From: Caleb Crome <caleb@crome.org>
To: timur@tabi.org, nicoleotsuka@gmail.com, Xiubo.Lee@gmail.com,
lgirdwood@gmail.com, broonie@kernel.org, perex@perex.cz,
tiwai@suse.com, alsa-devel@alsa-project.org,
linuxppc-dev@lists.ozlabs.org, linux-kernel@vger.kernel.org
Cc: Caleb Crome <caleb@crome.org>
Subject: [PATCH] ASoC: fsl_ssi: Set watermark and maxburst settings to eliminate DMA xruns on imx processors
Date: Mon, 18 Jan 2016 10:36:45 -0800 [thread overview]
Message-ID: <1453142205-83135-1-git-send-email-caleb@crome.org> (raw)
Change watermark settings for imx processors. The previous setting
was fifo_depth-2, which causes silent but deadly xruns when using more
than 2 channels. This patch sets the watermark & maxburst
settings to
DMA, imx processors: 4
DMA, mpc8610 : fifo_depth - 2
FIQ, all : fifo_depth
This leaves mpc8610 with identically the same settings as before this
patch, and changes the IMX processors to have a watermark setting of
4, meaning when there are 4 or more empty spaces in the TX FIFO, to
start a DMA burst. And similarly, when there are 4 or more
filled spaces in the RX fifo, to start a DMA burst.
The tradeoff with this patch is that on IMX processors, the DMA will
be more reliable (not have DMA slips), epseically when running more
than 2 channels, at the expense of more DMA transactions. Interrupt
frequency is unaffected.
Signed-off-by: Caleb Crome <caleb@crome.org>
---
sound/soc/fsl/fsl_ssi.c | 54 ++++++++++++++++++++++++++++++-------------------
1 file changed, 33 insertions(+), 21 deletions(-)
diff --git a/sound/soc/fsl/fsl_ssi.c b/sound/soc/fsl/fsl_ssi.c
index 95d2392..5080c29 100644
--- a/sound/soc/fsl/fsl_ssi.c
+++ b/sound/soc/fsl/fsl_ssi.c
@@ -220,6 +220,10 @@ struct fsl_ssi_soc_data {
* @dbg_stats: Debugging statistics
*
* @soc: SoC specific data
+ *
+ * @fifo_watermark: the fifo level to signal the DMA for more words
+ * @dma_maxburst: the maximum number of words to send to the fifo
+ * in a DMA burst
*/
struct fsl_ssi_private {
struct regmap *regs;
@@ -257,6 +261,9 @@ struct fsl_ssi_private {
struct fsl_ssi_dbg dbg_stats;
const struct fsl_ssi_soc_data *soc;
+
+ u32 fifo_watermark;
+ u32 dma_maxburst;
};
/*
@@ -985,21 +992,7 @@ static int _fsl_ssi_set_dai_fmt(struct device *dev,
regmap_write(regs, CCSR_SSI_SRCR, srcr);
regmap_write(regs, CCSR_SSI_SCR, scr);
- /*
- * Set the watermark for transmit FIFI 0 and receive FIFO 0. We don't
- * use FIFO 1. We program the transmit water to signal a DMA transfer
- * if there are only two (or fewer) elements left in the FIFO. Two
- * elements equals one frame (left channel, right channel). This value,
- * however, depends on the depth of the transmit buffer.
- *
- * We set the watermark on the same level as the DMA burstsize. For
- * fiq it is probably better to use the biggest possible watermark
- * size.
- */
- if (ssi_private->use_dma)
- wm = ssi_private->fifo_depth - 2;
- else
- wm = ssi_private->fifo_depth;
+ wm = ssi_private->soc.fifo_watermark;
regmap_write(regs, CCSR_SSI_SFCSR,
CCSR_SSI_SFCSR_TFWM0(wm) | CCSR_SSI_SFCSR_RFWM0(wm) |
@@ -1307,12 +1300,8 @@ static int fsl_ssi_imx_probe(struct platform_device *pdev,
dev_dbg(&pdev->dev, "could not get baud clock: %ld\n",
PTR_ERR(ssi_private->baudclk));
- /*
- * We have burstsize be "fifo_depth - 2" to match the SSI
- * watermark setting in fsl_ssi_startup().
- */
- ssi_private->dma_params_tx.maxburst = ssi_private->fifo_depth - 2;
- ssi_private->dma_params_rx.maxburst = ssi_private->fifo_depth - 2;
+ ssi_private->dma_params_tx.maxburst = ssi_private->dma_maxburst;
+ ssi_private->dma_params_rx.maxburst = ssi_private->dma_maxburst;
ssi_private->dma_params_tx.addr = ssi_private->ssi_phys + CCSR_SSI_STX0;
ssi_private->dma_params_rx.addr = ssi_private->ssi_phys + CCSR_SSI_SRX0;
@@ -1465,6 +1454,29 @@ static int fsl_ssi_probe(struct platform_device *pdev)
/* Older 8610 DTs didn't have the fifo-depth property */
ssi_private->fifo_depth = 8;
+ if (ssi_private->use_dma) {
+ /*
+ * using DMA, set both watermark & maxburst
+ * imx parts need a value of 4 to keep up with the
+ * fastest data rates.
+ * older non-imx parts keep the old values of
+ * fifo_depth-2.
+ * maxburst must be <= fifo_watermark;
+ * and must be even if dual fifo is used.
+ */
+ if (ssi_private->soc->imx) {
+ ssi_private->fifo_watermark = 4;
+ ssi_private->dma_maxburst = 4;
+ } else {
+ ssi_private->fifo_watermark =
+ ssi_private->fifo_depth - 2;
+ ssi_private->dma_maxburst =
+ ssi_private->fifo_depth - 2;
+ }
+ } else
+ /* using FIQ. Keep settings what they were originally */
+ ssi_private->fifo_watermark = ssi_private->fifo_depth;
+
dev_set_drvdata(&pdev->dev, ssi_private);
if (ssi_private->soc->imx) {
--
2.1.4
next reply other threads:[~2016-01-18 18:37 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-01-18 18:36 Caleb Crome [this message]
2016-01-18 18:59 ` [PATCH] ASoC: fsl_ssi: Set watermark and maxburst settings to eliminate DMA xruns on imx processors kbuild test robot
2016-01-18 19:25 ` Caleb Crome
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1453142205-83135-1-git-send-email-caleb@crome.org \
--to=caleb@crome.org \
--cc=Xiubo.Lee@gmail.com \
--cc=alsa-devel@alsa-project.org \
--cc=broonie@kernel.org \
--cc=lgirdwood@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=nicoleotsuka@gmail.com \
--cc=perex@perex.cz \
--cc=timur@tabi.org \
--cc=tiwai@suse.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox