Re: Atomicity in KMS panic notifier

[Takashi Iwai <tiwai@suse.de>]

At Tue, 6 May 2014 15:53:32 +0200,
David Herrmann wrote:
> 
> Hi
> 
> On Tue, May 6, 2014 at 3:38 PM, Takashi Iwai <tiwai@suse.de> wrote:
> > At Tue, 6 May 2014 15:32:21 +0200,
> > David Herrmann wrote:
> >> fbcon is called through the VT or fbdev layer, which is called by
> >> bust_spinlocks(1) via either unblank_screen() or console_unblank().
> >
> > You mean bust_spinlocks(0), right?
> >
> > void __attribute__((weak)) bust_spinlocks(int yes)
> > {
> >         if (yes) {
> >                 ++oops_in_progress;
> >         } else {
> > #ifdef CONFIG_VT
> >                 unblank_screen();
> > #endif
> >                 console_unblank();
> >                 if (--oops_in_progress == 0)
> >                         wake_up_klogd();
> >         }
> > }
> >
> > bust_spinlocks(0) is called after the notifier chain, and it's almost
> > at the end of panic().
> 
> Yes, it's called _after_ the panic-handlers but _before_
> console_unlock() (see console_unblank() in printk.c). Therefore, we
> call into set_config() before the serial drivers get the panic-message
> (flushed via console_unlock()). If the serial drivers (or whatever you
> use for debugging) register their own panic-handlers, then they're
> fine of course.

Thanks for clarification.  I see it's at the sensible place.

FWIW, the problem I'm tackling now is the blockage of other panic
notifiers due to drm_fb.  For example, pvpanic isn't executed reliably
because of this when a KMS (either cirrus or qxl) driver is loaded.
So, for me, it's fine that the system stalls after that point :)


Takashi