[LU-18899] o2iblnd: graceful handling of RDMA_CM_EVENT_ESTABLISHED

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.17.0
• Affects Version/s: Lustre 2.17.0
• Labels: lnet o2iblnd
• Severity: 3
• Assignee: Serguei Smirnov
• Reporter: Serguei Smirnov

Description

When handling CM events, o2iblnd expects RDMA_CM_EVENT_ESTABLISHED to be received when the connection is either in IBLND_CONN_PASSIVE_WAIT or IBLND_CONN_ACTIVE_CONNECT state, otherwise LBUG occurs.

Modify kiblnd_cm_callback() to handle this gracefully: report an error with relevant details and ignore the event rather than just crash.

This has been reported in the field with RoCE v2 setup.

 

 

Comments

[Peter Jones]

Merged for 2.17

[Gerrit Updater]

"Oleg Drokin <[e-mail]>" merged in patch https://review.whamcloud.com/c/fs/lustre-release/+/58711/
Subject: LU-18899 o2iblnd: handle RDMA_CM_EVENT_ESTABLISHED gracefully
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 64d7a7e42022d070cb15e79a0d1655e8df4eb33d

[Gerrit Updater]

"Serguei Smirnov <[e-mail]>" uploaded a new patch: https://review.whamcloud.com/c/fs/lustre-release/+/58711
Subject: LU-18899 o2iblnd: handle RDMA_CM_EVENT_ESTABLISHED gracefully
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: ca806d00aaee5c0594baaa1b099a7d6b39f23e5c