osc/rdma: two bug fixes #9358
Merged
osc/rdma: two bug fixes #9358
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
Currenntly, in function allocate_state_shared,
"module->use_memory_registation" is set to false when all MPI ranks
are on same instances (local_size == global_size).
This is harmful when there is only one btl in use, in which case the selected
btl should determine whether memory registration should be used. For
example, btl/ofi uses memory registration even on same instance.
This is unnecessary when three are two btls in use, in which case
btls that uses memory registration have been excluded in function
ompi_osc_rdma_query_alternate_btls.
Therefore, this commit removes the setting of
module->use_memory_registration
in allocate_state_shared.
Signed-off-by: Wei Zhang <[email protected]>
In function "allocate_state_shared", the peer->state_endpoint was copied from 1st peer (a.k.a local_leader). However, state_endpoint and state_btl_index of the 1st peer was not set, causing all peers' state_endpoint being NULL. This patch addresses the issue by setting 1st peer's state_endpoint and state_btl_index from its data_endpoint and data_btl_index. Signed-off-by: Wei Zhang <[email protected]>
bwbarrett
approved these changes
Sep 7, 2021
|
The GitHub Action CI failure is because I did not add a description in the PR. After I did that, the GitHub Action CI passed. |
|
Is there anything else needed for this PR to get merged? |
|
I'd like @hjelmn to comment in case I was totally off in my review, even if he doesn't do a full review. |
|
@hjelmn Do you have time to take a look? |
|
@hjelmn Can you please take a look at this PR? |
hjelmn
approved these changes
Sep 14, 2021
awlauria
approved these changes
Sep 14, 2021
|
Thanks! I will open a PR to back port to v5.0.x. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
fix two bugs in
allocate_state_sharedin osc/rdma/osc_rdma_component.c