v4.1.x: osc/pt2pt: Some fixes #9247
Merged
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
* To prevent deadlock force all previously requested locks to complete before starting a new lock. Otherwise this can lead to deadlock if the locks are processes in arbitrary order. Signed-off-by: Mark Allen <[email protected]> (cherry picked from commit 20b4618)
* The `opal_list_remove_first` is protected by the accumulation lock not the module lock. So make sure to use the same lock to protect the `opal_list_append` to prevent current access, which can cause a segv and/or lost enties in the list. Signed-off-by: Joshua Hursey <[email protected]> (cherry picked from commit 130d196)
* The `ompi_osc_pt2pt_sync_pscw_peer` call needs ot be protected when
called, and the `sync` structure need to be protected in the `start`
function.
- This prevents the `ompi_osc_pt2pt_sync_pscw_peer` function from
processing while `start` is also in progress on the same window in
two different threads (e.g., `progress` and `main` thread)
- This seems to happen when the 'main' thread is part way through
the `start` function then the progress thread starts processing
the `post` message received from another peer for this window.
Both functions try to access the `peer_list` portion of the
structure and a NULL is stepped on in the `ompi_osc_pt2pt_sync_pscw_peer`
function.
- This patch locks the `sync` structure for the duration of the `start`
and ensures that whenever `ompi_osc_pt2pt_sync_pscw_peer` is called
the caller is holding the `sync->lock`. This provides exclusivity
to this structure ensuring that the latter function sees a fully
updated structure.
Signed-off-by: Joshua Hursey <[email protected]>
(cherry picked from commit ab49b8a)
jjhursey
approved these changes
Aug 16, 2021
Member
|
bot:ompi:retest |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 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.
Some misc. fixes that haven't been brought back.