Re: random bug generator (was: 2.6.21-git4 Scheduler, NOHZ, VFS bugs)

[Kay Sievers <kay.sievers@vrfy.org>]

On Sat, 2007-05-05 at 20:04 +0200, Michal Piotrowski wrote:
> Hi,
> 
> Michal Piotrowski napisał(a):
> > On 04/05/07, Andrew Morton <akpm@linux-foundation.org> wrote:
> >> On Fri, 04 May 2007 18:20:51 +0200 Michal Piotrowski
> >> <michal.k.k.piotrowski@gmail.com> wrote:
> >>
> >> > I ran this script tree times,
> >> >
> >> > #! /bin/sh
> >> >
> >> > for i in `find /sys/ -type f`
> >> > do
> >> >     echo "wyświetlam $i"
> >> >     sudo cat $i > /dev/null
> >> > done
> >> >
> >> > First run - scheduler bug
> >> > Second run - NOHZ bug
> >> > Third - VFS bug
> >> >
> >> > Hmmmm...
> >> >
> >> > [93298.252601] BUG: at /mnt/md0/devel/linux-git/kernel/sched.c:3241
> >> add_preempt_count()
> >> > [93298.260334]  [<c0105039>] show_trace_log_lvl+0x1a/0x2f
> >> > [93298.265507]  [<c0105720>] show_trace+0x12/0x14
> >> > [93298.269974]  [<c01057d2>] dump_stack+0x16/0x18
> >> > [93298.274434]  [<c011d18a>] add_preempt_count+0x89/0x8b
> >> > [93298.279501]  [<c0126492>] irq_enter+0xd/0x2e
> >> > [93298.283788]  [<c0114c1e>] smp_apic_timer_interrupt+0x2a/0x84
> >> > [93298.289458]  [<c0104b2b>] apic_timer_interrupt+0x33/0x38
> >> > [93298.294783]  [<c0257b79>] show_uevent+0x58/0xcd
> >> > [93298.299329]  =======================
> >> > [93390.468056] NOHZ: local_softirq_pending 22
> >> > [93447.105850] NOHZ: local_softirq_pending 22
> >> > [93450.332884] BUG: unable to handle kernel paging request at
> >> virtual address 3e343c0c
> >> > [93450.340626]  printing eip:
> >> > [93450.343333] c018e2bc
> >> > [93450.345520] *pde = 00000000
> >> > [93450.348314] Oops: 0000 [#1]
> >>
> >> Nice.  What was the last file which it read before crashing?
> >>
> >> Are you able to consistently crash it by reading just that file?
> > 
> > Random files, random bugs (usually it starts with a slab corruption).
> > 
> > git-bisect will give an answer.
> 
> So this is a random bug generator...
> 
> git-bisect bad
> 16574dccd8f62dc1b585325f8a6a0aab10047ed8 is first bad commit
> commit 16574dccd8f62dc1b585325f8a6a0aab10047ed8
> Author: Kay Sievers <kay.sievers@vrfy.org>
> Date:   Fri Apr 6 01:40:38 2007 +0200
> 
>     Driver core: make uevent-environment available in uevent-file
> 
>     This allows sysfs to show the environment variables that are available
>     if the uevent happens.  This lets userspace not have to cache all of
>     this information as the kernel already knows it.
> 
>     Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
>     Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
> 
> :040000 040000 34152258985968083c04ab003ced54d73e92a790 c3301e8cd7705c293817c6747b3cc59d6cbf14bb M      drivers
> 
> http://www.stardust.webpages.pl/files/tbf/bitis-gabonica/2.6.21-git4/git-config

Greg fixed this:
  http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=c7308c81a8220ab68eebfadde37db881a2800064

Kay