } + } + // The way forward has to be the one that works. Telling the operator to + // re-run elevated is the advice that produced the unusable state. + if strings.Contains(strings.ToLower(err.Error()), "elevated terminal") { + t.Errorf("the refusal still points at an elevated terminal, which does not help: %v", err) + } + if !strings.Contains(err.Error(), windowsSandboxIdentityEnv) { + t.Errorf("the refusal does not name the opt-out: %v", err) + } +} + +// THE REGRESSION, STATED DIRECTLY: an elevated caller must be refused too. +// +// This is the case the old check let through, and it is the only one that +// creates durable machine state. A gate that consults the current process +// cannot express it, which is why the seam takes no token and no argument. +func TestSetupRefusalDoesNotDependOnTheCallersPrivileges(t *testing.T) { + optIn := true + options := WindowsSandboxSetupArgsOptions{ + CommandCWD: t.TempDir(), + SandboxHome: t.TempDir(), + PrincipalOptIn: &optIn, + } + + // Whatever this process holds, twice, must give the same answer. + first, firstErr := BuildWindowsSandboxSetupArgs(options) + second, secondErr := BuildWindowsSandboxSetupArgs(options) + if firstErr == nil || secondErr == nil { + t.Fatalf("provisioning succeeded: %v / %v", first, second) + } + if firstErr.Error() != secondErr.Error() { + t.Errorf("the refusal is not stable across calls:\n %v\n %v", firstErr, secondErr) + } +} + +// AND THE GATE IS ONLY FOR THE OPTED-IN PATH. Without the opt-in there is no +// principal to provision, so the ordinary restricted-token sandbox, which needs +// none of those privileges, must still set up. +func TestSetupWithoutTheOptInIsUnaffected(t *testing.T) { + optOut := false + if _, err := BuildWindowsSandboxSetupArgs(WindowsSandboxSetupArgsOptions{ + CommandCWD: t.TempDir(), + SandboxHome: t.TempDir(), + PrincipalOptIn: &optOut, + }); err != nil { + t.Errorf("the restricted-token sandbox was refused for a privilege it does not need: %v", err) + } +} + +// allowPrincipalLaunchForTest opens the gate for tests that are about argument +// PLUMBING rather than about whether a principal can be launched. The opt-in +// still has to round-trip correctly for `zero doctor` and for the day a broker +// lands, and those assertions should not be deleted just because the entry +// point is closed today. +func allowPrincipalLaunchForTest(t *testing.T) { + t.Helper() + previous := windowsPrincipalLaunchAvailable + windowsPrincipalLaunchAvailable = func() error { return nil } + t.Cleanup(func() { windowsPrincipalLaunchAvailable = previous }) +} diff --git a/internal/sandbox/windows_read_capability_test.go b/internal/sandbox/windows_read_capability_test.go new file mode 100644 index 000000000..a3e9f513d --- /dev/null +++ b/internal/sandbox/windows_read_capability_test.go @@ -0,0 +1,135 @@ +package sandbox + +import "testing" + +// readCapabilityProfile is the shape that selects the strict token: a restricted +// filesystem with DenyRead set. Without DenyRead the token stays +// WRITE_RESTRICTED and reads never reach the restricted-SID check at all, so +// every assertion below would hold vacuously. +func readCapabilityProfile(readRoot string) PermissionProfile { + return PermissionProfile{ + FileSystem: FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{Root: `C:\workspace`}}, + ReadRoots: []string{`C:\`, readRoot}, + DenyRead: []string{`C:\Users\someone\.aws\credentials`}, + }, + Network: NetworkPolicy{Mode: NetworkAllow}, + } +} + +// A READ ROOT MUST CARRY A RESTRICTING SID. +// +// On the active-principal path the command runs on a strict token, because +// DenyRead drops WRITE_RESTRICTED and Windows then applies the restricted-SID +// check to reads as well as writes. Read roots were granted only to the +// principal's account SID, which windowsPrincipalJailSIDs removes from the +// restricting set on purpose, so the normal check passed and the restricted +// check matched nothing. The effect was not a narrower sandbox but an unusable +// one: no read root readable, up to and including the executable being launched. +// +// The custom read root is the case a workspace-only assertion would miss: an SDK +// or toolchain directory outside the workspace that the command must read. +func TestCapabilityPlanGrantsReadRootsToARestrictingSID(t *testing.T) { + home := t.TempDir() + custom := `C:\sdk\toolchain` + config := WindowsSandboxCommandConfig{ + SandboxHome: home, + CommandCWD: `C:\workspace`, + WorkspaceRoots: []string{`C:\workspace`}, + PermissionProfile: readCapabilityProfile(custom), + } + plan, err := BuildWindowsACLPlan(config) + if err != nil { + t.Fatalf("BuildWindowsACLPlan: %v", err) + } + readSID, err := WindowsReadAllowSID(home) + if err != nil { + t.Fatalf("WindowsReadAllowSID: %v", err) + } + if readSID == "" { + t.Fatal("no read capability SID, so this test proves nothing") + } + + for _, want := range []string{`C:\`, custom} { + found := false + for _, entry := range plan.Entries { + if entry.Action == WindowsACLAllowRead && + windowsCapabilityPathKey(entry.Path) == windowsCapabilityPathKey(want) && + entry.Capability == readSID { + found = true + break + } + } + if !found { + t.Errorf("read root %s has no allow-read entry for the read capability SID, so a strict token cannot read it", want) + } + } +} + +// The read grant must not reopen the deny list. +// +// The read roots start at the filesystem root, so the capability that makes them +// readable also covers every DenyRead carveout underneath. If the deny entries +// name only the write capabilities, the carveout stays readable through the read +// capability and DenyRead silently stops meaning anything — which is the whole +// reason the strict token is selected in the first place. +func TestDenyReadCoversTheReadCapabilitySID(t *testing.T) { + home := t.TempDir() + config := WindowsSandboxCommandConfig{ + SandboxHome: home, + CommandCWD: `C:\workspace`, + WorkspaceRoots: []string{`C:\workspace`}, + PermissionProfile: readCapabilityProfile(`C:\sdk\toolchain`), + } + plan, err := BuildWindowsACLPlan(config) + if err != nil { + t.Fatalf("BuildWindowsACLPlan: %v", err) + } + readSID, err := WindowsReadAllowSID(home) + if err != nil { + t.Fatalf("WindowsReadAllowSID: %v", err) + } + + denied := false + sawDeny := false + for _, entry := range plan.Entries { + if entry.Action != WindowsACLDenyRead { + continue + } + sawDeny = true + if entry.Capability == readSID { + denied = true + } + } + if !sawDeny { + t.Fatal("plan has no deny-read entries at all, so this test proves nothing") + } + if !denied { + t.Error("no deny-read entry names the read capability SID, so every DenyRead path stays readable through the read grant") + } +} + +// Without DenyRead the token stays WRITE_RESTRICTED, reads skip the restricted +// check, and the read grant would be ACEs that buy nothing — on roots as broad +// as the filesystem root. Both halves of the protocol decide this from the +// profile alone, so setup and the command cannot disagree about whether the +// entries exist. +func TestReadCapabilityIsAbsentWithoutDenyRead(t *testing.T) { + profile := readCapabilityProfile(`C:\sdk\toolchain`) + profile.FileSystem.DenyRead = nil + plan, err := BuildWindowsACLPlan(WindowsSandboxCommandConfig{ + SandboxHome: t.TempDir(), + CommandCWD: `C:\workspace`, + WorkspaceRoots: []string{`C:\workspace`}, + PermissionProfile: profile, + }) + if err != nil { + t.Fatalf("BuildWindowsACLPlan: %v", err) + } + for _, entry := range plan.Entries { + if entry.Action == WindowsACLAllowRead { + t.Fatalf("allow-read entry %s present without DenyRead, where reads are never restricted", entry.Path) + } + } +} diff --git a/internal/sandbox/windows_read_grant_refusal_windows_test.go b/internal/sandbox/windows_read_grant_refusal_windows_test.go new file mode 100644 index 000000000..3963aaef8 --- /dev/null +++ b/internal/sandbox/windows_read_grant_refusal_windows_test.go @@ -0,0 +1,164 @@ +//go:build windows + +package sandbox + +import ( + "bytes" + "path/filepath" + "strings" + "testing" +) + +// denyReadSetupConfig builds the production shape: read everywhere, and a +// denyRead that selects the fully restricted token. +func denyReadSetupConfig(t *testing.T) WindowsSandboxSetupConfig { + t.Helper() + workspace := t.TempDir() + volumeRoot := filepath.VolumeName(workspace) + string(filepath.Separator) + if !isWindowsVolumeRoot(volumeRoot) { + t.Fatalf("SETUP INVALID: %q is not recognised as a volume root", volumeRoot) + } + return WindowsSandboxSetupConfig{ + SandboxHome: t.TempDir(), + WorkspaceRoots: []string{workspace}, + PermissionProfile: PermissionProfile{ + FileSystem: FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{Root: workspace}}, + ReadRoots: []string{volumeRoot}, + DenyRead: []string{filepath.Join(workspace, "secrets")}, + }, + }, + } +} + +// ELEVATED SETUP CAN WRITE A VOLUME ROOT'S DACL, WHICH IS WHY IT MUST NOT. +// +// A profile that configures denyRead selects a fully restricted token, and that +// token applies its restricted-SID check to reads, so the plan grants the read +// capability at every read root. Production profiles seed ReadRoots with the bare +// filesystem root. +// +// The applier marks allow entries on a directory inheritable and calls +// SetSecurityInfo, and Windows propagates inheritable ACEs onto existing +// children. So applying that one entry is not a change to a sandbox-owned +// object: it walks and rewrites DACL inheritance across unrelated system, +// application and user trees on the drive, with the result depending on which +// descendants happen to be locked or exclusively open. +// +// It is not even sufficient afterwards. A bare root resolves on one volume, and +// the executables and libraries a command needs can live on another without +// being read roots of their own, so the strict token can still fail before its +// executable starts. +// +// The unelevated tier already refused, for the narrower reason that it lacks the +// rights. This pins the tier that HAS the rights, and pins that it refuses BEFORE +// the first mutation. +func TestElevatedSetupRefusesAVolumeRootReadGrant(t *testing.T) { + config := denyReadSetupConfig(t) + + // SETUP: the plan really reaches the volume root, or the refusal below asserts + // nothing. + plan, err := BuildWindowsACLPlan(config.commandConfig()) + if err != nil { + t.Fatalf("BuildWindowsACLPlan: %v", err) + } + if got := windowsPlanVolumeRootGrant(plan); got == "" { + t.Fatalf("SETUP INVALID: this profile carries no volume-root entry: %+v", plan.Entries) + } + + previousElevated := windowsProcessIsElevatedFn + previousApply := applyWindowsACLPlanFn + t.Cleanup(func() { + windowsProcessIsElevatedFn = previousElevated + applyWindowsACLPlanFn = previousApply + }) + windowsProcessIsElevatedFn = func() bool { return true } + applied := false + applyWindowsACLPlanFn = func(WindowsACLPlan) (func() error, error) { + applied = true + return func() error { return nil }, nil + } + + var stderr bytes.Buffer + if code := runWindowsSandboxSetup(config, &stderr); code == 0 { + t.Fatal("elevated setup accepted a plan that rewrites DACL inheritance across the whole volume") + } + if applied { + t.Fatal("the plan was applied before the refusal, so the volume-wide edit already happened") + } + // Since #1006 the upfront DenyRead gate speaks first, before setup reaches + // the volume-root refusal below it, so the message is that gate's. What the + // reader needs from either is the same: that DenyRead is the cause and that + // removing it is the fix. + message := stderr.String() + for _, want := range []string{"DenyRead", "Remove DenyRead"} { + if !strings.Contains(message, want) { + t.Errorf("the refusal does not mention %q, so the reader cannot act on it:\n%s", want, message) + } + } +} + +// And a profile with no denyRead still sets up: it needs no volume-root grant, so +// refusing it would take the Windows sandbox away from everyone. +func TestElevatedSetupStillAcceptsAWriteJailOnlyProfile(t *testing.T) { + workspace := t.TempDir() + config := WindowsSandboxSetupConfig{ + SandboxHome: t.TempDir(), + WorkspaceRoots: []string{workspace}, + PermissionProfile: PermissionProfile{ + FileSystem: FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{Root: workspace}}, + ReadRoots: []string{workspace}, + }, + }, + } + plan, err := BuildWindowsACLPlan(config.commandConfig()) + if err != nil { + t.Fatalf("BuildWindowsACLPlan: %v", err) + } + if refusal := WindowsACLPlanReadGrantRefusal(plan); refusal != "" { + t.Fatalf("a write-jail-only profile was refused: %s", refusal) + } +} + +// AND A NARROWED READ LIST IS REFUSED TOO. +// +// The first version of this guard looked for a volume root, which is a symptom +// of the production profile rather than the thing that is unsafe. +// permissionProfileReadRoots happens to seed the bare filesystem root, so that +// check covered production by coincidence. A denyRead profile with a narrowed +// read list still put an inheritable ACE on C:\Windows and sailed through. +func TestSetupRefusesAReadGrantOnANarrowedReadList(t *testing.T) { + workspace := t.TempDir() + config := WindowsSandboxCommandConfig{ + SandboxHome: t.TempDir(), + CommandCWD: workspace, + WorkspaceRoots: []string{workspace}, + PermissionProfile: PermissionProfile{ + FileSystem: FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{Root: workspace}}, + // No volume root anywhere in this list. + ReadRoots: []string{workspace, `C:\Windows`}, + DenyRead: []string{filepath.Join(workspace, "secrets")}, + }, + }, + } + plan, err := BuildWindowsACLPlan(config) + if err != nil { + t.Fatalf("BuildWindowsACLPlan: %v", err) + } + // SETUP: no volume root, or this is the case the old guard already caught. + if got := windowsPlanVolumeRootGrant(plan); got != "" { + t.Fatalf("SETUP INVALID: the plan reaches the volume root at %q", got) + } + refusal := WindowsACLPlanReadGrantRefusal(plan) + if refusal == "" { + t.Fatal("a denyRead profile with a narrowed read list was accepted; elevated setup would put an inheritable read ACE on the Windows directory") + } + if !strings.Contains(refusal, `C:\Windows`) { + t.Errorf("the refusal does not name the directory at fault: %s", refusal) + } +} diff --git a/internal/sandbox/windows_role_inventory_windows_test.go b/internal/sandbox/windows_role_inventory_windows_test.go new file mode 100644 index 000000000..5a2140b1e --- /dev/null +++ b/internal/sandbox/windows_role_inventory_windows_test.go @@ -0,0 +1,68 @@ +//go:build windows + +package sandbox + +import "testing" + +func containsRole(roles []windowsSandboxRole, want windowsSandboxRole) bool { + for _, role := range roles { + if role == want { + return true + } + } + return false +} + +// THE LEGACY ROLE BELONGS TO EXACTLY ONE OF THE TWO INVENTORIES. +// +// Both directions are real failures that have nearly happened here. +// +// In the retirable list: teardown retires legacy, and the is-it-still-installed +// guard used to ask only the live pair. An upgraded machine still holding the +// untagged pre-split account was therefore reported clean, and the opted-out +// marker claimed a teardown that had not happened. +// +// Out of the live list: provisioning legacy would create that pre-split account +// fresh on every setup of a machine that had already been upgraded past it. +func TestTheLegacyRoleIsRetirableButNeverProvisioned(t *testing.T) { + if containsRole(windowsSandboxLiveRoles, windowsSandboxRoleLegacy) { + t.Error("the legacy role is in the provisioning inventory; setup would recreate the pre-split account on every run") + } + if !containsRole(windowsSandboxRetirableRoles, windowsSandboxRoleLegacy) { + t.Error("the legacy role is not in the retirable inventory; teardown and the installed check would both miss the pre-split account") + } + for _, role := range windowsSandboxLiveRoles { + if !containsRole(windowsSandboxRetirableRoles, role) { + t.Errorf("live role %q is not retirable, so teardown would leave it behind", role) + } + } +} + +// And the installed check consults the retirable inventory, so a machine still +// holding only the legacy account is not reported clean. +func TestPrincipalIsInstalledSeesTheLegacyAccount(t *testing.T) { + prev := lookupWindowsSandboxIdentityFn + t.Cleanup(func() { lookupWindowsSandboxIdentityFn = prev }) + + var asked []windowsSandboxRole + lookupWindowsSandboxIdentityFn = func(_ string, role windowsSandboxRole) (windowsSandboxIdentity, error) { + asked = append(asked, role) + if role == windowsSandboxRoleLegacy { + return windowsSandboxIdentity{}, nil + } + return windowsSandboxIdentity{}, errWindowsSandboxIdentityUnavailable + } + + workspace := t.TempDir() + config := WindowsSandboxCommandConfig{ + SandboxHome: t.TempDir(), + CommandCWD: workspace, + WorkspaceRoots: []string{workspace}, + } + if !windowsSandboxPrincipalIsInstalled(config) { + t.Errorf("a machine still holding the legacy account was reported clean; roles asked: %v", asked) + } + if !containsRole(asked, windowsSandboxRoleLegacy) { + t.Errorf("the legacy role was never looked up: %v", asked) + } +} diff --git a/internal/sandbox/windows_role_test.go b/internal/sandbox/windows_role_test.go new file mode 100644 index 000000000..42f094b23 --- /dev/null +++ b/internal/sandbox/windows_role_test.go @@ -0,0 +1,26 @@ +package sandbox + +import "testing" + +// Doctor and the runtime must name the SAME principal. +// +// The helper this replaces existed to be that single rule and said so, and the +// drift it was meant to prevent happened anyway: dual-role provisioning made the +// runtime use an offline principal under NetworkDeny while doctor still reported +// the old restricted-token standdown, so an operator was told their reads were +// unconfined while they were confined. Both sides now call this, and the test +// pins the mapping rather than either caller's copy of it. +func TestPrincipalRoleForNetwork(t *testing.T) { + for mode, want := range map[NetworkMode]string{ + NetworkAllow: "online", + NetworkDeny: "offline", + // Unset and unrecognized lose the network rather than keeping it, matching + // what the runtime does with a mode it does not know. + "": "offline", + "not-a-mode": "offline", + } { + if got := WindowsSandboxPrincipalRoleForNetwork(mode); got != want { + t.Errorf("role for %q = %q, want %q", mode, got, want) + } + } +} diff --git a/internal/sandbox/windows_runner.go b/internal/sandbox/windows_runner.go index 400ecad58..f51183cc9 100644 --- a/internal/sandbox/windows_runner.go +++ b/internal/sandbox/windows_runner.go @@ -173,6 +173,12 @@ type WindowsSandboxCommandConfig struct { Env map[string]string SandboxLevel WindowsSandboxLevel Command []string + // CallerSID names the Windows user whose sandbox principal this config plans + // against. Empty — the case for every ordinary command — means the process + // running it, which is already the invoking user. It is set only by the + // elevated setup half, which runs as somebody else under over-the-shoulder UAC + // and would otherwise provision an account the caller can never find. + CallerSID string } func BuildWindowsSandboxCommandArgs(options WindowsSandboxCommandArgsOptions) ([]string, error) { @@ -340,15 +346,41 @@ func windowsRestrictedTokenCommandPlan(execRequest SandboxExecutionRequest, poli level := WindowsSandboxLevelRestrictedToken if execRequest.EnforcementLevel == EnforcementUnelevated { level = WindowsSandboxLevelUnelevated + // THE UNELEVATED PLAN GRANTS BOTH RUNTIME CANDIDATES, SO BOTH HAVE TO EXIST. + // Elevated setup creates them before it applies; this tier has no setup + // step, and preparation below creates only the candidate this process + // selects. With a fresh usable cache the unused fallback was absent, the + // runner refused its own plan with "windows ACL target does not exist", + // and the command never launched, on every retry. Derived and created here + // in the parent, where TEMP is still the operator's: the runner's TEMP is + // already redirected into the runtime tree, so it cannot name the + // temp-derived candidate correctly itself. + if err := ensureMissingWindowsSandboxRuntimeCandidates([]string{execRequest.WorkspaceRoot}); err != nil { + return CommandPlan{}, err + } } args, err := BuildWindowsSandboxCommandArgs(WindowsSandboxCommandArgsOptions{ - SandboxHome: sandboxHome, - CommandCWD: spec.Dir, - WorkspaceRoots: []string{execRequest.WorkspaceRoot}, - PermissionProfile: execRequest.PermissionProfile, - Env: childEnv, - SandboxLevel: level, - Command: append([]string{spec.Name}, spec.Args...), + SandboxHome: sandboxHome, + CommandCWD: spec.Dir, + WorkspaceRoots: []string{execRequest.WorkspaceRoot}, + // Augmented HERE, in the parent, because this is the last process whose + // TEMP is the operator's. + // + // The runner cannot derive the runtime candidates itself: the child env it + // runs with has TEMP and TMP pointed at the sandbox runtime temp, so + // os.TempDir() inside it returns the redirected value and the temp-derived + // candidate comes out rooted under the runtime tree instead of under the + // real temp. Setup, whose TEMP is untouched, derived the other spelling — + // same number of roots, different paths — and every command died on + // "permission roots or deny lists changed" with two 12-entry plans that + // disagreed. + PermissionProfile: windowsSandboxProfileWithRuntime( + execRequest.PermissionProfile, + []string{execRequest.WorkspaceRoot}, + ), + Env: childEnv, + SandboxLevel: level, + Command: append([]string{spec.Name}, spec.Args...), }) if err != nil { return CommandPlan{}, err @@ -448,6 +480,20 @@ type WindowsCapabilitySIDs struct { // commands include it and are blocked; online (approved) commands omit it and // reach the network — both still write-jailed by the capability SIDs. Offline string `json:"offline,omitempty"` + // ReadAllow carries the principal's READ grants, and exists because a strict + // (non-WRITE_RESTRICTED) token runs the restricted-SID check over reads too. + // + // A profile with DenyRead drops WRITE_RESTRICTED, so every read has to satisfy + // the restricting set as well as the normal token. Read roots were granted + // only to the principal's own account SID, which the write jail deliberately + // keeps OUT of that set, so the normal check passed and the restricted check + // matched nothing: the whole disk became unreadable, down to the executable + // the command was trying to start. + // + // Synthetic and held by nobody. An ACE naming it grants access only to a token + // the sandbox itself composed, and restricting SIDs can only narrow a token, + // never widen it, so publishing this grant on the read roots hands out nothing. + ReadAllow string `json:"readAllow,omitempty"` } func ResolveWindowsSandboxHome(env map[string]string) (string, error) { @@ -484,8 +530,13 @@ func LoadOrCreateWindowsCapabilitySIDs(sandboxHome string) (WindowsCapabilitySID // Back-compat: an older (schema 1) file has no offline-marker SID. // Mint one and persist so the setup helper and the runner agree on a // single value for the WFP filter scope across processes. - if caps.Offline == "" { - caps.Offline = randomWindowsCapabilitySID() + if caps.Offline == "" || caps.ReadAllow == "" { + if caps.Offline == "" { + caps.Offline = randomWindowsCapabilitySID() + } + if caps.ReadAllow == "" { + caps.ReadAllow = randomWindowsCapabilitySID() + } caps.SchemaVersion = windowsCapabilitySIDSchemaVersion if err := saveWindowsCapabilitySIDs(path, caps); err != nil { return WindowsCapabilitySIDs{}, err @@ -500,6 +551,7 @@ func LoadOrCreateWindowsCapabilitySIDs(sandboxHome string) (WindowsCapabilitySID SchemaVersion: windowsCapabilitySIDSchemaVersion, ReadOnly: randomWindowsCapabilitySID(), Offline: randomWindowsCapabilitySID(), + ReadAllow: randomWindowsCapabilitySID(), WorkspaceByRoot: map[string]string{}, WritableRootByPath: map[string]string{}, } @@ -692,3 +744,19 @@ func randomWindowsCapabilitySID() string { } return fmt.Sprintf("S-1-5-21-%d-%d-%d-%d", words[0], words[1], words[2], words[3]) } + +// WindowsReadAllowSID returns the sandbox home's read-capability SID, minting and +// persisting it on first use. Both halves of the setup protocol ask for it: the +// capability ACL plan grants it on every read root, and the principal's strict +// token carries it as a restricting SID, so the read allow-list and the +// restriction come from one value instead of two that can drift. +func WindowsReadAllowSID(sandboxHome string) (string, error) { + caps, err := LoadOrCreateWindowsCapabilitySIDs(sandboxHome) + if err != nil { + return "", err + } + if strings.TrimSpace(caps.ReadAllow) == "" { + return "", errors.New("windows sandbox read-capability SID is missing") + } + return caps.ReadAllow, nil +} diff --git a/internal/sandbox/windows_runtime_ace_windows.go b/internal/sandbox/windows_runtime_ace_windows.go new file mode 100644 index 000000000..28fffb763 --- /dev/null +++ b/internal/sandbox/windows_runtime_ace_windows.go @@ -0,0 +1,99 @@ +//go:build windows + +package sandbox + +import ( + "fmt" + "os" + "strings" + "unsafe" + + "golang.org/x/sys/windows" +) + +// verifyWindowsRuntimeRootCapability confirms the runtime roots this command is +// about to write to still carry the capability ACE setup gave them. +// +// THE MARKER ATTESTS TO A PLAN, NOT TO AN OBJECT. cleanupSandboxRuntimeRoots +// removes inactive sibling workspace roots on an age and count policy. When that +// workspace runs again, command-side preparation recreates the SAME deterministic +// pathname as the ordinary caller, and the new directory inherits from its +// parent: it does not carry the capability SID ACE that elevated setup applied to +// the object that used to be there. ValidateWindowsSandboxSetupMarker still +// passes, because it fingerprints pathnames and actions rather than the identity +// of the ACL-bearing object, so the restricted child launched and then failed its +// cache, temp and package-cache writes with a bare ACCESS_DENIED and nothing +// pointing at setup. +// +// Cleanup treats these directories as disposable cache state; setup and its +// marker treat their DACL as durable provisioned state. Checking the ACE here is +// what reconciles the two: the command refuses with the remedy instead of +// launching into a tree it cannot write. +func verifyWindowsRuntimeRootCapability(config WindowsSandboxCommandConfig) error { + runtime := config.PermissionProfile.Runtime + if runtime == nil { + return nil + } + root := strings.TrimSpace(runtime.Root) + if root == "" { + return nil + } + sid, err := windowsCapabilitySIDForWriteRoot(config, root) + if err != nil { + return fmt.Errorf("resolve the sandbox runtime capability for %s: %w", root, err) + } + present, err := windowsPathGrantsCapability(root, sid) + if err != nil { + if os.IsNotExist(err) { + return fmt.Errorf("the sandbox runtime root %s does not exist, so setup's provisioning is gone — run `zero sandbox setup` from an elevated (Administrator) terminal", root) + } + return fmt.Errorf("inspect the sandbox runtime root %s: %w", root, err) + } + if !present { + return fmt.Errorf("the sandbox runtime root %s no longer carries the capability grant setup applied to it, "+ + "which happens when the directory was reclaimed and recreated by an ordinary run; "+ + "every sandboxed write into it would fail with access denied — run `zero sandbox setup` from an elevated (Administrator) terminal", root) + } + return nil +} + +// windowsPathGrantsCapability reports whether path's DACL carries an allow ACE +// for sid. Read off the object, because the whole point is that the pathname +// says nothing about which object now answers to it. +func windowsPathGrantsCapability(path