Re: swsusp regression [Was: 2.6.17-mm1]

[Pavel Machek <pavel@ucw.cz>]

Hi!

> > > > 2.6.17 wasn't supposed to oops.  Do you have details on this?
> > > > 
> > > For some reason, unknown to me, the oops won't display on the serial
> > > link :(.
> > 
> > Serial console is currently broken by suspend, resume. _But_ I have a
> > patch I'd like you to try.... pretty please?
> > 
> Sure :)... I applied it but the output went to the laptop's screen anyway...

Do you need some kernel command line options? This is s2ram, do I
recall it correctly?

> > That is not an oops, rather a kernel BUG(). Can you just remove
> > might_sleep line and see what happens?
> > 
> > Unfortunately, backtrace does not tell me which notifier chain did
> > that :-(. Are you using audit or something like that?
> > 
> > /*
> >  * lock for reading
> >  */
> > static inline void down_read(struct rw_semaphore *sem)
> > {
> >         might_sleep();
> > ~~~~~~~~~~~~~~~~~~~~~~
> >         rwsemtrace(sem,"Entering down_read");
> >         __down_read(sem);
> >         rwsemtrace(sem,"Leaving down_read");
> > }
> > 
> Here's the (hand copied) dump when the might_sleep is removed:
> 
> esi: 00000003  edit: 00000000 ebp: f6cb9eb8  esp: f6cb9ea4
> ds: 007b  es: 007b  ss: 0068
> Process bash (pid: 9402, threadinfo=f6cb8000 task=f7a5c570)

This stack lines are not really interesting (can you comment them from
sources?) and they make interesting info scroll away :-(. Or maybe
vga=1 can help?

> Stack: c0229b71 00000046 00000000 00000286 c0383ca7 f6cb9ecc c013b242 00000003
>         00000000 00000003 f6cb9ee0 c013b2e8 00000003 c0436890 f6c9a003 f6cb9f08
>         c013b481 00000003 00000003 00000246 c1788b00 00000003 c04368a0 c043692c
> Call Trace:
>  <c0103eea> show_stack_log_lvl+0x92/0xb7  <c0104100> show_registers+0x1a3/0x21b
>  <c0104319> die+0x117/0x230  <c03627a6> do_page_fault+0x39c/0x72a
>  <c0103b2f> error_code+0x4f/0x54  <c013b242> suspend_enter+0x2f/0x52
>  <c013b2e8> enter_state+0x4b/0x8d  <c013b481> state_store+0xa0/0xa2
>  <c01a5151> subsys_attr_store+0x37/0x41  <c01a53d2> flush_write_buffer+0x3c/0x46
>  <c01a5443> sysfs_write_file+0x67/0x8b  <c0166bb6> vfs_write+0x1b9/0x1be
>  <c0166c7b> sys_write+0x4b/0x75  <c010300f> sysenter_past_esp+0x54/0x75
> 
> Code: 05 c4 42 43 c0 31 43 43 c0 c3 8b 2d 68 6e 54 c0 8b 1d 60 6e 54 c0 8b 35 6c 6e 54 c0 8b 3d 70 6d 54 c0 ff 35 74 6e 54 c0 9d c3 90 <e8> 6d 38 ea ff e8 a2 ff ff ff 6a 03 e8 ec b6 de ff 83 c4 04 c3
> EIP: [c043431c>] do_suspend_lowlevel+0x0/0x15 SS:ESP 0068:f6cb6ea4
  ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Ha, wait a moment, this is interesting line. Can you trace down which
instruction causes this?

We recently changed pagetable handling during swsusp, perhaps thats
it? It went to Linus few minutes ago...
									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html