aio_kqueue_test: Fix CID 1558429
AcceptedPublic
Actions

Authored by dab on Sun, Jan 5, 7:59 PM.

Details

Reviewers

ngie
vangyzen
asomers

Summary

Fix a Coverity error in the aio_kqueue_test that could theoretically
(but probably not realistically) cause over-indexing an array.

Test Plan

Ran test on local system; passed.

Diff Detail

Repository

rG FreeBSD src repository

Lint

Lint Passed

Unit

No Test Coverage

Build Status

Buildable 61512
Build 58396: arc lint + arc unit

Event Timeline

dab created this revision.Sun, Jan 5, 7:59 PM

dab held this revision as a draft.

Herald added subscribers: badger, asomers, imp. · View Herald TranscriptSun, Jan 5, 7:59 PM

Harbormaster completed remote builds in B61511: Diff 148794.Sun, Jan 5, 7:59 PM

dab edited the summary of this revision. (Show Details)Sun, Jan 5, 8:01 PM

Typo correction

Harbormaster completed remote builds in B61512: Diff 148795.Sun, Jan 5, 8:05 PM

dab published this revision for review.Sun, Jan 5, 8:05 PM

I think this is OK. But it would also be fine just to put an assert(j < max_queue_per_proc before the array dereference.

This revision is now accepted and ready to land.Mon, Jan 6, 1:59 PM

In D48328#1102276, @asomers wrote:

I think this is OK. But it would also be fine just to put an assert(j < max_queue_per_proc before the array dereference.

That was actually my first inclination, but didn't know if it was better to have the explicit message and let the test fail gracefully rather than have it abort on an assert.
Doing an assert would certainly be more straightforward and perhaps more clearly indicate that this is a scenario that should not ever happen.
I think I'll take your suggestion and revise this to use the assert method.

Revision Contents
Changeset List

Path

Size

tests/

sys/

aio/

aio_kqueue_test.c

9 lines

Diff 148795

View Options

tests/sys/aio/aio_kqueue_test.c

aio_kqueue_test: Fix CID 1558429AcceptedPublicActions