string, sid string) (bool, error) { + wanted, err := windows.StringToSid(sid) + if err != nil { + return false, fmt.Errorf("parse the capability SID %q: %w", sid, err) + } + handle, _, err := openWindowsACLTarget(path) + if err != nil { + return false, err + } + defer func() { _ = windows.CloseHandle(handle) }() + descriptor, err := windows.GetSecurityInfo(handle, windows.SE_FILE_OBJECT, windows.DACL_SECURITY_INFORMATION) + if err != nil { + return false, err + } + dacl, _, err := descriptor.DACL() + if err != nil || dacl == nil { + return false, err + } + // GetAce hands back a generic header that is reinterpreted here, which is + // sound only for the fixed-layout ACE types: an object ACE carries Flags and + // two GUIDs ahead of the trustee, so SidStart would land mid-structure. A type + // this does not recognise is skipped rather than decoded into a nonsense SID. + for index := uint32(0); index < uint32(dacl.AceCount); index++ { + var ace *windows.ACCESS_ALLOWED_ACE + if err := windows.GetAce(dacl, index, &ace); err != nil { + continue + } + if ace.Header.AceType != windows.ACCESS_ALLOWED_ACE_TYPE { + continue + } + aceSID := (*windows.SID)(unsafe.Pointer(&ace.SidStart)) + if aceSID.Equals(wanted) { + return true, nil + } + } + return false, nil +} diff --git a/internal/sandbox/windows_runtime_candidate_dacl_windows.go b/internal/sandbox/windows_runtime_candidate_dacl_windows.go new file mode 100644 index 000000000..d68de35e6 --- /dev/null +++ b/internal/sandbox/windows_runtime_candidate_dacl_windows.go @@ -0,0 +1,108 @@ +//go:build windows + +package sandbox + +import ( + "errors" + "fmt" + "os" + + "golang.org/x/sys/windows" +) + +// SETUP'S PREPARATION IS A MUTATION, AND IT HAPPENS BEFORE THE TRANSACTION. +// +// ensureWindowsSandboxRuntimeCandidates calls EnsurePrivateDir on every +// candidate, existing ones included, and on Windows that replaces the DACL with +// a protected owner-and-SYSTEM descriptor. A root a previous setup already +// granted the capability SID on loses that grant right there, before the ACL +// transaction has snapshotted anything. If setup then fails on the second +// candidate, or the transaction fails and restores its own baseline, the old +// grant is gone: the marker still says setup succeeded and every command fails +// runtime-capability verification. So the DACLs are snapshotted before the +// preparation touches them and put back on any failure that follows, which +// makes the preparation part of the recoverable operation without giving up +// the ownership and no-follow checks EnsurePrivateDir performs. +type windowsRuntimeCandidateDACL struct { + Path string + Descriptor *windows.SECURITY_DESCRIPTOR + Protected bool + TargetID windowsFileIdentity +} + +// snapshotWindowsRuntimeCandidateDACLs records the DACL and identity of every +// candidate that exists. A missing candidate has nothing to preserve. +func snapshotWindowsRuntimeCandidateDACLs(roots []string) ([]windowsRuntimeCandidateDACL, error) { + var snapshots []windowsRuntimeCandidateDACL + for _, root := range roots { + handle, _, err := openWindowsACLTarget(root) + if err != nil { + if errors.Is(err, os.ErrNotExist) { + continue + } + return nil, fmt.Errorf("snapshot runtime root %s before setup: %w", root, err) + } + snapshot, err := snapshotWindowsRuntimeCandidateDACL(handle, root) + _ = windows.CloseHandle(handle) + if err != nil { + return nil, err + } + snapshots = append(snapshots, snapshot) + } + return snapshots, nil +} + +func snapshotWindowsRuntimeCandidateDACL(handle windows.Handle, root string) (windowsRuntimeCandidateDACL, error) { + descriptor, err := windows.GetSecurityInfo(handle, windows.SE_FILE_OBJECT, windows.DACL_SECURITY_INFORMATION) + if err != nil { + return windowsRuntimeCandidateDACL{}, fmt.Errorf("read runtime root DACL for %s: %w", root, err) + } + control, _, err := descriptor.Control() + if err != nil { + return windowsRuntimeCandidateDACL{}, fmt.Errorf("read runtime root descriptor control for %s: %w", root, err) + } + identity, err := windowsIdentityOfHandle(handle) + if err != nil { + return windowsRuntimeCandidateDACL{}, fmt.Errorf("identify runtime root %s: %w", root, err) + } + return windowsRuntimeCandidateDACL{ + Path: root, + Descriptor: descriptor, + Protected: control&windows.SE_DACL_PROTECTED != 0, + TargetID: identity, + }, nil +} + +// restoreWindowsRuntimeCandidateDACLs puts each recorded DACL back on the +// object it was read from, protection state included, and refuses an object +// that is no longer the one recorded. +func restoreWindowsRuntimeCandidateDACLs(snapshots []windowsRuntimeCandidateDACL) error { + var errs []error + for _, snapshot := range snapshots { + dacl, _, err := snapshot.Descriptor.DACL() + if err != nil { + errs = append(errs, fmt.Errorf("read recorded DACL for %s: %w", snapshot.Path, err)) + continue + } + handle, _, err := openWindowsACLTarget(snapshot.Path) + if err != nil { + errs = append(errs, fmt.Errorf("re-open runtime root %s to restore its DACL: %w", snapshot.Path, err)) + continue + } + got, identityErr := windowsIdentityOfHandle(handle) + if identityErr != nil || got != snapshot.TargetID { + _ = windows.CloseHandle(handle) + errs = append(errs, fmt.Errorf("refusing to restore the DACL of %s: it is no longer the object setup found", snapshot.Path)) + continue + } + flags := windows.SECURITY_INFORMATION(windows.DACL_SECURITY_INFORMATION | windows.UNPROTECTED_DACL_SECURITY_INFORMATION) + if snapshot.Protected { + flags = windows.DACL_SECURITY_INFORMATION | windows.PROTECTED_DACL_SECURITY_INFORMATION + } + if err := windows.SetSecurityInfo(handle, windows.SE_FILE_OBJECT, flags, nil, nil, dacl, nil); err != nil { + errs = append(errs, fmt.Errorf("restore the DACL of %s: %w", snapshot.Path, err)) + } + _ = windows.CloseHandle(handle) + } + return errors.Join(errs...) +} diff --git a/internal/sandbox/windows_runtime_candidate_dacl_windows_test.go b/internal/sandbox/windows_runtime_candidate_dacl_windows_test.go new file mode 100644 index 000000000..6645e9b9a --- /dev/null +++ b/internal/sandbox/windows_runtime_candidate_dacl_windows_test.go @@ -0,0 +1,135 @@ +//go:build windows + +package sandbox + +import ( + "bytes" + "errors" + "os" + "path/filepath" + "testing" + + "golang.org/x/sys/windows" +) + +// runtimeRootGrants reports whether path's DACL carries an allow ACE for sid. +func runtimeRootGrants(t *testing.T, path string, sid string) bool { + t.Helper() + want, err := windows.StringToSid(sid) + if err != nil { + t.Fatal(err) + } + descriptor, err := windows.GetNamedSecurityInfo(path, windows.SE_FILE_OBJECT, windows.DACL_SECURITY_INFORMATION) + if err != nil { + t.Fatalf("GetNamedSecurityInfo %s: %v", path, err) + } + dacl, _, err := descriptor.DACL() + if err != nil { + t.Fatal(err) + } + if dacl == nil { + return false + } + for index := uint16(0); index < dacl.AceCount; index++ { + var ace *windows.ACCESS_ALLOWED_ACE + if err := windows.GetAce(dacl, uint32(index), &ace); err != nil { + t.Fatalf("GetAce %d: %v", index, err) + } + if ace.Header.AceType != windows.ACCESS_ALLOWED_ACE_TYPE { + continue + } + if got, ok := windowsAceSID(ace); ok && got.Equals(want) { + return true + } + } + return false +} + +// A FAILED SETUP RETRY LEAVES THE EARLIER SETUP'S GRANT IN PLACE. +// +// Preparing the runtime candidates rewrote an existing root's DACL before the +// ACL transaction had a baseline, so a retry that then failed on the second +// candidate, or in the apply, returned with the working grant gone and the old +// marker still claiming success. The DACLs are snapshotted before preparation +// and restored on every later failure. +func TestSetupRetryFailurePreservesAnExistingRuntimeGrant(t *testing.T) { + withWindowsSetupSeams(t, windowsSetupSeams{elevated: true}) + isolateSandboxRuntimeRoots(t) + config := windowsSetupTestConfig(t, false) + candidates := windowsSandboxRuntimeCandidates(config.WorkspaceRoots) + if len(candidates) != 2 { + t.Fatalf("SETUP INVALID: want a cache-derived and a fallback candidate, got %v", candidates) + } + granted, fallback := candidates[0], candidates[1] + + // A previous, successful setup: the first candidate exists and grants the + // capability SID. Guests stands in for it here. + if err := ensureRuntimeCandidateDir(granted); err != nil { + t.Fatalf("prepare the first candidate: %v", err) + } + rollback, err := applyWindowsACLPlan(WindowsACLPlan{Entries: []WindowsACLEntry{{ + Action: WindowsACLAllowWrite, Path: granted, Capability: windowsACLTestDenySID, + }}}) + if err != nil { + t.Fatalf("grant the first candidate: %v", err) + } + t.Cleanup(func() { _ = rollback() }) + if !runtimeRootGrants(t, granted, windowsACLTestDenySID) { + t.Fatal("SETUP INVALID: the grant did not land on the first candidate") + } + + t.Run("second candidate fails validation", func(t *testing.T) { + // A file where the fallback candidate directory should be: EnsurePrivateDir + // refuses it, after it has already rewritten the first candidate. + if err := os.MkdirAll(filepath.Dir(fallback), 0o700); err != nil { + t.Fatal(err) + } + if err := os.WriteFile(fallback, []byte("in the way"), 0o600); err != nil { + t.Fatal(err) + } + t.Cleanup(func() { _ = os.Remove(fallback) }) + + var stderr bytes.Buffer + if code := runWindowsSandboxSetup(config, &stderr); code == 0 { + t.Fatalf("setup succeeded with a file at the fallback candidate: %s", stderr.String()) + } + if !runtimeRootGrants(t, granted, windowsACLTestDenySID) { + t.Fatalf("the failed retry stripped the earlier grant from %s: %s", granted, stderr.String()) + } + }) + + t.Run("apply fails", func(t *testing.T) { + if err := os.Remove(fallback); err != nil && !os.IsNotExist(err) { + t.Fatal(err) + } + previous := applyWindowsACLPlanFn + applyWindowsACLPlanFn = func(WindowsACLPlan) (func() error, error) { + return nil, errors.New("injected apply failure") + } + t.Cleanup(func() { applyWindowsACLPlanFn = previous }) + + var stderr bytes.Buffer + if code := runWindowsSandboxSetup(config, &stderr); code == 0 { + t.Fatalf("setup succeeded through an injected apply failure: %s", stderr.String()) + } + if !runtimeRootGrants(t, granted, windowsACLTestDenySID) { + t.Fatalf("the failed apply stripped the earlier grant from %s: %s", granted, stderr.String()) + } + }) +} + +// CONTROL: fresh roots still get created and setup still succeeds. +func TestSetupCreatesFreshRuntimeCandidates(t *testing.T) { + withWindowsSetupSeams(t, windowsSetupSeams{elevated: true}) + isolateSandboxRuntimeRoots(t) + config := windowsSetupTestConfig(t, false) + var stderr bytes.Buffer + if code := runWindowsSandboxSetup(config, &stderr); code != 0 { + t.Fatalf("setup on fresh roots failed: %s", stderr.String()) + } + for _, root := range windowsSandboxRuntimeCandidates(config.WorkspaceRoots) { + if info, err := os.Lstat(root); err != nil || !info.IsDir() { + t.Fatalf("candidate %s was not created: %v", root, err) + } + } +} diff --git a/internal/sandbox/windows_runtime_reclaim_windows_test.go b/internal/sandbox/windows_runtime_reclaim_windows_test.go new file mode 100644 index 000000000..1edce596e --- /dev/null +++ b/internal/sandbox/windows_runtime_reclaim_windows_test.go @@ -0,0 +1,104 @@ +//go:build windows + +package sandbox + +import ( + "os" + "path/filepath" + "strings" + "testing" +) + +// CLEANUP AND SETUP DISAGREE ABOUT WHO OWNS THE RUNTIME DIRECTORY. +// +// cleanupSandboxRuntimeRoots reclaims inactive sibling workspace roots on an age +// and count policy, treating them as disposable cache state. Setup and its marker +// treat their DACL as durable provisioned state. When the reclaimed workspace runs +// again, command-side preparation recreates the SAME deterministic pathname as +// the ordinary caller, and the new directory inherits from its parent without the +// capability SID ACE elevated setup applied to the object that used to be there. +// +// The marker still validates, because it fingerprints pathnames and actions +// rather than the identity of the ACL-bearing object. So the restricted child +// launched and then failed its cache, temp and package-cache writes with a bare +// access-denied and nothing pointing at setup. +// +// Driven by applying a real plan, reclaiming the directory the way cleanup does, +// recreating the pathname the way an ordinary run does, and asking the check the +// command now makes. DACL edits on a user-owned directory need no Administrator +// rights, so this runs unelevated. +// +// Not covered here: an actual write performed by a real capability token, which +// needs a provisioned sandbox this box cannot create. The ACE presence is the +// fact the command consumes, and it is what this pins. +func TestRuntimeRootReclaimedAndRecreatedLosesItsCapability(t *testing.T) { + cacheRoot := t.TempDir() + previous := sandboxUserCacheDir + t.Cleanup(func() { sandboxUserCacheDir = previous }) + sandboxUserCacheDir = func() (string, error) { return cacheRoot, nil } + + workspace := t.TempDir() + runtimeRoot, ok := deterministicSandboxRuntimeRoot(canonicalSandboxWorkspaceRoot(workspace), canonicalSandboxWorkspaceRoot(cacheRoot)) + if !ok { + t.Fatalf("SETUP INVALID: no deterministic runtime root for %s", workspace) + } + if err := os.MkdirAll(runtimeRoot, 0o700); err != nil { + t.Fatal(err) + } + + config := WindowsSandboxCommandConfig{ + SandboxHome: t.TempDir(), + CommandCWD: workspace, + WorkspaceRoots: []string{workspace}, + PermissionProfile: PermissionProfile{ + Runtime: &SandboxRuntime{Root: runtimeRoot}, + FileSystem: FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{ + {Root: workspace}, + {Root: runtimeRoot}, + }, + }, + }, + } + + plan, err := BuildWindowsACLPlan(config) + if err != nil { + t.Fatalf("BuildWindowsACLPlan: %v", err) + } + if _, err := applyWindowsACLPlan(plan); err != nil { + t.Skipf("cannot apply an ACL plan here: %v", err) + } + + // A provisioned machine passes, or the assertion below would be satisfied by a + // check that refuses unconditionally. + if err := verifyWindowsRuntimeRootCapability(config); err != nil { + t.Fatalf("SETUP INVALID: a freshly provisioned runtime root was rejected: %v", err) + } + + // Cleanup reclaims it, and the next ordinary run recreates the same pathname. + if err := os.RemoveAll(runtimeRoot); err != nil { + t.Fatalf("reclaim the runtime root the way cleanup does: %v", err) + } + if err := os.MkdirAll(runtimeRoot, 0o700); err != nil { + t.Fatalf("recreate the pathname the way an ordinary run does: %v", err) + } + + err = verifyWindowsRuntimeRootCapability(config) + if err == nil { + t.Fatal("a recreated runtime root with no capability grant was accepted; the command would launch and then fail every sandboxed write with nothing explaining why") + } + message := err.Error() + if !strings.Contains(message, runtimeRoot) { + t.Errorf("the refusal does not name the root: %v", err) + } + if !strings.Contains(strings.ToLower(message), "sandbox setup") { + t.Errorf("the refusal does not point at the remedy: %v", err) + } + + // And the marker on its own still says everything is fine, which is exactly + // why the object has to be checked separately. + if _, statErr := os.Stat(filepath.Dir(runtimeRoot)); statErr != nil { + t.Fatalf("the runtime parent vanished: %v", statErr) + } +} diff --git a/internal/sandbox/windows_secret_removal_windows_test.go b/internal/sandbox/windows_secret_removal_windows_test.go new file mode 100644 index 000000000..fd0ca9dc6 --- /dev/null +++ b/internal/sandbox/windows_secret_removal_windows_test.go @@ -0,0 +1,63 @@ +//go:build windows + +package sandbox + +import ( + "errors" + "os" + "path/filepath" + "testing" +) + +// A secret this account cannot delete must be distinguishable from one that +// simply would not delete. +// +// Teardown carries on past the first and aborts on the second. Before the +// distinction existed, a secret owned by another administrator's setup stopped +// teardown at its very first step, leaving the account, its logon rights, its +// ACEs and its ledger all installed. The operator asked for removal, got an +// error, and kept a working principal. +func TestRemovingAnUndeletableSecretIsDistinguishable(t *testing.T) { + // A non-empty directory standing in for a secret: os.Remove refuses it, which + // gives a real removal failure without needing a second administrator. + path := filepath.Join(t.TempDir(), "secret") + if err := os.Mkdir(path, 0o700); err != nil { + t.Fatalf("seed: %v", err) + } + if err := os.WriteFile(filepath.Join(path, "occupant"), []byte("x"), 0o600); err != nil { + t.Fatalf("seed occupant: %v", err) + } + + err := removeWindowsSandboxSecret(path) + if err == nil { + t.Fatal("a secret that could not be removed reported success") + } + // This is NOT the foreign-owner case, so teardown must still treat it as + // fatal rather than shrugging and carrying on. + if errors.Is(err, errWindowsSandboxSecretNotOurs) { + t.Errorf("an ordinary removal failure was classified as a foreign owner: %v", err) + } +} + +// Absence is success. Teardown runs on machines that never provisioned a +// principal, and re-running it must converge rather than fail. +func TestRemovingAMissingSecretIsNotAnError(t *testing.T) { + if err := removeWindowsSandboxSecret(filepath.Join(t.TempDir(), "never-existed")); err != nil { + t.Fatalf("removing a missing secret reported an error: %v", err) + } +} + +// The ordinary case still works, so the tolerance above did not turn removal +// into a no-op. +func TestRemovingOurOwnSecretDeletesIt(t *testing.T) { + path := filepath.Join(t.TempDir(), "secret") + if err := os.WriteFile(path, []byte("x"), 0o600); err != nil { + t.Fatalf("seed: %v", err) + } + if err := removeWindowsSandboxSecret(path); err != nil { + t.Fatalf("removeWindowsSandboxSecret: %v", err) + } + if _, err := os.Stat(path); !errors.Is(err, os.ErrNotExist) { + t.Errorf("the secret survived removal: stat err = %v", err) + } +} diff --git a/internal/sandbox/windows_setup.go b/internal/sandbox/windows_setup.go index ac7366a7c..3d3143797 100644 --- a/internal/sandbox/windows_setup.go +++ b/internal/sandbox/windows_setup.go @@ -11,17 +11,121 @@ import ( "path/filepath" "sort" "strings" + + "github.com/Gitlawb/zero/internal/peermsg" ) const WindowsSandboxSetupName = "zero-windows-sandbox-setup.exe" -const windowsSandboxSetupMarkerSchemaVersion = 5 +// Bumped to 7 when the single principal was split into offline and online role +// accounts. THE NAMES CHANGED, so a marker written by the previous version is +// not evidence that this version's setup has run: it validates by comparing +// serialized plans and hashes, and neither names an account. Left at 6, an +// upgraded machine kept a valid marker, setup was never re-run, and the runner +// found neither "zero-sbx-d" nor "zero-sbx-n" and fell back to the +// restricted-token backend with no read confinement, silently. +// +// Bumping it makes that installation report as out of date, which is what sends +// the operator back through elevated setup. Setup then provisions both roles and +// retires the untagged predecessor (see removeWindowsSandboxPrincipalsForSetup). +const windowsSandboxSetupMarkerSchemaVersion = 7 + +// windowsSandboxIdentityEnv opts a machine into the principal backend while it +// is still experimental. Provisioning is inert without it, so an existing +// install keeps the restricted-token behaviour until someone turns this on. +// +// Lives here, beside the setup protocol rather than beside the Windows-only +// runtime, because the opt-in is part of that protocol: it has to be readable on +// every platform so the setup args and the marker can carry it. +const windowsSandboxIdentityEnv = "ZERO_WINDOWS_SANDBOX_IDENTITY" + +// windowsSandboxIdentityEnabled reports whether the principal backend is opted +// into. An explicit entry in env is authoritative; otherwise the process +// environment decides. +func windowsSandboxIdentityEnabled(env map[string]string) bool { + if value, ok := env[windowsSandboxIdentityEnv]; ok { + return strings.TrimSpace(value) == "1" + } + return strings.TrimSpace(os.Getenv(windowsSandboxIdentityEnv)) == "1" +} + +// WindowsSandboxPrincipalOptIn resolves the principal opt-in for callers outside +// this package (the `zero sandbox setup` CLI and `zero doctor`), so both sides +// of the setup protocol read the opt-in the same way. Pass nil to consult the +// current process environment. +func WindowsSandboxPrincipalOptIn(env map[string]string) bool { + return windowsSandboxIdentityEnabled(env) +} + +func windowsSandboxPrincipalOptInValue(optIn bool) string { + if optIn { + return "1" + } + return "0" +} type WindowsSandboxSetupArgsOptions struct { SandboxHome string CommandCWD string WorkspaceRoots []string PermissionProfile PermissionProfile + // PrincipalOptIn is the caller's principal opt-in, serialized into the setup + // args. Elevated setup runs in its own process — a UAC-elevated one whose + // environment is not the caller's — so it must be told the value rather than + // left to sample an environment nobody set. + // + // Tri-state on purpose. nil means "this caller did not resolve the opt-in", + // and BuildWindowsSandboxSetupArgs then resolves it from the environment of + // the process building the args — which is the caller's own process, the one + // place where the ambient value is the value the operator typed. A plain bool + // could not say that: its zero value asserts "opted out", so every caller that + // simply did not know about this field would serialize `--sandbox-principal 0` + // while the command half still resolved the opt-in from its environment. Under + // a machine-wide opt-in the two halves would then disagree and marker + // validation would refuse every command — the same silent-disagreement bug + // this flag exists to remove, re-created one layer up. + // + // Set it only to override the environment (a caller holding a command's Env + // map, or a test pinning a value); leave it nil to mean "whatever this shell + // says", which is what `zero sandbox setup` and `zero doctor` want. + PrincipalOptIn *bool + // CallerSID identifies the Windows user setup was invoked BY, serialized into + // the args for the same reason PrincipalOptIn is: elevated setup runs in its + // own process, and under over-the-shoulder UAC (or `runas`) that process + // belongs to a DIFFERENT administrator than the caller. + // + // Everything a principal needs is scoped to the invoking user — the account + // name, its ACL ledger and its DPAPI secret all key off that identity — so a + // helper that sampled its own token would provision an account for the + // administrator who typed the credentials while every later command looked for + // one named after the ordinary user, found nothing, fell back to the restricted + // token, and left an admin-keyed account nothing would ever reclaim. + // + // Empty means "resolve from the process building the args", which IS the + // caller: these args are built before the UAC boundary is crossed. + CallerSID string +} + +// callerSID resolves the invoking user, in the caller's own process, before the +// args cross the UAC boundary. Empty when the token cannot be read (and on +// non-Windows builds), which leaves the elevated helper on its pre-existing +// behaviour of sampling its own token rather than inventing an identity. +func (options WindowsSandboxSetupArgsOptions) callerSID() string { + if sid := strings.TrimSpace(options.CallerSID); sid != "" { + return sid + } + return windowsCurrentUserSID() +} + +// principalOptIn resolves the tri-state. It runs inside +// BuildWindowsSandboxSetupArgs, i.e. in the caller's process, before the args +// cross the UAC boundary — so an unset caller still ships an explicit 0|1 that +// the elevated helper can trust. +func (options WindowsSandboxSetupArgsOptions) principalOptIn() bool { + if options.PrincipalOptIn != nil { + return *options.PrincipalOptIn + } + return windowsSandboxIdentityEnabled(nil) } type WindowsSandboxSetupConfig struct { @@ -29,6 +133,11 @@ type WindowsSandboxSetupConfig struct { CommandCWD string WorkspaceRoots []string PermissionProfile PermissionProfile + PrincipalOptIn bool + // CallerSID is the invoking user this setup is provisioning FOR. See + // WindowsSandboxSetupArgsOptions.CallerSID. Empty means the caller did not + // say, and the helper falls back to its own token. + CallerSID string } type WindowsSandboxSetupMarker struct { @@ -43,6 +152,21 @@ type WindowsSandboxSetupMarker struct { NetworkInfraHash string `json:"networkInfraHash"` OfflineFilterSID string `json:"offlineFilterSid"` NetworkFilters int `json:"networkFilters"` + // PrincipalOptIn records whether the run that wrote this marker provisioned a + // sandbox principal. Without it the two halves each sampled their own + // environment and could disagree silently — see + // ValidateWindowsSandboxSetupMarker. + PrincipalOptIn bool `json:"principalOptIn"` + // PrincipalPlanHash fingerprints the PRINCIPAL ACL plan, which ACLPlanHash + // above does not cover: that one hashes BuildWindowsACLPlan, the + // capability-SID plan, while principal grants are built separately by + // buildWindowsPrincipalACLPlan from the same profile. + // + // Without it, narrowing or removing a principal read root left setup looking + // current, so the old AllowRead ACEs stayed on disk with nothing to notice + // they no longer matched the policy. Empty when the principal backend is not + // opted into, which keeps the marker stable for the default install. + PrincipalPlanHash string `json:"principalPlanHash,omitempty"` } func WindowsSandboxSetupMarkerPath(sandboxHome string) string { @@ -66,6 +190,43 @@ func BuildWindowsSandboxSetupArgs(options WindowsSandboxSetupArgsOptions) ([]str if len(workspaceRoots) == 0 { workspaceRoots = []string{commandCWD} } + // THE PRINCIPAL BACKEND HAS NO LAUNCH PATH YET, SO IT IS NOT PROVISIONED. + // + // Launching a command as a separate account needs SeAssignPrimaryTokenPrivilege + // and SeIncreaseQuotaPrivilege. An ordinary unelevated process holds neither, + // and the command runs LATER, from the caller, not from setup. So the + // advertised lifecycle -- provision once from an elevated terminal, then run + // ordinary commands -- cannot execute: the elevated setup process holds the + // privileges and the process that needs them does not. + // + // An earlier version of this check tested the CALLING process's token, which + // is the wrong lifetime. It refused unelevated setup, which was never the + // dangerous case, and PASSED elevated setup, which is exactly the case that + // creates a local account, its password, its logon-right assignments, + // workspace ACEs, the recovery ledger and network filter state for a backend + // no later command can use. Durable machine state serving something that + // cannot run is the outcome the check existed to prevent, and it produced it. + // + // Requiring every sandboxed command to run elevated would make the privileges + // line up, and is deliberately not done: it is a worse boundary than the + // restricted token this would replace, and it is not the lifecycle the feature + // is for. Until a reviewed launch mechanism exists -- a bootstrap or an + // authenticated broker that carries launch authority across the UAC boundary + // -- provisioning stays unavailable rather than half-working. + // + // Everything below the launch path (the plans, the ledger, the ACL and filter + // work) still builds and is still tested; it is the provisioning entry point + // that is closed. See #662, which this PR is foundation for and does not + // close. + if options.principalOptIn() { + if err := windowsPrincipalLaunchAvailable(); err != nil { + return