fix(security): re-verify PR head SHA in performance workflow

[causten]

Summary

• Re-check PR headRefOid via gh pr view after ok-to-test gate to close TOCTOU window

JIRA

• ROCM-26604

Test plan

• Exercise performance workflow on labeled external PR

Made with Cursor

[Copilot]

Pull request overview

This PR strengthens the performance workflow’s external PR security gate and updates manual-dispatch repository defaults.

Changes:

• Rechecks the PR head SHA after the ok-to-test gate.
• Updates repository defaults to internal repositories.
• Documents the mutable reusable workflow reference.

Suppressed comments (2)

.github/workflows/performance.yaml:181

• The reusable workflow remains referenced by the mutable @testing branch, while this pull_request_target path passes BENCHMARK_UTILS_READ_TOKEN to it. Anyone able to change that branch, or a compromised upstream maintainer account, could alter the workflow to exfiltrate the token or run arbitrary work on the privileged runner. Pin this reference to a reviewed full commit SHA before relying on this workflow as a security boundary; the added comment currently defers that mitigation.

    # Pin to immutable commit once available in migraphx-benchmark; @testing is mutable.

.github/workflows/performance.yaml:100

• security_gate has no checkout, so gh pr view cannot resolve the current repository unless --repo is supplied; this new command will fail on the external-PR path. Also, although it requests labels, jq discards them, so removing ok-to-test after dispatch still lets the stale event-label check pass. Scope the query to ${{ github.repository }} and require the current labels to contain ok-to-test before accepting the SHA.

          PR_HEAD=$(gh pr view ${{ github.event.pull_request.number }} --json headRefOid,labels --jq '.headRefOid')
          if [[ "$PR_HEAD" != "${{ github.event.pull_request.head.sha }}" ]]; then

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[gh-app-migraphx-bot-pr-write]

Test	Batch	New Rate (c502fe)	Old Rate (3a503c)*	Diff	Status
torchvision-resnet50	64	3,324.59	3,264.92	1.83%	✅
torchvision-resnet50_fp16	64	7,873.53	7,548.67	4.30%	✅
torchvision-densenet121	32	2,489.96	2,483.99	0.24%	✅
torchvision-densenet121_fp16	32	5,026.06	5,004.24	0.44%	✅
torchvision-inceptionv3	32	2,070.22	2,058.51	0.57%	✅
torchvision-inceptionv3_fp16	32	4,460.17	4,416.99	0.98%	✅
cadene-inceptionv4	16	817.88	820.61	-0.33%	✅
cadene-resnext64x4	16	781.78	782.78	-0.13%	✅
slim-mobilenet	64	8,379.87	8,386.36	-0.08%	✅
slim-nasnetalarge	64	228.36	228.86	-0.22%	✅
slim-resnet50v2	64	3,238.93	3,180.91	1.82%	✅
bert-mrpc-onnx	8	1,170.58	1,168.84	0.15%	✅
bert-mrpc-tf	1	498.46	498.63	-0.03%	✅
pytorch-examples-wlang-gru	1	492.00	473.35	3.94%	✅
pytorch-examples-wlang-lstm	1	411.79	384.83	7.01%	🔆
torchvision-resnet50_1	1	1,046.86	1,046.63	0.02%	✅
cadene-dpn92_1	1	449.25	437.32	2.73%	✅
cadene-resnext101_1	1	366.54	365.89	0.18%	✅
onnx-taau-downsample	1	847.18	844.09	0.37%	✅
dlrm-criteoterabyte	1	32.28	32.42	-0.44%	✅
dlrm-criteoterabyte_fp16	1	51.58	51.80	-0.43%	✅
agentmodel	1	15,099.19	9,209.12	63.96%	🔆
unet_fp16	2	58.27	58.80	-0.90%	✅
resnet50v1_fp16	1	1,437.88	1,366.11	5.25%	🔆
resnet50v1_int8	1	1,764.27	1,883.96	-6.35%	🔴
bert_base_cased_fp16	64	1,098.09	1,098.16	-0.01%	✅
bert_large_uncased_fp16	32	345.72	345.59	0.04%	✅
bert_large_fp16	1	206.64	206.59	0.03%	✅
distilgpt2_fp16	16	2,085.27	2,092.89	-0.36%	✅
yolov5s	1	551.93	558.33	-1.15%	✅
tinyllama	1	45.84	45.83	0.01%	✅
vicuna-fastchat	1	44.25	44.20	0.12%	✅
whisper-tiny-encoder	1	413.26	411.87	0.34%	✅
whisper-tiny-decoder	1	410.23	408.48	0.43%	✅
llama2_7b	1	20.87	20.84	0.16%	✅
qwen1.5-7b	1	23.64	23.58	0.27%	✅
phi3-3.8b	1	26.77	26.72	0.21%	✅
llama3-8b	1	21.72	21.80	-0.39%	✅
whisper-large-encoder	1	10.17	10.18	-0.06%	✅
whisper-large-decoder	1	105.88	105.30	0.55%	✅
mistral-7b	1	23.75	23.78	-0.12%	✅
FLUX.1-schnell	1	764.20	755.22	1.19%	✅

