Skip to contentSkip to conversation

ASoC: arizona: Add delay for output disable

PATCH Mark Brown 2 participants 4 messages Jan 19, 2015 1 message not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): 1421682586-6303-1-git-send-email-ckeepax@opensource.wolfsonmicro.com

Mark Brown

author
#1 lore raw

Review

Quoted 10 lines · Charles Keepax
(quoted diff) +10 -0
+	case SND_SOC_DAPM_POST_PMD:
+		switch (w->shift) {
+		case ARIZONA_OUT1L_ENA_SHIFT:
+		case ARIZONA_OUT1R_ENA_SHIFT:
+		case ARIZONA_OUT2L_ENA_SHIFT:
+		case ARIZONA_OUT2R_ENA_SHIFT:
+		case ARIZONA_OUT3L_ENA_SHIFT:
+		case ARIZONA_OUT3R_ENA_SHIFT:
+			udelay(750);
+			break;

That's a really quite long udelay() especially given that as the code is written it's going to be cumalative over all the outputs being torn down so typically will be 1.5ms for headphones (not sure if that's the intention or not). msleep() would be more friendly than udelay() and it'd also be good to arrange to coalesce the delays. Perhaps set a flag in _PRE_PMD and check if a delay is needed in the clock teardown? Or better yet record a timestamp and then figure out if a delay is needed when tearing down the clock though that is probably overengineering.

#2 lore raw
Quoted 13 lines · Mark Brown

Review

Quoted 10 lines · Charles Keepax
(quoted diff) +10 -0
+	case SND_SOC_DAPM_POST_PMD:
+		switch (w->shift) {
+		case ARIZONA_OUT1L_ENA_SHIFT:
+		case ARIZONA_OUT1R_ENA_SHIFT:
+		case ARIZONA_OUT2L_ENA_SHIFT:
+		case ARIZONA_OUT2R_ENA_SHIFT:
+		case ARIZONA_OUT3L_ENA_SHIFT:
+		case ARIZONA_OUT3R_ENA_SHIFT:
+			udelay(750);
+			break;
Quoted 8 lines

That's a really quite long udelay() especially given that as the code is written it's going to be cumalative over all the outputs being torn down so typically will be 1.5ms for headphones (not sure if that's the intention or not). msleep() would be more friendly than udelay() and it'd also be good to arrange to coalesce the delays. Perhaps set a flag in _PRE_PMD and check if a delay is needed in the clock teardown? Or better yet record a timestamp and then figure out if a delay is needed when tearing down the clock though that is probably overengineering.

Similar to the power up the delay needs to be cumulative as the write sequences will queue themselves up if the last one is still running when the next one is started. It is fairly long for a udelay I will replace with an msleep(1) instead.

Thanks,
Charles

author
#3 lore raw
Quoted 4 lines · Charles Keepax

Similar to the power up the delay needs to be cumulative as the write sequences will queue themselves up if the last one is still running when the next one is started. It is fairly long for a udelay I will replace with an msleep(1) instead.

Given that you are needing to delay per output it's probably going to be better to do the coalescing of the delays - record how much delay is going to be needed when stopping the clock then do that at once when doing that. It's going to be more friendly for the rest of the system to see a single delay than to delay, schedule the thread again and then almost immediately enter another delay.

#4 lore raw
Quoted 7 lines · Mark Brown
Quoted 4 lines · Charles Keepax

Similar to the power up the delay needs to be cumulative as the write sequences will queue themselves up if the last one is still running when the next one is started. It is fairly long for a udelay I will replace with an msleep(1) instead.

Quoted 6 lines

Given that you are needing to delay per output it's probably going to be better to do the coalescing of the delays - record how much delay is going to be needed when stopping the clock then do that at once when doing that. It's going to be more friendly for the rest of the system to see a single delay than to delay, schedule the thread again and then almost immediately enter another delay.

That makes sense I will look at adding that.

Thanks,
Charles