This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
Re: How to get the detailed kernel stack trace if specified fuction takes too long to finish?
- From: liang xie <xieliang007 at gmail dot com>
- To: Josh Stone <jistone at redhat dot com>
- Cc: systemtap at sourceware dot org
- Date: Thu, 3 Apr 2014 10:47:10 +0800
- Subject: Re: How to get the detailed kernel stack trace if specified fuction takes too long to finish?
- Authentication-results: sourceware.org; auth=none
- References: <CADu=CFo=AtrBJLacPTa_6EfQA7vao=bSuT2Yns45x-hWYfmWZA at mail dot gmail dot com> <5335A3A6 dot 8010300 at redhat dot com> <CADu=CFrnaoiWcD8Jt8uFd1Q087Kyp28gKukKH_JD7craJeF3xw at mail dot gmail dot com> <533C3B93 dot 6060404 at redhat dot com>
Yeh, systemtap-1.8 works for me, thanks! now i can get strace trace
like this with the above scheduler script:
long sys_write in tid 11892
0xffffffff814ed0a8 : thread_return+0x6d6/0x77e [kernel] (inexact)
0xffffffff812c4a41 : intel_idle+0xc1/0x170 [kernel] (inexact)
0xffffffff81097c6d : sched_clock_cpu+0xcd/0x110 [kernel] (inexact)
0xffffffff81009e3e : cpu_idle+0xee/0x110 [kernel] (inexact)
0xffffffff814e5f23 : start_secondary+0x202/0x245 [kernel] (inexact)
long sys_write in tid 16716
0xffffffff814ed0a8 : thread_return+0x6d6/0x77e [kernel] (inexact)
0xffffffff812c4a41 : intel_idle+0xc1/0x170 [kernel] (inexact)
0xffffffff81097c6d : sched_clock_cpu+0xcd/0x110 [kernel] (inexact)
0xffffffff81009e3e : cpu_idle+0xee/0x110 [kernel] (inexact)
0xffffffff814e5f23 : start_secondary+0x202/0x245 [kernel] (inexact)
long sys_write in tid 16716
0xffffffff814ed0a8 : thread_return+0x6d6/0x77e [kernel] (inexact)
0xffffffff812c4a41 : intel_idle+0xc1/0x170 [kernel] (inexact)
0xffffffff81097c6d : sched_clock_cpu+0xcd/0x110 [kernel] (inexact)
0xffffffff81009e3e : cpu_idle+0xee/0x110 [kernel] (inexact)
0xffffffff814e5f23 : start_secondary+0x202/0x245 [kernel] (inexact)
long sys_write in tid 17191
0xffffffff814ed0a8 : thread_return+0x6d6/0x77e [kernel] (inexact)
0xffffffff812c4a41 : intel_idle+0xc1/0x170 [kernel] (inexact)
0xffffffff81097c6d : sched_clock_cpu+0xcd/0x110 [kernel] (inexact)
0xffffffff81009e3e : cpu_idle+0xee/0x110 [kernel] (inexact)
0xffffffff814e5f23 : start_secondary+0x202/0x245 [kernel] (inexact)
But still no interesting filesystem related stack trace be found
On Thu, Apr 3, 2014 at 12:32 AM, Josh Stone <jistone@redhat.com> wrote:
> On 04/02/2014 03:25 AM, liang xie wrote:
>> I tried the last scheduler script, the output just like this:
>> # stap --all-modules y.stp
>> WARNING: missing unwind/symbol data for module 'uprobes'
>> long sys_write in tid 10162
>> Systemtap probe: kernel.trace("sched_switch")
>> long sys_write in tid 11510
>> Systemtap probe: kernel.trace("sched_switch")
>> long sys_write in tid 11288
>> Systemtap probe: kernel.trace("sched_switch")
>>
>> I could not figure out why "print_backtrace()" didn't output any log:)
>
> Ah, this "Systemtap probe:" message is fairly old.
>
>>>> My OS:2.6.32-220.el6.x86_64
>>>> # stap -V
>>>> Systemtap translator/driver (version 1.6/0.152 non-git sources)
>
> Is it possible for you to update your systemtap rpm? 1.6 was released
> with RHEL 6.2 (presumably CentOS 6.2 as well), and we've since released
> systemtap-1.7 in 6.3, 1.8 in 6.4, and 2.3 in 6.5. I believe at least
> systemtap-1.8 will work better for you here.
>
> Josh