make check failing on times with -m32

[Dmitry V. Levin]

On Mon, Apr 18, 2016 at 12:09:55AM +0100, Steve McIntyre wrote:
> On Thu, Jan 14, 2016 at 02:31:01AM +0000, Steve McIntyre wrote:
> >
> >Coincidentally, I've just been doing exactly that already this
> >evening. I'm in the middle of hacking on a script to run tests
> >periodically on the various Debian porter boxes, so should be able to
> >give you some better and more timely results in future too.
> >
> >Results from VERBOSE=1 shortly.
> 
> Wow, this has taken a while to get back to - sorry... :-(
> 
> However, I now have stuff in place to run a HEAD build automatically
> for all the Debian (release) arches nightly and push logs out. See
> 
>   http://www.einval.com/debian/strace/build-logs/
> 
> for the logs, split by architecture. Please let me know if this is
> useful - I'm open to suggestions for changes.

Some simple ideas:
- keep git quiet: specify -q option to git clone;
- print key information about the build environment, e.g.
uname -rm, version of kernel headers, libc version, gcc version.

> In the meantime, I'm still struggling to get strace 4.11 to pass test
> suite checks for all our arches. I've backported fixes for test suite
> failures in lseek and llseek and now armel and armhf build OK. But
> pselect6() is failing due to a mismatch in output - see
> pselect6-NSIG.patch attached.

Thanks, fixed.

> I still have outstanding failures for:
> 
> mips/mipsel: prctl-seccomp-filter-v

This is somewhat similar to lseek/llseek case: libc does something
on its own that has to be filtered out.  Should be fixed now.

> 	     prctl-seccomp-strict

I'm aware of kernel issues with both prctl(PR_SET_SECCOMP, SECCOMP_MODE_STRICT)
and seccomp(SECCOMP_SET_MODE_STRICT) on x32 and mips, but it looks strange
when seccomp works while prctl fails.

> 	     pselect6 *

Should be fixed now.

> 	     rt_sigsuspend

This might be a big endian bug in the test.
It passes on parisc, though.

> mips64el:    bpf
> 	     execveat-v
> 	     execveat
> 	     fcntl
> 	     fstat
> 	     lstat
> 	     membarrier
> 	     mlock2
> 	     newfstatat
> 	     pselect6 *
> 	     stat
> 	     userfaultfd

Something is wrong with strace support on this architecture if it cannot
decode syscall numbers properly.


-- 
ldv
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <http://lists.strace.io/pipermail/strace-devel/attachments/20160418/30ab1faa/attachment.bin>