ci: declare contents:read on Flash Attention Benchmark workflow#176
Open
arpitjain099 wants to merge 1 commit into
Open
ci: declare contents:read on Flash Attention Benchmark workflow#176arpitjain099 wants to merge 1 commit into
arpitjain099 wants to merge 1 commit into
Conversation
The benchmark-flash-attn job clones this repo, clones Dao-AILab/flash-attention into a sibling path, and runs benchmarks on linux.dgx.b200.8. No GitHub API write beyond actions/checkout. Style matches the per-job permissions blocks already used by tritonbench.yml, vllm-benchmark.yml, and the other hardened workflows in this repo (typically id-token:write + contents:read). inductor.yml is the other workflow without permissions, but it's a caller of pytorch/test-infra/.github/workflows/linux_job.yml@main, so adding permissions there would intersect with the reusable workflow's grant. Signed-off-by: Arpit Jain <arpitjain099@gmail.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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The
Flash Attention Benchmarkworkflow runs nightly onlinux.dgx.b200.8, checks out this repo +Dao-AILab/flash-attention, and runs the benchmark scripts. There's no GitHub API write beyondactions/checkout.This patch pins the job to
permissions: contents: read, matching the per-job permission blocks already declared bytritonbench.yml,tritonbench-bisect.yml,vllm-benchmark.yml,vllm-ci-test.yml,vllm-profiling.yml,sglang-benchmark.yml, andpytorch-bisect.yaml.With explicit scope:
Token-Permissionscheck passes for this fileactions/checkout(cf. tj-actions/changed-files CVE-2025-30066) stays boxed in read-onlyinductor.ymlis the other workflow without an explicitpermissions:block, but it calls the reusablepytorch/test-infra/.github/workflows/linux_job.yml@mainworkflow. Declaring permissions on the caller would intersect with the callee's grant, so I've left it out of this PR.No behavioural change.