Hi,
While running kernel self tests on ppc64, tm/tm-signal-context-chk-vsx
tests fails with Oops message.
I was able to reproduce only twice out of 20 runs on next-20170324 only.
so it is difficult to bisect the commit causing the issue.
Machine Type; Power8 Power VM LPAR
Kernel version : 4.11.0-rc3-next-20170324
tm-signal-msr-r[14322]: bad frame in rt_sigreturn: 00003fffd38b0920 nip
00003fff8be2f094 lr 00003fff8bfe04d8
tm-signal-stack[14328]: bad frame in setup_rt_frame: 0000000000000000
nip 0000000010000d44 lr 0000000010000d28
Bad kernel stack pointer 3ffffc1e8180 at c00000000000b860
Oops: Bad kernel stack pointer, sig: 6 [#1]
SMP NR_CPUS=2048
NUMA
pSeries
Dumping ftrace buffer:
(ftrace buffer empty)
Modules linked in: binfmt_misc(E) rtc_generic(E) pseries_rng(E)
vmx_crypto(E) autofs4(E)
CPU: 0 PID: 14619 Comm: tm-signal-conte Tainted: G E
4.11.0-rc3-next-20170324-autotest #1
task: c000000767608400 task.stack: c0000005fa2f4000
NIP: c00000000000b860 LR: 0000000010001bb4 CTR: 0000000000000000
REGS: c00000000ecafd40 TRAP: 0700 Tainted: G E
(4.11.0-rc3-next-20170324-autotest)
MSR: 8000000102a03031 <SF,VEC,VSX,FP,ME,IR,DR,LE,TM[E]>
CR: 42000822 XER: 00000000
CFAR: c00000000000b7b4 SOFTE: 0
GPR00: 0000000000000025 00003ffffc1e8180 0000000010028300
0000000000000000
GPR04: 000000000000000a 00000000000000b0 0000000010020010
0000000000000000
GPR08: 00000000000000f8 0000000000000000 0000000000000000
0000000000000000
GPR12: 0000000000000000 00003fff7c98c440 0000000000000000
0000000000000000
GPR16: 0000000000000000 0000000000000000 0000000000000000
0000000000000000
GPR20: 0000000000000000 0000000000000000 0000000000000000
0000000000000000
GPR24: 0000000000000000 0000000000000000 0000000000000000
00003fff7c97f948
GPR28: 0000000010020010 000000001002041c 000000000000391b
0000000000004638
NIP [c00000000000b860] fast_exception_return+0x90/0x98
LR [0000000010001bb4] 0x10001bb4
Call Trace:
Instruction dump:
7c40e3a6 e9a100d8 7c7b03a6 e84101a0 7c4ff120 e8410170 7c5a03a6 e8010070
e8410080 e8610088 e8810090 e8210078 <4c000024> 48000000 e8610178
88ed023b
---[ end trace baea4a94af54b1b4 ]---
Steps to recreate:
-------------------
make -C tools/testing/selftests/powerpc/tm run_tests
test: tm_signal_context_chk_fpu
tags: git_version:unknown
!! child died by signal 6
failure: tm_signal_context_chk_fpu
selftests: tm-signal-context-chk-fpu [FAIL]
test: tm_signal_context_chk_vsx
tags: git_version:unknown
!! child died by signal 6
failure: tm_signal_context_chk_vsx
selftests: tm-signal-context-chk-vsx [FAIL]
Config file is attached.
--
Regard's
Abdul Haleem
IBM Linux Technology Centre
Abdul Haleem <[email protected]> writes:
> Hi,
>
> While running kernel self tests on ppc64, tm/tm-signal-context-chk-vsx
> tests fails with Oops message.
>
> I was able to reproduce only twice out of 20 runs on next-20170324 only.
> so it is difficult to bisect the commit causing the issue.
Can you try mainline as of this commit:
605df8d674ac ("selftests/powerpc: Replace stxvx and lxvx with stxvd2x/lxvd2x")
https://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux.git/commit/?id=605df8d674ac65e044a0bf4998b28c2f350b7f9e
cheers
On Tue, 2017-03-28 at 21:00 +1100, Michael Ellerman wrote:
> Abdul Haleem <[email protected]> writes:
>
> > Hi,
> >
> > While running kernel self tests on ppc64, tm/tm-signal-context-chk-vsx
> > tests fails with Oops message.
> >
> > I was able to reproduce only twice out of 20 runs on next-20170324 only.
> > so it is difficult to bisect the commit causing the issue.
>
> Can you try mainline as of this commit:
>
> 605df8d674ac ("selftests/powerpc: Replace stxvx and lxvx with stxvd2x/lxvd2x")
>
> https://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux.git/commit/?id=605df8d674ac65e044a0bf4998b28c2f350b7f9e
>
> cheers
>
Trace is not reproducible on mainline with above commit.
Cyril was able to reproduce it and is working on it.
--
Regard's
Abdul Haleem
IBM Linux Technology Centre
On Mon, 2017-04-03 at 14:28 +0530, Abdul Haleem wrote:
> On Tue, 2017-03-28 at 21:00 +1100, Michael Ellerman wrote:
> > Abdul Haleem <[email protected]> writes:
> >
> > > Hi,
> > >
> > > While running kernel self tests on ppc64, tm/tm-signal-context-chk-vsx
> > > tests fails with Oops message.
> > >
> > > I was able to reproduce only twice out of 20 runs on next-20170324 only.
> > > so it is difficult to bisect the commit causing the issue.
> >
> > Can you try mainline as of this commit:
> >
> > 605df8d674ac ("selftests/powerpc: Replace stxvx and lxvx with stxvd2x/lxvd2x")
> >
> > https://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux.git/commit/?id=605df8d674ac65e044a0bf4998b28c2f350b7f9e
> >
> > cheers
> >
>
>
> Trace is not reproducible on mainline with above commit.
>
> Cyril was able to reproduce it and is working on it.
>
>
>
>
cc Cyril Bur <[email protected]>
--
Regard's
Abdul Haleem
IBM Linux Technology Centre