ACK: [Disco/Eoan] Revert "selftests/seccomp: Catch garbage on SECCOMP_IOCTL_NOTIF_RECV"

Andrea Righi andrea.righi at canonical.com
Fri Apr 10 12:07:38 UTC 2020


On Fri, Apr 10, 2020 at 09:00:06AM -0300, Thadeu Lima de Souza Cascardo wrote:
> BugLink: https://bugs.launchpad.net/bugs/1862588
> 
> This reverts upstream commit e4ab5ccc357b978999328fadae164e098c26fa40.
> 
> Running linux/tools/testing/selftests/seccomp/seccomp_bpf will fail
> with:
> 
> seccomp_bpf.c:3149:global.user_notification_basic:Expected -1
> (18446744073709551615) == ret (0)
> seccomp_bpf.c:3150:global.user_notification_basic:Expected EINVAL (22)
> == errno (0)
> global.user_notification_basic: Test failed at step #3
> [ FAIL ] global.user_notification_basic
> 
> The test is checking that the given structure which the kernel will write
> to is all zeroes and would return EINVAL otherwise. It's doing it because
> it wants userspace to have the possibility in the future to give data there
> indicating support for an extension that might be developed in the future.
> 
> As the test is there right now, not applying the breaking uABI fix might
> cause us to miss applications that would break in future kernels. As the
> backport for that is prone for more regression potential, we are deciding
> to revert the new test.
> 
> Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo at canonical.com>

Makes sense to me.

Acked-by: Andrea Righi <andrea.righi at canonical.com>



More information about the kernel-team mailing list