Re: echoaudio.c, Revision 1.6

[Takashi Iwai <tiwai@suse.de>]

At Mon, 14 Mar 2005 09:19:28 +0100 (MET),
Clemens Ladisch wrote:
> 
> Giuliano Pochini wrote:
> > I'm a bit confused about the latest changes to pci/echoaudio/echoaudio.c
> >
> > The CVS interface says: "Use spin_lock_irqsave() instead of spin_lock() in
> > places where we could be interrupted by another hardware interrupt that
> > could call the rawmidi trigger callback that could try to take the same
> > lock."
> >
> > But in snd_echo_interrupt() we are running in hard irq context, so we know
> > irqs are already disabled.
> 
> Only the sound card interrupt itself is disabled.  It is architecture-
> specific whether other interrupts are disabled.
> 
> > Then: "Additionally, remove locking code that is no longer needed now that
> > the trigger callback is no longer called recursively from the rawmidi
> > "event" handler."
> >
> > Your patch does not remove anything.
> 
> It removed code in some other files.
> 
> > It only replaces spin_lock() in pcm_trigger() which is a stupid
> > bug of mine, sorry.
> 
> I think this wasn't a bug because the trigger callback is already
> called with interrupts disabled, AFAICS.  I'm planning to revert most
> of these locking changes when I've proven that it's safe.

I fixed snd_pcm_suspend*() functions to do in irq disabled.
So, all calls of trigger and pointer callbacks are atomic now.


Takashi


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click