All of lore.kernel.org
 help / color / mirror / Atom feed
* pxa2xx-pcm and dma_request
@ 2009-05-12 21:06 Robert Jarzmik
  2009-05-12 22:06 ` Mark Brown
  2009-05-13  4:43 ` Dmitry Eremin-Solenikov
  0 siblings, 2 replies; 3+ messages in thread
From: Robert Jarzmik @ 2009-05-12 21:06 UTC (permalink / raw)
  To: Mark Brown, dbaryshkov; +Cc: alsa-devel

Hi Mark and Dimitry,

I've been playing for some time with DMA on the PXA series. I was wondering if
the behaviour of pxa2xx-pcm is correct.

As far as I understand, in an asoc context, the __pxa2xx_pcm_open() is called,
while in "arm context", pxa2xx_pcm_open() is called, which calls
__pxa2xx_pcm_open().

This was introduced by commit a6d77317678148c973bb0131cc5a3a772f756d23 I think.

One main difference is that in "arm context", pxa_dma_request() is called, while
in "asoc context", it is not.

I'd like one of you to cross-check, as I must admit I'm not following all the
impacts of the pcm code split-up.

Cheers.

--
Robert

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: pxa2xx-pcm and dma_request
  2009-05-12 21:06 pxa2xx-pcm and dma_request Robert Jarzmik
@ 2009-05-12 22:06 ` Mark Brown
  2009-05-13  4:43 ` Dmitry Eremin-Solenikov
  1 sibling, 0 replies; 3+ messages in thread
From: Mark Brown @ 2009-05-12 22:06 UTC (permalink / raw)
  To: Robert Jarzmik; +Cc: dbaryshkov, alsa-devel

On Tue, May 12, 2009 at 11:06:22PM +0200, Robert Jarzmik wrote:

> One main difference is that in "arm context", pxa_dma_request() is called, while
> in "asoc context", it is not.

> I'd like one of you to cross-check, as I must admit I'm not following all the
> impacts of the pcm code split-up.

AFAICT this isn't a result of the code unification but rather is a case
of the code unification mirroring exactly the pre-merge state.  It's all
pre-git but my best guess would be that the ASoC version was done before
the DMA API was introduced - I expect the ASoC variant ought to be
updated to use the DMA API but I've not looked closely enough to be
certain.

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: pxa2xx-pcm and dma_request
  2009-05-12 21:06 pxa2xx-pcm and dma_request Robert Jarzmik
  2009-05-12 22:06 ` Mark Brown
@ 2009-05-13  4:43 ` Dmitry Eremin-Solenikov
  1 sibling, 0 replies; 3+ messages in thread
From: Dmitry Eremin-Solenikov @ 2009-05-13  4:43 UTC (permalink / raw)
  To: Robert Jarzmik; +Cc: alsa-devel, Mark Brown

2009/5/13 Robert Jarzmik <robert.jarzmik@free.fr>:
> Hi Mark and Dimitry,
>
> I've been playing for some time with DMA on the PXA series. I was wondering if
> the behaviour of pxa2xx-pcm is correct.
>
> As far as I understand, in an asoc context, the __pxa2xx_pcm_open() is called,
> while in "arm context", pxa2xx_pcm_open() is called, which calls
> __pxa2xx_pcm_open().
>
> This was introduced by commit a6d77317678148c973bb0131cc5a3a772f756d23 I think.
>
> One main difference is that in "arm context", pxa_dma_request() is called, while
> in "asoc context", it is not.

This was done by intent as before code merge plain ARM driver used DMA while
ASoC didn't. That was done mainly because I didn't have all the hw
variants (only
pxa255 at hand) and didn't want to introduce any points of possible failure.
Now merging DMA code into generic code path would be a nice feature though.

> I'd like one of you to cross-check, as I must admit I'm not following all the
> impacts of the pcm code split-up.

The initial impact of the merge should be void, as it was an
_unification_ of code
presented in ASoC and non-ASoC drivers.

-- 
With best wishes
Dmitry

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2009-05-13  4:43 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2009-05-12 21:06 pxa2xx-pcm and dma_request Robert Jarzmik
2009-05-12 22:06 ` Mark Brown
2009-05-13  4:43 ` Dmitry Eremin-Solenikov

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.