[DNM] UCT/IB/MLX5: add wqe token and failover arm - #11684
Open
jeynmann wants to merge 18 commits into
Open
Conversation
Rebuild the transport implementation on top of the API-only failover_ext_uct branch: - align naming to the finalized "purge" API (outstanding purge callback, params and field mask), dropping the intermediate "extract"/flags shape - switch the mlx5 external plugin path from qp_query to ep_query, with RX token derivation performed by the iface query - re-add the uct_ep_outstanding_purge()/uct_ep_failover_enable() prototypes and the uct_ib_mlx5_ext_ep_outstanding_purge() dispatch wrapper - update gtests (test_ib_mlx5_ext, test_uct_outstanding_purge) accordingly
jeynmann
force-pushed
the
failover_ext_uct_wqe_meta
branch
from
July 23, 2026 16:21
a11c3c9 to
b2770dc
Compare
jeynmann
force-pushed
the
failover_ext_uct_wqe_meta
branch
2 times, most recently
from
July 27, 2026 04:16
dd9adc1 to
42bf8e2
Compare
# Conflicts: # src/uct/ib/mlx5/rc/rc_mlx5_iface.c
jeynmann
force-pushed
the
failover_ext_uct_wqe_meta
branch
from
July 27, 2026 10:12
738e18e to
35acf5a
Compare
jeynmann
changed the title
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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?
Save tokens on the tx side and add support for failover arm.
Why?
Need to keep tokens on the tx side so the plugin can purge outstanding operations using tokens received from the remote side.
Need to arm the wqe so failover and tx progress do not act on the wqe and pending operations together.
How?
Add an ep param to indicate that failover is enabled (failover mode + iface cap support).
Allocate and set tokens only when ep failover is enabled.
Add an arm api to let failover arm the ep.