[RFC] ASoC: pcm: Trigger all commands on error

Markus Pargmann mpa at pengutronix.de
Wed Oct 9 09:05:53 EDT 2013


On Tue, Oct 08, 2013 at 11:44:57AM +0100, Mark Brown wrote:
> On Tue, Oct 08, 2013 at 11:48:39AM +0200, Markus Pargmann wrote:
> > On Tue, Oct 08, 2013 at 10:34:13AM +0100, Mark Brown wrote:
> 
> > > I meant just unwinding the things done in the trigger call.
> 
> > You mean unwinding when we detect the error? If we return the
> > error code from the trigger function, the pcm_native.c will still
> > trigger a STOP command which is executed on all components.
> 
> Sure, but it's another way of getting consistency and does seem a bit
> more obvious.

It is still not consistent as we get an additional STOP after cleanly
unwinding the START:

snd_pcm_do_start()
    soc_pcm_trigger() # START
        codec_dai->driver->ops->trigger() # START
	platform->driver->ops->trigger() # Assuming an error here

	unwind:
        codec_dai->driver->ops->trigger() # STOP

A failed snd_pcm_do_start() is detected, so undo_start is called:

snd_pcm_undo_start()
    soc_pcm_trigger() # STOP
        codec_dai->driver->ops->trigger() # STOP
	platform->driver->ops->trigger() # STOP
	cpu_dai->driver->ops->trigger() # STOP

Now all component trigger functions are called without a matching START
command.

> 
> > > > Another possibility is to explicitly allow multiple _STOPs. Then I could
> > > > fix the driver to store the clock state. But I would prefer a generic
> > > > solution.
> 
> > > You're going to have to handle this anyway to be robust - what happens
> > > if it's the clock enable that fails?
> 
> > If clk_enable fails, we can return an error code in the mxs-saif trigger
> > function. Of course the error handling is missing here, but it is not
> > necessary to store the state of the clock to handle errors properly.
> 
> Right, but then _STOP will be called without the clock having been
> successfully enabled so you'll have an unbalanced clk_disable().

_STOP is only called as long as the error handling is not working
properly. Otherwise STOP should not be called for a failed START.

Regards,

Markus

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |



More information about the linux-arm-kernel mailing list