[LTP] [PATCH v2] syscalls/execveat01: new test to verify execveat unlinked fd

[Jan Stancek]

----- Original Message -----
> Jan Stancek <jstancek@redhat.com> 於 2018年9月11日 週二 下午11:24寫道：
> > We should fix this for release. If EINVAL is expected behaviour,
> > test should give PASS.
> >
> > It's failing with 4.18:
> >
> > tst_test.c:1063: INFO: Timeout per run is 0h 05m 00s
> > execveat03.c:71: FAIL: execveat() returned unexpected errno: EINVAL
> >
> 
> Hi Jan,
> The fix of original regression, ie. EINVAL returns in execveat03 test case,
> has been added to 4.18-stable and 4.14-stable:
> https://www.spinics.net/lists/stable/msg255906.html
> https://www.spinics.net/lists/stable/msg255724.html
> If give EINVAL as PASS, the case will PASS with or without the regression
> fix. I'm not sure which behavior of execveat03 is more appropriate.
> Please advise.

Hi Eddie,

thanks for the links. I think I confused myself with the comment in testcase.
I thought EINVAL is expected behavior with the fix.

For now, I only added comment with commit id of the fix to testcase.

Thanks,
Jan

> 
> Thanks,
> Eddie
>