Regressions detected 🔴

* No develop baseline was found for this PR's branch point; compared against the latest available develop run instead.

[gh-app-migraphx-bot-pr-write]

Test	Status	Result
bert-mrpc-onnx	✅	PASSED: MIGraphX meets tolerance
bert-mrpc-tf	❌	ERROR - check error output traceback Traceback (most recent call last): File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 377, in main() File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 313, in main import tensorflow as tf File "/usr/local/lib/python3.12/dist-packages/tensorflow/init.py", line 40, in from tensorflow.python import pywrap_tensorflow as _pywrap_tensorflow # pylint: disable=unused-import ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/usr/local/lib/python3.12/dist-packages/tensorflow/python/pywrap_tensorflow.py", line 37, in self_check.preload_check() File "/usr/local/lib/python3.12/dist-packages/tensorflow/python/platform/self_check.py", line 63, in preload_check from tensorflow.python.platform import _pywrap_cpu_feature_guard ImportError: libnuma.so.1: cannot open shared object file: No such file or directory
pytorch-examples-wlang-gru	✅	PASSED: MIGraphX meets tolerance
pytorch-examples-wlang-lstm	✅	PASSED: MIGraphX meets tolerance
dlrm-criteoterabyte	✅	PASSED: MIGraphX meets tolerance
agentmodel	✅	PASSED: MIGraphX meets tolerance
unet	✅	PASSED: MIGraphX meets tolerance
resnet50v1	✅	PASSED: MIGraphX meets tolerance
bert_base_cased_fp16	✅	PASSED: MIGraphX meets tolerance
bert_large_uncased_fp16	🔴	FAILED: MIGraphX is not within tolerance - check verbose output
bert_large	✅	PASSED: MIGraphX meets tolerance
yolov5s	✅	PASSED: MIGraphX meets tolerance
tinyllama	✅	PASSED: MIGraphX meets tolerance
vicuna-fastchat	✅	PASSED: MIGraphX meets tolerance
whisper-tiny-encoder	✅	PASSED: MIGraphX meets tolerance
whisper-tiny-decoder	✅	PASSED: MIGraphX meets tolerance
distilgpt2_fp16	🔴	FAILED: MIGraphX is not within tolerance - check verbose output
llama2_7b	✅	PASSED: MIGraphX meets tolerance
qwen1.5-7b	✅	PASSED: MIGraphX meets tolerance
phi3-3.8b	✅	PASSED: MIGraphX meets tolerance
llama3-8b	✅	PASSED: MIGraphX meets tolerance
whisper-large-encoder	✅	PASSED: MIGraphX meets tolerance
whisper-large-decoder	✅	PASSED: MIGraphX meets tolerance
mistral-7b	✅	PASSED: MIGraphX meets tolerance
FLUX.1-schnell	✅	PASSED: MIGraphX meets tolerance