mbox series

[SRU,B/D/E/F/U,0/1] Fix seccomp_bpf on powerpc

Message ID 20200630170608.1270030-1-cascardo@canonical.com
Headers show
Series Fix seccomp_bpf on powerpc | expand

Message

Thadeu Lima de Souza Cascardo June 30, 2020, 5:06 p.m. UTC
BugLink: https://bugs.launchpad.net/bugs/1885757

The patch has just been submitted upstream, still pending review.

[Impact]
seccomp_bpf test fails on powerpc and ends up being disabled on some series, or
causing engineers to waste their time verifying the same failures are the only
ones we see every kernel release.

[Test case]
Run the test and notice there are no more failures.

[Regression potential]
We may break the test on different architectures. That doesn't break users,
though, as the changes are only on tests. It has been tested at least on
ppc64el and amd64.

Thadeu Lima de Souza Cascardo (1):
  UBUNTU: SAUCE: selftests/seccomp: fix ptrace tests on powerpc

 tools/testing/selftests/seccomp/seccomp_bpf.c | 24 +++++++++++++++----
 1 file changed, 20 insertions(+), 4 deletions(-)

Comments

Andrea Righi July 1, 2020, 7:09 a.m. UTC | #1
On Tue, Jun 30, 2020 at 02:06:06PM -0300, Thadeu Lima de Souza Cascardo wrote:
> BugLink: https://bugs.launchpad.net/bugs/1885757
> 
> The patch has just been submitted upstream, still pending review.

Maybe we include this link in the patches:
https://lore.kernel.org/linux-kselftest/20200630164739.1268222-1-cascardo@canonical.com/T/#u

> 
> [Impact]
> seccomp_bpf test fails on powerpc and ends up being disabled on some series, or
> causing engineers to waste their time verifying the same failures are the only
> ones we see every kernel release.
> 
> [Test case]
> Run the test and notice there are no more failures.
> 
> [Regression potential]
> We may break the test on different architectures. That doesn't break users,
> though, as the changes are only on tests. It has been tested at least on
> ppc64el and amd64.

The code makes sense to me and the impact of the change is very
isolated, considering that it affects only the kernel tests. It has been
also tested on ppc64el and amd64, so:

Acked-by: Andrea Righi <andrea.righi@canonical.com>
Juerg Haefliger July 1, 2020, 7:54 a.m. UTC | #2
On Tue, 30 Jun 2020 14:06:06 -0300
Thadeu Lima de Souza Cascardo <cascardo@canonical.com> wrote:

> BugLink: https://bugs.launchpad.net/bugs/1885757
> 
> The patch has just been submitted upstream, still pending review.
> 
> [Impact]
> seccomp_bpf test fails on powerpc and ends up being disabled on some series, or
> causing engineers to waste their time verifying the same failures are the only
> ones we see every kernel release.
> 
> [Test case]
> Run the test and notice there are no more failures.
> 
> [Regression potential]
> We may break the test on different architectures. That doesn't break users,
> though, as the changes are only on tests. It has been tested at least on
> ppc64el and amd64.
> 
> Thadeu Lima de Souza Cascardo (1):
>   UBUNTU: SAUCE: selftests/seccomp: fix ptrace tests on powerpc
> 
>  tools/testing/selftests/seccomp/seccomp_bpf.c | 24 +++++++++++++++----
>  1 file changed, 20 insertions(+), 4 deletions(-)


This code makes my head spin but given the description it seems to make sense.
Any particular reason why this wasn't tested on more architectures?

It's 'only' selftests, so:

Acked-by: Juerg Haefliger <juergh@canonical.com>
Thadeu Lima de Souza Cascardo July 1, 2020, 10:05 a.m. UTC | #3
On Wed, Jul 01, 2020 at 09:54:42AM +0200, Juerg Haefliger wrote:
> On Tue, 30 Jun 2020 14:06:06 -0300
> Thadeu Lima de Souza Cascardo <cascardo@canonical.com> wrote:
> 
> > BugLink: https://bugs.launchpad.net/bugs/1885757
> > 
> > The patch has just been submitted upstream, still pending review.
> > 
> > [Impact]
> > seccomp_bpf test fails on powerpc and ends up being disabled on some series, or
> > causing engineers to waste their time verifying the same failures are the only
> > ones we see every kernel release.
> > 
> > [Test case]
> > Run the test and notice there are no more failures.
> > 
> > [Regression potential]
> > We may break the test on different architectures. That doesn't break users,
> > though, as the changes are only on tests. It has been tested at least on
> > ppc64el and amd64.
> > 
> > Thadeu Lima de Souza Cascardo (1):
> >   UBUNTU: SAUCE: selftests/seccomp: fix ptrace tests on powerpc
> > 
> >  tools/testing/selftests/seccomp/seccomp_bpf.c | 24 +++++++++++++++----
> >  1 file changed, 20 insertions(+), 4 deletions(-)
> 
> 
> This code makes my head spin but given the description it seems to make sense.
> Any particular reason why this wasn't tested on more architectures?
> 

No easy access to other architectures, and I wanted to get at least powerpc
done with. I will tackle other architectures in the future. s390 should be
next. Then, arm64.

Cascardo.

> It's 'only' selftests, so:
> 
> Acked-by: Juerg Haefliger <juergh@canonical.com>
Kleber Souza July 1, 2020, 4:33 p.m. UTC | #4
On 2020-06-30 19:06, Thadeu Lima de Souza Cascardo wrote:
> BugLink: https://bugs.launchpad.net/bugs/1885757
> 
> The patch has just been submitted upstream, still pending review.
> 
> [Impact]
> seccomp_bpf test fails on powerpc and ends up being disabled on some series, or
> causing engineers to waste their time verifying the same failures are the only
> ones we see every kernel release.
> 
> [Test case]
> Run the test and notice there are no more failures.
> 
> [Regression potential]
> We may break the test on different architectures. That doesn't break users,
> though, as the changes are only on tests. It has been tested at least on
> ppc64el and amd64.
> 
> Thadeu Lima de Souza Cascardo (1):
>   UBUNTU: SAUCE: selftests/seccomp: fix ptrace tests on powerpc
> 
>  tools/testing/selftests/seccomp/seccomp_bpf.c | 24 +++++++++++++++----
>  1 file changed, 20 insertions(+), 4 deletions(-)
> 

Applied to bionic/linux, disco/linux, eoan/linux and focal/linux with adding
the URL to the upstream thread as suggested by Andrea.

Thanks,
Kleber
Seth Forshee July 1, 2020, 9:05 p.m. UTC | #5
On Tue, Jun 30, 2020 at 02:06:06PM -0300, Thadeu Lima de Souza Cascardo wrote:
> BugLink: https://bugs.launchpad.net/bugs/1885757
> 
> The patch has just been submitted upstream, still pending review.
> 
> [Impact]
> seccomp_bpf test fails on powerpc and ends up being disabled on some series, or
> causing engineers to waste their time verifying the same failures are the only
> ones we see every kernel release.
> 
> [Test case]
> Run the test and notice there are no more failures.
> 
> [Regression potential]
> We may break the test on different architectures. That doesn't break users,
> though, as the changes are only on tests. It has been tested at least on
> ppc64el and amd64.

Applied to unstable/master-next, thanks!