Repository navigation
[Bugfix][Diffusion] Restore supports_packed_mask_free on teacache FakeBackend - #5997
Merged
hsliuustc0106 merged 2 commits intoAug 10, 2026
Merged
Conversation
|
Codex usage limits have been reached for code reviews. Please check with the admins of this repo to increase the limits by adding credits. |
|
This PR appears to belong to: docs/design/module/cache_management.md. Module owners: @Isotr0py @princepride @SamitHuang Please take a look when you have a chance. If you would like an automated review, mention @vllm-omni-review-bot in a comment. |
david6666666
approved these changes
Aug 10, 2026
david6666666
enabled auto-merge (squash)
August 10, 2026 12:47
david6666666
added
merge-test
hsliuustc0106
disabled auto-merge
August 10, 2026 14:20
Signed-off-by: brandneway <gyuan4892@gmail.com>
khairulkabir1661
pushed a commit
to khairulkabir1661/vllm-omni
that referenced
this pull request
Sep 25, 2026
…eBackend (vllm-project#5997) Signed-off-by: brandneway <gyuan4892@gmail.com> Co-authored-by: Hongsheng Liu <liuhongsheng4@huawei.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
4 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Context
#5891 added
supports_packed_mask_freetoAttentionBackendand calls it in theH3 packed-attention path (
minimax_h3_transformer.py:421). TheFakeBackendintests/diffusion/cache/test_teacache_extractors.pydid not implement it, so after#5891 merged, 7
TestMiniMaxH3Extractortests fail withAttributeError: type object 'FakeBackend' has no attribute 'supports_packed_mask_free',leaving
buildkite/vllm-omnired on main.Fix
Add the classmethod to
FakeBackendreturningFalse(the abstract-backend default →masked path), matching the fakes already in
test_minimax_h3_contract.py.Verification
AttributeError resolved (traceback now advances past the attention call). These tests
exercise NPU ops (npu_rms_norm) that require real NPU hardware, so they're validated in CI.