Re: 2.6.18-rc4-mm1

[john stultz]

On Mon, 2006-08-14 at 21:27 +0200, Michal Piotrowski wrote:
> On 14/08/06, john stultz <johnstul@xxxxxxxxxx> wrote:
> > On Mon, 2006-08-14 at 21:01 +0200, Michal Piotrowski wrote:
> > > Andrew Morton wrote:
> > > > On Mon, 14 Aug 2006 16:02:52 +0200
> > > > "Michal Piotrowski" <michal.k.k.piotrowski@xxxxxxxxx> wrote:
> > > >> On 13/08/06, Andrew Morton <akpm@xxxxxxxx> wrote:
> > > >>> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.18-rc4/2.6.18-rc4-mm1/
> > > >>>
> > > >> Aug 14 15:35:10 ltg01-fedora kernel: BUG: unable to handle kernel
> > > >> paging request at virtual address fffeffbf
> > > >> Aug 14 15:35:10 ltg01-fedora kernel:  printing eip:
> > > >> Aug 14 15:35:10 ltg01-fedora kernel: c013d539
> > > >> Aug 14 15:35:10 ltg01-fedora kernel: *pde = 00004067
> > > >> Aug 14 15:35:10 ltg01-fedora kernel: *pte = 00000000
> > > >> Aug 14 15:35:10 ltg01-fedora kernel: Oops: 0000 [#1]
> > > >> Aug 14 15:35:10 ltg01-fedora kernel: 4K_STACKS PREEMPT SMP
> > > >> Aug 14 15:35:10 ltg01-fedora kernel: last sysfs file:
> > > >> /devices/platform/i2c-9191/9191-0290/temp2_input
> > > >> Aug 14 15:35:10 ltg01-fedora kernel: Modules linked in: ipv6 w83627hf
> > > >> hwmon_vid hwmon i2c_isa af_packet ip_conntrack_netbios
> > > >> _ns ipt_REJECT xt_state ip_conntrack nfnetlink xt_tcpudp
> > > >> iptable_filter ip_tables x_tables cpufreq_userspace p4_clockmod spe
> > > >> edstep_lib binfmt_misc thermal processor fan container evdev
> > > >> snd_intel8x0 snd_ac97_codec snd_ac97_bus snd_seq_dummy snd_seq_
> > > >> oss snd_seq_midi_event snd_seq snd_seq_device snd_pcm_oss
> > > >> snd_mixer_oss snd_pcm sk98lin snd_timer skge snd soundcore snd_pag
> > > >> e_alloc ide_cd intel_agp agpgart cdrom i2c_i801 rtc unix
> > > >> Aug 14 15:35:10 ltg01-fedora kernel: CPU:    1
> > > >> Aug 14 15:35:10 ltg01-fedora kernel: EIP:    0060:[<c013d539>]    Not
> > > >> tainted VLI
> > > >> Aug 14 15:35:10 ltg01-fedora kernel: EFLAGS: 00210286   (2.6.18-rc4-mm1 #97)
> > > >> Aug 14 15:35:10 ltg01-fedora kernel: EIP is at futex_wake+0x9c/0xcb
> > > >> Aug 14 15:35:10 ltg01-fedora kernel: eax: 0808c000   ebx: c0670a60
> > > >> ecx: d3a1dfa2   edx: fffeffbf
> > > >> Aug 14 15:35:10 ltg01-fedora kernel: esi: 00000000   edi: fffeffbf
> > > >> ebp: f4896f64   esp: f4896f40
> > > >> Aug 14 15:35:10 ltg01-fedora kernel: ds: 007b   es: 007b   ss: 0068
> > > >> Aug 14 15:35:10 ltg01-fedora kernel: Process firefox-bin (pid: 2210,
> > > >> ti=f4896000 task=f3d180f0 task.ti=f4896000)
> > > >> Aug 14 15:35:10 ltg01-fedora kernel: Stack: c0670a80 00000001 0808c000
> > > >> f4302e74 00000044 ffffffe7 0808c044 bf8f35b0
> > > >> Aug 14 15:35:10 ltg01-fedora kernel:        00000000 f4896f7c c013ed39
