avm@xxxxxx wrote:
> From: Arun Mani <a0270733@xxxxxxxxxxxxxxxxxxxxxxxxxx>
>
> Signed-off-by: Arun Mani <a0270733@xxxxxxxxxxxxxxxxxxxxxxxxxx>
> ---
> sound/soc/davinci/davinci-i2s.c | 2 ++
> 1 files changed, 2 insertions(+), 0 deletions(-)
>
> diff --git a/sound/soc/davinci/davinci-i2s.c b/sound/soc/davinci/davinci-i2s.c
> index b1ea52f..61b1763 100644
> --- a/sound/soc/davinci/davinci-i2s.c
> +++ b/sound/soc/davinci/davinci-i2s.c
> @@ -104,10 +104,12 @@ enum {
>
> static struct davinci_pcm_dma_params davinci_i2s_pcm_out = {
> .name = "I2S PCM Stereo out",
> + .data_type = 2, //Initialize the data type for playback to avoid divide by zero
You should have ".acnt = 2" also for consistency.
> };
>
> static struct davinci_pcm_dma_params davinci_i2s_pcm_in = {
> .name = "I2S PCM Stereo in",
> + .data_type = 2, //Initialize the data type for playback to avoid divide by zero
> };
>
> struct davinci_mcbsp_dev {
And if the 1st stream is an 8-bit audio stream, will it initialize it incorrectly ???
Of course, we have formats = SNDRV_PCM_FMTBIT_S16_LE, currently so it's not an issue yet...
I don't see how data_type is not being set in davinci_i2s_hw_params
before being used in davinci_pcm_prepare.
Can prepare be called before hw_params ?
There is a "return -EINVAL" in davinci_i2s_hw_params with a
"printk(KERN_WARNING "davinci-i2s: unsupported PCM format\n")" before it.
Do you see this message in your log ?
Thanks
Troy
_______________________________________________
Alsa-devel mailing list
Alsa-devel@xxxxxxxxxxxxxxxx
http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
