[Bug runtime/20279] New: in parallel testsuite mode, we're getting SIGABORTs

[Bug runtime/20279] New: in parallel testsuite mode, we're getting SIGABORTs

[dsmith at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=20279

            Bug ID: 20279
           Summary: in parallel testsuite mode, we're getting SIGABORTs
           Product: systemtap
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: runtime
          Assignee: systemtap at sourceware dot org
          Reporter: dsmith at redhat dot com
  Target Milestone: ---

When running the testsuite in parallel mode, I see several of the following
messages on the console:

[ 5619.908020] rm[61446]: unhandled signal 11 at 00003fffdc83ff90 nip
00003fffa8a51f38 lr 00003fffa8a34008 code 30001
[ 6611.824811] bz5274[54850]: unhandled signal 11 at 0000000000000000 nip
0000000000000000 lr 0000000000000000 code 30001

On RHEL7 ppc64, I see around 7 of these during a full testsuite run. I don't
see these errors when the testsuite is run in non-parallel mode.

I'm not 100% sure it is related, but the testsuite also got hung at the end of
this run, waiting on a 'loop' process (from unprivileged_probes.exp) to be
killed.

-- 
You are receiving this mail because:
You are the assignee for the bug.

[Bug runtime/20279] in parallel testsuite mode, we're getting SIGABORTs

[dsmith at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=20279

--- Comment #1 from David Smith <dsmith at redhat dot com> ---
For reference sake, here are the 7 messages:

[ 1206.152228] times[1039]: unhandled signal 11 at ffffffffffffffff nip
00003fffa18147b4 lr 0000000010000700 code 30001
[ 1568.066789] times[18596]: unhandled signal 11 at ffffffff nip 0fd970d8 lr
1000050c code 30001
[ 2194.143278] times[9533]: unhandled signal 11 at ffffffffffffffff nip
00003fff8d1347b4 lr 0000000010000700 code 30001
[ 2409.046690] times[30608]: unhandled signal 11 at ffffffff nip 0fd970d8 lr
1000050c code 30001
[ 5619.908020] rm[61446]: unhandled signal 11 at 00003fffdc83ff90 nip
00003fffa8a51f38 lr 00003fffa8a34008 code 30001
[ 6611.824811] bz5274[54850]: unhandled signal 11 at 0000000000000000 nip
0000000000000000 lr 0000000000000000 code 30001
[ 7880.699568] stap[21140]: unhandled signal 11 at 0000000000000000 nip
0000000000000000 lr 00003fff7b2cd188 code 30001

So, this happened to several different exes: times, rm, bz5274, and stap. I'm
also unsure of why the kernel reported this on the console, I don't believe it
normally does that.

-- 
You are receiving this mail because:
You are the assignee for the bug.

[Bug runtime/20279] in parallel testsuite mode, we're getting SIGABORTs

[jistone at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=20279

Josh Stone <jistone at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jistone at redhat dot com

--- Comment #2 from Josh Stone <jistone at redhat dot com> ---
FWIW, signal 11 is SIGSEGV.

I do seem to recall that the rlimit test triggers these on purpose.  I would
not expect that to affect others tests running in parallel though.

-- 
You are receiving this mail because:
You are the assignee for the bug.

[Bug runtime/20279] in parallel testsuite mode, we're getting SIGSEGVs

[dsmith at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=20279

David Smith <dsmith at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|in parallel testsuite mode, |in parallel testsuite mode,
                   |we're getting SIGABORTs     |we're getting SIGSEGVs

--- Comment #3 from David Smith <dsmith at redhat dot com> ---
(In reply to Josh Stone from comment #2)
> FWIW, signal 11 is SIGSEGV.

Sigh. That's what I get for relying on my memory.

> I do seem to recall that the rlimit test triggers these on purpose.  I would
> not expect that to affect others tests running in parallel though.

When I run the rlimit.exp test by itself, I don't see any of those messages on
the console. That test will involve stap calling getrlimit/setrlimit, but that
should only affect that particular stap pid (and its descendants), not other
previous or future stap processes.

There is another test, bad-code.exp, that purposely sends a SIGSEGV, but when
that test is run I don't see the message on the console.

I need to poke around in the kernel and see why it prints these messages - why
these SIGSGVs are different than regular SIGSEGVs.

-- 
You are receiving this mail because:
You are the assignee for the bug.

[Bug runtime/20279] in parallel testsuite mode, we're getting SIGSEGVs

[jistone at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=20279

--- Comment #4 from Josh Stone <jistone at redhat dot com> ---
This exact "unhandled signal" message appears to be powerpc only, from
_exception() in arch/powerpc/kernel/traps.c.  It looks like there are a few
calls in there with SIGSEGV.

-- 
You are receiving this mail because:
You are the assignee for the bug.