Page MenuHomeFreeBSD

nvme: limit I/O queue trackers to Identify MAXCMD
ClosedPublic

Authored by seuros on Aug 24 2026, 6:20 PM.
Tags
None
Referenced Files
F173195474: D59154.diff
Thu, Sep 24, 5:46 AM
Unknown Object (File)
Tue, Sep 22, 7:04 PM
Unknown Object (File)
Sat, Sep 19, 3:23 PM
Unknown Object (File)
Sat, Sep 19, 1:31 PM
Unknown Object (File)
Thu, Sep 17, 6:59 AM
Unknown Object (File)
Tue, Sep 15, 5:52 AM
Unknown Object (File)
Tue, Sep 15, 2:56 AM
Unknown Object (File)
Mon, Sep 14, 1:07 PM
Subscribers

Details

Summary

MAXCMD is the maximum number of commands the controller processes at
one time for a particular I/O queue. Allocating more trackers than
that only queues commands inside the controller. Controllers
reporting 0 are unaffected.

Approved by: ngie (co-mentor)

Diff Detail

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

Event Timeline

imp added inline comments.
sys/dev/nvme/nvme_ctrlr.c
200

This printf should be behind bootverbose, I think.

This revision is now accepted and ready to land.Aug 25 2026, 6:57 PM
sys/dev/nvme/nvme_ctrlr.c
200

I tested it with around 30 models, but not all.

I want to keep it like this for few weeks so we can identify the disks that are quirky.

After that we can remove it or guard it.

sys/dev/nvme/nvme_ctrlr.c
200

OK.

ngie added a subscriber: ngie.

Approved by: ngie (co-mentor)

seuros marked 2 inline comments as done.