Re: latest -git: suspend: unable to handle kernel paging request (was Re: no_console_suspend doesn't work?)

["Pekka Enberg" <penberg@cs.helsinki.fi>]

Hi Rafael,

On Fri, Aug 22, 2008 at 1:10 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
>> I don't know the suspend code at all but the bogus pointer coming from
>> hibernation_snapshot() seems suspicious.
>
> Which one?

This trace here (which has IDE too):

> > Call Trace:
> >  [<c0683402>] ? wait_for_completion+0x12/0x20
>
> Here the block layer is passing a pointer to a non-slab page to
> kmem_cache_free(). Assuming we can rely on the stack trace, of
> course...
>
> >  [<c018976e>] ? mempool_free_slab+0xe/0x10
> >  [<c036a7b0>] ? blk_end_sync_rq+0x0/0x30
> >  [<c01b3c9c>] ? kmem_cache_free+0x5c/0x200
> >  [<c018976e>] ? mempool_free_slab+0xe/0x10
> >  [<c018979c>] ? mempool_free+0x2c/0x90
> >  [<c0367252>] ? __blk_put_request+0x62/0x90
> >  [<c036764c>] ? blk_put_request+0x2c/0x50
> >  [<c0470732>] ? generic_ide_resume+0xa2/0xf0     <-- IDE
> >  [<c03f767e>] ? device_resume+0x32e/0x380
> >  [<c01687d1>] ? hibernation_snapshot+0xa1/0x220 <--- Suspend
> >  [<c013b57b>] ? printk+0x1b/0x20
> >  [<c0168a30>] ? hibernate+0xe0/0x180
> >  [<c01674e0>] ? state_store+0x0/0xd0
> >  [<c016759f>] ? state_store+0xbf/0xd0
> >  [<c01674e0>] ? state_store+0x0/0xd0
> >  [<c03749d4>] ? kobj_attr_store+0x24/0x30
> >  [<c01f8ef2>] ? sysfs_write_file+0xa2/0x100
> >  [<c01ba946>] ? vfs_write+0x96/0x130
> >  [<c01f8e50>] ? sysfs_write_file+0x0/0x100
> >  [<c01bae8d>] ? sys_write+0x3d/0x70
> >  [<c0104f3b>] ? sysenter_do_call+0x12/0x3f

But like I said, the stack traces seem more unreliable than usual so I
could be totally wrong here.