Page MenuHomeFreeBSD

nvme/nda: Fail all nvme I/Os after controller fails
ClosedPublic

Authored by imp on Jul 28 2021, 10:33 PM.
Tags
None
Referenced Files
Unknown Object (File)
Tue, Jun 24, 3:35 AM
Unknown Object (File)
Sat, Jun 21, 5:32 PM
Unknown Object (File)
Fri, Jun 20, 7:29 AM
Unknown Object (File)
Thu, Jun 19, 6:01 PM
Unknown Object (File)
Sun, Jun 15, 9:44 PM
Unknown Object (File)
Thu, Jun 12, 9:03 PM
Unknown Object (File)
Mon, Jun 9, 4:46 PM
Unknown Object (File)
Jun 1 2025, 3:50 AM
Subscribers

Details

Summary

Once the controller has failed, fail all I/O w/o sending it to the
device. The reset of the nvme driver won't schedule any I/O to the
failed device, and the controller is in an indeterminate state and can't
accept I/O. Fail both at the top end of the sim and the bottom
end. Don't bother queueing up the I/O for failure in a different task.

Sponsored by: Netflix

Diff Detail

Repository
rG FreeBSD src repository
Lint
Lint Not Applicable
Unit
Tests Not Applicable

Event Timeline

imp requested review of this revision.Jul 28 2021, 10:33 PM
sys/dev/nvme/nvme_qpair.c
1080–1081

Should this comment get updated as nvme_qpair_manual_complete_request() isn't posting to a task?

sys/dev/nvme/nvme_sim.c
262

I'm not seeing a definition for CAM_REQ_DEV_NOT_THERE. Should this be CAM_DEV_NOT_THERE?

update per review / chuck

imp marked 2 inline comments as done.Aug 2 2021, 10:15 PM
This revision is now accepted and ready to land.Aug 2 2021, 10:52 PM
sys/dev/nvme/nvme_sim.c
264

It is not a problem, but is it really needed?