Re: [RFC PATCH] audit-testsuite: improve our chances of losing records in lost_reset
On 2018-12-13 18:23, Paul Moore wrote:
On Thu, Dec 13, 2018 at 6:17 PM Paul Moore
<paul(a)paul-moore.com> wrote:
> If the point of the lost_reset test is to flood the system with audit
> records, why are we restricting ourselves with a filter? Let's log
> everything.
>
> tests/lost_reset/test | 4 ++--
This resolves a problem I've been seeing with a fairly stubborn
system. I would be very curious to hear if any others were
experiencing problems with the lost_reset that this patch resolves.
This saddistic test would have brought the system to its knees before
the queue rework. This still works fine on my system, so this is
progress. I suspect that something weird was going on with your test
such that it wasn't getting the right PID. I don't know how this is
possible, but I have seen this same thing recently on the FUSE umount2
test where I needed to bump up the PID by one to catch the right task.
paul moore