> > > >> 00000001 0808c044 7fffffff bf8f35b0 f4896fb4
> > > >> Aug 14 15:35:10 ltg01-fedora kernel:        c013ee84 7fffffff bf8f35b0
> > > >> 00000000 bf8f3528 00000000 f4896fa8 00000000
> > > >> Aug 14 15:35:10 ltg01-fedora kernel: Call Trace:
> > > >> Aug 14 15:35:10 ltg01-fedora kernel:  [<c013ed39>] do_futex+0x3c/0x92
> > > >> Aug 14 15:35:10 ltg01-fedora kernel:  [<c013ee84>] sys_futex+0xf5/0x101
> > > >> Aug 14 15:35:11 ltg01-fedora kernel:  [<c010312d>] sysenter_past_esp+0x56/0x8d
> > > >> Aug 14 15:35:11 ltg01-fedora kernel:  [<b7f1d410>] 0xb7f1d410
> > > >> Aug 14 15:35:11 ltg01-fedora kernel:  [<c0103fc1>] show_trace_log_lvl+0x12/0x22
> > > >> Aug 14 15:35:11 ltg01-fedora kernel:  [<c0104067>] show_stack_log_lvl+0x87/0x8f
> > > >> Aug 14 15:35:11 ltg01-fedora kernel:  [<c0104203>] show_registers+0x151/0x1d8
> > > >> Aug 14 15:35:11 ltg01-fedora kernel:  [<c010444d>] die+0x120/0x1f0
> > > >> Aug 14 15:35:11 ltg01-fedora kernel:  [<c01185cf>] do_page_fault+0x49d/0x580
> > > >> Aug 14 15:35:11 ltg01-fedora kernel:  [<c0303ba9>] error_code+0x39/0x40
> > > >> Aug 14 15:35:11 ltg01-fedora kernel:  [<c013ed39>] do_futex+0x3c/0x92
> > > >> Aug 14 15:35:11 ltg01-fedora kernel:  [<c013ee84>] sys_futex+0xf5/0x101
> > > >> Aug 14 15:35:11 ltg01-fedora kernel:  [<c010312d>] sysenter_past_esp+0x56/0x8d
> > > >> Aug 14 15:35:11 ltg01-fedora kernel:  =======================
> > > >> Aug 14 15:35:11 ltg01-fedora kernel: Code: 45 e8 39 41 04 75 22 8b 45
> > > >> ec 39 41 08 75 1a 83 7a 48 00 74 07 be ea ff ff ff eb
> > > >> 16 89 d0 46 e8 00 fd ff ff 3b 75 e0 7d 09 89 fa <8b> 3f 3b 55 dc eb c0
> > > >> 89 d8 e8 bd 60 1c 00 89 e0 25 00 f0 ff ff
> > > >> Aug 14 15:35:11 ltg01-fedora kernel: EIP: [<c013d539>]
> > > >> futex_wake+0x9c/0xcb SS:ESP 0068:f4896f40
> > > >
> > > > This is worrisome.  Is it reproducible?
> > >
> > > I don't know how to reproduce it, but it happened second time today.
> > >
> > > >  If so, reverting
> > > > futex_handle_fault-always-fails.patch and retesting would be useful.
> > >
> > > I reverted this patch.
> >
> > Just to be clear, the issue has shown itself without the patch?
> 
> No, it hasn't.

Thanks for the clarification. 

> > Or is
> > that not the case?
> 
> In current -mm only futex_handle_fault-always-fails.patch changes
> kernel/futex.c, so IMHO it's very probable that there is something
> wrong with this patch.

Yea, I'd believe that. I'm newish to the futex code, and as I said w/
the patch I'm not 100% sure its the right fix, but it does resolve an
issue we found.

I'll take another look to see if I futzed something up or if the fix is
just incomplete, although some more familiar eyes with the code would
probably help here.

thanks
-john

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/