Re: [Bug runtime/6772] New: probing all functions of kernel/sched.c causes system crash on 2.6.27-rc1

[Srikar Dronamraju <srikar at linux dot vnet dot ibm dot com>]

Srini, 

I think I saw this couple of times on 0.6.2 version of SystemTap too esp
when run on livecd.  It seems to crash only when this script is one of
the first few scripts to be run. If the script is run after running good
number of scripts, the problem seems to vanish.  However I was not able
to reproduce it when running with KVM. 

-Srikar

* srinivasa at in dot ibm dot com <sourceware-bugzilla@sourceware.org> [2008-07-29 13:18:54]:

> Environment: Kernel:2.6.27-rc1, systemtap:20080726 snapshot elfutils:-0.131
> Arch:ppc64
> 
> Probing all functions of kernel/sched.c causes system crash on 2.6.27-rc1.
> 
> Some of the observations I made are listed below
> 1) I haven't seen this bug in previous kernel releases(2.6.26-rc's)
> 2) I tried disabling CONFIG_GROUP_SCHED, but still problem persists.
> 3) Iam able reproduce this bug only on ppc64 and not on other archs like x86_64.
> 4) Iam trying to isolate the problem to one single function in kernel/sched.c,
> but Iam not successful till now.
> 
> 
> I tried this script.
> ============================
> ./stap -vvvv -e 'probe kernel.function("*@kernel/sched.c") { printf("hello\n")}'
> 
> =======================
> llm27lp1.in.ibm.com login: cpu 0x0: Vector: d00 (Single Step) at [c000000000682f50]
>     pc: c00000000006e6b8: .scheduler_tick+0xf0/0x168
>     lr: c00000000006e6b8: .scheduler_tick+0xf0/0x168
>     sp: c0000000006831d0
>    msr: 8000000040101032
>   current = 0xc0000000005c3080
>   paca    = 0xc0000000006f3480
>     pid   = 0, comm = swapper
>         ld      r2,40(r1)
> enter ? for help
> [c0000000006831d0] c00000000006e610 .scheduler_tick+0x48/0x168 (unreliable)
> [c000000000683260] c00000000007efcc .update_process_times+0x60/0x8c
> [c0000000006832f0] c000000000096f74 .tick_periodic+0xc0/0xec
> [c000000000683380] c000000000096fd4 .tick_handle_periodic+0x34/0xb0
> [c000000000683420] c000000000025150 .timer_interrupt+0xb0/0x104
> [c0000000006834b0] c000000000003718 decrementer_common+0x118/0x180
> --- Exception: 901 (Decrementer) at c00000000000bcc8
> .raw_local_irq_restore+0x3c/0x40
> [link register   ] d000000000835cc8 .enter_kprobe_probe+0x90/0x1948
> [stap_6f5152cafd0831436846cf37ddd0b96b_34101]
> [c0000000006837a0] d000000000835e48 .enter_kprobe_probe+0x210/0x1948
> [stap_6f5152cafd0831436846cf37ddd0b96b_34101] (unreliable)
> [c000000000683840] c000000000403424 .kprobe_exceptions_notify+0x288/0x4b4
> [c0000000006838f0] c0000000004046d8 .notifier_call_chain+0x5c/0xcc
> [c000000000683990] c000000000091560 .notify_die+0x3c/0x50
> [c000000000683a30] c0000000004029c4 .program_check_exception+0x124/0x684
> [c000000000683af0] c00000000000500c program_check_common+0x10c/0x180
> --- Exception: 700 (Program Check) at c0000000003fef3c .schedule+0x0/0x8e4
> [link register   ] c000000000012af8 .cpu_idle+0x1d8/0x1e0
> [c000000000683de0] c000000000012a28 .cpu_idle+0x108/0x1e0 (unreliable)
> [c000000000683e60] c00000000041007c
> [c000000000683ee0] c000000000520bb4 .start_kernel+0x448/0x470
> [c000000000683f90] c000000000008568 .start_here_common+0x3c/0x54
> 
> -- 
>            Summary: probing all functions of kernel/sched.c causes system
>                     crash on 2.6.27-rc1
>            Product: systemtap
>            Version: unspecified
>             Status: NEW
>           Severity: normal
>           Priority: P2
>          Component: runtime
>         AssignedTo: systemtap at sources dot redhat dot com
>         ReportedBy: srinivasa at in dot ibm dot com
> 
> 
> http://sourceware.org/bugzilla/show_bug.cgi?id=6772
> 
> ------- You are receiving this mail because: -------
> You are the assignee for the bug, or are watching the assignee.