Re: 100% iowait on one of cpus in current -git

From: Maxim Levitsky
Date: Mon Oct 22 2007 - 08:22:19 EST


On Monday 22 October 2007 13:19:08 Fengguang Wu wrote:
> On Mon, Oct 22, 2007 at 12:58:11PM +0200, Maxim Levitsky wrote:
> > On Monday 22 October 2007 12:55:25 Fengguang Wu wrote:
> > > On Mon, Oct 22, 2007 at 12:40:24PM +0200, Maxim Levitsky wrote:
> > > > On Monday 22 October 2007 12:22:10 Peter Zijlstra wrote:
> > > > > > [ 673.365631] pdflush D c21bdecc 0 221 2
> > > > > > [ 673.365635] c21bdee0 00000046 00000002 c21bdecc c21bdec4 00000000 c21b3000 00000002
> > > > > > [ 673.365643] c0134892 c21b3164 c1e00200 00000001 c7109280 c21bdec0 c03ff849 c21bdef0
> > > > > > [ 673.365650] 00052974 00000000 000000ff 00000000 00000000 00000000 c21bdef0 000529dc
> > > > > > [ 673.365657] Call Trace:
> > > > > > [ 673.365659] [<c03fd728>] schedule_timeout+0x48/0xc0
> > > > > > [ 673.365663] [<c03fd50e>] io_schedule_timeout+0x5e/0xb0
> > > > > > [ 673.365667] [<c0170d11>] congestion_wait+0x71/0x90
> > > > > > [ 673.365671] [<c016b92e>] wb_kupdate+0x9e/0xf0
> > > > > > [ 673.365675] [<c016beb2>] pdflush+0x102/0x1d0
> > > > > > [ 673.365679] [<c013fa82>] kthread+0x42/0x70
> > > > > > [ 673.365683] [<c01050df>] kernel_thread_helper+0x7/0x18
> > > > > >
> > > > >
> > > > > That looks more like the inode writeback patches from Wu than the per
> > > > > bdi dirty stuff. The later typically hangs in balance_dirty_pages().
> > > > >
> > > > >
> > > > >
> > > >
> > > > Yes, you are right,
> > > >
> > > > both revisions 1f7d6668c29b1dfa307a44844f9bb38356fc989b and 3e26c149c358529b1605f8959341d34bc4b880a3 work fine
> > > > But I didn't pay attention that those are before f4a1c2bce002f683801bcdbbc9fd89804614fb6b.
> > > > So, back to the drawing board.... :-)
> > > >
> > > > Will test revision 2e6883bdf49abd0e7f0d9b6297fc3be7ebb2250b, just after writeback patches.
> > >
> > > Thank you. I'll try if I can reproduce it locally...
> > >
> > > Fengguang
> > >
> > >
> >
> > Bingo,
> >
> > Revision 2e6883bdf49abd0e7f0d9b6297fc3be7ebb2250b shows this bug.
> >
> > I will now bisect to find exact patch that caused this bug,
>
> This one is most relevant:
>
> 1f7decf6d9f06dac008b8d66935c0c3b18e564f9
> writeback: introduce writeback_control.more_io to indicate more io
Exactly.
>
> Still compiling the kernel...
>
> Thank you,
> Fengguang
>
>
Hi,


I Bisected this bug to exactly this commit:

2e6883bdf49abd0e7f0d9b6297fc3be7ebb2250b
writeback: introduce writeback_control.more_io to indicate more io

Reverting it and compiling latest git shows no more signs of that bug.
Thanks,
Best regards,
Maxim Levitsky
-
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/