DEVICE/API: Fix segfault when releasing multiple exported batches on same endpoint #893
Conversation
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
…same endpoint Co-Authored-By: Roey Azran <[email protected]> Signed-off-by: Michal Shalev <[email protected]>
michal-shalev
requested review from
brminich,
gleon99 and
yosefe
as code owners
|
👋 Hi michal-shalev! Thank you for contributing to ai-dynamo/nixl. Your PR reviewers will review your contribution then trigger the CI to test your changes. 🚀 |
ovidiusm
reviewed
Oct 11, 2025
src/utils/ucx/ucx_utils.h
Outdated
| public: | ||
| void err_cb(ucp_ep_h ucp_ep, ucs_status_t status); | ||
| void* exported_batch{nullptr}; | ||
| std::map<nixlUcxReq, ucp_batch_h> exported_batches; |
There was a problem hiding this comment.
Can we use unordered map?
Signed-off-by: Michal Shalev <[email protected]>
michal-shalev
force-pushed
the
fix-multiple-batch-export-segfault
branch
2 times, most recently
from
b7ee68e to
0fa6861
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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Outdated suggestions cannot be applied.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.
What?
Fix segmentation fault when releasing multiple exported batches on the same UCX endpoint.
Why?
Single
exported_batchvariable was overwritten by subsequent exports, causingreleaseBatch()to use wrong batch handle. This caused crashes during elastic training when removing ranks with multiple active batches.How?
std::map<nixlUcxReq, ucp_batch_h>to track multiple batch handlesmultipleExportReleaseSameEndpointthat reproduces the bug