disambiguate msleep KASSERT diagnostics
ClosedPublic
Actions

Authored by emaste on Nov 21 2016, 3:32 PM.

Details

Reviewers

kib
jhb

Commits

rS312311: disambiguate msleep KASSERT diagnostics

Summary

I encountered "panic: msleep" while investigating an issue, which led to @kib's change in D8421. I applied the change here locally while investigating, because four different cases would result in the same panic string. There may be a better way to disambiguate these cases though; what do you think?

Diff Detail

Lint

Lint Skipped

Unit

Tests Skipped

Event Timeline

emaste updated this revision to Diff 22407.Nov 21 2016, 3:32 PM

emaste retitled this revision from to disambiguate msleep KASSERT diagnostics.

emaste updated this object.

emaste edited the test plan for this revision. (Show Details)

emaste added reviewers: kib, jhb.

emaste added a subscriber: kib.

kib accepted this revision.Nov 21 2016, 3:42 PM

kib edited edge metadata.

kib added inline comments.

sys/kern/kern_synch.c
155	What about this line ? Although I think that we cannot really get this asserted even for extremely broken kernels.
157	Might be say 'current thread is not running' instead of td.

This revision is now accepted and ready to land.Nov 21 2016, 3:42 PM

jhb accepted this revision.Nov 22 2016, 6:18 PM

jhb edited edge metadata.

jhb added inline comments.

sys/kern/kern_synch.c
155	Yeah, that is old (in 4.x p was NULL in the idle loop, but that was also all in assembly). It can probably go away. As bde@ would point out it will panic shortly anyway via a trap when td is dereferenced.
157	Or "curthread not running". I think the existing message is ok as well.