nil, fmt.Errorf("refusing to provision a sandbox principal: %w", err) + } + } + // Augmented here, in the caller's shell, before the args cross into the + // elevated helper. Same reason the opt-in and caller SID are: the value has to + // be resolved where the environment is the operator's. + options.PermissionProfile = WindowsSandboxProfileWithRuntimeRoots(options.PermissionProfile, workspaceRoots) profileJSON, err := json.Marshal(options.PermissionProfile) if err != nil { return nil, fmt.Errorf("marshal windows sandbox setup permission profile: %w", err) @@ -74,6 +235,16 @@ func BuildWindowsSandboxSetupArgs(options WindowsSandboxSetupArgsOptions) ([]str "--sandbox-home", sandboxHome, "--command-cwd", commandCWD, "--permission-profile", string(profileJSON), + // Always explicit, never omitted-means-false: the elevated helper must be + // able to tell "the caller wants no principal" from "an older caller said + // nothing", and only the first of those is safe to run silently. + "--sandbox-principal", windowsSandboxPrincipalOptInValue(options.principalOptIn()), + } + // Omitted rather than sent empty when the caller's identity cannot be read: + // absence means "this caller did not say", which the helper answers by + // sampling its own token exactly as it did before this flag existed. + if callerSID := options.callerSID(); callerSID != "" { + args = append(args, "--caller-sid", callerSID) } for _, root := range workspaceRoots { args = append(args, "--workspace-root", root) @@ -117,6 +288,31 @@ func ParseWindowsSandboxSetupArgs(args []string) (WindowsSandboxSetupConfig, err } profileJSON = strings.TrimSpace(value) index = next + case "--sandbox-principal": + value, next, err := nextWindowsSandboxFlagValue(args, index) + if err != nil { + return WindowsSandboxSetupConfig{}, err + } + switch strings.TrimSpace(value) { + case "1": + config.PrincipalOptIn = true + case "0": + config.PrincipalOptIn = false + default: + // Refused rather than treated as off: a value this helper cannot read + // is a caller it does not understand, and guessing "no principal" + // there would provision a weaker sandbox than the caller asked for + // while reporting success. + return WindowsSandboxSetupConfig{}, fmt.Errorf("invalid --sandbox-principal %q, want 0 or 1", value) + } + index = next + case "--caller-sid": + value, next, err := nextWindowsSandboxFlagValue(args, index) + if err != nil { + return WindowsSandboxSetupConfig{}, err + } + config.CallerSID = strings.TrimSpace(value) + index = next default: return WindowsSandboxSetupConfig{}, fmt.Errorf("unknown windows sandbox setup flag %q", arg) } @@ -152,22 +348,49 @@ func RunWindowsSandboxSetup(args []string, stderr io.Writer) int { return runWindowsSandboxSetup(config, stderr) } +// commandConfig is the command-shaped view the setup half plans against. Its Env +// carries the opt-in the caller serialized into the setup args, so every +// downstream windowsSandboxIdentityEnabled call — the gate that decides whether +// elevated setup provisions a principal at all — reads the caller's intent +// rather than sampling the elevated helper's own environment, which UAC does not +// inherit from the shell the user typed in. func (config WindowsSandboxSetupConfig) commandConfig() WindowsSandboxCommandConfig { return WindowsSandboxCommandConfig{ - SandboxHome: config.SandboxHome, - CommandCWD: config.CommandCWD, - WorkspaceRoots: cloneStrings(config.WorkspaceRoots), + SandboxHome: config.SandboxHome, + CommandCWD: config.CommandCWD, + WorkspaceRoots: cloneStrings(config.WorkspaceRoots), + // Taken as given. This computes the hash BOTH halves compare, and it runs + // in whichever process is asking — including the command runner, whose TEMP + // points at the sandbox runtime temp. Deriving the candidates here made the + // answer depend on the caller's environment, which is the one thing a + // shared fingerprint cannot afford. WindowsSandboxProfileWithRuntimeRoots + // is applied by the callers whose TEMP is still the operator's. PermissionProfile: config.PermissionProfile, + Env: map[string]string{windowsSandboxIdentityEnv: windowsSandboxPrincipalOptInValue(config.PrincipalOptIn)}, SandboxLevel: WindowsSandboxLevelRestrictedToken, + // Carried through for the same reason as the opt-in above: every principal + // name, ledger and secret path the setup half plans is derived from the + // invoking user, and that is the caller's identity, not this process's. + CallerSID: config.CallerSID, } } +// WindowsSandboxSetupConfigFromCommand is how a command asks "was setup run for +// what I need?". It carries the command's own opt-in so marker validation can +// compare it against what setup actually provisioned. func WindowsSandboxSetupConfigFromCommand(config WindowsSandboxCommandConfig) WindowsSandboxSetupConfig { return WindowsSandboxSetupConfig{ - SandboxHome: config.SandboxHome, - CommandCWD: config.CommandCWD, - WorkspaceRoots: cloneStrings(config.WorkspaceRoots), + SandboxHome: config.SandboxHome, + CommandCWD: config.CommandCWD, + WorkspaceRoots: cloneStrings(config.WorkspaceRoots), + // NOT augmented here. This runs inside the command runner, whose TEMP and + // TMP are the sandbox runtime's, so deriving the temp-based candidate here + // yields a path no other process agrees on. The parent already folded the + // candidates into the profile it serialized, back when TEMP was still the + // operator's, so take the profile as given. PermissionProfile: config.PermissionProfile, + PrincipalOptIn: windowsSandboxIdentityEnabled(config.Env), + CallerSID: config.CallerSID, } } @@ -195,16 +418,63 @@ func BuildWindowsSandboxSetupMarker(config WindowsSandboxSetupConfig) (WindowsSa if len(infraPlan.IdentitySIDs) > 0 { offlineSID = infraPlan.IdentitySIDs[0] } + principalHash, err := windowsPrincipalPlanFingerprint(config) + if err != nil { + return WindowsSandboxSetupMarker{}, err + } return WindowsSandboxSetupMarker{ - SchemaVersion: windowsSandboxSetupMarkerSchemaVersion, - ACLPlanHash: hash, - ACLPlanEntries: len(plan.Entries), - NetworkInfraHash: infraHash, - OfflineFilterSID: offlineSID, - NetworkFilters: len(infraPlan.Filters), + SchemaVersion: windowsSandboxSetupMarkerSchemaVersion, + ACLPlanHash: hash, + ACLPlanEntries: len(plan.Entries), + NetworkInfraHash: infraHash, + OfflineFilterSID: offlineSID, + NetworkFilters: len(infraPlan.Filters), + PrincipalOptIn: config.PrincipalOptIn, + PrincipalPlanHash: principalHash, }, nil } +// windowsPrincipalPlanFingerprint hashes the principal ACL plan so a change to +// principal read or write roots invalidates setup. +// +// The SID is a fixed placeholder rather than the real principal's, deliberately. +// The account is recreated with a fresh SID on reprovision, so hashing the real +// one would make the fingerprint change every time the account is rebuilt even +// though the GRANTED PATHS are identical, and every command would then rerun +// setup. What must invalidate the marker is the set of paths and actions, which +// is exactly what this captures. +// +// Returns empty when the principal backend is not opted into, so the default +// install's marker is unchanged. +func windowsPrincipalPlanFingerprint(config WindowsSandboxSetupConfig) (string, error) { + if !config.PrincipalOptIn { + return "", nil + } + filesystem := config.commandConfig().PermissionProfile.FileSystem + // EVERY FIELD apply and teardown pass, or the fingerprint describes a + // different plan than the one that gets applied. This omitted DenyWrite while + // both buildWindowsPrincipalACLPlan call sites in windows_identity_runtime_windows.go + // passed it, so a change to the policy's deny-write paths moved the applied + // plan and left this hash where it was. The capability ACLPlanHash happens to + // cover the same paths today, which is what kept it from being a live hole + // and also what would have kept it invisible. + plan, err := buildWindowsPrincipalACLPlan(windowsPrincipalACLInput{ + PrincipalSID: windowsPrincipalFingerprintSID, + WriteRoots: filesystem.WriteRoots, + ReadRoots: filesystem.ReadRoots, + DenyRead: filesystem.DenyRead, + DenyWrite: filesystem.DenyWrite, + }) + if err != nil { + return "", fmt.Errorf("fingerprint windows principal ACL plan: %w", err) + } + return WindowsACLPlanHash(plan) +} + +// windowsPrincipalFingerprintSID is a placeholder trustee used only for hashing. +// It never reaches an ACE. +const windowsPrincipalFingerprintSID = "S-1-0-0" + func WriteWindowsSandboxSetupMarker(config WindowsSandboxSetupConfig) (WindowsSandboxSetupMarker, error) { marker, err := BuildWindowsSandboxSetupMarker(config) if err != nil { @@ -259,8 +529,56 @@ func ValidateWindowsSandboxSetupMarker(config WindowsSandboxSetupConfig) error { if actual.SchemaVersion != expected.SchemaVersion { return fmt.Errorf("windows sandbox setup is out of date: schema %d, want %d", actual.SchemaVersion, expected.SchemaVersion) } + // The two halves of the protocol run in different processes, so they can + // disagree about the principal opt-in. Refuse the command rather than pick a + // winner. + // + // The direction that matters is the first one: the opt-in is on, setup never + // provisioned an account, and the runtime's lookup declines with a nil error — + // so without this the command runs on the restricted token, which does not + // confine reads, while the operator believes a principal is isolating them. + // A sandbox that is weaker than advertised has to be loud. + // + // The reverse is refused too. It is not the dangerous direction — the command + // gets the well-worn restricted token it asked for — but setup did create a + // local account and grant it ACEs on the workspace, and letting commands run + // as if that had not happened leaves nothing to reconcile it. Both directions + // clear the same way: run `zero sandbox setup` again with the environment you + // actually want. + if actual.PrincipalOptIn != expected.PrincipalOptIn { + if expected.PrincipalOptIn { + return fmt.Errorf("windows sandbox setup is out of date: %s=1 asks for a sandbox principal, but setup provisioned none — "+ + "re-run `zero sandbox setup` from an elevated (Administrator) terminal with %s=1, or unset it to use the restricted-token sandbox", + windowsSandboxIdentityEnv, windowsSandboxIdentityEnv) + } + // "principals", plural: a workspace gets an offline and an online account, + // and the opt-out path retires the legacy one alongside them. The singular + // described the pre-split world and understated what the re-run does. + return fmt.Errorf("windows sandbox setup is out of date: setup provisioned sandbox principals, but %s is not set for this command — "+ + "set %s=1, or re-run `zero sandbox setup` from an elevated (Administrator) terminal without it to retire them", + windowsSandboxIdentityEnv, windowsSandboxIdentityEnv) + } if actual.ACLPlanHash != expected.ACLPlanHash || actual.ACLPlanEntries != expected.ACLPlanEntries { - return errors.New("windows sandbox setup is out of date: permission roots or deny lists changed") + // Say WHAT disagrees. The bare "permission roots or deny lists changed" + // left an operator with a command that refuses to run and no way to tell + // which side is wrong, or even whether the marker belongs to this + // workspace at all. The entry counts separate the two shapes this takes: + // equal counts mean the same roots spelled differently, unequal counts + // mean one side has roots the other has never heard of. + return fmt.Errorf("windows sandbox setup is out of date: permission roots or deny lists changed "+ + "(marker %s has %d entries, hash %s; this command expects %d entries, hash %s) — "+ + "re-run `zero sandbox setup` from an elevated (Administrator) terminal", + path, actual.ACLPlanEntries, shortWindowsACLPlanHash(actual.ACLPlanHash), + expected.ACLPlanEntries, shortWindowsACLPlanHash(expected.ACLPlanHash)) + } + // The capability-SID plan above and the principal plan are built separately + // from the same profile, so the hash above does not cover principal grants. + // Without this check, removing a principal read root left setup looking + // current and the stale AllowRead ACE in place, which is the opposite of + // what narrowing a policy is supposed to do. + if actual.PrincipalPlanHash != expected.PrincipalPlanHash { + return errors.New("windows sandbox setup is out of date: sandbox principal grants changed — " + + "re-run `zero sandbox setup` from an elevated (Administrator) terminal so the old grants are revoked") } // Mode-agnostic: validate the provisioned infrastructure, never the // per-command network mode — so an approved (allow) network command and an @@ -310,3 +628,266 @@ func canonicalWindowsACLEntries(entries []WindowsACLEntry) []WindowsACLEntry { }) return out } + +// windowsSandboxRuntimeCandidates returns every runtime root setup provisions. +// +// BOTH candidates, not the one this process would select. sandboxRuntimeRootFor +// prefers the cache-derived root and falls back to the temp-derived one when the +// first would land inside the workspace or its lease cannot be taken, and that +// choice is made per process. Setup that granted only its own choice left the +// other unprovisioned, so a command that fell back wrote to a tree with no ACE +// on it. Both are deterministic now, so setup can cover both and command +// selection lands on a provisioned root either way. +func windowsSandboxRuntimeCandidates(workspaceRoots []string) []string { + workspaceRoot := "" + for _, candidate := range workspaceRoots { + if trimmed := strings.TrimSpace(candidate); trimmed != "" { + workspaceRoot = canonicalSandboxWorkspaceRoot(trimmed) + break + } + } + if workspaceRoot == "" || workspaceRoot == "." { + return nil + } + var roots []string + if cacheRoot, err := sandboxUserCacheDir(); err == nil { + if cacheRoot = canonicalSandboxWorkspaceRoot(cacheRoot); cacheRoot != "" && cacheRoot != "." { + if root, ok := deterministicSandboxRuntimeRoot(workspaceRoot, cacheRoot); ok { + roots = append(roots, root) + } + } + } + if root, err := fallbackSandboxRuntimeRoot(workspaceRoot); err == nil { + roots = append(roots, root) + } + return roots +} + +// windowsSandboxProfileWithRuntime adds the runtime candidates as write roots. +// +// Applied on BOTH sides of the setup protocol, which is the whole point. The +// marker fingerprints the capability ACL plan built from this profile, while +// every command reaches the Windows runner having already had +// permissionProfileWithRuntime append the root it selected. Setup fingerprinted +// the bare profile and the command presented an augmented one, so a marker +// written seconds earlier was rejected with "permission roots or deny lists +// changed" and no command could run at all. +// +// Adding the full candidate set on both sides makes the two hashes agree without +// the command having to know which root setup happened to pick, and it puts the +// runtime roots into the CAPABILITY plan as well. That second part matters since +// the principal command runs on a WRITE_RESTRICTED token restricted to the +// capability SIDs: a runtime root carrying only the principal ACE satisfies the +// normal token and fails the restricted check, so cache and temp writes were +// denied even once the marker agreed. +func windowsSandboxProfileWithRuntime(profile PermissionProfile, workspaceRoots []string) PermissionProfile { + candidates := windowsSandboxRuntimeCandidates(workspaceRoots) + if len(candidates) == 0 { + return profile + } + existing := make(map[string]struct{}, len(profile.FileSystem.WriteRoots)) + for _, root := range profile.FileSystem.WriteRoots { + existing[windowsCapabilityPathKey(root.Root)] = struct{}{} + } + writeRoots := append([]WritableRoot{}, profile.FileSystem.WriteRoots...) + for _, candidate := range candidates { + if _, ok := existing[windowsCapabilityPathKey(candidate)]; ok { + continue + } + existing[windowsCapabilityPathKey(candidate)] = struct{}{} + writeRoots = append(writeRoots, WritableRoot{Root: candidate}) + } + profile.FileSystem.WriteRoots = writeRoots + return profile +} + +// ensureWindowsSandboxRuntimeCandidates creates every runtime root setup grants. +// +// Paired with windowsSandboxProfileWithRuntime: that function puts the candidates +// into the ACL plan, and this one makes them exist. Splitting the two is what +// broke elevated setup once already, because the capability plan refuses to +// materialize a write root and fails the whole run on a path that is merely +// absent. Whenever one of these grows a candidate, so must the other. +// +// Called on the setup side only. A command must never create these: setup is the +// gate that decides which trees the sandbox may write to, and a command that +// created its own root would be granting itself one. +func ensureWindowsSandboxRuntimeCandidates(workspaceRoots []string) error { + for _, root := range windowsSandboxRuntimeCandidates(workspaceRoots) { + if err := ensureRuntimeCandidateDir(root); err != nil { + return fmt.Errorf("create sandbox runtime root %s: %w", root, err) + } + } + return nil +} + +// ensureRuntimeCandidateDir creates one candidate without following a reparse +// point into anything Zero does not own. +// +// VALIDATION BEFORE THE PRIVILEGED MUTATION, NOT AFTER IT. This runs as +// Administrator, and os.MkdirAll follows links: the invoking user can prepare +// either candidate's ancestry, so a junction planted at a not-yet-created +// component below the cache directory, or at the fallback anchor, made elevated +// setup create the tree beneath a redirected, Administrator-writable target. The +// ACL applier's own no-follow check runs later and correctly spots the +// redirection, but by then the privileged create has already happened, outside +// the applier's rollback boundary. A leaf check after MkdirAll would leave the +// same gap. +// +// EnsurePrivateDir is the descent that does not have it: handle-relative, +// no-follow, refusing any component that is a link and any leaf this user does +// not own. It is given a PHYSICAL base first, because it walks from the volume +// root and a redirected cache or TEMP above the owned tail is the operator's +// business, not ours -- the same pairing the fallback anchor uses, and the same +// mistake that refused every fallback on macOS when the parent was left +// unresolved. +// ensureMissingWindowsSandboxRuntimeCandidates creates, with the same ownership +// and no-follow checks setup uses, every runtime candidate that does not exist +// yet, and leaves an existing one exactly as it is. The unelevated tier calls +// this on every command, and an existing candidate carries the capability ACE +// the previous command applied: re-running EnsurePrivateDir on it would strip +// that grant, fail the marker's capability verification, and reapply the plan +// on every launch. An existing candidate is still validated by the plan apply, +// which opens it without following a reparse point. +func ensureMissingWindowsSandboxRuntimeCandidates(workspaceRoots []string) error { + for _, root := range windowsSandboxRuntimeCandidates(workspaceRoots) { + info, err := os.Lstat(root) + switch { + case err == nil: + // Left alone, but not unexamined: a file or a reparse point at a + // candidate's name is not a runtime root, and the apply would happily + // put an ACE on either. Setup's full validation refuses both shapes + // too; this is the same answer without the DACL rewrite. + if !info.IsDir() || info.Mode()&os.ModeSymlink != 0 || info.Mode()&os.ModeIrregular != 0 { + return fmt.Errorf("sandbox runtime root %s exists but is not a plain directory", root) + } + continue + case errors.Is(err, os.ErrNotExist): + if err := ensureRuntimeCandidateDir(root); err != nil { + return fmt.Errorf("create sandbox runtime root %s: %w", root, err) + } + default: + return fmt.Errorf("inspect sandbox runtime root %s: %w", root, err) + } + } + return nil +} + +func ensureRuntimeCandidateDir(root string) error { + base, ok := runtimeCandidateBase(root) + if !ok { + return fmt.Errorf("candidate %s does not sit beneath a known base, so its trust boundary cannot be established", root) + } + tail, err := filepath.Rel(base, root) + if err != nil { + return fmt.Errorf("locate %s beneath %s: %w", root, base, err) + } + // A resolve failure keeps the unresolved base deliberately: EnsurePrivateDir + // is still the fail-closed check, and refusing there names the real component + // rather than hiding it behind a resolver error. + physical := base + if resolved, resolveErr := physicalDir(base); resolveErr == nil && resolved != "" { + physical = resolved + } + return peermsg.EnsurePrivateDir(filepath.Join(physical, tail)) +} + +// runtimeCandidateBase returns the ancestor of a candidate that belongs to the +// operator rather than to Zero. Everything below it was created by us and has +// no business being a link. +func runtimeCandidateBase(root string) (string, bool) { + // The fallback candidate is /zero-runtime-/v1/, and the + // anchor's own parent is the operator's temp directory. + if isFallbackRuntimeRoot(root) { + return filepath.Dir(fallbackRuntimeAnchor()), true + } + // The cache candidate is /zero/runtime/v1/. + if cacheRoot, err := sandboxUserCacheDir(); err == nil { + if cleaned := canonicalSandboxWorkspaceRoot(cacheRoot); cleaned != "" && cleaned != "." && pathWithinRoot(cleaned, root) { + return cleaned, true + } + } + return "", false +} + +// shortWindowsACLPlanHash trims a plan hash for a human-facing error. Twelve hex +// characters is plenty to tell two plans apart by eye, and the full 64 buries the +// rest of the message. +func shortWindowsACLPlanHash(hash string) string { + hash = strings.TrimSpace(hash) + if hash == "" { + return "(none)" + } + if len(hash) > 12 { + return hash[:12] + } + return hash +} + +// WindowsSandboxProfileWithRuntimeRoots folds the sandbox runtime roots into a +// permission profile, for callers that build a setup config outside this package. +// +// Call it ONLY from a process whose TEMP and TMP are the operator's. The +// temp-derived candidate reads os.TempDir(), and the sandbox points those +// variables at its own runtime temp for anything it launches, so a process on the +// far side of that redirection derives a path no other process agrees on. The +// command runner is exactly such a process: it takes the profile it is handed. +func WindowsSandboxProfileWithRuntimeRoots(profile PermissionProfile, workspaceRoots []string) PermissionProfile { + return windowsSandboxProfileWithRuntime(profile, workspaceRoots) +} + +// WindowsSandboxPrincipalRoleForNetwork names the principal a command in this +// network mode runs as. +// +// Exported and portable so `zero doctor` and the Windows runtime answer from ONE +// rule. The helper this replaces existed for that reason and said so, and the +// thing it was protecting against happened anyway once dual-role provisioning +// changed the behaviour under NetworkDeny: the runtime started using an offline +// principal while doctor still reported the old restricted-token standdown, so +// operators were told their reads were unconfined when they were confined. +// windowsSandboxRoleForNetwork delegates here rather than repeating the switch. +// +// Anything that is not an EXACT allow is offline: a mode this does not recognise +// should lose the network, not keep it. +// +// Deliberately not normalized first. NormalizeNetworkMode case-folds, so routing +// through it made "ALLOW" select the online principal where the runtime required +// an exact match and failed closed to offline. Sharing one rule is only an +// improvement if it shares the STRICTER one; a shared rule that quietly widened +// the runtime would be worse than the drift it fixes. +func WindowsSandboxPrincipalRoleForNetwork(network NetworkMode) string { + if network == NetworkAllow { + return "online" + } + return "offline" +} + +// errWindowsPrincipalLaunchUnavailable is why provisioning refuses. +// +// Stated as one sentence about the MECHANISM rather than about the caller's +// token, because the caller's token was never the question: the process that +// needs the launch privileges is a later, ordinary one, and no supported path +// carries launch authority to it. A message about the current process invited +// the operator to re-run from an elevated terminal, which passes such a check +// and still cannot run a command. +var errWindowsPrincipalLaunchUnavailable = errors.New( + "launching a command as a separate account needs SeAssignPrimaryTokenPrivilege and SeIncreaseQuotaPrivilege, " + + "which an ordinary command process does not hold; elevated setup cannot grant them to it, and no bootstrap " + + "or broker exists yet to carry that authority across. The restricted-token sandbox is unaffected: unset " + + windowsSandboxIdentityEnv + " to use it", +) + +// windowsPrincipalLaunchAvailable reports whether a provisioned principal could +// actually be launched by the process that will run the command. +// +// ONE SEAM, DEFAULTING TO UNAVAILABLE. The answer does not depend on the token +// of whoever calls setup, so this takes no argument: it is a statement about +// which mechanisms exist, not about the current process. When a bootstrap or +// broker lands, this is the single place that learns to say yes, and the plans, +// ledger and ACL work below it are already built and tested against it. +// +// It stays a var so the argument-plumbing tests can exercise the opt-in path +// without depending on a capability no machine currently has. +var windowsPrincipalLaunchAvailable = func() error { + return errWindowsPrincipalLaunchUnavailable +} diff --git a/internal/sandbox/windows_setup_caller_windows_test.go b/internal/sandbox/windows_setup_caller_windows_test.go new file mode 100644 index 000000000..31cb6816d --- /dev/null +++ b/internal/sandbox/windows_setup_caller_windows_test.go @@ -0,0 +1,326 @@ +//go:build windows + +package sandbox + +import ( + "bytes" + "errors" + "os" + "strings" + "testing" +) + +// windowsSetupSeams stubs every external effect of the setup transaction so a +// test can drive its control flow on an ordinary machine. Each seam defaults to +// a benign success, and a test overrides only the one it is about. +type windowsSetupSeams struct { + elevated bool + // principalStillInstalled is what the account lookup reports AFTER retirement + // ran. It is the whole of the opt-out branch's fatality rule, so it is stated + // separately from retireErr rather than inferred from it. + principalStillInstalled bool + aclRollback func() error + retireErr error + provisionErr error + networkErr error + markerErr error + rollbackCalled *bool + markerWritten *bool +} + +func withWindowsSetupSeams(t *testing.T, seams windowsSetupSeams) { + t.Helper() + + originalElevated := windowsProcessIsElevatedFn + originalLock := acquireWindowsSandboxSetupLockFn + originalACL := applyWindowsACLPlanFn + originalPrincipal := setupWindowsSandboxPrincipalFn + originalRetire := removeWindowsSandboxPrincipalsForSetupFn + originalGroupSID := resolveWindowsSandboxOfflineGroupSIDHook + originalLookup := lookupWindowsSandboxIdentityFn + originalNetwork := applyWindowsNetworkPlanFn + originalMarker := writeWindowsSandboxSetupMarkerFn + t.Cleanup(func() { + lookupWindowsSandboxIdentityFn = originalLookup + windowsProcessIsElevatedFn = originalElevated + acquireWindowsSandboxSetupLockFn = originalLock + applyWindowsACLPlanFn = originalACL + setupWindowsSandboxPrincipalFn = originalPrincipal + removeWindowsSandboxPrincipalsForSetupFn = originalRetire + resolveWindowsSandboxOfflineGroupSIDHook = originalGroupSID + applyWindowsNetworkPlanFn = originalNetwork + writeWindowsSandboxSetupMarkerFn = originalMarker + }) + + // The launch seam says no on every machine today; these fixtures test the + // provisioning flow behind it, so they open it and restore it after. + originalLaunch := windowsPrincipalLaunchAvailable + windowsPrincipalLaunchAvailable = func() error { return nil } + t.Cleanup(func() { windowsPrincipalLaunchAvailable = originalLaunch }) + windowsProcessIsElevatedFn = func() bool { return seams.elevated } + // A zero lock is safe to release: release() tolerates a nil handle and + // runtime.UnlockOSThread is a no-op when the thread was never pinned. + acquireWindowsSandboxSetupLockFn = func(string) (*windowsSandboxSetupLock, error) { + return &windowsSandboxSetupLock{}, nil + } + applyWindowsACLPlanFn = func(WindowsACLPlan) (func() error, error) { + return func() error { + if seams.rollbackCalled != nil { + *seams.rollbackCalled = true + } + if seams.aclRollback != nil { + return seams.aclRollback() + } + return nil + }, nil + } + setupWindowsSandboxPrincipalFn = func(WindowsSandboxCommandConfig) (func() error, error) { + if seams.provisionErr != nil { + return nil, seams.provisionErr + } + return func() error { return nil }, nil + } + removeWindowsSandboxPrincipalsForSetupFn = func(WindowsSandboxCommandConfig) error { + return seams.retireErr + } + // Dual-role setup asserts the block filters name the offline group, which a + // stubbed provisioning never creates. + resolveWindowsSandboxOfflineGroupSIDHook = func() (string, error) { return "S-1-5-32-999", nil } + lookupWindowsSandboxIdentityFn = func(string, windowsSandboxRole) (windowsSandboxIdentity, error) { + if seams.principalStillInstalled { + return windowsSandboxIdentity{Username: "zero-sbx-stub"}, nil + } + return windowsSandboxIdentity{}, errWindowsSandboxIdentityUnavailable + } + applyWindowsNetworkPlanFn = func(WindowsNetworkPlan) error { return seams.networkErr } + writeWindowsSandboxSetupMarkerFn = func(config WindowsSandboxSetupConfig) (WindowsSandboxSetupMarker, error) { + if seams.markerWritten != nil { + *seams.markerWritten = true + } + if seams.markerErr != nil { + return WindowsSandboxSetupMarker{}, seams.markerErr + } + return BuildWindowsSandboxSetupMarker(config) + } +} + +func windowsSetupTestConfig(t *testing.T, optIn bool) WindowsSandboxSetupConfig { + t.Helper() + root := t.TempDir() + return WindowsSandboxSetupConfig{ + SandboxHome: t.TempDir(), + CommandCWD: root, + WorkspaceRoots: []string{root}, + PermissionProfile: PermissionProfile{ + FileSystem: FileSystemPolicy{Kind: FileSystemRestricted, WriteRoots: []WritableRoot{{Root: root}}}, + Network: NetworkPolicy{Mode: NetworkDeny}, + }, + PrincipalOptIn: optIn, + } +} + +// windowsForeignSID is a well-formed account SID that no test machine's own +// process can be running as. +const windowsForeignSID = "S-1-5-21-1111111111-2222222222-3333333333-1001" + +// The finding: opting out printed the retirement failure and carried on to write +// an opted-out marker and exit 0. +// +// That is the one outcome this branch must never produce. Teardown is +// idempotent, so a failure here means a real account, secret, logon right, ACE +// set and ledger are still installed — and the marker now says none of them +// exist, so nothing afterwards will ever look for them again. +// +// Asserting on the marker as well as the exit code is deliberate: an exit code +// alone would still pass if the marker had already been written, and the marker +// is what makes the residue permanent. +func TestOptOutRetirementFailureFailsSetupAndWritesNoMarker(t *testing.T) { + markerWritten := false + rollbackCalled := false + withWindowsSetupSeams(t, windowsSetupSeams{ + elevated: true, + retireErr: errors.New("sandbox principal zero-sbx-abc could not be deleted"), + principalStillInstalled: true, + markerWritten: &markerWritten, + rollbackCalled: &rollbackCalled, + }) + + var stderr bytes.Buffer + config := windowsSetupTestConfig(t, false) + if code := runWindowsSandboxSetup(config, &stderr); code == 0 { + t.Fatal("setup reported success after failing to retire an existing sandbox principal, so the operator is told the machine is clean while the account is still installed") + } + if markerWritten { + t.Error("setup wrote an opted-out marker after the retirement failed, which is what makes the leftover principal permanently invisible") + } + if !rollbackCalled { + t.Error("setup failed without undoing the ACL grants it had already applied") + } + if !strings.Contains(stderr.String(), "could not be deleted") { + t.Errorf("the underlying retirement error must reach the operator, got %q", stderr.String()) + } + if _, err := os.Stat(WindowsSandboxSetupMarkerPath(config.SandboxHome)); !os.IsNotExist(err) { + t.Errorf("no marker file should exist on disk after a failed setup, stat error = %v", err) + } +} + +// The other side of that rule, and a regression I introduced fixing the first +// one: failing on ANY retirement error is wrong. +// +// removeWindowsSandboxPrincipalForSetup deliberately carries on past a secret +// file owned by another administrator's setup (errWindowsSandboxSecretNotOurs +// goes into secretErr and is reported at the end via errors.Join) so one +// unlinkable file cannot strand the account, its logon rights and its ACEs. The +// account is therefore already gone when that error is returned, and treating it +// as fatal made `zero sandbox setup` fail permanently on such a machine, taking +// the DEFAULT restricted-token sandbox down with it over inert residue. +// +// Setup must complete, and must still say what was left behind. +func TestOptOutSucceedsWhenOnlyInertResidueSurvivesRetirement(t *testing.T) { + markerWritten := false + rollbackCalled := false + withWindowsSetupSeams(t, windowsSetupSeams{ + elevated: true, + // What teardown reports when the only leftover is another administrator's + // secret file: the account, its rights, its ACEs and its ledger are gone. + retireErr: errWindowsSandboxSecretNotOurs, + principalStillInstalled: false, + markerWritten: &markerWritten, + rollbackCalled: &rollbackCalled, + }) + + var stderr bytes.Buffer + if code := runWindowsSandboxSetup(windowsSetupTestConfig(t, false), &stderr); code != 0 { + t.Fatalf("setup failed over residue left behind by an already-retired principal, which makes the default sandbox unsetupable on this machine: %s", stderr.String()) + } + if !markerWritten { + t.Error("setup that completed must write its marker") + } + if rollbackCalled { + t.Error("setup rolled back its ACL work over residue it had already reported as survivable") + } + if !strings.Contains(stderr.String(), "residue") { + t.Errorf("the leftover must still be named for whoever cleans it up, got %q", stderr.String()) + } +} + +// The opt-out path still succeeds when retirement succeeds. Without this the +// test above would pass against a branch that always failed. +func TestOptOutSucceedsWhenRetirementSucceeds(t *testing.T) { + markerWritten := false + withWindowsSetupSeams(t, windowsSetupSeams{elevated: true, markerWritten: &markerWritten}) + + var stderr bytes.Buffer + if code := runWindowsSandboxSetup(windowsSetupTestConfig(t, false), &stderr); code != 0 { + t.Fatalf("opting out with nothing to retire must succeed, got %d: %s", code, stderr.String()) + } + if !markerWritten { + t.Error("a successful setup must write its marker") + } +} + +// The other half of the caller-identity finding: a principal provisioned by an +// administrator who is not the caller can never be used, because its password is +// sealed to the user that stores it. Refuse before anything is created rather +// than leave an account, a secret and a grant nobody can reach. +func TestPrincipalSetupRefusesWhenElevatedAsAnotherUser(t *testing.T) { + markerWritten := false + withWindowsSetupSeams(t, windowsSetupSeams{elevated: true, markerWritten: &markerWritten}) + + config := windowsSetupTestConfig(t, true) + config.CallerSID = windowsForeignSID + if current := windowsCurrentUserSID(); current == "" || strings.EqualFold(current, config.CallerSID) { + t.Skipf("cannot distinguish the caller from this process (current SID %q)", current) + } + + var stderr bytes.Buffer + if code := runWindowsSandboxSetup(config, &stderr); code == 0 { + t.Fatal("setup provisioned a sandbox principal for a caller whose secret it cannot seal, so every later command would find the account and fail to unseal its password") + } + if markerWritten { + t.Error("setup wrote a marker for a principal it refused to provision") + } + if !strings.Contains(stderr.String(), "different Windows user") { + t.Errorf("the refusal must say what is wrong and how to clear it, got %q", stderr.String()) + } +} + +// The refusal is scoped to the principal opt-in. The default restricted-token +// sandbox stores no secret and names no account after the user, so it works +// perfectly well across this boundary and must keep doing so. +func TestRestrictedTokenSetupIsUnaffectedByACrossUserElevation(t *testing.T) { + withWindowsSetupSeams(t, windowsSetupSeams{elevated: true}) + + config := windowsSetupTestConfig(t, false) + config.CallerSID = windowsForeignSID + + var stderr bytes.Buffer + if code := runWindowsSandboxSetup(config, &stderr); code != 0 { + t.Fatalf("the restricted-token sandbox does not depend on the invoking user and must still set up, got %d: %s", code, stderr.String()) + } +} + +// An unknown caller SID is treated as a match. That is the pre-existing +// behaviour for an older caller or a token query that failed, and refusing on +// absence would break setup on machines where nothing is wrong. +func TestPrincipalSetupProceedsWhenTheCallerIsUnknown(t *testing.T) { + withWindowsSetupSeams(t, windowsSetupSeams{elevated: true}) + + var stderr bytes.Buffer + if code := runWindowsSandboxSetup(windowsSetupTestConfig(t, true), &stderr); code != 0 { + t.Fatalf("an unstated caller must not block setup, got %d: %s", code, stderr.String()) + } +} + +// The account name, its ledger and its ACEs are all derived from this key, and +// under over-the-shoulder UAC the elevated helper is a different user than the +// caller. Keying off the caller is what makes setup provision the account the +// caller's next command will actually look for. +func TestPrincipalKeyFollowsTheCallerNotTheCurrentProcess(t *testing.T) { + roots := []string{`C:\ws\shared`} + self := windowsSandboxPrincipalKey(WindowsSandboxCommandConfig{WorkspaceRoots: roots}) + caller := windowsSandboxPrincipalKey(WindowsSandboxCommandConfig{WorkspaceRoots: roots, CallerSID: windowsForeignSID}) + if self == caller { + t.Fatal("the principal key ignored the caller SID, so elevated setup would provision an account named after the administrator who typed the credentials") + } + if again := windowsSandboxPrincipalKey(WindowsSandboxCommandConfig{WorkspaceRoots: roots, CallerSID: windowsForeignSID}); again != caller { + t.Error("the caller-scoped key is not stable, so setup and the command half would derive different accounts") + } +} + +// THE HELPER REFUSES WHAT THE BUILDER REFUSES. An elevated same-caller run that +// hands the helper `--sandbox-principal 1` directly must stop before any +// account, secret or ACL effect while no launch path exists, exactly as +// BuildWindowsSandboxSetupArgs would have stopped it from being serialized. +func TestPrincipalSetupRefusesAtTheHelperWhileLaunchIsUnavailable(t *testing.T) { + withWindowsSetupSeams(t, windowsSetupSeams{elevated: true}) + windowsPrincipalLaunchAvailable = func() error { return errWindowsPrincipalLaunchUnavailable } + + provisioned := false + setupWindowsSandboxPrincipalFn = func(WindowsSandboxCommandConfig) (func() error, error) { + provisioned = true + return func() error { return nil }, nil + } + applied := false + applyWindowsACLPlanFn = func(WindowsACLPlan) (func() error, error) { + applied = true + return func() error { return nil }, nil + } + + var stderr bytes.Buffer + if code := runWindowsSandboxSetup(windowsSetupTestConfig(t, true), &stderr); code == 0 { + t.Fatalf("setup provisioned a principal that nothing can launch: %s", stderr.String()) + } + if !strings.Contains(stderr.String(), "refusing to provision a sandbox principal") { + t.Fatalf("refusal did not say why: %s", stderr.String()) + } + if provisioned || applied { + t.Fatalf("the refusal came after an effect: provisioned=%v aclApplied=%v", provisioned, applied) + } + + // CONTROL: the opt-out path does not consult launch availability at all. + stderr.Reset() + if code := runWindowsSandboxSetup(windowsSetupTestConfig(t, false), &stderr); code != 0 { + t.Fatalf("setup without a principal must not depend on launch availability, got %d: %s", code, stderr.String()) + } +} diff --git a/internal/sandbox/windows_setup_lock_windows.go b/internal/sandbox/windows_setup_lock_windows.go new file mode 100644 index 000000000..1cbce8b28 --- /dev/null +++ b/internal/sandbox/windows_setup_lock_windows.go @@ -0,0 +1,193 @@ +//go:build windows + +package sandbox + +import ( + "fmt" + "runtime" + "strings" + "time" + "unsafe" + + "golang.org/x/sys/windows" +) + +// Serializing elevated setup across processes. +// +// WHY THIS EXISTS. Setup is not one write, it is a transaction: it rotates the +// principal's password, stores the secret, read-modify-writes the ACL ledger, +// installs WFP filters and finally writes the marker. Two setups for the same +// workspace running at once interleave those steps, and the interleavings are +// not merely untidy. The account ends up with process B's password while the +// stored secret is process A's, so every later command fails to log on. The +// ledger's read-modify-write loses one process's path set, so ACEs that should +// have been revoked stay on disk with nothing recording them. +// +// Atomic individual writes do not help. Each step being all-or-nothing is +// orthogonal to two processes interleaving BETWEEN steps, which is the actual +// failure. The lock has to span the whole transaction. +// +// Worth recording: the reference implementation this was compared against has +// the identical bug. Its password rotation and secret write hold no lock at +// all, its named mutex guards only an unrelated read-ACL-only mode, and the +// singleflight that looks like a fix is process-local and gives no cross-process +// guarantee. There was nothing to port, so this is built. + +// windowsSetupLockTimeout bounds the wait. Setup is interactive and elevated, +// so a minute is generous for a legitimate concurrent run and short enough that +// a stuck holder gets reported rather than hung on. A var rather than a const +// only so a contention test does not have to wait a real minute to prove the +// two setups exclude each other. +var windowsSetupLockTimeout = 60 * time.Second + +// windowsSetupLockSDDL grants full control to Administrators and SYSTEM and +// nobody else, with the DACL protected so no inherited ACE widens it. +// +// Load-bearing rather than tidy. A sandbox principal able to open this object +// could hold it and stall every future setup, or create it first and squat the +// name so real setup believes it holds a lock it does not. Only accounts that +// could already run setup may touch it. +// +// A var so a test can exercise the mutex wait path without elevation. Setup +// itself always runs elevated, so unelevated callers cannot even open the +// object under this DACL, which would make the contention test unrunnable +// anywhere it matters. TestSetupLockIsAdministratorsOnly pins the production +// value so loosening it needs an explicit change here. +var windowsSetupLockSDDL = "D:P(A;;GA;;;BA)(A;;GA;;;SY)" + +const ( + // windowsWaitTimeout is WAIT_TIMEOUT. x/sys exports WAIT_ABANDONED, + // WAIT_OBJECT_0 and WAIT_FAILED at this version but not this one. + windowsWaitTimeout = 0x00000102 + + // windowsSetupLockKeyChars keeps the object name short while staying + // collision-free in practice: the key is already a hash, so a prefix of it + // distinguishes workspaces. + windowsSetupLockKeyChars = 32 +) + +// windowsSandboxSetupLock is a held cross-process setup lock. Release it once. +type windowsSandboxSetupLock struct { + handle windows.Handle + released bool + // abandoned records that the previous holder died without releasing, which + // means it stopped somewhere inside the transaction. + abandoned bool +} + +// windowsSandboxSetupLockName derives the object name for a workspace. +// +// Per-workspace rather than machine-wide because every artifact the lock +// protects is per-workspace: this account, this secret, this ledger. A global +// lock would serialize unrelated workspaces for no reason. +// +// Global rather than Local because the protected state is not session-scoped. +// The local account, its LSA logon rights and the WFP filters are machine +// state, so two setups in different logon sessions must still exclude each +// other. Creating a Global object needs SeCreateGlobalPrivilege, which the +// elevated setup helper has and which is the only caller. +func windowsSandboxSetupLockName(workspaceKey string) string { + key := strings.Map(func(r rune) rune { + switch { + case r >= 'a' && r <= 'z', r >= '0' && r <= '9': + return r + case r >= 'A' && r <= 'Z': + return r + ('a' - 'A') + default: + return -1 + } + }, workspaceKey) + if key == "" { + key = "default" + } + if len(key) > windowsSetupLockKeyChars { + key = key[:windowsSetupLockKeyChars] + } + return `Global\ZeroSandboxSetup-` + key +} + +// acquireWindowsSandboxSetupLock blocks until this process owns the setup lock +// for a workspace, or fails with an actionable error. +// +// The OS thread is pinned for the lock's lifetime. Win32 mutex ownership is +// thread-affine, so a goroutine that migrated between the wait and the release +// would call ReleaseMutex from a thread that does not own the object, leaving +// the mutex held until the process exits. That is a deadlock for every later +// setup on the machine, so the pinning is not optional. +func acquireWindowsSandboxSetupLock(workspaceKey string) (*windowsSandboxSetupLock, error) { + descriptor, err := windows.SecurityDescriptorFromString(windowsSetupLockSDDL) + if err != nil { + return nil, fmt.Errorf("build the sandbox setup lock security descriptor: %w", err) + } + attributes := windows.SecurityAttributes{SecurityDescriptor: descriptor} + attributes.Length = uint32(unsafe.Sizeof(attributes)) + + name := windowsSandboxSetupLockName(workspaceKey) + namePtr, err := windows.UTF16PtrFromString(name) + if err != nil { + return nil, fmt.Errorf("encode the sandbox setup lock name: %w", err) + } + // initialOwner false, then wait explicitly. Asking for ownership at creation + // would take the lock only when this process created the object, so the + // create and open paths would differ and only one of them would ever wait. + handle, err := windows.CreateMutex(&attributes, false, namePtr) + if handle == 0 { + return nil, fmt.Errorf("open the sandbox setup lock %s: %w", name, err) + } + runtime.LockOSThread() + + lock := &windowsSandboxSetupLock{handle: handle} + state, waitErr := windows.WaitForSingleObject(handle, uint32(windowsSetupLockTimeout/time.Millisecond)) + switch state { + case windows.WAIT_OBJECT_0: + return lock, nil + case windows.WAIT_ABANDONED: + // The previous holder died inside the transaction. We own the mutex now. + // Proceeding is right rather than refusing: setup is idempotent and + // re-running it is exactly how a half-finished transaction is repaired, + // whereas refusing would leave the machine stuck in that half state with + // no way forward. + lock.abandoned = true + return lock, nil + case windowsWaitTimeout: + lock.releaseThreadAndHandle() + return nil, fmt.Errorf("another `zero sandbox setup` is already running for this workspace and did not finish within %s; "+ + "wait for it to finish, or look for a stuck elevated setup process before re-running", windowsSetupLockTimeout) + default: + lock.releaseThreadAndHandle() + if waitErr != nil { + return nil, fmt.Errorf("wait for the sandbox setup lock %s: %w", name, waitErr) + } + return nil, fmt.Errorf("wait for the sandbox setup lock %s returned an unexpected state %#x", name, state) + } +} + +// Abandoned reports that the previous holder crashed mid-transaction, so this +// run is repairing a half-finished setup rather than starting a clean one. +func (lock *windowsSandboxSetupLock) Abandoned() bool { + return lock != nil && lock.abandoned +} + +// release drops ownership and unpins the thread. Safe to call twice, so a defer +// and an explicit call cannot double-release. +func (lock *windowsSandboxSetupLock) release() { + if lock == nil || lock.released { + return + } + lock.released = true + // Release before closing. Closing a still-held mutex leaves it abandoned for + // the next waiter, which is recoverable but reports a crash that never + // happened. + _ = windows.ReleaseMutex(lock.handle) + lock.releaseThreadAndHandle() +} + +// releaseThreadAndHandle unpins the thread and drops the handle without +// releasing ownership, for the paths that never acquired it. +func (lock *windowsSandboxSetupLock) releaseThreadAndHandle() { + if lock.handle != 0 { + _ = windows.CloseHandle(lock.handle) + lock.handle = 0 + } + runtime.UnlockOSThread() +} diff --git a/internal/sandbox/windows_setup_lock_windows_test.go b/internal/sandbox/windows_setup_lock_windows_test.go new file mode 100644 index 000000000..89d7fe6fd --- /dev/null +++ b/internal/sandbox/windows_setup_lock_windows_test.go @@ -0,0 +1,214 @@ +//go:build windows + +package sandbox + +import ( + "fmt" + "strings" + "testing" + "time" +) + +// A unique key per test run, so a leftover object from an earlier run cannot +// make these pass or fail for the wrong reason. +func lockTestKey(t *testing.T) string { + t.Helper() + return fmt.Sprintf("t%dx%d", time.Now().UnixNano(), len(t.Name())) +} + +// The production DACL is Administrators only, so an unelevated test process +// cannot OPEN an existing lock and never reaches the mutex wait at all. These +// tests are about the wait, so they relax the DACL for themselves. +// TestSetupLockIsAdministratorsOnly pins the real value. +func withOpenableLock(t *testing.T) { + t.Helper() + previous := windowsSetupLockSDDL + windowsSetupLockSDDL = "D:P(A;;GA;;;WD)" + t.Cleanup(func() { windowsSetupLockSDDL = previous }) +} + +func withShortLockTimeout(t *testing.T, d time.Duration) { + t.Helper() + previous := windowsSetupLockTimeout + windowsSetupLockTimeout = d + t.Cleanup(func() { windowsSetupLockTimeout = previous }) +} + +// THE POINT OF THE WHOLE FILE. A second setup for the same workspace must not +// get in while the first holds the lock. +// +// Without this, two elevated setups interleave: the account ends up with one +// process's password while the stored secret is the other's, and every later +// command fails to log on. The second acquisition runs on its own goroutine +// because Win32 mutex ownership is per THREAD and acquire pins the thread, so a +// same-thread re-entry would succeed recursively and prove nothing. +func TestSecondSetupIsExcludedWhileTheFirstHoldsTheLock(t *testing.T) { + withShortLockTimeout(t, 300*time.Millisecond) + withOpenableLock(t) + key := lockTestKey(t) + + first, err := acquireWindowsSandboxSetupLock(key) + if err != nil { + t.Fatalf("first acquisition: %v", err) + } + defer first.release() + + result := make(chan error, 1) + go func() { + second, err := acquireWindowsSandboxSetupLock(key) + if err == nil { + second.release() + } + result <- err + }() + + select { + case err := <-result: + if err == nil { + t.Fatal("a second setup acquired the lock while the first still held it, so the transaction is not serialized") + } + // It has to say what to do, since the operator's next question is always + // whether to wait or to go looking for a stuck process. + if !strings.Contains(err.Error(), "already running") { + t.Errorf("contention error does not explain itself: %v", err) + } + case <-time.After(10 * time.Second): + t.Fatal("the second acquisition neither succeeded nor timed out, so the bounded wait is not bounded") + } +} + +// Releasing must actually hand the lock on. A lock that excludes forever is a +// worse failure than one that never excluded: the machine could never be set up +// again without a reboot. +func TestReleasingTheLockLetsTheNextSetupIn(t *testing.T) { + withShortLockTimeout(t, 5*time.Second) + withOpenableLock(t) + key := lockTestKey(t) + + first, err := acquireWindowsSandboxSetupLock(key) + if err != nil { + t.Fatalf("first acquisition: %v", err) + } + first.release() + + result := make(chan error, 1) + go func() { + second, err := acquireWindowsSandboxSetupLock(key) + if err == nil { + second.release() + } + result <- err + }() + + select { + case err := <-result: + if err != nil { + t.Fatalf("the lock was not released, so no later setup could ever run: %v", err) + } + case <-time.After(15 * time.Second): + t.Fatal("acquiring a released lock hung") + } +} + +// Different workspaces must not block each other. A machine-wide lock would +// serialize unrelated setups for no reason. +func TestDifferentWorkspacesDoNotExcludeEachOther(t *testing.T) { + withShortLockTimeout(t, 5*time.Second) + withOpenableLock(t) + base := lockTestKey(t) + + first, err := acquireWindowsSandboxSetupLock(base + "alpha") + if err != nil { + t.Fatalf("first workspace: %v", err) + } + defer first.release() + + result := make(chan error, 1) + go func() { + second, err := acquireWindowsSandboxSetupLock(base + "beta") + if err == nil { + second.release() + } + result <- err + }() + + select { + case err := <-result: + if err != nil { + t.Fatalf("a second workspace was blocked by an unrelated one: %v", err) + } + case <-time.After(15 * time.Second): + t.Fatal("a second workspace hung behind an unrelated lock") + } +} + +// Release must be safe to call twice, since it is both deferred and reachable +// explicitly. A double ReleaseMutex would fail and, worse, a double +// UnlockOSThread panics the runtime. +func TestReleasingTwiceIsSafe(t *testing.T) { + withShortLockTimeout(t, 5*time.Second) + withOpenableLock(t) + lock, err := acquireWindowsSandboxSetupLock(lockTestKey(t)) + if err != nil { + t.Fatalf("acquire: %v", err) + } + lock.release() + lock.release() +} + +// The object name has to be per workspace and stable, since that is what makes +// the exclusion scoped correctly rather than machine-wide or per-run. +func TestLockNameIsPerWorkspaceAndStable(t *testing.T) { + first := windowsSandboxSetupLockName("workspace-one") + again := windowsSandboxSetupLockName("workspace-one") + other := windowsSandboxSetupLockName("workspace-two") + + if first != again { + t.Errorf("the same workspace derived two names, %q then %q", first, again) + } + if first == other { + t.Errorf("two workspaces share the lock name %q, so unrelated setups would serialize", first) + } + // Global rather than Local: the account, its logon rights and the WFP + // filters are machine state, so setups in different logon sessions must + // still exclude each other. + if !strings.HasPrefix(first, `Global\`) { + t.Errorf("lock name %q is not in the Global namespace, so it would not span logon sessions", first) + } + // A backslash past the namespace prefix is rejected by the object manager, + // so a key that smuggled one in would fail at CreateMutex. + if strings.Contains(strings.TrimPrefix(first, `Global\`), `\`) { + t.Errorf("lock name %q contains a separator past its namespace prefix", first) + } +} + +// The production lock must stay reachable only by accounts that could already +// run setup. A sandbox principal able to open it could hold it and stall every +// future setup, or squat the