fix(recomp): start VCALLMSR from CMSAR0 instead of indexing vi[] - #268
Open
drakolordx7 wants to merge 1 commit into
Open
drakolordx7 wants to merge 1 commit into
drakolordx7 wants to merge 1 commit into
Conversation
VCALLMSR was generated as
uint16_t instr_index = ctx->vi[27] & 0x1FF;
but vi[] only holds the 16 VU0 integer registers (uint16_t vi[16] in
R5900Context). "27" is the VU0 control register number of CMSAR0 (the
$vi27 that CTC2 writes, VU0_CR_CMSAR0), so the generated code read 22
bytes past the end of the array, into neighbouring context fields, and
never saw the value the game had stored with CTC2. The CTC2 side already
stores it in ctx->vu0_cmsar0.
Read ctx->vu0_cmsar0 instead; the instruction has no register operand.
Generated code has to be regenerated to pick the fix up.
A game that calls VCALLMSR from a helper (CTC2 $vi27, then VCALLMSR) hit
this; it had been worked around in the runtime. This fixes it at the source,
as also described in issue ran-j#262. The new code generator test fails without
the change.
Made by drakolord and assisted with Claude Code.
Scotho
added a commit
to Scotho/socom-unzipped
that referenced
this pull request
Oct 3, 2026
…] (upstream PS2Recomp #268) VCALLMSR was generated as ctx->vi[27] & 0x1FF on a uint16_t vi[16]: 27 is the CMSAR0 control-register number, so the read went past the array and never saw the value CTC2 stores in ctx->vu0_cmsar0 (our CTC2/CFC2 use that field too). The generator now reads ctx->vu0_cmsar0. Ported from upstream PS2Recomp PR #268 by drako (drakolordx7), the fix and its test as written there. Upstream: ran-j/PS2Recomp#268, fixing ran-j/PS2Recomp#262 (open, not merged on 2026-10-03; no merge commit; ported from its head commit a0f89eb3d4ffa70a4695fa89cc444bf4743a845b). Same licence (GPL-3.0). The issue's reporter later noted the register file really has 32 VI slots (16..31 the control registers) and proposed vi[32] instead; that wider change is not taken here. SOCOM II's generated code emits no VCALLMSR (no vu0StartMicroProgram in recomp/output), so codegen is unchanged for the game; it takes effect at the next recomp. LATER row 83; research/85 section 4. Co-authored-by: drako <98249188+drakolordx7@users.noreply.github.com> Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fixes #262.
VCALLMSRwas generated asctx->vi[27] & 0x1FF, butvi[]has 16 entries (uint16_t vi[16]). 27 is the CMSAR0 control-register number, so the read went past the array and never saw the valueCTC2stored inctx->vu0_cmsar0. The generator now readsctx->vu0_cmsar0.Games need their generated code regenerated to pick this up. A new code generator test fails without the change;
ps2x_tests437/437.Made by drakolord and assisted with Claude Code.