[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20240402113320.GO946323@nvidia.com>
Date: Tue, 2 Apr 2024 08:33:20 -0300
From: Jason Gunthorpe <jgg@...dia.com>
To: Shuah Khan <skhan@...uxfoundation.org>
Cc: Joao Martins <joao.m.martins@...cle.com>,
Mirsad Todorovac <mirsad.todorovac@....unizg.hr>,
iommu@...ts.linux.dev, Kevin Tian <kevin.tian@...el.com>,
Shuah Khan <shuah@...nel.org>, linux-kselftest@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [BUG] seltests/iommu: runaway ./iommufd consuming 99% CPU after
a failed assert()
On Wed, Mar 27, 2024 at 06:05:46PM -0600, Shuah Khan wrote:
> ASSERT_*() is supposed to exit the test right away. If this
> isn't happening it needs to be debugged.
We know it doesn't work in setup/teardown functions, you can see it in
the code it jumps back and does the teardown again in an infinite
loop.
If mising assert and ASSERT causes loops that is also a bug, we can't
guarentee that no libraries linked into this (like glibc) doesn't call
assert.
Jason
Powered by blists - more mailing lists