From: Raffaele Recalcati <[email protected]>
i2s_fast_clock switch can be used to have a better approximate
frequency.
The waveform will be not symmetric.
This patch has been developed against the
http://git.kernel.org/pub/scm/linux/kernel/git/khilman/linux-davinci.git
git tree and has been tested on bmx board (similar to dm365 evm, but using
uda1345 as external audio codec).
Signed-off-by: Raffaele Recalcati <[email protected]>
Signed-off-by: Davide Bonfanti <[email protected]>
---
arch/arm/mach-davinci/include/mach/asp.h | 7 +++++++
sound/soc/davinci/davinci-i2s.c | 24 +++++++++++++++++++-----
2 files changed, 26 insertions(+), 5 deletions(-)
diff --git a/arch/arm/mach-davinci/include/mach/asp.h b/arch/arm/mach-davinci/include/mach/asp.h
index 314570d..2d8f8af 100644
--- a/arch/arm/mach-davinci/include/mach/asp.h
+++ b/arch/arm/mach-davinci/include/mach/asp.h
@@ -63,6 +63,13 @@ struct snd_platform_data {
unsigned sram_size_playback;
unsigned sram_size_capture;
+ /*
+ * This define works when both clock and FS are output for the cpu
+ * and makes clock very fast (FS is not symmetrical, but sampling
+ * frequency is better approximated
+ */
+ bool i2s_fast_clock;
+
/* To be used when cpu gets clock from external pin */
int clk_input_pin;
diff --git a/sound/soc/davinci/davinci-i2s.c b/sound/soc/davinci/davinci-i2s.c
index e478be9..0af5bae 100644
--- a/sound/soc/davinci/davinci-i2s.c
+++ b/sound/soc/davinci/davinci-i2s.c
@@ -150,6 +150,7 @@ struct davinci_mcbsp_dev {
*/
unsigned enable_channel_combine:1;
+ bool i2s_fast_clock;
unsigned int fmt;
int clk_div;
int clk_input_pin;
@@ -443,11 +444,23 @@ static int davinci_i2s_hw_params(struct snd_pcm_substream *substream,
DAVINCI_MCBSP_SRGR_CLKSM;
srgr |= DAVINCI_MCBSP_SRGR_FWID(mcbsp_word_length *
8 - 1);
- /* symmetric waveforms */
- clk_div = freq / (mcbsp_word_length * 16) /
- params->rate_num * params->rate_den;
- srgr |= DAVINCI_MCBSP_SRGR_FPER(mcbsp_word_length *
- 16 - 1);
+ if (dev->i2s_fast_clock) {
+ clk_div = 256;
+ do {
+ framesize = (freq / (--clk_div)) /
+ params->rate_num *
+ params->rate_den;
+ } while (((framesize < 33) || (framesize > 4095)) &&
+ (clk_div));
+ clk_div--;
+ srgr |= DAVINCI_MCBSP_SRGR_FPER(framesize - 1);
+ } else {
+ /* symmetric waveforms */
+ clk_div = freq / (mcbsp_word_length * 16) /
+ params->rate_num * params->rate_den;
+ srgr |= DAVINCI_MCBSP_SRGR_FPER(mcbsp_word_length *
+ 16 - 1);
+ }
clk_div &= 0xFF;
srgr |= clk_div;
break;
@@ -643,6 +656,7 @@ static int davinci_i2s_probe(struct platform_device *pdev)
pdata->sram_size_playback;
dev->dma_params[SNDRV_PCM_STREAM_CAPTURE].sram_size =
pdata->sram_size_capture;
+ dev->i2s_fast_clock = pdata->i2s_fast_clock;
dev->clk_input_pin = pdata->clk_input_pin;
}
dev->clk = clk_get(&pdev->dev, NULL);
--
1.7.0.4
On Wed, Jun 30, 2010 at 03:47:58PM +0200, Raffaele Recalcati wrote:
> + /*
> + * This define works when both clock and FS are output for the cpu
> + * and makes clock very fast (FS is not symmetrical, but sampling
> + * frequency is better approximated
> + */
> + bool i2s_fast_clock;
I'm having a hard time following the description here - which clock is
being made very fast? The output clocks, which are the ones people can
observe, will presumably not suddenly start running very fast.
It's probably better to rename this option to reflect the actual
function (trading off between frequency accuracy and mark/space ratio)
rather than the way it's implemented internally.
> - srgr |= DAVINCI_MCBSP_SRGR_FPER(mcbsp_word_length *
> - 16 - 1);
> + if (dev->i2s_fast_clock) {
> + clk_div = 256;
> + do {
> + framesize = (freq / (--clk_div)) /
> + params->rate_num *
> + params->rate_den;
> + } while (((framesize < 33) || (framesize > 4095)) &&
> + (clk_div));
> + clk_div--;
> + srgr |= DAVINCI_MCBSP_SRGR_FPER(framesize - 1);
> + } else {
> + /* symmetric waveforms */
> + clk_div = freq / (mcbsp_word_length * 16) /
> + params->rate_num * params->rate_den;
> + srgr |= DAVINCI_MCBSP_SRGR_FPER(mcbsp_word_length *
> + 16 - 1);
> + }
Hrm. This doesn't really correspond to your commit message at all.
Your commit message makes it sound like you've changed something about
the clocking setup of the device, such as adding another clock source,
but what you've actually done here is change the method used to
calculate the divider.
I'm *guessing* that the actual effect of your change is that you will
normally end up selecting a very much higher bit clock than would
otherwise be the case. It strikes me that there must be a better
algorithm for the calculation - for example, working up from the minimum
clock rate - which will give the same results as we currently have where
the driver is already generating accurate rates.
On Wed, Jun 30, 2010 at 03:47:58PM +0200, Raffaele Recalcati wrote:
> From: Raffaele Recalcati <[email protected]>
>
> i2s_fast_clock switch can be used to have a better approximate
Oh, and please look at the CC list for these posts. You're CCing *very*
widely and at least one address ([email protected]) actually bounces.