Re: [PATCH 3/3] imx-ssi: Use a hrtimer in FIQ mode

[Valentin Longchamp <valentin.longchamp@epfl.ch>]

Sascha Hauer wrote:
> Using a regular timer results in poll times < 1 jiffie with small
> buffers, so we loaded the timer with the actual jiffie value. We can
> be more accurate using a hrtimer. Also, we have to call
> snd_pcm_period_elapsed after playing period_bytes and not
> runtime->period_size (which is in samples and not in bytes).
> 
> Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
> ---
>  sound/soc/imx/imx-pcm-fiq.c |   45 ++++++++++++++++++++----------------------
>  1 files changed, 21 insertions(+), 24 deletions(-)
> 

Sorry to bother you guys in your discussion (most of which I don't 
understand because of my poor alsa knowledge).

I just wanted to tell that I have given this patch a spin on a 
mx31moboard system. imx-ssi now works much better (the audio does not 
loop on one buffer anymore as it did with the earlier timer version) in 
various configuration (speaker-test with various rates up to 48 kHz as 
well as aplay with a wav file, which had never worked for me) and the 
CPU usage is much lower BUT I have experienced some deadlocks completely 
hanging the system looping on what should be the last buffer of the wave 
file.

Val

-- 
Valentin Longchamp, PhD Student, EPFL-STI-LSRO1
valentin.longchamp@epfl.ch, Phone: +41216937827
http://people.epfl.ch/valentin.longchamp
MEB3494, Station 9, CH-1015 Lausanne