name so real setup believes it holds a lock it +// does not. +func TestSetupLockIsAdministratorsOnly(t *testing.T) { + // BA is Administrators, SY is SYSTEM, and the leading P protects the DACL so + // no inherited ACE can widen it. + for _, want := range []string{"D:P", "(A;;GA;;;BA)", "(A;;GA;;;SY)"} { + if !strings.Contains(windowsSetupLockSDDL, want) { + t.Errorf("lock SDDL %q is missing %q", windowsSetupLockSDDL, want) + } + } + // WD is Everyone. Granting it here would let any sandboxed process stall + // setup for the whole machine. + if strings.Contains(windowsSetupLockSDDL, ";WD)") { + t.Errorf("lock SDDL %q grants Everyone, so any process could hold the setup lock", windowsSetupLockSDDL) + } +} + +// An empty or unusable key must still produce a valid name rather than a bare +// prefix, which every such workspace would then share. +func TestLockNameHandlesAnEmptyKey(t *testing.T) { + name := windowsSandboxSetupLockName("") + if name == `Global\ZeroSandboxSetup-` { + t.Fatal("an empty key produced a bare prefix, so every such workspace would share one lock") + } + if !strings.HasPrefix(name, `Global\ZeroSandboxSetup-`) { + t.Errorf("unexpected name %q", name) + } +} diff --git a/internal/sandbox/windows_setup_other.go b/internal/sandbox/windows_setup_other.go index a18eb2c9b..4a794cbaf 100644 --- a/internal/sandbox/windows_setup_other.go +++ b/internal/sandbox/windows_setup_other.go @@ -11,3 +11,9 @@ func runWindowsSandboxSetup(config WindowsSandboxSetupConfig, stderr io.Writer) fmt.Fprintln(stderr, WindowsSandboxSetupName+": Windows sandbox setup is only available on Windows") return 1 } + +// windowsCurrentUserSID has no meaning off Windows. Returning empty keeps +// BuildWindowsSandboxSetupArgs from emitting a --caller-sid nobody could +// interpret, which is also what a Windows caller does when its own token cannot +// be read. +func windowsCurrentUserSID() string { return "" } diff --git a/internal/sandbox/windows_setup_principal_fingerprint_test.go b/internal/sandbox/windows_setup_principal_fingerprint_test.go new file mode 100644 index 000000000..3f724955f --- /dev/null +++ b/internal/sandbox/windows_setup_principal_fingerprint_test.go @@ -0,0 +1,112 @@ +package sandbox + +import ( + "encoding/json" + "os" + "path/filepath" + "strings" + "testing" +) + +func principalFingerprintConfig(sandboxHome string, readRoots []string) WindowsSandboxSetupConfig { + workspace := filepath.FromSlash("/ws/project") + return WindowsSandboxSetupConfig{ + SandboxHome: sandboxHome, + CommandCWD: workspace, + WorkspaceRoots: []string{workspace}, + PrincipalOptIn: true, + PermissionProfile: PermissionProfile{ + FileSystem: FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{Root: workspace}}, + ReadRoots: readRoots, + }, + }, + } +} + +// Narrowing a principal's read roots MUST invalidate setup. +// +// ACLPlanHash covers BuildWindowsACLPlan, the capability-SID plan. Principal +// grants are built separately by buildWindowsPrincipalACLPlan from the same +// profile, so before this the marker was blind to them: remove a read root and +// setup still looked current while the AllowRead ACE stayed on disk. Narrowing +// a policy has to be able to take access away. +func TestSetupMarkerInvalidatesWhenPrincipalReadRootsShrink(t *testing.T) { + home := t.TempDir() + wide, err := BuildWindowsSandboxSetupMarker(principalFingerprintConfig(home, []string{ + filepath.FromSlash("/ws/project"), + filepath.FromSlash("/ws/extra-read"), + })) + if err != nil { + t.Fatalf("build wide marker: %v", err) + } + narrow, err := BuildWindowsSandboxSetupMarker(principalFingerprintConfig(home, []string{ + filepath.FromSlash("/ws/project"), + })) + if err != nil { + t.Fatalf("build narrow marker: %v", err) + } + + if wide.PrincipalPlanHash == "" { + t.Fatal("no principal fingerprint recorded while opted in, so principal grants are unfingerprinted") + } + if wide.PrincipalPlanHash == narrow.PrincipalPlanHash { + t.Error("dropping a principal read root did not change the fingerprint, so stale AllowRead ACEs survive a narrowed policy") + } +} + +// A stale principal fingerprint must be refused through the real validator, +// which reads the marker off disk, with a message that says what to do. +func TestValidateRefusesAChangedPrincipalFingerprint(t *testing.T) { + home := t.TempDir() + config := principalFingerprintConfig(home, []string{filepath.FromSlash("/ws/project")}) + if _, err := WriteWindowsSandboxSetupMarker(config); err != nil { + t.Fatalf("write marker: %v", err) + } + if err := ValidateWindowsSandboxSetupMarker(config); err != nil { + t.Fatalf("a freshly written marker did not validate, so this test cannot isolate the fingerprint: %v", err) + } + + // Rewrite only the principal fingerprint, the way a policy change would. + path := WindowsSandboxSetupMarkerPath(home) + raw, err := os.ReadFile(path) + if err != nil { + t.Fatalf("read marker: %v", err) + } + var marker map[string]any + if err := json.Unmarshal(raw, &marker); err != nil { + t.Fatalf("parse marker: %v", err) + } + marker["principalPlanHash"] = "stale-hash-from-an-earlier-policy" + rewritten, err := json.Marshal(marker) + if err != nil { + t.Fatalf("marshal marker: %v", err) + } + if err := os.WriteFile(path, rewritten, 0o600); err != nil { + t.Fatalf("rewrite marker: %v", err) + } + + err = ValidateWindowsSandboxSetupMarker(config) + if err == nil { + t.Fatal("a marker whose principal grants no longer match the policy was accepted") + } + if !strings.Contains(err.Error(), "principal grants changed") { + t.Errorf("refused for the wrong reason: %v", err) + } +} + +// The default install must be untouched: opted out means no fingerprint, so the +// marker does not churn for the overwhelming majority of users. +func TestSetupMarkerHasNoPrincipalFingerprintWhenOptedOut(t *testing.T) { + config := principalFingerprintConfig(t.TempDir(), []string{filepath.FromSlash("/ws/project")}) + config.PrincipalOptIn = false + + marker, err := BuildWindowsSandboxSetupMarker(config) + if err != nil { + t.Fatalf("build marker: %v", err) + } + if marker.PrincipalPlanHash != "" { + t.Errorf("principal fingerprint %q recorded while opted out", marker.PrincipalPlanHash) + } +} diff --git a/internal/sandbox/windows_setup_runtime_root_test.go b/internal/sandbox/windows_setup_runtime_root_test.go new file mode 100644 index 000000000..d291efceb --- /dev/null +++ b/internal/sandbox/windows_setup_runtime_root_test.go @@ -0,0 +1,292 @@ +package sandbox + +import ( + "os" + "strings" + "testing" +) + +// runtimeRootTestConfig is the shape every command reaches the Windows runner +// with: a restricted filesystem rooted at the workspace, which is what makes the +// runtime root necessary in the first place. +func runtimeRootTestConfig(t *testing.T) WindowsSandboxCommandConfig { + t.Helper() + // REDIRECT THE AMBIENT CACHE PRODUCER, not only the obvious inputs. + // + // The workspace and sandbox home below are test-owned, but + // windowsSandboxRuntimeCandidates still derives a candidate from + // sandboxUserCacheDir, and the runtime-root test creates every candidate and + // registers os.RemoveAll cleanup for them. That reaches into the real user + // cache: it fails outright on a read-only home, and on an ordinary developer + // or CI account it creates and then deletes a path outside the test's + // storage. The workspace hash makes a collision unlikely; it does not make + // somebody else's directory test-owned. + cache := t.TempDir() + previousCache := sandboxUserCacheDir + sandboxUserCacheDir = func() (string, error) { return cache, nil } + t.Cleanup(func() { sandboxUserCacheDir = previousCache }) + + workspace := t.TempDir() + return WindowsSandboxCommandConfig{ + SandboxHome: t.TempDir(), + CommandCWD: workspace, + WorkspaceRoots: []string{workspace}, + PermissionProfile: PermissionProfile{ + FileSystem: FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{Root: workspace}}, + }, + Network: NetworkPolicy{Mode: NetworkDeny}, + }, + } +} + +// A marker written by a fresh setup must accept the ordinary command, and the +// ordinary command is the RUNTIME-AUGMENTED one: Engine.run calls +// permissionProfileWithRuntime before the Windows runner ever sees the profile, +// so the profile presented at validation always carries the selected runtime +// root as an extra write root. +// +// Setup used to fingerprint the bare profile. The extra write root changed the +// ACL plan, the plan hash changed with it, and validation rejected a marker +// written seconds earlier with "permission roots or deny lists changed" — so on +// a restricted filesystem no command could run at all, including the very +// command that had just been set up for. +// +// Asserted for BOTH candidates because which one a process selects is not fixed: +// sandboxRuntimeRootFor prefers the cache-derived root and falls back to the +// temp-derived one, and a marker that only accepts the preferred root bricks +// every machine that falls back. +func TestWindowsSandboxSetupMarkerAcceptsRuntimeAugmentedCommand(t *testing.T) { + config := runtimeRootTestConfig(t) + // The setup half, as BuildWindowsSandboxSetupArgs prepares it in the + // operator's shell before the elevated helper ever runs. + setup := WindowsSandboxSetupConfigFromCommand(config) + setup.PermissionProfile = WindowsSandboxProfileWithRuntimeRoots(setup.PermissionProfile, config.WorkspaceRoots) + if _, err := WriteWindowsSandboxSetupMarker(setup); err != nil { + t.Fatalf("WriteWindowsSandboxSetupMarker: %v", err) + } + + candidates := windowsSandboxRuntimeCandidates(config.WorkspaceRoots) + if len(candidates) == 0 { + t.Fatal("windowsSandboxRuntimeCandidates returned none, so this test proves nothing") + } + for _, candidate := range candidates { + augmented := config + // The command half, in the same order the real path builds it: the engine + // appends the SELECTED root, then the Windows plan folds in the candidate + // set before serializing the profile to the runner. + augmented.PermissionProfile = WindowsSandboxProfileWithRuntimeRoots( + permissionProfileWithRuntime(config.PermissionProfile, SandboxRuntime{Root: candidate}), + config.WorkspaceRoots, + ) + err := ValidateWindowsSandboxSetupMarker(WindowsSandboxSetupConfigFromCommand(augmented)) + if err != nil { + t.Fatalf("ValidateWindowsSandboxSetupMarker with runtime root %s: %v", candidate, err) + } + } + + // THE RUNNER'S TEMP IS NOT THE OPERATOR'S. + // + // sandboxRuntimeEnvironment points TMPDIR/TMP/TEMP at the sandbox runtime + // temp for everything the sandbox launches, and the command runner inherits + // that env. While the runner derived the candidate set itself, os.TempDir() + // there returned the redirected value, so it produced a temp-derived root + // under the runtime tree while setup produced one under the real temp: two + // plans with the SAME entry count and different hashes, and every sandboxed + // command refused to run with "permission roots or deny lists changed". + // + // Validation must not move when that variable does. + // Augmented FIRST, standing in for the parent, whose TEMP is still real. + runner := config + runner.PermissionProfile = WindowsSandboxProfileWithRuntimeRoots( + permissionProfileWithRuntime(config.PermissionProfile, SandboxRuntime{Root: candidates[0]}), + config.WorkspaceRoots, + ) + // Only THEN does the environment become the runner's. Anything downstream of + // this line that re-derives a runtime root gets the redirected answer, which + // is precisely the defect: validation has to be settled before here. + t.Setenv("TEMP", t.TempDir()) + t.Setenv("TMP", os.Getenv("TEMP")) + if err := ValidateWindowsSandboxSetupMarker(WindowsSandboxSetupConfigFromCommand(runner)); err != nil { + t.Fatalf("ValidateWindowsSandboxSetupMarker with a redirected TEMP: %v", err) + } + + // The guard has to still bite, or the test above passes for the wrong reason + // — a validator that accepts everything would satisfy it too. + changed := config + changed.PermissionProfile.FileSystem.DenyRead = []string{`C:\workspace\secret`} + if err := ValidateWindowsSandboxSetupMarker(WindowsSandboxSetupConfigFromCommand(changed)); err == nil { + t.Fatal("ValidateWindowsSandboxSetupMarker accepted a changed deny list, so it no longer detects drift") + } else if !strings.Contains(err.Error(), "out of date") { + t.Fatalf("ValidateWindowsSandboxSetupMarker changed error = %v, want out of date", err) + } +} + +// The runtime root needs BOTH sides of the write-restricted grant. +// +// A principal command runs on a token restricted to the capability SIDs, and a +// WRITE_RESTRICTED token grants a write only when the normal token check AND the +// restricting-SID check both pass. The runtime root used to be appended to the +// principal plan alone, so it carried the account ACE and no capability ACE: the +// normal check passed, the restricted check found nothing, and every cache and +// temp write was denied even once the marker agreed. +// +// This asserts the capability side, which is the side that was missing. Both +// candidates again, for the same reason as above. +func TestWindowsSandboxRuntimeRootsAreInTheCapabilityPlan(t *testing.T) { + config := runtimeRootTestConfig(t) + candidates := windowsSandboxRuntimeCandidates(config.WorkspaceRoots) + if len(candidates) == 0 { + t.Fatal("windowsSandboxRuntimeCandidates returned none, so this test proves nothing") + } + + setup := WindowsSandboxSetupConfigFromCommand(config) + setup.PermissionProfile = WindowsSandboxProfileWithRuntimeRoots(setup.PermissionProfile, config.WorkspaceRoots) + plan, err := BuildWindowsACLPlan(setup.commandConfig()) + if err != nil { + t.Fatalf("BuildWindowsACLPlan: %v", err) + } + granted := make(map[string]struct{}, len(plan.Entries)) + for _, entry := range plan.Entries { + granted[windowsCapabilityPathKey(entry.Path)] = struct{}{} + } + for _, candidate := range candidates { + if _, ok := granted[windowsCapabilityPathKey(candidate)]; !ok { + t.Fatalf("capability ACL plan has no entry for runtime root %s; writes there fail the restricting-SID check", candidate) + } + } +} + +// EVERY write root the capability plan grants must exist by the time setup +// applies it. +// +// The capability plan deliberately refuses to materialize a write root, so a +// granted path that is merely absent fails the entire setup run with +// +// windows ACL target does not exist: ...\zero\runtime\v1\ +// +// which is exactly what an elevated run hit: the runtime candidates were added +// to the plan while only the selected one was ever created, and `zero sandbox +// setup` stopped working altogether. The two halves are separate functions, so +// nothing but this test stops one from growing a candidate without the other. +func TestWindowsSandboxSetupProvisionsEveryGrantedWriteRoot(t *testing.T) { + config := runtimeRootTestConfig(t) + candidates := windowsSandboxRuntimeCandidates(config.WorkspaceRoots) + if len(candidates) == 0 { + t.Fatal("windowsSandboxRuntimeCandidates returned none, so this test proves nothing") + } + // The workspace is unique per run, and the roots are derived from it, so these + // cannot pre-exist. Assert that rather than trust it: if they did, the test + // would pass with the provisioning step deleted. + for _, candidate := range candidates { + if _, err := os.Stat(candidate); err == nil { + t.Fatalf("runtime root %s already exists before provisioning, so this test proves nothing", candidate) + } + t.Cleanup(func() { _ = os.RemoveAll(candidate) }) + } + + if err := ensureWindowsSandboxRuntimeCandidates(config.WorkspaceRoots); err != nil { + t.Fatalf("ensureWindowsSandboxRuntimeCandidates: %v", err) + } + + setup := WindowsSandboxSetupConfigFromCommand(config) + setup.PermissionProfile = WindowsSandboxProfileWithRuntimeRoots(setup.PermissionProfile, config.WorkspaceRoots) + plan, err := BuildWindowsACLPlan(setup.commandConfig()) + if err != nil { + t.Fatalf("BuildWindowsACLPlan: %v", err) + } + for _, entry := range plan.Entries { + if entry.Action != WindowsACLAllowWrite || entry.Materialize { + continue + } + if _, err := os.Stat(entry.Path); err != nil { + t.Errorf("capability plan grants write on %s but nothing created it, so setup fails with "+ + "\"windows ACL target does not exist\": %v", entry.Path, err) + } + } +} + +// Both candidates are pure functions of the workspace root. Setup provisions the +// set and a later command selects from it in a different process, so a candidate +// that varied per process (a random or time-seeded fallback) would be granted by +// setup and never selected, or selected and never granted. +func TestWindowsSandboxRuntimeCandidatesAreDeterministic(t *testing.T) { + config := runtimeRootTestConfig(t) + first := windowsSandboxRuntimeCandidates(config.WorkspaceRoots) + if len(first) == 0 { + t.Fatal("windowsSandboxRuntimeCandidates returned none, so this test proves nothing") + } + second := windowsSandboxRuntimeCandidates(config.WorkspaceRoots) + if len(first) != len(second) { + t.Fatalf("candidate count = %d then %d, want stable", len(first), len(second)) + } + for i := range first { + if first[i] != second[i] { + t.Fatalf("candidate %d = %q then %q, want stable", i, first[i], second[i]) + } + } + + other := runtimeRootTestConfig(t) + otherCandidates := windowsSandboxRuntimeCandidates(other.WorkspaceRoots) + for _, candidate := range otherCandidates { + for _, mine := range first { + if candidate == mine { + t.Fatalf("workspaces %s and %s share runtime root %s, so one workspace's grant covers the other", + config.CommandCWD, other.CommandCWD, candidate) + } + } + } +} + +// EVERY ROOT THE SELECTOR CAN CHOOSE HAS TO BE IN THE SET SETUP GRANTS. +// +// The tests above all iterate whatever windowsSandboxRuntimeCandidates returns, +// so they hold for any self-consistent candidate set, including one missing an +// entry. That is the half this pins: sandboxRuntimeRootFor prefers the +// cache-derived root and falls back to the temp-derived one when the cache root +// is unusable, and prepareSandboxRuntime takes that fallback when the primary +// lease cannot be acquired. A candidate set carrying only the preferred root +// would be internally consistent and still leave the fallback ungranted, so a +// machine that fell back would be refused the write it needs. Reported by +// @anandh8x. +// +// Driven through the selector rather than by re-deriving the fallback here, so +// the assertion stays true if the selection rule changes. +func TestWindowsSandboxRuntimeCandidatesCoverEverySelectableRoot(t *testing.T) { + config := runtimeRootTestConfig(t) + workspace := canonicalSandboxWorkspaceRoot(config.WorkspaceRoots[0]) + candidates := windowsSandboxRuntimeCandidates(config.WorkspaceRoots) + if len(candidates) == 0 { + t.Fatal("windowsSandboxRuntimeCandidates returned none, so this test proves nothing") + } + granted := make(map[string]struct{}, len(candidates)) + for _, candidate := range candidates { + granted[windowsCapabilityPathKey(candidate)] = struct{}{} + } + + cacheRoot, err := sandboxUserCacheDir() + if err != nil { + t.Fatalf("sandboxUserCacheDir: %v", err) + } + // The ordinary selection, and then the selection a machine makes when the + // cache-derived root is not usable. Naming an unusable cache root reproduces + // that second case without depending on the host's real cache directory. + preferred, err := sandboxRuntimeRootFor(workspace, canonicalSandboxWorkspaceRoot(cacheRoot)) + if err != nil { + t.Fatalf("sandboxRuntimeRootFor(usable cache): %v", err) + } + fallback, err := sandboxRuntimeRootFor(workspace, workspace) + if err != nil { + t.Fatalf("sandboxRuntimeRootFor(unusable cache): %v", err) + } + if preferred == fallback { + t.Fatalf("both selections returned %s, so this test cannot tell the fallback from the preferred root", preferred) + } + for _, selected := range []string{preferred, fallback} { + if _, ok := granted[windowsCapabilityPathKey(selected)]; !ok { + t.Errorf("the selector can choose runtime root %s but setup grants only %v, "+ + "so a command that selects it is refused the write it was provisioned for", selected, candidates) + } + } +} diff --git a/internal/sandbox/windows_setup_test.go b/internal/sandbox/windows_setup_test.go index ecb76831a..8679f1ec9 100644 --- a/internal/sandbox/windows_setup_test.go +++ b/internal/sandbox/windows_setup_test.go @@ -5,6 +5,7 @@ import ( "encoding/json" "os" "path/filepath" + "runtime" "strings" "testing" ) @@ -187,6 +188,178 @@ func TestWindowsSandboxSetupMarkerRejectsOldSchema(t *testing.T) { } } +// The principal opt-in has to travel in the setup args, because the elevated +// half runs in its own process: a UAC-elevated helper does not inherit the +// environment of the shell that asked for setup. Sampling the ambient +// environment there let the two halves disagree, so the serialized value must +// win over the environment in BOTH directions. +func TestWindowsSandboxSetupPrincipalOptInSurvivesElevatedEnvironment(t *testing.T) { + allowPrincipalLaunchForTest(t) + profile := PermissionProfile{ + FileSystem: FileSystemPolicy{Kind: FileSystemRestricted, WriteRoots: []WritableRoot{{Root: `C:\workspace`}}}, + Network: NetworkPolicy{Mode: NetworkAllow}, + } + testCases := []struct { + name string + optIn bool + ambientEnv string + }{ + // The reported case: the caller's shell opted in, the elevated helper's + // environment has nothing. Without the serialized value setup provisions no + // principal and every later command silently falls back. + {name: "opted in, elevated environment empty", optIn: true, ambientEnv: ""}, + // The mirror: the elevated helper happens to have a machine-wide opt-in the + // caller did not ask for. Setup must not create an account on its own say-so. + {name: "opted out, elevated environment opted in", optIn: false, ambientEnv: "1"}, + } + for _, testCase := range testCases { + t.Run(testCase.name, func(t *testing.T) { + t.Setenv(windowsSandboxIdentityEnv, testCase.ambientEnv) + optIn := testCase.optIn + args, err := BuildWindowsSandboxSetupArgs(WindowsSandboxSetupArgsOptions{ + SandboxHome: t.TempDir(), + CommandCWD: `C:\workspace`, + WorkspaceRoots: []string{`C:\workspace`}, + // The parent folds in the runtime roots before the runner sees the profile, + // so a command half that skips it no longer matches the setup half. + PermissionProfile: WindowsSandboxProfileWithRuntimeRoots(profile, []string{`C:\workspace`}), + PrincipalOptIn: &optIn, + }) + if err != nil { + t.Fatalf("BuildWindowsSandboxSetupArgs: %v", err) + } + config, err := ParseWindowsSandboxSetupArgs(args) + if err != nil { + t.Fatalf("ParseWindowsSandboxSetupArgs: %v", err) + } + if config.PrincipalOptIn != testCase.optIn { + t.Fatalf("parsed PrincipalOptIn = %v, want %v", config.PrincipalOptIn, testCase.optIn) + } + // This is the value the elevated setup gate actually reads before it + // decides to provision an account. + if got := windowsSandboxIdentityEnabled(config.commandConfig().Env); got != testCase.optIn { + t.Fatalf("elevated setup opt-in = %v, want %v (ambient %s=%q must not decide)", + got, testCase.optIn, windowsSandboxIdentityEnv, testCase.ambientEnv) + } + }) + } +} + +// A setup helper that cannot read the opt-in must refuse rather than default to +// "no principal": provisioning less than the caller asked for and reporting +// success is the silent downgrade this protocol exists to prevent. +func TestParseWindowsSandboxSetupArgsRejectsUnreadablePrincipalOptIn(t *testing.T) { + allowPrincipalLaunchForTest(t) + optIn := true + args, err := BuildWindowsSandboxSetupArgs(WindowsSandboxSetupArgsOptions{ + SandboxHome: t.TempDir(), + CommandCWD: `C:\workspace`, + WorkspaceRoots: []string{`C:\workspace`}, + PermissionProfile: PermissionProfile{FileSystem: FileSystemPolicy{Kind: FileSystemRestricted}, Network: NetworkPolicy{Mode: NetworkDeny}}, + PrincipalOptIn: &optIn, + }) + if err != nil { + t.Fatalf("BuildWindowsSandboxSetupArgs: %v", err) + } + for index, arg := range args { + if arg == "--sandbox-principal" { + args[index+1] = "yes" + } + } + if _, err := ParseWindowsSandboxSetupArgs(args); err == nil || !strings.Contains(err.Error(), "--sandbox-principal") { + t.Fatalf("ParseWindowsSandboxSetupArgs error = %v, want rejection of the unreadable opt-in", err) + } +} + +// Setup and the commands that follow it run in separate processes, so they can +// disagree about the opt-in. The marker records what setup provisioned and the +// command refuses on a mismatch — most of all when the command opted in and +// setup did not, because the runtime's principal lookup declines with a nil +// error and the command would otherwise run on the read-unconfined +// restricted-token backend while the operator believes a principal is isolating +// it. +func TestWindowsSandboxSetupMarkerRejectsPrincipalOptInMismatch(t *testing.T) { + // Neutral ambient environment: the disagreement under test is between the two + // recorded intents, not between either of them and this process. + t.Setenv(windowsSandboxIdentityEnv, "") + profile := PermissionProfile{ + FileSystem: FileSystemPolicy{Kind: FileSystemRestricted, WriteRoots: []WritableRoot{{Root: `C:\workspace`}}}, + Network: NetworkPolicy{Mode: NetworkAllow}, + } + command := func(home string, env map[string]string) WindowsSandboxCommandConfig { + return WindowsSandboxCommandConfig{ + SandboxHome: home, + CommandCWD: `C:\workspace`, + WorkspaceRoots: []string{`C:\workspace`}, + // The parent folds in the runtime roots before the runner sees the profile, + // so a command half that skips it no longer matches the setup half. + PermissionProfile: WindowsSandboxProfileWithRuntimeRoots(profile, []string{`C:\workspace`}), + Env: env, + SandboxLevel: WindowsSandboxLevelRestrictedToken, + Command: []string{"cmd.exe", "/c", "echo"}, + } + } + testCases := []struct { + name string + setupOptIn bool + commandEnv map[string]string + wantError string + }{ + { + name: "command opts in, setup provisioned no principal", + setupOptIn: false, + commandEnv: map[string]string{windowsSandboxIdentityEnv: "1"}, + wantError: "asks for a sandbox principal, but setup provisioned none", + }, + { + name: "setup provisioned a principal, command opts out", + setupOptIn: true, + commandEnv: map[string]string{windowsSandboxIdentityEnv: "0"}, + wantError: "setup provisioned sandbox principals", + }, + } + for _, testCase := range testCases { + t.Run(testCase.name, func(t *testing.T) { + home := t.TempDir() + setupConfig := WindowsSandboxSetupConfig{ + SandboxHome: home, + CommandCWD: `C:\workspace`, + WorkspaceRoots: []string{`C:\workspace`}, + // The parent folds in the runtime roots before the runner sees the profile, + // so a command half that skips it no longer matches the setup half. + PermissionProfile: WindowsSandboxProfileWithRuntimeRoots(profile, []string{`C:\workspace`}), + PrincipalOptIn: testCase.setupOptIn, + } + marker, err := WriteWindowsSandboxSetupMarker(setupConfig) + if err != nil { + t.Fatalf("WriteWindowsSandboxSetupMarker: %v", err) + } + // Assert the setup half recorded what it was told before trusting what + // the command half makes of it. + if marker.PrincipalOptIn != testCase.setupOptIn { + t.Fatalf("marker PrincipalOptIn = %v, want %v", marker.PrincipalOptIn, testCase.setupOptIn) + } + // An agreeing command still validates, so the refusal below is about the + // disagreement and not about the marker being unusable. + agreeing := command(home, map[string]string{windowsSandboxIdentityEnv: windowsSandboxPrincipalOptInValue(testCase.setupOptIn)}) + if err := ValidateWindowsSandboxSetupMarker(WindowsSandboxSetupConfigFromCommand(agreeing)); err != nil { + t.Fatalf("agreeing command must validate against its own setup: %v", err) + } + + err = ValidateWindowsSandboxSetupMarker(WindowsSandboxSetupConfigFromCommand(command(home, testCase.commandEnv))) + if err == nil { + t.Fatalf("disagreeing command validated the marker, want refusal") + } + if !strings.Contains(err.Error(), testCase.wantError) { + t.Fatalf("validate error = %v, want it to contain %q", err, testCase.wantError) + } + if !strings.Contains(err.Error(), "zero sandbox setup") { + t.Fatalf("validate error = %v, want the remedy to name `zero sandbox setup`", err) + } + }) + } +} + func TestWindowsSandboxSetupConfigFromCommandPreservesProfileInputs(t *testing.T) { command := WindowsSandboxCommandConfig{ SandboxHome: t.TempDir(), @@ -212,6 +385,111 @@ func TestWindowsSandboxSetupConfigFromCommandPreservesProfileInputs(t *testing.T } } +// Serializing the opt-in makes it a field every caller of +// BuildWindowsSandboxSetupArgs could get wrong, so the field is a tri-state and +// its unset meaning is load-bearing: "consult the environment", never "opted +// out". The command half still resolves the opt-in from the process environment +// when its own Env carries no explicit entry, so if an unset setup caller +// asserted false instead, the two halves would disagree under a machine-wide +// opt-in and marker validation would refuse every command — safe, but it bricks +// the caller, and it re-creates the very disagreement this flag removes. That is +// exactly what the existing smoke callers +// (runner_windows_integration_test.go:43 and :52) do: they never set the field. +// +// This test runs on every GOOS and pins the unset default in both ambient +// states, so getting it backwards is a test failure here rather than a surprise +// on a real elevated machine. +func TestWindowsSandboxSetupArgsUnsetPrincipalOptInConsultsEnvironment(t *testing.T) { + allowPrincipalLaunchForTest(t) + profile := PermissionProfile{ + FileSystem: FileSystemPolicy{Kind: FileSystemRestricted, WriteRoots: []WritableRoot{{Root: `C:\workspace`}}}, + Network: NetworkPolicy{Mode: NetworkAllow}, + } + // The command half as an ambient caller declares it: no explicit entry in Env, + // so it resolves the opt-in from the environment. + command := func(home string) WindowsSandboxCommandConfig { + return WindowsSandboxCommandConfig{ + SandboxHome: home, + CommandCWD: `C:\workspace`, + WorkspaceRoots: []string{`C:\workspace`}, + // The parent folds in the runtime roots before the runner sees the profile, + // so a command half that skips it no longer matches the setup half. + PermissionProfile: WindowsSandboxProfileWithRuntimeRoots(profile, []string{`C:\workspace`}), + SandboxLevel: WindowsSandboxLevelRestrictedToken, + Command: []string{"cmd.exe", "/c", "echo"}, + } + } + // setupMarkerFor runs the full caller path — build args, cross the (simulated) + // UAC boundary by re-parsing them, write the marker — so what is asserted is + // what an elevated helper would actually have provisioned. + setupMarkerFor := func(t *testing.T, home string, optIn *bool) WindowsSandboxSetupMarker { + t.Helper() + args, err := BuildWindowsSandboxSetupArgs(WindowsSandboxSetupArgsOptions{ + SandboxHome: home, + CommandCWD: `C:\workspace`, + WorkspaceRoots: []string{`C:\workspace`}, + // Deliberately NOT pre-augmented. BuildWindowsSandboxSetupArgs folds the + // runtime roots in itself, and passing them in here would hide it from + // this test if it ever stopped. + PermissionProfile: profile, + PrincipalOptIn: optIn, + }) + if err != nil { + t.Fatalf("BuildWindowsSandboxSetupArgs: %v", err) + } + config, err := ParseWindowsSandboxSetupArgs(args) + if err != nil { + t.Fatalf("ParseWindowsSandboxSetupArgs: %v", err) + } + marker, err := WriteWindowsSandboxSetupMarker(config) + if err != nil { + t.Fatalf("WriteWindowsSandboxSetupMarker: %v", err) + } + return marker + } + + for _, ambient := range []string{"1", ""} { + name := "machine-wide opt-in" + if ambient == "" { + name = "no opt-in" + } + t.Run(name, func(t *testing.T) { + t.Setenv(windowsSandboxIdentityEnv, ambient) + want := ambient == "1" + + // An unset caller must provision what the environment says. Assert the + // setup half recorded that before trusting the agreement below: a marker + // that recorded the wrong thing could still "agree" if the command half + // were broken in the same direction. + home := t.TempDir() + marker := setupMarkerFor(t, home, nil) + if marker.PrincipalOptIn != want { + t.Fatalf("unset caller recorded PrincipalOptIn = %v, want %v (ambient %s=%q decides)", + marker.PrincipalOptIn, want, windowsSandboxIdentityEnv, ambient) + } + if err := ValidateWindowsSandboxSetupMarker(WindowsSandboxSetupConfigFromCommand(command(home))); err != nil { + t.Fatalf("an unset caller must agree with the ambient command half: %v", err) + } + + // And an explicit value still overrides the environment in both + // directions, or the tri-state would have no third state. + override := !want + overrideHome := t.TempDir() + overrideMarker := setupMarkerFor(t, overrideHome, &override) + if overrideMarker.PrincipalOptIn != override { + t.Fatalf("explicit caller recorded PrincipalOptIn = %v, want %v", overrideMarker.PrincipalOptIn, override) + } + err := ValidateWindowsSandboxSetupMarker(WindowsSandboxSetupConfigFromCommand(command(overrideHome))) + if err == nil { + t.Fatalf("an explicit opt-in of %v validated against an ambient command half of %v, want refusal", override, want) + } + if !strings.Contains(err.Error(), windowsSandboxIdentityEnv) { + t.Fatalf("validate error = %v, want it to name %s", err, windowsSandboxIdentityEnv) + } + }) + } +} + func TestWindowsACLPlanHashIsStableAcrossEntryOrder(t *testing.T) { left, err := WindowsACLPlanHash(WindowsACLPlan{Entries: []WindowsACLEntry{ {Action: WindowsACLDenyRead, Path: `C:\workspace\secret`, Capability: "S-1-5-21-3", Materialize: true}, @@ -233,3 +511,165 @@ func TestWindowsACLPlanHashIsStableAcrossEntryOrder(t *testing.T) { t.Fatalf("ACL plan hashes differ: %q vs %q", left, right) } } + +// The caller's identity has to survive the UAC boundary for the same reason the +// opt-in does: elevated setup is a separate process, and under over-the-shoulder +// elevation it belongs to a different administrator. Every principal name, its +// ACL ledger and its DPAPI secret are scoped to the invoking user, so a helper +// left to sample its own token provisions for the wrong one. +// +// Runs on every GOOS. The identity is resolved before the boundary, so what is +// pinned here is the transport, which is where it would silently go missing. +func TestWindowsSandboxSetupArgsCarryTheCallerIdentity(t *testing.T) { + profile := PermissionProfile{ + FileSystem: FileSystemPolicy{Kind: FileSystemRestricted, WriteRoots: []WritableRoot{{Root: `C:\workspace`}}}, + Network: NetworkPolicy{Mode: NetworkDeny}, + } + const callerSID = "S-1-5-21-1111111111-2222222222-3333333333-1001" + args, err := BuildWindowsSandboxSetupArgs(WindowsSandboxSetupArgsOptions{ + // Pinned: this fixture tests caller-identity transport, not principal mode. + // Left unset it inherits ZERO_WINDOWS_SANDBOX_IDENTITY from the host and + // stops at the launch refusal before reaching the assertions below. + PrincipalOptIn: new(bool), + SandboxHome: `C:\home`, + CommandCWD: `C:\workspace`, + WorkspaceRoots: []string{`C:\workspace`}, + // The parent folds in the runtime roots before the runner sees the profile, + // so a command half that skips it no longer matches the setup half. + PermissionProfile: WindowsSandboxProfileWithRuntimeRoots(profile, []string{`C:\workspace`}), + CallerSID: callerSID, + }) + if err != nil { + t.Fatalf("BuildWindowsSandboxSetupArgs: %v", err) + } + config, err := ParseWindowsSandboxSetupArgs(args) + if err != nil { + t.Fatalf("ParseWindowsSandboxSetupArgs: %v", err) + } + if config.CallerSID != callerSID { + t.Fatalf("the caller identity did not survive the setup args: got %q, want %q", config.CallerSID, callerSID) + } + // And on to the command-shaped view the setup half actually plans against — + // the only place it changes what gets provisioned. + if got := config.commandConfig().CallerSID; got != callerSID { + t.Errorf("commandConfig dropped the caller identity: got %q, want %q", got, callerSID) + } +} + +// A caller that cannot state its identity must not send an empty flag. Absence +// means "this caller did not say", which the helper answers by sampling its own +// token exactly as it did before the flag existed; an empty value would have to +// be guessed at instead. +func TestWindowsSandboxSetupArgsOmitAnUnknownCallerIdentity(t *testing.T) { + args, err := BuildWindowsSandboxSetupArgs(WindowsSandboxSetupArgsOptions{ + PrincipalOptIn: new(bool), // transport fixture; see TestWindowsSandboxSetupArgsCarryTheCallerIdentity + SandboxHome: `C:\home`, + CommandCWD: `C:\workspace`, + WorkspaceRoots: []string{`C:\workspace`}, + PermissionProfile: PermissionProfile{ + FileSystem: FileSystemPolicy{Kind: FileSystemRestricted}, + Network: NetworkPolicy{Mode: NetworkDeny}, + }, + // Left unset. On non-Windows builds windowsCurrentUserSID also returns + // empty, which is the same case a Windows caller hits when its token query + // fails. + }) + if err != nil { + t.Fatalf("BuildWindowsSandboxSetupArgs: %v", err) + } + if runtime.GOOS != "windows" { + for _, arg := range args { + if arg == "--caller-sid" { + t.Fatalf("an unresolvable caller identity was still serialized: %v", args) + } + } + } + config, err := ParseWindowsSandboxSetupArgs(args) + if err != nil { + t.Fatalf("ParseWindowsSandboxSetupArgs: %v", err) + } + if runtime.GOOS != "windows" && config.CallerSID != "" { + t.Errorf("CallerSID = %q, want empty so the helper falls back to its own token", config.CallerSID) + } +} + +// THE SMOKE'S COMMAND SIDE HAS TO PLAN WHAT ITS SETUP SIDE PLANNED. +// +// The real elevated smoke is gated behind an environment variable and an +// administrator, so this pins the portable half of what it does: the exact +// setup-builder, parser, marker and direct-command sequence with principal mode +// unset. The setup builder folds runtime roots into the profile; a command built +// from the bare profile plans fewer entries than the marker attests to and is +// refused before it runs. Built from the same augmented profile, the two agree. +func TestRealSmokeCommandProfileAgreesWithSetup(t *testing.T) { + root := t.TempDir() + profile := PermissionProfile{ + FileSystem: FileSystemPolicy{ + Kind: FileSystemRestricted, + ReadRoots: []string{root}, + WriteRoots: []WritableRoot{{Root: root, ProtectedMetadataNames: []string{".git", ".zero", ".agents"}}}, + IncludePlatformRoots: true, + AllowTemp: true, + }, + Network: NetworkPolicy{Mode: NetworkDeny}, + } + setupArgs, err := BuildWindowsSandboxSetupArgs(WindowsSandboxSetupArgsOptions{ + SandboxHome: filepath.Join(root, ".zero-sandbox"), + CommandCWD: root, + WorkspaceRoots: []string{root}, + PermissionProfile: profile, + PrincipalOptIn: new(bool), + }) + if err != nil { + t.Fatalf("BuildWindowsSandboxSetupArgs: %v", err) + } + setupConfig, err := ParseWindowsSandboxSetupArgs(setupArgs) + if err != nil { + t.Fatalf("ParseWindowsSandboxSetupArgs: %v", err) + } + setupPlan, err := BuildWindowsACLPlan(setupConfig.commandConfig()) + if err != nil { + t.Fatalf("setup plan: %v", err) + } + + commandPlanFor := func(commandProfile PermissionProfile) WindowsACLPlan { + t.Helper() + args, err := BuildWindowsSandboxCommandArgs(WindowsSandboxCommandArgsOptions{ + SandboxHome: filepath.Join(root, ".zero-sandbox"), + CommandCWD: root, + WorkspaceRoots: []string{root}, + PermissionProfile: commandProfile, + SandboxLevel: WindowsSandboxLevelRestrictedToken, + Command: []string{"cmd.exe", "/c", "ver"}, + }) + if err != nil { + t.Fatalf("BuildWindowsSandboxCommandArgs: %v", err) + } + config, err := ParseWindowsSandboxCommandArgs(args) + if err != nil { + t.Fatalf("ParseWindowsSandboxCommandArgs: %v", err) + } + plan, err := BuildWindowsACLPlan(config) + if err != nil { + t.Fatalf("command plan: %v", err) + } + return plan + } + + bare := commandPlanFor(profile) + if len(bare.Entries) == len(setupPlan.Entries) { + t.Fatalf("SETUP INVALID: the bare profile already plans %d entries like setup does, so this test cannot tell the two apart", len(bare.Entries)) + } + augmented := commandPlanFor(WindowsSandboxProfileWithRuntimeRoots(profile, []string{root})) + setupHash, err := WindowsACLPlanHash(setupPlan) + if err != nil { + t.Fatal(err) + } + commandHash, err := WindowsACLPlanHash(augmented) + if err != nil { + t.Fatal(err) + } + if setupHash != commandHash { + t.Fatalf("setup plans %d entries and the augmented command plans %d, with different hashes; the smoke would be refused at the marker", len(setupPlan.Entries), len(augmented.Entries)) + } +} diff --git a/internal/sandbox/windows_setup_windows.go b/internal/sandbox/windows_setup_windows.go index a71f78cae..d3df91b7a 100644 --- a/internal/sandbox/windows_setup_windows.go +++ b/internal/sandbox/windows_setup_windows.go @@ -3,12 +3,33 @@ package sandbox import ( + "errors" "fmt" "io" + "strings" "golang.org/x/sys/windows" ) +// Seams for the setup transaction's external effects. +// +// Every step below needs something a test machine cannot be asked for: an +// elevated token, a machine-global mutex, real DACLs, the WFP engine, a local +// account. That left the ORDER of those steps and the handling of each failure +// — which is where this function's bugs actually live — with no coverage at all. +// Each var defaults to its production function and is only ever reassigned by a +// test. removeWindowsSandboxPrincipalForSetupFn and applyWindowsACLPlanFn +// already existed in windows_identity_windows.go for the same reason; the rest +// follow them, and the ACL call below now goes through that one so a single stub +// covers both plans. +var ( + windowsProcessIsElevatedFn = windowsProcessIsElevated + acquireWindowsSandboxSetupLockFn = acquireWindowsSandboxSetupLock + applyWindowsNetworkPlanFn = applyWindowsNetworkPlan + setupWindowsSandboxPrincipalFn = setupWindowsSandboxPrincipal + writeWindowsSandboxSetupMarkerFn = WriteWindowsSandboxSetupMarker +) + func runWindowsSandboxSetup(config WindowsSandboxSetupConfig, stderr io.Writer) int { // Do not provision DenyRead ACLs for a token mode that cannot launch normal // tools with DenyRead under the narrow restricting-SID set (PR #640). @@ -19,29 +40,222 @@ func runWindowsSandboxSetup(config WindowsSandboxSetupConfig, stderr io.Writer) // Applying the WFP network filters and workspace ACLs requires Administrator // rights; without them WFP fails deep inside with a raw ACCESS_DENIED (0x5). // Check up front and return an actionable message instead. - if !windowsProcessIsElevated() { + if !windowsProcessIsElevatedFn() { fmt.Fprintln(stderr, WindowsSandboxSetupName+": Administrator rights are required. Re-run `zero sandbox setup` from an elevated (Run as administrator) terminal.") return 1 } + // Serialize the WHOLE transaction, not the individual writes. + // + // Setup rotates the principal's password, stores the secret, + // read-modify-writes the ACL ledger, installs WFP filters and writes the + // marker. Two setups for this workspace at once interleave those steps: the + // account ends up with one process's password while the stored secret is the + // other's, so every later command fails to log on, and the ledger's + // read-modify-write silently loses a path set. Each write being individually + // atomic does nothing about interleaving BETWEEN writes, which is the actual + // failure, so the lock is taken here and held to the end. + // + // Immediately after the elevation check, because acquiring a Global object + // needs the rights that check just confirmed. + lock, err := acquireWindowsSandboxSetupLockFn(windowsSandboxWorkspaceKey(config.WorkspaceRoots)) + if err != nil { + fmt.Fprintln(stderr, WindowsSandboxSetupName+": "+err.Error()) + return 1 + } + defer lock.release() + if lock.Abandoned() { + // The previous holder died somewhere inside the transaction, so this + // machine may be half set up. Setup is idempotent and re-running it is + // the repair, but say so: a silent recovery hides that something crashed + // while holding the machine's sandbox state open. + fmt.Fprintf(stderr, "%s: a previous setup for this workspace exited without finishing; re-running to repair it.\n", + WindowsSandboxSetupName) + } + + // Before anything is provisioned: a principal this helper could not hand back + // to the caller must not be created at all. + if config.PrincipalOptIn { + // THE SAME REFUSAL THE ARGUMENT BUILDER MAKES, AT THE BOUNDARY THAT MUTATES. + // BuildWindowsSandboxSetupArgs refuses to serialize a principal opt-in while + // no launch path exists, but both shipped entrypoints reach this function + // through the parser, and handwritten `--sandbox-principal 1` arrived here + // with nothing in the way of account, secret and ACL provisioning for a + // principal nothing can run. Arguments are an input format, not proof that + // the builder approved them. + if err := windowsPrincipalLaunchAvailable(); err != nil { + fmt.Fprintln(stderr, WindowsSandboxSetupName+": refusing to provision a sandbox principal: "+err.Error()) + return 1 + } + if err := assertWindowsSetupRunsAsCaller(config.CallerSID); err != nil { + fmt.Fprintln(stderr, WindowsSandboxSetupName+": "+err.Error()) + return 1 + } + } + + // Create the runtime roots BEFORE building the plan that grants them. + // + // The capability plan deliberately does not materialize a write root: an + // entry whose path is absent is a typo or a stale config, and inventing the + // tree would silently grant write on a directory nobody asked for. So every + // write root in the plan has to exist by the time it is applied, and the + // runtime roots are the only ones setup owns rather than receives. + // + // Both of them, not the one this process would select. Setup grants the whole + // candidate set so a later command that falls back still lands on a + // provisioned tree, and granting a path setup never created fails the entire + // run with "windows ACL target does not exist". + // + // And recoverable. Preparing an EXISTING candidate rewrites its DACL, which + // strips the grant a previous setup put there, before the ACL transaction + // below has a baseline to restore. Record what was there first and put it + // back on every failure between here and a completed apply. + candidateDACLs, err := snapshotWindowsRuntimeCandidateDACLs(windowsSandboxRuntimeCandidates(config.WorkspaceRoots)) + if err != nil { + fmt.Fprintln(stderr, WindowsSandboxSetupName+": "+err.Error()) + return 1 + } + restoreCandidateDACLs := func() { + if err := restoreWindowsRuntimeCandidateDACLs(candidateDACLs); err != nil { + fmt.Fprintln(stderr, WindowsSandboxSetupName+": restoring runtime root permissions after the failure: "+err.Error()) + } + } + if err := ensureWindowsSandboxRuntimeCandidates(config.WorkspaceRoots); err != nil { + fmt.Fprintln(stderr, WindowsSandboxSetupName+": "+err.Error()) + restoreCandidateDACLs() + return 1 + } + plan, err := BuildWindowsACLPlan(config.commandConfig()) if err != nil { fmt.Fprintln(stderr, WindowsSandboxSetupName+": "+err.Error()) + restoreCandidateDACLs() return 1 } - // Always provision the mode-INDEPENDENT infrastructure: the outbound block - // filters scoped to the offline-marker SID. Runtime gates network per command - // by whether the token carries that SID, so one setup serves both modes. - networkPlan, err := BuildWindowsNetworkInfraPlan(config.commandConfig()) + // Before the first mutation, and on this tier too. Elevated setup CAN write a + // volume root's DACL, which is exactly why it must not: the grant is + // inheritable, so applying it rewrites permissions across the drive rather than + // changing one sandbox-owned object. + if refusal := WindowsACLPlanReadGrantRefusal(plan); refusal != "" { + fmt.Fprintln(stderr, WindowsSandboxSetupName+": "+refusal) + restoreCandidateDACLs() + return 1 + } + rollback, err := applyWindowsACLPlanFn(plan) if err != nil { fmt.Fprintln(stderr, WindowsSandboxSetupName+": "+err.Error()) + restoreCandidateDACLs() return 1 } - rollback, err := applyWindowsACLPlan(plan) + // The transaction's own baseline is the post-preparation DACL, so a later + // failure that rolls the apply back still needs the pre-preparation one put + // back after it. + appliedRollback := rollback + rollback = func() error { + err := appliedRollback() + restoreCandidateDACLs() + return err + } + // Provision this workspace's sandbox principal, when opted in. A principal is + // a separate local account, so it is created only on an explicit opt-in: it + // is visible in `net user`, and account creation is exactly the kind of thing + // endpoint protection and enterprise policy object to. Without the opt-in the + // capability-SID backend above is the whole of setup, unchanged. + if windowsSandboxIdentityEnabled(config.commandConfig().Env) { + principalRollback, err := setupWindowsSandboxPrincipalFn(config.commandConfig()) + if err != nil { + if rollbackErr := rollback(); rollbackErr != nil { + fmt.Fprintf(stderr, "%s: %v; rollback failed: %v\n", WindowsSandboxSetupName, err, rollbackErr) + return 1 + } + fmt.Fprintln(stderr, WindowsSandboxSetupName+": "+err.Error()) + return 1 + } + // Fold the principal into the existing rollback so every later failure + // path undoes it too, rather than each one having to remember. + aclRollback := rollback + rollback = func() error { + principalErr := principalRollback() + aclErr := aclRollback() + if principalErr != nil { + return principalErr + } + return aclErr + } + } else if err := removeWindowsSandboxPrincipalsForSetupFn(config.commandConfig()); err != nil { + // Opting out has to actually retire the principal, because that is what + // we tell people it does. ValidateWindowsSandboxSetupMarker sends an + // operator here in as many words: re-run setup from an elevated terminal + // without the opt-in to retire the principal. Until now this branch did + // not exist, so the marker flipped to opted-out while the account, its + // secret, its logon rights, its ACEs and its ledger all stayed exactly + // where they were. The instruction was a lie, and the leftovers were + // invisible, because nothing afterwards looks for a principal it believes + // was never provisioned. + // + // Fatal only when the ACCOUNT survived, which is a narrower rule than + // either of the two this branch has had. + // + // It first printed and carried on, which was wrong: a marker that says "no + // principal" while the account is still installed is a lie nothing + // afterwards can detect, because nothing looks for a principal it believes + // was never provisioned. + // + // Then it failed on any error at all, which was wrong in the other + // direction. removeWindowsSandboxPrincipalForSetup deliberately continues + // past a secret file owned by ANOTHER administrator's setup and reports it + // at the end, so that one unlinkable file does not strand the account, its + // logon rights and its ACEs. Treating that report as fatal made the whole + // default sandbox unsetupable on such a machine, permanently, over inert + // residue: the account is already gone and the leftover is a file with a + // DACL we do not own. + // + // So the question is not whether teardown reported a problem, it is + // whether the thing the marker is about to deny is still there. Asked of + // the account rather than inferred from the error, because the error is + // a join and its parts are not separable at this distance. + if windowsSandboxPrincipalIsInstalled(config.commandConfig()) { + if rollbackErr := rollback(); rollbackErr != nil { + fmt.Fprintf(stderr, "%s: opted out, but the existing sandbox principal is still installed: %v; rollback failed: %v\n", + WindowsSandboxSetupName, err, rollbackErr) + return 1 + } + fmt.Fprintf(stderr, "%s: opted out, but the existing sandbox principal is still installed: %v\n", + WindowsSandboxSetupName, err) + return 1 + } + // Retired, with something inert left over. Named rather than swallowed: + // whoever has to clean it up needs to know it is there. + fmt.Fprintf(stderr, "%s: opted out; the sandbox principal was retired, but some of its residue could not be removed: %v\n", + WindowsSandboxSetupName, err) + } + // Built AFTER any principal provisioning, not before. The block filters are + // keyed to the offline group as well as the offline-marker SID, and that group + // is created as part of provisioning: planning first would install filters + // that name only the marker, leaving every offline principal with an open + // network while looking correctly set up. + // + // Mode-INDEPENDENT by design. Which identity a command runs under is what + // selects allow or deny, so one setup serves both modes. + networkPlan, err := BuildWindowsNetworkInfraPlan(config.commandConfig()) if err != nil { + if rollbackErr := rollback(); rollbackErr != nil { + fmt.Fprintf(stderr, "%s: %v; rollback failed: %v\n", WindowsSandboxSetupName, err, rollbackErr) + return 1 + } fmt.Fprintln(stderr, WindowsSandboxSetupName+": "+err.Error()) return 1 } - if err := applyWindowsNetworkPlan(networkPlan); err != nil { + // Refuse to install filters that would not cover the principals just + // provisioned. The ordering above is what makes them cover it, and nothing at + // this call site shows that, so assert it rather than trusting it to survive + // a later refactor. Installing anyway would leave a machine reporting a + // successful setup while every offline principal had an open network. + // Gated on the SAME opt-in that decides whether principals were provisioned + // at all (line 36). The offline group is created inside provisioning, so on + // an opt-out machine it legitimately does not exist, and asserting there + // would refuse a setup that has nothing to get wrong. + if err := assertWindowsNetworkPlanCoversOfflineGroup(networkPlan, resolveWindowsSandboxOfflineGroupSIDHook, + windowsSandboxIdentityEnabled(config.commandConfig().Env)); err != nil { if rollbackErr := rollback(); rollbackErr != nil { fmt.Fprintf(stderr, "%s: %v; rollback failed: %v\n", WindowsSandboxSetupName, err, rollbackErr) return 1 @@ -49,7 +263,29 @@ func runWindowsSandboxSetup(config WindowsSandboxSetupConfig, stderr io.Writer) fmt.Fprintln(stderr, WindowsSandboxSetupName+": "+err.Error()) return 1 } - if _, err := WriteWindowsSandboxSetupMarker(config); err != nil { + // Install the machine's plan, not this home's. The filters are global and are + // replaced wholesale on every setup, so an opted-out run here would otherwise + // strip the offline group SID that another workspace's principals depend on + // and hand them egress under a NetworkDeny profile. networkPlan itself is left + // alone because it is what this home's marker fingerprints. + planToInstall, planErr := WindowsNetworkPlanForApply(networkPlan, resolveWindowsSandboxOfflineGroupSIDHook) + if planErr != nil { + if rollbackErr := rollback(); rollbackErr != nil { + fmt.Fprintf(stderr, "%s: %v; rollback failed: %v\n", WindowsSandboxSetupName, planErr, rollbackErr) + return 1 + } + fmt.Fprintln(stderr, WindowsSandboxSetupName+": "+planErr.Error()) + return 1 + } + if err := applyWindowsNetworkPlanFn(planToInstall); err != nil { + if rollbackErr := rollback(); rollbackErr != nil { + fmt.Fprintf(stderr, "%s: %v; rollback failed: %v\n", WindowsSandboxSetupName, err, rollbackErr) + return 1 + } + fmt.Fprintln(stderr, WindowsSandboxSetupName+": "+err.Error()) + return 1 + } + if _, err := writeWindowsSandboxSetupMarkerFn(config); err != nil { if rollbackErr := rollback(); rollbackErr != nil { fmt.Fprintf(stderr, "%s: %v; rollback failed: %v\n", WindowsSandboxSetupName, err, rollbackErr) return 1 @@ -60,6 +296,94 @@ func runWindowsSandboxSetup(config WindowsSandboxSetupConfig, stderr io.Writer) return 0 } +// windowsSandboxPrincipalIsInstalled reports whether this workspace's principal +// account still resolves. +// +// Fails CLOSED. Only an outright "not provisioned" counts as gone; a name that +// resolves to something unexpected, or a lookup that fails for its own reasons, +// is treated as still installed. The consumer uses this to decide whether an +// opted-out marker would be a lie, and the expensive mistake there is claiming a +// principal is gone when nobody actually checked. +// A workspace now has TWO principals, and either one still installed makes the +// answer yes. Checking a single role would let an opted-out marker claim the +// workspace is clean while the other account is still on the machine, which is +// the exact lie this guard exists to prevent. +func windowsSandboxPrincipalIsInstalled(config WindowsSandboxCommandConfig) bool { + // EVERY RETIRABLE ROLE, not just the live pair. Teardown retires legacy too, + // so an upgraded machine still holding the untagged pre-split account would + // otherwise be reported clean here while that account was still installed, + // and the opted-out marker would be exactly the lie this guard prevents. + key := windowsSandboxPrincipalKey(config) + for _, role := range windowsSandboxRetirableRoles { + if _, err := lookupWindowsSandboxIdentityFn(key, role); !errors.Is(err, errWindowsSandboxIdentityUnavailable) { + return true + } + } + return false +} + +// assertWindowsSetupRunsAsCaller refuses to provision a sandbox principal when +// this elevated helper belongs to a different Windows user than the caller that +// launched it. +// +// A principal is unusable across that boundary, and not for a reason more +// plumbing can fix. Its password is sealed with CryptProtectData, which derives +// its key from the CALLING user, so a blob written here is decryptable only by +// this administrator; the caller's next command would find the account, find the +// secret file, and fail to unseal it. Threading the caller's SID through the +// setup args fixes the account NAME, the ledger and the ACEs — everything that +// is merely named after the invoking user — but it cannot re-key DPAPI for a +// user whose token this process does not hold. +// +// So the answer is to say so before anything exists, rather than provision an +// account, a secret and a grant that resolve to a sandbox nobody can log into. +// +// Same-account elevation is unaffected: a UAC consent prompt splits the caller's +// token but leaves the user SID identical, which is the ordinary path. An +// unknown caller SID (an older caller, or a token query that failed) is treated +// as a match: that is the pre-existing behaviour, and refusing on absence would +// break setup on machines where nothing is wrong. +// +// What this does NOT do, stated plainly because the name suggests otherwise: +// it cannot fire on any path Zero itself takes. Zero never elevates the helper. +// runSandboxSetupHelper (internal/cli/app.go) is a plain exec.Command with no +// token work, and runWindowsSandboxSetup refuses unless the process is ALREADY +// elevated, so the operator supplies elevation by opening an elevated terminal. +// Both halves then run in that terminal: BuildWindowsSandboxSetupArgs resolves +// the caller SID from the very process that later becomes the helper, so the two +// are equal by construction and this returns nil every time. +// +// It is kept rather than deleted because it is a correctness assertion, not a +// no-op. It becomes load-bearing the moment anyone adds a ShellExecute "runas" +// path, which is the natural way this command grows, and it already covers a +// helper .exe invoked directly with hand-written args. What it must not do is be +// mistaken for a live defence against over-the-shoulder elevation today. +// +// The residual hazard that boundary was supposed to describe is real but has a +// different shape: an operator who elevates a terminal as a DIFFERENT +// administrator runs both halves as that administrator, so setup provisions into +// THAT account's sandbox home. Their own unelevated session resolves its own +// home, finds no marker, and is told to run setup. Loud, not silent, and not a +// weakened sandbox. +// +// Scoped to the opt-in by its caller, deliberately. The default restricted-token +// sandbox stores no secret and names no account after the user, so it works +// perfectly well across this boundary and must keep doing so. +func assertWindowsSetupRunsAsCaller(callerSID string) error { + caller := strings.TrimSpace(callerSID) + if caller == "" { + return nil + } + current := windowsCurrentUserSID() + if current == "" || strings.EqualFold(current, caller) { + return nil + } + return fmt.Errorf("this elevated setup is running as a different Windows user (%s) than the one that started it (%s), "+ + "and a sandbox principal's password is sealed to the user that stores it, so the account provisioned here could never be used. "+ + "Re-run `zero sandbox setup` from a terminal elevated as your own account, or unset %s to use the restricted-token sandbox", + current, caller, windowsSandboxIdentityEnv) +} + // windowsProcessIsElevated reports whether the current process runs with an // elevated (Administrator) token. On any error obtaining the token it returns // true so the setup proceeds and surfaces the real WFP/ACL error rather than a diff --git a/internal/sandbox/windows_stale_ace_windows_test.go b/internal/sandbox/windows_stale_ace_windows_test.go new file mode 100644 index 000000000..21497aff9 --- /dev/null +++ b/internal/sandbox/windows_stale_ace_windows_test.go @@ -0,0 +1,274 @@ +//go:build windows + +package sandbox + +import ( + "os" + "path/filepath" + "testing" + "unsafe" + + "golang.org/x/sys/windows" +) + +// applyWindowsACLPlan merges into the existing DACL, so narrowing a policy and +// re-running setup used to leave the wider ACEs in place next to the new ones. +// The principal kept access the current policy no longer grants — a silent +// widening of the sandbox produced by tightening it. +func TestRevokeDropsStalePrincipalACEsBeforeReapply(t *testing.T) { + root := t.TempDir() + kept := filepath.Join(root, "kept") + dropped := filepath.Join(root, "dropped") + for _, dir := range []string{kept, dropped} { + if err := os.MkdirAll(dir, 0o700); err != nil { + t.Fatalf("mkdir %s: %v", dir, err) + } + } + // Guests: a real, resolvable trustee that the test process is not a member + // of, so the ACEs below are observable without affecting this process. + principal := "S-1-5-32-546" + + // First setup: both roots writable. + wide, err := buildWindowsPrincipalACLPlan(windowsPrincipalACLInput{ + PrincipalSID: principal, + WriteRoots: []WritableRoot{{Root: kept}, {Root: dropped}}, + }) + if err != nil { + t.Fatalf("wide plan: %v", err) + } + if _, err := applyWindowsACLPlan(wide); err != nil { + t.Fatalf("apply wide plan: %v", err) + } + if !hasACEForTrustee(t, dropped, principal) { + t.Fatal("precondition: the wide plan should have granted the root it is about to lose") + } + + // Policy narrows: "dropped" is no longer a write root. + narrow, err := buildWindowsPrincipalACLPlan(windowsPrincipalACLInput{ + PrincipalSID: principal, + WriteRoots: []WritableRoot{{Root: kept}}, + }) + if err != nil { + t.Fatalf("narrow plan: %v", err) + } + // Revocation has to cover the paths the OLD plan touched, not just the new + // one — the whole point is the path that left the policy. + if _, err := revokeWindowsPrincipalACEs(principal, windowsACLPlanPaths(wide)); err != nil { + t.Fatalf("revoke: %v", err) + } + if _, err := applyWindowsACLPlan(narrow); err != nil { + t.Fatalf("apply narrow plan: %v", err) + } + + if hasACEForTrustee(t, dropped, principal) { + t.Error("principal kept its grant on a root the narrowed policy removed") + } + if !hasACEForTrustee(t, kept, principal) { + t.Error("revocation also dropped the grant the narrowed policy still wants") + } +} + +// Revoking a path that was never created is cleanup with nothing to clean, not +// an error — setup would otherwise fail on any carveout git has not made yet. +func TestRevokeIgnoresPathsThatDoNotExist(t *testing.T) { + missing := filepath.Join(t.TempDir(), "never-created") + if _, err := revokeWindowsPrincipalACEs("S-1-5-32-546", []string{missing}); err != nil { + t.Fatalf("revoke over a missing path: %v", err) + } +} + +func hasACEForTrustee(t *testing.T, path string, trustee string) bool { + t.Helper() + descriptor, err := windows.GetNamedSecurityInfo(path, windows.SE_FILE_OBJECT, windows.DACL_SECURITY_INFORMATION) + if err != nil { + t.Fatalf("GetNamedSecurityInfo(%s): %v", path, err) + } + dacl, _, err := descriptor.DACL() + if err != nil { + t.Fatalf("DACL(%s): %v", path, err) + } + if dacl == nil { + return false + } + want, err := windows.StringToSid(trustee) + if err != nil { + t.Fatalf("StringToSid(%s): %v", trustee, err) + } + // Deny ACEs count here as much as allow ACEs: revocation is by trustee and + // drops both, so an assertion that only saw allows would call a leftover + // deny "revoked". + for index := uint32(0); index < uint32(dacl.AceCount); index++ { + var header *windows.ACE_HEADER + if err := windows.GetAce(dacl, index, (**windows.ACCESS_ALLOWED_ACE)(unsafe.Pointer(&header))); err != nil { + continue + } + ace := (*windows.ACCESS_ALLOWED_ACE)(unsafe.Pointer(header)) + switch ace.Header.AceType { + case windows.ACCESS_ALLOWED_ACE_TYPE, windows.ACCESS_DENIED_ACE_TYPE: + default: + continue + } + if (*windows.SID)(unsafe.Pointer(&ace.SidStart)).Equals(want) { + return true + } + } + return false +} + +// The mechanism working is not the same as the production path using it. This +// pins the call site and its ORDER: revocation is only worth anything if it +// runs before the plan that re-adds the current grants. +func TestApplyPrincipalACLsRevokesBeforeApplying(t *testing.T) { + prevApply := applyWindowsACLPlanFn + t.Cleanup(func() { applyWindowsACLPlanFn = prevApply }) + + var actions []WindowsACLAction + applyWindowsACLPlanFn = func(plan WindowsACLPlan) (func() error, error) { + if len(plan.Entries) > 0 { + actions = append(actions, plan.Entries[0].Action) + } + return func() error { return nil }, nil + } + + workspace := t.TempDir() + filesystem := FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{Root: workspace}}, + } + if _, err := applyWindowsPrincipalACLs(t.TempDir(), "zero-sbx-test", "S-1-5-32-546", filesystem, filesystem.WriteRoots); err != nil { + t.Fatalf("applyWindowsPrincipalACLs: %v", err) + } + + if len(actions) != 2 { + t.Fatalf("saw %d ACL plans (%v), want a revocation then the grants", len(actions), actions) + } + if actions[0] != windowsACLRevoke { + t.Errorf("first plan was %q, want the trustee revocation to go first", actions[0]) + } + if actions[1] == windowsACLRevoke { + t.Error("second plan was another revocation; the current grants were never applied") + } +} + +// The revocation's rollback has to be returned, not discarded. +// +// Discarding it was justified on the grounds that the only failure path from +// applyWindowsPrincipalACLs removes the principal outright, so restoring ACEs +// for a doomed account would be pointless. That holds only for a principal the +// run CREATED. #812 keeps an ADOPTED principal alive on failure rather than +// destroying a working account someone else provisioned — and then the discarded +// snapshot left it logged-on and unable to reach its own workspace, with its +// previous ACEs revoked and the new ones rolled back. +func TestApplyPrincipalACLsRollbackRestoresTheRevokedACEs(t *testing.T) { + prevApply := applyWindowsACLPlanFn + t.Cleanup(func() { applyWindowsACLPlanFn = prevApply }) + + var applied []WindowsACLAction + var reverted []WindowsACLAction + applyWindowsACLPlanFn = func(plan WindowsACLPlan) (func() error, error) { + action := WindowsACLAction("") + if len(plan.Entries) > 0 { + action = plan.Entries[0].Action + } + applied = append(applied, action) + return func() error { + reverted = append(reverted, action) + return nil + }, nil + } + + workspace := t.TempDir() + filesystem := FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{Root: workspace}}, + } + rollback, err := applyWindowsPrincipalACLs(t.TempDir(), "zero-sbx-test", "S-1-5-32-546", filesystem, filesystem.WriteRoots) + if err != nil { + t.Fatalf("applyWindowsPrincipalACLs: %v", err) + } + if len(applied) != 2 || applied[0] != windowsACLRevoke { + t.Fatalf("applied %v, want a revocation then the grants", applied) + } + if err := rollback(); err != nil { + t.Fatalf("rollback: %v", err) + } + + // Both halves must unwind, and in reverse order: the grant comes off first, + // then the ACEs the revocation removed go back. + if len(reverted) != 2 { + t.Fatalf("rollback reverted %v, want both the grant and the revocation", reverted) + } + if reverted[0] == windowsACLRevoke { + t.Error("rollback undid the revocation before the grant; the grant would survive") + } + if reverted[1] != windowsACLRevoke { + t.Errorf("rollback never restored the revoked ACEs, got %v", reverted) + } +} + +// THE LEDGER IS PART OF THE ROLLBACK, and nothing pinned it. +// +// applyWindowsPrincipalACLs narrows the ledger before the whole setup +// transaction has succeeded. If network setup or the marker write then fails, +// rollback restores the DACLs, and the ledger has to come back with them: a +// ledger that still records only the narrowed path set leaves the restored +// grants unnamed, so later cleanup cannot find them. +// +// The neighbouring test asserts only the ORDER of the two ACL reverts and never +// reads the ledger afterwards, so the restore could be deleted and the suite +// would stay green. +func TestApplyPrincipalACLsRollbackRestoresTheLedger(t *testing.T) { + prevApply := applyWindowsACLPlanFn + t.Cleanup(func() { applyWindowsACLPlanFn = prevApply }) + applyWindowsACLPlanFn = func(WindowsACLPlan) (func() error, error) { + return func() error { return nil }, nil + } + + sandboxHome := t.TempDir() + const username = "zero-sbx-test" + + // A wider path set than the new plan will record, standing in for a previous + // policy's grants. + stale := []string{filepath.Join(t.TempDir(), "previously-granted")} + if err := writeWindowsPrincipalACLLedger(sandboxHome, username, stale); err != nil { + t.Fatalf("seed the ledger: %v", err) + } + + workspace := t.TempDir() + filesystem := FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{Root: workspace}}, + } + rollback, err := applyWindowsPrincipalACLs(sandboxHome, username, "S-1-5-32-546", filesystem, filesystem.WriteRoots) + if err != nil { + t.Fatalf("applyWindowsPrincipalACLs: %v", err) + } + + narrowed, ok := readWindowsPrincipalACLLedger(sandboxHome, username) + if !ok { + t.Fatal("no ledger after apply") + } + if slicesContainsString(narrowed, stale[0]) { + t.Fatalf("the ledger still names the old path after apply, so this test is not exercising the restore: %v", narrowed) + } + + if err := rollback(); err != nil { + t.Fatalf("rollback: %v", err) + } + restored, ok := readWindowsPrincipalACLLedger(sandboxHome, username) + if !ok { + t.Fatal("no ledger after rollback") + } + if !slicesContainsString(restored, stale[0]) { + t.Errorf("rollback restored the DACLs but left the narrowed ledger %v; the path %q it put back is now unnamed, so cleanup cannot find it", restored, stale[0]) + } +} + +func slicesContainsString(values []string, want string) bool { + for _, value := range values { + if value == want { + return true + } + } + return false +} diff --git a/internal/sandbox/windows_stale_secret_windows_test.go b/internal/sandbox/windows_stale_secret_windows_test.go new file mode 100644 index 000000000..222b8fde3 --- /dev/null +++ b/internal/sandbox/windows_stale_secret_windows_test.go @@ -0,0 +1,99 @@ +//go:build windows + +package sandbox + +import ( + "errors" + "strings" + "testing" + + "golang.org/x/sys/windows" +) + +// "Absent beats stale" is the invariant the rollback's secret removal exists to +// keep: the command path treats a missing secret as not-provisioned and falls +// back, while a stale one fails the logon and reports a broken sandbox. +// +// So when the removal itself fails after a password rotation, the invariant was +// NOT restored — a credential for a password that no longer works is still on +// disk. Swallowing that error claims otherwise, and the operator finds out on +// the next command instead of from the setup that broke it. +func TestProvisionSurfacesFailedStaleSecretCleanup(t *testing.T) { + for name, testCase := range map[string]struct { + rotate bool + removeErr error + wantInErr string + wantRemove bool + }{ + "rotation happened and the stale secret cannot be removed": { + rotate: true, removeErr: errors.New("access is denied"), + wantRemove: true, wantInErr: "stale and could not be removed", + }, + "rotation happened and cleanup succeeds": { + rotate: true, wantRemove: true, + }, + } { + t.Run(name, func(t *testing.T) { + prevProvision := provisionWindowsSandboxIdentityFn + prevRemove := removeWindowsSandboxSecretFn + prevReset := resetWindowsSandboxUserPasswordFn + prevGrant := grantWindowsSandboxLogonRightsFn + prevWrite := writeWindowsSandboxSecretFn + t.Cleanup(func() { + provisionWindowsSandboxIdentityFn = prevProvision + removeWindowsSandboxSecretFn = prevRemove + resetWindowsSandboxUserPasswordFn = prevReset + grantWindowsSandboxLogonRightsFn = prevGrant + writeWindowsSandboxSecretFn = prevWrite + }) + + sid, err := windows.CreateWellKnownSid(windows.WinLocalSystemSid) + if err != nil { + t.Fatal(err) + } + // created=false so the run ADOPTS an account and rotation applies. + provisionWindowsSandboxIdentityFn = func(string, windowsSandboxRole) (windowsSandboxIdentity, string, bool, error) { + return windowsSandboxIdentity{Username: "zero-sbx-test", SID: sid}, "pw", false, nil + } + grantWindowsSandboxLogonRightsFn = func(*windows.SID) error { return nil } + removed := false + removeWindowsSandboxSecretFn = func(string) error { + removed = true + return testCase.removeErr + } + // Rotate, then fail immediately after so undo runs with rotated=true. + resetWindowsSandboxUserPasswordFn = func(string, string) error { + if testCase.rotate { + return nil + } + return errors.New("no rotation") + } + + // The only step after rotation; failing it is what drives undo with + // rotated=true, which is the state the invariant is about. + writeWindowsSandboxSecretFn = func(string, string) error { + return errors.New("secret store refused") + } + + config := WindowsSandboxCommandConfig{ + SandboxHome: t.TempDir(), + CommandCWD: `C:\ws`, + WorkspaceRoots: []string{`C:\ws`}, + } + _, _, err = provisionWindowsSandboxPrincipalForSetup(config, windowsSandboxRoleOffline) + + if removed != testCase.wantRemove { + t.Fatalf("stale secret removal attempted = %v, want %v", removed, testCase.wantRemove) + } + if testCase.wantInErr == "" { + return + } + if err == nil { + t.Fatal("a failed stale-secret cleanup was swallowed") + } + if !strings.Contains(err.Error(), testCase.wantInErr) { + t.Fatalf("error = %q, want it to mention %q", err, testCase.wantInErr) + } + }) + } +} diff --git a/internal/sandbox/windows_token_windows.go b/internal/sandbox/windows_token_windows.go index 1cf360948..77c8863bb 100644 --- a/internal/sandbox/windows_token_windows.go +++ b/internal/sandbox/windows_token_windows.go @@ -48,6 +48,54 @@ func (sid windowsLocalSID) close() { } } +// restrictWindowsTokenForCapabilitySIDs applies the same write jail to an +// arbitrary base token that createWindowsRestrictedTokenForCapabilitySIDs +// applies to the calling process's own. +// +// The sandbox principal path needs this. A LogonUser token is a full token for +// that account: the ACL plan can deny it at named paths, but it cannot revoke +// what the account's ambient memberships already grant, so an opted-in command +// could still write any path whose DACL admits BUILTIN\Users, Authenticated +// Users, or NT AUTHORITY\BATCH - C:\Users\Public\Documents being the obvious +// one - regardless of the profile's write roots. +// +// The caller must include the principal's OWN SID among the capability SIDs. +// The plan grants the workspace to that SID rather than to a capability SID, so +// without it the restricted-SID check has nothing to match and the principal +// loses its own workspace: a jail that locks out the inmate and no one else. +func restrictWindowsTokenForCapabilitySIDs(base windows.Token, capabilitySIDStrings []string, writeRestricted bool) (windows.Token, error) { + capabilitySIDs, err := parseWindowsCapabilitySIDs(capabilitySIDStrings) + if err != nil { + return 0, err + } + defer func() { + for _, sid := range capabilitySIDs { + sid.close() + } + }() + return createWindowsRestrictedTokenFromBase(base, capabilitySIDs, writeRestricted) +} + +// parseWindowsCapabilitySIDs converts SID strings, closing what it already +// allocated if one fails to parse. +func parseWindowsCapabilitySIDs(values []string) ([]windowsLocalSID, error) { + if len(values) == 0 { + return nil, errors.New("windows restricted token requires at least one capability SID") + } + parsed := make([]windowsLocalSID, 0, len(values)) + for _, value := range values { + sid, err := newWindowsLocalSID(value) + if err != nil { + for _, existing := range parsed { + existing.close() + } + return nil, fmt.Errorf("parse windows capability SID %q: %w", value, err) + } + parsed = append(parsed, sid) + } + return parsed, nil +} + func createWindowsRestrictedTokenForCapabilitySIDs(capabilitySIDStrings []string, writeRestricted bool) (windows.Token, error) { if len(capabilitySIDStrings) == 0 { return 0, errors.New("windows restricted token requires at least one capability SID") diff --git a/internal/sandbox/windows_unelevated.go b/internal/sandbox/windows_unelevated.go index b7cb05f53..6b354b2b1 100644 --- a/internal/sandbox/windows_unelevated.go +++ b/internal/sandbox/windows_unelevated.go @@ -146,3 +146,7 @@ func recordWindowsUnelevatedAppliedPlan(sandboxHome string, applied WindowsUnele } return nil } + +// windowsACLPlanDeniedPath lives in windows_command_runner_windows.go, beside +// its only caller. Defining it here, in the portable file, would make it dead +// code on every non-Windows build and fail the static analysis gate. diff --git a/internal/sandbox/windows_unelevated_candidates_windows_test.go b/internal/sandbox/windows_unelevated_candidates_windows_test.go new file mode 100644 index 000000000..db74eff01 --- /dev/null +++ b/internal/sandbox/windows_unelevated_candidates_windows_test.go @@ -0,0 +1,105 @@ +//go:build windows + +package sandbox + +import ( + "os" + "path/filepath" + "strings" + "testing" +) + +// THE UNELEVATED PLAN CAN BE APPLIED ON FRESH ROOTS. +// +// The unelevated command's plan grants both runtime candidates, but nothing on +// that tier created the one this process does not select, so with a fresh +// usable cache the runner refused its own plan with "windows ACL target does +// not exist" and the command never launched. The parent now creates whichever +// candidates are missing, with setup's own ownership and no-follow checks, +// before the runner applies. Driven through the real planner and the real +// runner-side apply, on fresh roots, with the preferred candidate selected and +// with the fallback selected. +func TestUnelevatedPlanAppliesOnFreshRuntimeRoots(t *testing.T) { + for name, cacheInsideWorkspace := range map[string]bool{"preferred candidate": false, "fallback selected": true} { + t.Run(name, func(t *testing.T) { + isolateSandboxRuntimeRoots(t) + config := testWindowsUnelevatedCommandConfig(t) + if cacheInsideWorkspace { + original := sandboxUserCacheDir + sandboxUserCacheDir = func() (string, error) { return filepath.Join(config.WorkspaceRoots[0], ".cache"), nil } + t.Cleanup(func() { sandboxUserCacheDir = original }) + } + config.PermissionProfile = windowsSandboxProfileWithRuntime(config.PermissionProfile, config.WorkspaceRoots) + candidates := windowsSandboxRuntimeCandidates(config.WorkspaceRoots) + if len(candidates) == 0 { + t.Fatal("SETUP INVALID: no runtime candidates") + } + for _, root := range candidates { + if _, err := os.Lstat(root); err == nil { + t.Fatalf("SETUP INVALID: candidate %s already exists", root) + } + } + + // What the parent does for this tier before the runner starts. + if err := ensureMissingWindowsSandboxRuntimeCandidates(config.WorkspaceRoots); err != nil { + t.Fatalf("ensureMissingWindowsSandboxRuntimeCandidates: %v", err) + } + for _, root := range candidates { + if info, err := os.Lstat(root); err != nil || !info.IsDir() { + t.Fatalf("candidate %s was not created: %v", root, err) + } + } + // What the runner does: build and apply the plan for real, as this user. + if err := ensureWindowsUnelevatedSetup(config); err != nil { + t.Fatalf("the unelevated plan could not be applied on fresh roots: %v", err) + } + }) + } +} + +// AN EXISTING CANDIDATE IS LEFT AS IT IS. +// +// This runs on every command, and an existing candidate carries the capability +// ACE the previous command applied. Re-preparing it would strip that grant and +// force a reapply on every launch. +func TestEnsureMissingCandidatesLeavesAnExistingOneAlone(t *testing.T) { + isolateSandboxRuntimeRoots(t) + config := testWindowsUnelevatedCommandConfig(t) + candidates := windowsSandboxRuntimeCandidates(config.WorkspaceRoots) + if len(candidates) != 2 { + t.Fatalf("SETUP INVALID: want two candidates, got %v", candidates) + } + existing := candidates[0] + if err := ensureRuntimeCandidateDir(existing); err != nil { + t.Fatal(err) + } + rollback, err := applyWindowsACLPlan(WindowsACLPlan{Entries: []WindowsACLEntry{{ + Action: WindowsACLAllowWrite, Path: existing, Capability: windowsACLTestDenySID, + }}}) + if err != nil { + t.Fatalf("grant the existing candidate: %v", err) + } + t.Cleanup(func() { _ = rollback() }) + + if err := ensureMissingWindowsSandboxRuntimeCandidates(config.WorkspaceRoots); err != nil { + t.Fatal(err) + } + if !runtimeRootGrants(t, existing, windowsACLTestDenySID) { + t.Fatalf("preparing the missing candidates rewrote the DACL of the existing one at %s", existing) + } + if info, err := os.Lstat(candidates[1]); err != nil || !info.IsDir() { + t.Fatalf("the missing candidate %s was not created: %v", candidates[1], err) + } + // A file wearing a candidate's name is refused by name, since the apply + // would otherwise put an ACE on it without complaint. + if err := os.RemoveAll(candidates[1]); err != nil { + t.Fatal(err) + } + if err := os.WriteFile(candidates[1], []byte("in the way"), 0o600); err != nil { + t.Fatal(err) + } + err = ensureMissingWindowsSandboxRuntimeCandidates(config.WorkspaceRoots) + if err == nil || !strings.Contains(err.Error(), candidates[1]) || !strings.Contains(err.Error(), "not a plain directory") { + t.Fatalf("a file standing in for a candidate was not refused by name: %v", err) + } +} diff --git a/internal/sandbox/windows_unelevated_denied_windows_test.go b/internal/sandbox/windows_unelevated_denied_windows_test.go new file mode 100644 index 000000000..8d9266eb3 --- /dev/null +++ b/internal/sandbox/windows_unelevated_denied_windows_test.go @@ -0,0 +1,111 @@ +//go:build windows + +package sandbox + +import ( + "errors" + "os" + "path/filepath" + "testing" + + "golang.org/x/sys/windows" +) + +// denyWriteDACTarget returns a disposable directory this process cannot re-DACL. +// +// A plain deny ACE is not enough: the owner keeps an implicit