[PATCH] ASoC: ssm2518: Fix off-by-one error by ffs()

[PATCH] ASoC: ssm2518: Fix off-by-one error by ffs()

[Takashi Iwai]

ffs() returns the bit position from 1, while the ssm2158 driver code
assumes it being 0-based.  Also, the bit mask computation of the two
channel slots are incorrect; it must have worked just casually.

Signed-off-by: Takashi Iwai <tiwai@suse.de>
---

I found it casully while checking ab8500 bug.
The patch is untested, so please review/test first, and give Cc to
stable if needed.

 sound/soc/codecs/ssm2518.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/sound/soc/codecs/ssm2518.c b/sound/soc/codecs/ssm2518.c
index 95aed552139a..cc8debce752f 100644
--- a/sound/soc/codecs/ssm2518.c
+++ b/sound/soc/codecs/ssm2518.c
@@ -549,13 +549,13 @@ static int ssm2518_set_tdm_slot(struct snd_soc_dai *dai, unsigned int tx_mask,
 		right_slot = 0;
 	} else {
 		/* We assume the left channel < right channel */
-		left_slot = ffs(tx_mask);
-		tx_mask &= ~(1 << tx_mask);
+		left_slot = __ffs(tx_mask);
+		tx_mask &= ~(1 << left_slot);
 		if (tx_mask == 0) {
 			right_slot = left_slot;
 		} else {
-			right_slot = ffs(tx_mask);
-			tx_mask &= ~(1 << tx_mask);
+			right_slot = __ffs(tx_mask);
+			tx_mask &= ~(1 << right_slot);
 		}
 	}
 
-- 
1.8.5.2

Re: [PATCH] ASoC: ssm2518: Fix off-by-one error by ffs()

[Lars-Peter Clausen]

On 12/23/2013 11:42 AM, Takashi Iwai wrote:
> ffs() returns the bit position from 1, while the ssm2158 driver code
> assumes it being 0-based.  Also, the bit mask computation of the two
> channel slots are incorrect; it must have worked just casually.
> 
> Signed-off-by: Takashi Iwai <tiwai@suse.de>

Looks good, thanks.

Acked-by: Lars-Peter Clausen <lars@metafoo.de>

Re: [PATCH] ASoC: ssm2518: Fix off-by-one error by ffs()

[Mark Brown]

[-- Attachment #1.1: Type: text/plain, Size: 288 bytes --]

On Mon, Dec 23, 2013 at 11:42:55AM +0100, Takashi Iwai wrote:
> ffs() returns the bit position from 1, while the ssm2158 driver code
> assumes it being 0-based.  Also, the bit mask computation of the two
> channel slots are incorrect; it must have worked just casually.

Applied, thanks.

[-- Attachment #1.2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

[-- Attachment #2: Type: text/plain, Size: 0 bytes --]