WRITE_DAC that +// defeats it, and the apply succeeds. OWNER_RIGHTS (S-1-3-4) is what replaces +// that implicit grant, so the DACL becomes the whole story and the access check +// actually fails. Probed rather than reasoned; the deny-only shape was measured +// to still let the mutation through. +func denyWriteDACTarget(t *testing.T) string { + t.Helper() + target := filepath.Join(t.TempDir(), "denied") + if err := os.MkdirAll(target, 0o700); err != nil { + t.Fatal(err) + } + ownerRights, err := windows.StringToSid("S-1-3-4") + if err != nil { + t.Fatalf("parse OWNER RIGHTS: %v", err) + } + // READ_CONTROL only: the applier can read the descriptor and then fails on the + // write, which is the exact production shape this pins. + dacl, err := windows.ACLFromEntries([]windows.EXPLICIT_ACCESS{{ + AccessPermissions: windows.READ_CONTROL | windows.SYNCHRONIZE | windows.FILE_GENERIC_READ, + AccessMode: windows.GRANT_ACCESS, + Inheritance: windows.NO_INHERITANCE, + Trustee: windows.TRUSTEE{ + TrusteeForm: windows.TRUSTEE_IS_SID, + TrusteeType: windows.TRUSTEE_IS_WELL_KNOWN_GROUP, + TrusteeValue: windows.TrusteeValueFromSID(ownerRights), + }, + }}, nil) + if err != nil { + t.Fatalf("build the deny-WRITE_DAC descriptor: %v", err) + } + if err := windows.SetNamedSecurityInfo( + target, + windows.SE_FILE_OBJECT, + windows.DACL_SECURITY_INFORMATION|windows.PROTECTED_DACL_SECURITY_INFORMATION, + nil, nil, dacl, nil, + ); err != nil { + t.Skipf("cannot rig a deny-WRITE_DAC directory here: %v", err) + } + return target +} + +// The diagnostic reads a path back out of an error string produced two +// functions away. That coupling is invisible to the compiler, so it is pinned +// here by driving the REAL producer rather than by hand-writing the message: +// if openWindowsACLTarget ever rewords its error, this fails instead of the +// diagnostic silently going quiet and users losing the one clue they had. +// +// THE TARGET IS DISPOSABLE, AND THAT IS THE POINT. +// +// This used to apply the real plan to C:\Windows\System32 and use the result as +// its privilege probe: on success it skipped, having already mutated. The +// snapshot and the applied result were both discarded at the call site, so +// nothing could put the descriptor back. Any process that does hold WRITE_DAC +// there would leave an inheritable allow ACE for a synthetic SID on System32 and +// everything later created beneath it, with no code path anywhere to remove it. +// +// The rigged directory below produces the same access denial from the same +// producer without depending on a system path or on the test's own privilege +// level, so the coupling stays pinned and no run can leave residue outside +// t.TempDir(). +func TestDeniedPathIsRecoveredFromARealApplyFailure(t *testing.T) { + target := denyWriteDACTarget(t) + + _, _, err := applyWindowsACLPathGroup(windowsACLPathGroup{ + Path: target, + Entries: []WindowsACLEntry{{ + Action: WindowsACLAllowWrite, + Path: target, + Capability: testPrincipalSID, + }}, + }) + if err == nil { + t.Fatal("SETUP INVALID: the apply succeeded on a directory rigged to refuse WRITE_DAC, so the denial path was not exercised") + } + if !errors.Is(err, os.ErrPermission) { + t.Skipf("failed for a reason other than access denial, nothing to extract here: %v", err) + } + + got := windowsACLPlanDeniedPath(err) + if got == "" { + t.Fatalf("no path recovered from a real access-denied apply failure, so the operator is told only that something was denied: %v", err) + } + if got != target { + t.Errorf("recovered %q, want %q", got, target) + } +} + +// Anything that is not an access denial must return empty, so the caller falls +// back to the generic message rather than naming an innocent path. +func TestDeniedPathIgnoresUnrelatedErrors(t *testing.T) { + for _, err := range []error{nil, errors.New(`open windows ACL target C:\somewhere: disk full`), os.ErrNotExist} { + if got := windowsACLPlanDeniedPath(err); got != "" { + t.Errorf("recovered %q from %v, want empty", got, err) + } + } +} diff --git a/internal/sandbox/windows_unelevated_reclaim_windows_test.go b/internal/sandbox/windows_unelevated_reclaim_windows_test.go new file mode 100644 index 000000000..f2740ac8d --- /dev/null +++ b/internal/sandbox/windows_unelevated_reclaim_windows_test.go @@ -0,0 +1,144 @@ +//go:build windows + +package sandbox + +import ( + "os" + "testing" +) + +// THE UNELEVATED MARKER ATTESTS TO A PLAN, NOT TO THE OBJECT THE COMMAND USES. +// +// cleanupSandboxRuntimeRoots reclaims inactive sibling runtime roots on an age +// and count policy. When the reclaimed workspace runs again, ordinary +// preparation recreates the SAME deterministic pathname with the caller-private +// DACL and no capability ACE. The serialized plan is unchanged, so the cached +// marker matched and setup returned before anything looked at the directory. The +// restricted child still carried the capability SID; the new object simply did +// not grant it, so every temp, package-cache and build-cache write failed after +// launch with a bare access denial and nothing pointing at setup. +// +// The restricted-token tier got an object check in this branch already. This one +// covers the tier that actually reaches it without elevation, and it must REAPPLY +// rather than refuse: this tier owns its plan, and telling an ordinary user to +// run elevated setup is both unnecessary and unfollowable for someone with no +// Administrator account. +// +// Driven through ensureWindowsUnelevatedSetup, which is what the command runner +// calls, rather than through the verifier helper: the helper was already correct, +// and the defect was that this path never consulted it. DACL edits on a +// user-owned directory need no Administrator rights, so this runs unelevated. +func TestUnelevatedSetupRestoresAReclaimedRuntimeRoot(t *testing.T) { + cacheRoot := t.TempDir() + previous := sandboxUserCacheDir + t.Cleanup(func() { sandboxUserCacheDir = previous }) + sandboxUserCacheDir = func() (string, error) { return cacheRoot, nil } + + workspace := t.TempDir() + runtimeRoot, ok := deterministicSandboxRuntimeRoot(canonicalSandboxWorkspaceRoot(workspace), canonicalSandboxWorkspaceRoot(cacheRoot)) + if !ok { + t.Fatalf("SETUP INVALID: no deterministic runtime root for %s", workspace) + } + if err := os.MkdirAll(runtimeRoot, 0o700); err != nil { + t.Fatal(err) + } + + config := WindowsSandboxCommandConfig{ + SandboxHome: t.TempDir(), + CommandCWD: workspace, + WorkspaceRoots: []string{workspace}, + SandboxLevel: WindowsSandboxLevelUnelevated, + PermissionProfile: PermissionProfile{ + Runtime: &SandboxRuntime{Root: runtimeRoot}, + FileSystem: FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{ + {Root: workspace}, + {Root: runtimeRoot}, + }, + ReadRoots: []string{workspace}, + }, + }, + } + + // First run provisions and records the marker. + if err := ensureWindowsUnelevatedSetup(config); err != nil { + t.Skipf("cannot run unelevated setup here: %v", err) + } + sid, err := windowsCapabilitySIDForWriteRoot(config, runtimeRoot) + if err != nil { + t.Fatalf("resolve the runtime capability SID: %v", err) + } + granted, err := windowsPathGrantsCapability(runtimeRoot, sid) + if err != nil || !granted { + t.Skipf("SETUP: the first run did not grant the runtime root here (granted=%v err=%v)", granted, err) + } + + // Cleanup reclaims it; the next ordinary run recreates the same pathname. + if err := os.RemoveAll(runtimeRoot); err != nil { + t.Fatalf("reclaim the runtime root the way cleanup does: %v", err) + } + if err := os.MkdirAll(runtimeRoot, 0o700); err != nil { + t.Fatalf("recreate the pathname the way an ordinary run does: %v", err) + } + + // SETUP: the case is really reproduced, or the assertion below would hold for + // a run that never lost anything. + stillGranted, err := windowsPathGrantsCapability(runtimeRoot, sid) + if err != nil { + t.Fatalf("inspect the recreated runtime root: %v", err) + } + if stillGranted { + t.Fatal("SETUP INVALID: the recreated directory still carries the capability grant, so nothing was lost to restore") + } + + // The second run must notice and reapply rather than trust the marker. + if err := ensureWindowsUnelevatedSetup(config); err != nil { + t.Fatalf("the second run refused instead of restoring the reclaimed root: %v", err) + } + restored, err := windowsPathGrantsCapability(runtimeRoot, sid) + if err != nil { + t.Fatalf("inspect the restored runtime root: %v", err) + } + if !restored { + t.Fatal("the cached marker was accepted for a runtime object that no longer carries the grant; the command would launch and then fail every sandboxed write into it with nothing explaining why") + } +} + +// And an untouched runtime root still takes the cached fast path, so the object +// check does not turn every command into a full reapply. +func TestUnelevatedSetupStillTrustsAnIntactRuntimeRoot(t *testing.T) { + cacheRoot := t.TempDir() + previous := sandboxUserCacheDir + t.Cleanup(func() { sandboxUserCacheDir = previous }) + sandboxUserCacheDir = func() (string, error) { return cacheRoot, nil } + + workspace := t.TempDir() + runtimeRoot, ok := deterministicSandboxRuntimeRoot(canonicalSandboxWorkspaceRoot(workspace), canonicalSandboxWorkspaceRoot(cacheRoot)) + if !ok { + t.Fatalf("SETUP INVALID: no deterministic runtime root for %s", workspace) + } + if err := os.MkdirAll(runtimeRoot, 0o700); err != nil { + t.Fatal(err) + } + config := WindowsSandboxCommandConfig{ + SandboxHome: t.TempDir(), + CommandCWD: workspace, + WorkspaceRoots: []string{workspace}, + SandboxLevel: WindowsSandboxLevelUnelevated, + PermissionProfile: PermissionProfile{ + Runtime: &SandboxRuntime{Root: runtimeRoot}, + FileSystem: FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{Root: workspace}, {Root: runtimeRoot}}, + ReadRoots: []string{workspace}, + }, + }, + } + if err := ensureWindowsUnelevatedSetup(config); err != nil { + t.Skipf("cannot run unelevated setup here: %v", err) + } + if err := ensureWindowsUnelevatedSetup(config); err != nil { + t.Fatalf("a second run over an intact runtime root was refused: %v", err) + } +} diff --git a/internal/sandbox/windows_unelevated_tier_windows_test.go b/internal/sandbox/windows_unelevated_tier_windows_test.go new file mode 100644 index 000000000..e4b1a0037 --- /dev/null +++ b/internal/sandbox/windows_unelevated_tier_windows_test.go @@ -0,0 +1,102 @@ +//go:build windows + +package sandbox + +import ( + "path/filepath" + "strings" + "testing" +) + +// ONE PLAN, TWO TIERS, DIFFERENT AUTHORITY. +// +// Production profiles seed ReadRoots with the filesystem root, so a profile that +// configures DenyRead adds an allow-read ACE for the read capability at the +// volume root: the strict token that DenyRead selects applies the restricted-SID +// check to reads, so without it the command cannot even open its own executable. +// Elevated setup can write that DACL. An ordinary user cannot, and the common +// opener asks for WRITE_DAC on every entry, so the unelevated tier failed on that +// one root on every command, with a generic diagnosis and a remedy that did not +// apply. +// +// The real smoke test misses this because it substitutes a user-owned temporary +// directory for the production read root, which is exactly why this asserts on +// the production shape. +func TestUnelevatedSetupRefusesAVolumeRootReadGrant(t *testing.T) { + workspace := t.TempDir() + volumeRoot := filepath.VolumeName(workspace) + string(filepath.Separator) + if !isWindowsVolumeRoot(volumeRoot) { + t.Fatalf("SETUP INVALID: %q is not recognised as a volume root", volumeRoot) + } + + config := WindowsSandboxCommandConfig{ + SandboxHome: t.TempDir(), + CommandCWD: workspace, + WorkspaceRoots: []string{workspace}, + SandboxLevel: WindowsSandboxLevelUnelevated, + PermissionProfile: PermissionProfile{ + FileSystem: FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{Root: workspace}}, + // The production shape: read everywhere, and a denyRead that forces + // the strict token. + ReadRoots: []string{volumeRoot}, + DenyRead: []string{filepath.Join(workspace, "secrets")}, + }, + }, + } + + plan, err := BuildWindowsACLPlan(config) + if err != nil { + t.Fatalf("BuildWindowsACLPlan: %v", err) + } + // SETUP: the plan really does reach the volume root, or the refusal below + // would be asserting nothing. + if got := windowsPlanVolumeRootGrant(plan); got == "" { + t.Fatalf("SETUP INVALID: the plan carries no volume-root entry, so this profile does not reproduce the case: %+v", plan.Entries) + } + + err = ensureWindowsUnelevatedSetup(config) + if err == nil { + t.Fatal("unelevated setup accepted a plan it cannot apply; every command would fail later with a generic ACCESS_DENIED") + } + message := err.Error() + if !strings.Contains(message, volumeRoot) { + t.Errorf("the refusal does not name the volume root, so the reader cannot tell which entry is at fault: %v", err) + } + if !strings.Contains(message, "denyRead") { + t.Errorf("the refusal does not name the cause, so the reader cannot act on it: %v", err) + } + if !strings.Contains(message, "elevated") { + t.Errorf("the refusal does not name a remedy the reader can carry out: %v", err) + } +} + +// And a profile with no denyRead still runs on this tier: it needs no volume-root +// grant, so refusing it would take the unelevated sandbox away from everyone. +func TestUnelevatedSetupStillAcceptsAWriteJailOnlyProfile(t *testing.T) { + workspace := t.TempDir() + config := WindowsSandboxCommandConfig{ + SandboxHome: t.TempDir(), + CommandCWD: workspace, + WorkspaceRoots: []string{workspace}, + SandboxLevel: WindowsSandboxLevelUnelevated, + PermissionProfile: PermissionProfile{ + FileSystem: FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{Root: workspace}}, + ReadRoots: []string{workspace}, + }, + }, + } + plan, err := BuildWindowsACLPlan(config) + if err != nil { + t.Fatalf("BuildWindowsACLPlan: %v", err) + } + if got := windowsPlanVolumeRootGrant(plan); got != "" { + t.Fatalf("SETUP INVALID: a write-jail-only profile reached the volume root at %q", got) + } + if err := ensureWindowsUnelevatedSetup(config); err != nil && strings.Contains(err.Error(), "volume root") { + t.Fatalf("a profile needing no volume-root grant was refused by the tier check: %v", err) + } +} diff --git a/internal/sandbox/windows_workspace_canonical_windows_test.go b/internal/sandbox/windows_workspace_canonical_windows_test.go new file mode 100644 index 000000000..87e9ec5e8 --- /dev/null +++ b/internal/sandbox/windows_workspace_canonical_windows_test.go @@ -0,0 +1,320 @@ +//go:build windows + +package sandbox + +import ( + "os" + "path/filepath" + "strings" + "testing" +) + +// Setup grants the principal a runtime tree; every command derives that tree +// again from the workspace root. If the two normalize differently the grant +// lands somewhere nothing reads, and the only symptom is a bare ACCESS_DENIED +// on the first cache write. +// +// This needs no symlink and no privilege. Windows opens a path whatever its +// casing, and Engine.resolveCommandDir runs EvalSymlinks (runner.go) which +// canonicalizes it, while setup used to only Clean. +func TestSetupRuntimeRootMatchesTheCommandPathForADifferentlyCasedRoot(t *testing.T) { + base := t.TempDir() + workspace := filepath.Join(base, "MyWorkspace") + if err := os.MkdirAll(workspace, 0o700); err != nil { + t.Fatalf("mkdir: %v", err) + } + lowered := strings.ToLower(workspace) + if _, err := os.Stat(lowered); err != nil { + t.Skipf("filesystem is case-sensitive here: %v", err) + } + if lowered == workspace { + t.Skip("temp path has no case to differ on") + } + + // What the command path ends up keyed to, per resolveCommandDir. + commandRoot := lowered + if resolved, err := filepath.EvalSymlinks(filepath.Clean(lowered)); err == nil { + commandRoot = resolved + } + if commandRoot == lowered { + t.Skip("EvalSymlinks changed nothing on this host; no divergence to assert") + } + + // Drive the PRODUCTION derivation, not the helper. A test that called + // canonicalSandboxWorkspaceRoot directly would pass just as happily + // with setup still doing filepath.Clean, which is exactly the bug. + fromSetup, err := windowsSandboxRuntimeRootPath(WindowsSandboxCommandConfig{ + WorkspaceRoots: []string{lowered}, + CommandCWD: lowered, + }) + if err != nil { + t.Fatalf("windowsSandboxRuntimeRootPath: %v", err) + } + cacheRoot, err := sandboxUserCacheDir() + if err != nil { + t.Fatalf("sandboxUserCacheDir: %v", err) + } + fromCommand, err := sandboxRuntimeRootFor(commandRoot, filepath.Clean(cacheRoot)) + if err != nil { + t.Fatalf("sandboxRuntimeRootFor(command): %v", err) + } + if !grantedRuntimeRootsCover(fromSetup, fromCommand) { + t.Errorf("setup grants a runtime tree commands never use:\n setup: %s\n command: %s", fromSetup, fromCommand) + } +} + +// A root whose final segments do not exist still normalizes: the existing +// ancestor resolves and the missing remainder is re-appended. +// +// This asserted the whole cleaned path unchanged at first, which was the +// all-or-nothing behaviour the ancestor walk replaced. Windows CI failed it — +// correctly — because RUNNER~1 resolved to runneradmin while never-created +// stayed put, which is exactly the behaviour the walk exists to produce. +func TestCanonicalWorkspaceRootResolvesTheExistingAncestor(t *testing.T) { + parent := t.TempDir() + missing := filepath.Join(parent, "never-created", "deeper") + + got := canonicalSandboxWorkspaceRoot(missing) + want := filepath.Join(canonicalSandboxWorkspaceRoot(parent), "never-created", "deeper") + if got != want { + t.Errorf("canonical(%q) = %q, want %q", missing, got, want) + } + // The missing segments must survive rather than be dropped to the ancestor. + if !strings.HasSuffix(got, filepath.Join("never-created", "deeper")) { + t.Errorf("canonical(%q) = %q, lost the segments that do not exist yet", missing, got) + } + if canonicalSandboxWorkspaceRoot(" ") != "" { + t.Error("a blank root should stay blank, not become the process directory") + } +} + +// The pair has to agree, not just each side individually. CI caught this the +// hard way: canonicalizing only the setup side made setup and +// prepareSandboxRuntime disagree on a Windows runner, whose TEMP is an 8.3 +// short path that resolution expands. Lowercasing reproduces the same class of +// non-canonical spelling without needing a short name or any privilege. +func TestSetupAndPrepareRuntimeAgreeOnANonCanonicalRoot(t *testing.T) { + isolateSandboxRuntimeRoots(t) + base := t.TempDir() + workspace := filepath.Join(base, "MyWs") + if err := os.MkdirAll(workspace, 0o700); err != nil { + t.Fatalf("mkdir: %v", err) + } + lowered := strings.ToLower(workspace) + if _, err := os.Stat(lowered); err != nil { + t.Skipf("filesystem is case-sensitive here: %v", err) + } + + granted, err := setupWindowsSandboxRuntimeRoot(WindowsSandboxCommandConfig{ + WorkspaceRoots: []string{lowered}, + CommandCWD: lowered, + }) + if err != nil { + t.Fatalf("setupWindowsSandboxRuntimeRoot: %v", err) + } + state, release, err := prepareSandboxRuntime(lowered) + if err != nil { + t.Fatalf("prepareSandboxRuntime: %v", err) + } + if release != nil { + defer release() + } + if !grantedRuntimeRootsCover(granted, state.Root) { + t.Errorf("setup granted %q but commands write to %q", granted, state.Root) + } +} + +// The carveout shape has to survive a non-canonical root too. The first fix +// rebuilt the spec paths from the RESOLVED write root and compared them against +// subpaths that could not resolve (.git/config does not exist yet), so on a +// short-name or differently-cased path the match missed and .git/config went +// back to being created as a directory. +func TestGitConfigCarveoutShapeSurvivesANonCanonicalRoot(t *testing.T) { + base := t.TempDir() + workspace := filepath.Join(base, "MyWs") + if err := os.MkdirAll(workspace, 0o700); err != nil { + t.Fatalf("mkdir: %v", err) + } + lowered := strings.ToLower(workspace) + if _, err := os.Stat(lowered); err != nil { + t.Skipf("filesystem is case-sensitive here: %v", err) + } + + plan, err := buildWindowsPrincipalACLPlan(windowsPrincipalACLInput{ + PrincipalSID: "S-1-5-32-546", + WriteRoots: []WritableRoot{{ + Root: lowered, + ReadOnlySubpaths: gitMetadataWriteCarveouts(lowered), + }}, + }) + if err != nil { + t.Fatalf("buildWindowsPrincipalACLPlan: %v", err) + } + found := false + for _, entry := range plan.Entries { + if !strings.EqualFold(filepath.Base(entry.Path), "config") { + continue + } + found = true + if !entry.MaterializeFile { + t.Errorf(".git/config entry %q lost its file shape on a non-canonical root", entry.Path) + } + } + if !found { + t.Fatal("no .git/config entry in the plan") + } +} + +// Teardown must name the runtime tree without creating anything. The comment +// on windowsPrincipalTeardownPaths claimed that and it was false: the resolver +// it used ends in sandboxRuntimeRootFor, whose fallback calls os.MkdirTemp, so +// a workspace whose cache root sits inside it made setup's cleanup path create +// a fresh temp directory on its way out — and a useless one, since the fallback +// root is random per process and never matches what commands used. +func TestTeardownPathDerivationCreatesNothing(t *testing.T) { + workspace := t.TempDir() + // A PRIVATE OBSERVATION ROOT. The assertions below snapshot the temp + // directory and fail on anything that appears, and the shared one is written + // by every other package the runner is testing at the same time. It also + // keeps the fallback runtime root this test drives inside the fixture. + observed := filepath.Join(t.TempDir(), "observed") + if err := os.Mkdir(observed, 0o700); err != nil { + t.Fatal(err) + } + t.Setenv("TMP", observed) + t.Setenv("TEMP", observed) + // Force the branch that falls back: cache root inside the workspace. + original := sandboxUserCacheDir + sandboxUserCacheDir = func() (string, error) { return filepath.Join(workspace, ".cache"), nil } + t.Cleanup(func() { sandboxUserCacheDir = original }) + + before := tempDirEntryNames(t) + + // Drive the PRODUCTION teardown path, not the helper. Calling the resolver + // directly passes just as happily with the call site reverted to the one + // that creates. + paths, err := windowsPrincipalTeardownPaths(WindowsSandboxCommandConfig{ + WorkspaceRoots: []string{workspace}, + CommandCWD: workspace, + PermissionProfile: PermissionProfile{ + FileSystem: FileSystemPolicy{ + Kind: FileSystemRestricted, + WriteRoots: []WritableRoot{{Root: workspace}}, + }, + }, + }, "S-1-5-32-546") + if err != nil { + t.Fatalf("windowsPrincipalTeardownPaths: %v", err) + } + if len(paths) == 0 { + t.Error("teardown named no paths at all; the workspace root should still be revoked") + } + assertCreatedNothing(t, before, "teardown path derivation") + + // Setup's resolver must agree with teardown's, and create nothing either. + // + // This assertion has now been inverted twice, so the history is worth keeping. + // It first required setup to fall back to a usable tree. That was then wrong, + // because the fallback was os.MkdirTemp memoized only in-process: setup would + // grant an ACE on temp root A, the next command would derive root B and fail + // ACCESS_DENIED, and teardown would clean a third. So it was changed to demand + // setup report NOTHING here. + // + // That is what is wrong now. fallbackSandboxRuntimeRoot derives its path by + // hashing the workspace and creates nothing, so every process reaches the same + // answer. Commands in this exact layout therefore DO select it and redirect + // TMP, GOCACHE and the package caches into it, while setup granting nothing + // left both principals without an ACE on the one tree those writes land in. + // Reporting none is no longer the safe answer; it is the ACCESS_DENIED. + beforeSetup := tempDirEntryNames(t) + setupRoots, err := windowsSandboxRuntimeRootPath(WindowsSandboxCommandConfig{ + WorkspaceRoots: []string{workspace}, + CommandCWD: workspace, + }) + if err != nil { + t.Fatalf("windowsSandboxRuntimeRootPath: %v", err) + } + if len(setupRoots) == 0 { + t.Error("setup named no runtime root for a workspace whose cache-derived root is unusable; commands still select the stable fallback and would write there with no ACE") + } + // Purity is a property of RESOLUTION. Assert it here, before preparation, + // which is allowed to create the fallback anchor it selects; asserting after + // it blamed the resolver for preparation's legitimate directory. + assertCreatedNothing(t, beforeSetup, "setup path derivation") + // And it must be the tree a command actually picks, not merely some tree. + commandState, release, err := prepareSandboxRuntime(workspace) + if err != nil { + t.Fatalf("prepareSandboxRuntime: %v", err) + } + if release != nil { + defer release() + } + if !grantedRuntimeRootsCover(setupRoots, commandState.Root) { + t.Errorf("setup granted %q but commands write to %q", setupRoots, commandState.Root) + } +} + +// Setup and teardown must derive the SAME root in the ordinary case, since one +// grants the ACE the other revokes. This is the case the fix above must not +// break: reporting "no root" is only correct when the root is underivable. +func TestSetupAndTeardownDeriveTheSameRuntimeRoot(t *testing.T) { + workspace := t.TempDir() + cacheRoot := t.TempDir() // outside the workspace, so the derivation is usable + previous := sandboxUserCacheDir + sandboxUserCacheDir = func() (string, error) { return cacheRoot, nil } + t.Cleanup(func() { sandboxUserCacheDir = previous }) + + setupRoot, err := windowsSandboxRuntimeRootPath(WindowsSandboxCommandConfig{ + WorkspaceRoots: []string{workspace}, + CommandCWD: workspace, + }) + if err != nil { + t.Fatalf("windowsSandboxRuntimeRootPath: %v", err) + } + if len(setupRoot) == 0 { + t.Fatal("setup named no runtime root for a derivable workspace") + } + teardownRoot, ok := deterministicSandboxRuntimeRoot( + canonicalSandboxWorkspaceRoot(workspace), canonicalSandboxWorkspaceRoot(cacheRoot)) + if !ok { + t.Fatal("precondition: the deterministic root should be usable here") + } + // The cache-derived root is the one a command picks in this layout, so it has + // to be among what setup granted. Setup covering the fallback as well is the + // point rather than a discrepancy: whichever the command lands on is + // provisioned. + if !grantedRuntimeRootsCover(setupRoot, teardownRoot) { + t.Errorf("setup grants on %q but teardown revokes %q", setupRoot, teardownRoot) + } +} + +// tempDirEntryNames snapshots what the shared temp directory holds. +// +// NAMES, NOT A COUNT. The assertion is that deriving a path CREATES nothing, +// and a count cannot tell creation from removal: the shared temp root is also +// used by every other test binary running at the same time and by the OS, so a +// concurrent cleanup made the count fall and the test reported that naming the +// paths had "gained -1 entries". Comparing the sets answers the question that +// was actually being asked and is indifferent to anything disappearing. +func tempDirEntryNames(t *testing.T) map[string]struct{} { + t.Helper() + entries, err := os.ReadDir(os.TempDir()) + if err != nil { + t.Fatalf("read temp dir: %v", err) + } + names := make(map[string]struct{}, len(entries)) + for _, entry := range entries { + names[entry.Name()] = struct{}{} + } + return names +} + +// assertCreatedNothing reports entries that appeared since the snapshot. +func assertCreatedNothing(t *testing.T, before map[string]struct{}, what string) { + t.Helper() + for name := range tempDirEntryNames(t) { + if _, existed := before[name]; !existed { + t.Errorf("%s created %q in the temp directory; naming the paths must not create one", what, name) + } + } +}