ALSA: hda: Use standard waitqueue for RIRB wakeup

The HD-audio CORB/RIRB communication was programmed in a way that was
documented in the reference in decades ago, which is essentially a
polling in the waiter side.  It's working fine but costs CPU cycles on
some platforms that support only slow communications.  Also, for some
platforms that had unreliable communications, we put longer wait time
(2 ms), which accumulate quite long time if you execute many verbs in
a shot (e.g. at the initialization or resume phase).

This patch attempts to improve the situation by introducing the
standard waitqueue in the RIRB waiter side instead of polling.  The
test results on my machine show significant improvements.  The time
spent for "cat /proc/asound/card*/codec#*" were changed like:

* Intel SKL + Realtek codec
  before the patch:
   0.00user 0.04system 0:00.10elapsed 40.0%CPU
  after the patch:
   0.00user 0.01system 0:00.10elapsed 10.0%CPU

* Nvidia GP107GL + Nvidia HDMI codec
  before the patch:
   0.00user 0.00system 0:02.76elapsed 0.0%CPU
  after the patch:
   0.00user 0.00system 0:00.01elapsed 17.0%CPU

So, for Intel chips, the total time is same, while the total time is
greatly reduced (from 2.76 to 0.01s) for Nvidia chips.
The only negative data here is the increase of CPU time for Nvidia,
but this is the unavoidable cost for faster wakeups, supposedly.

Link: https://lore.kernel.org/r/20191210145727.22054-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>

1 files changed, 14 insertions, 15 deletions

diff --git a/sound/pci/hda/hda_controller.c b/sound/pci/hda/hda_controller.c
index 2f3b7a35f2d9..f30a053d981e 100644
--- a/sound/pci/hda/hda_controller.c
+++ b/sound/pci/hda/hda_controller.c
@@ -792,21 +792,25 @@ static int azx_rirb_get_response(struct hdac_bus *bus, unsigned int addr,
 	struct hda_bus *hbus = &chip->bus;
 	unsigned long timeout;
 	unsigned long loopcounter;
-	int do_poll = 0;
+	wait_queue_entry_t wait;
 	bool warned = false;
 
+	init_wait_entry(&wait, 0);
  again:
 	timeout = jiffies + msecs_to_jiffies(1000);
 
 	for (loopcounter = 0;; loopcounter++) {
 		spin_lock_irq(&bus->reg_lock);
-		if (bus->polling_mode || do_poll)
+		if (!bus->polling_mode)
+			prepare_to_wait(&bus->rirb_wq, &wait,
+					TASK_UNINTERRUPTIBLE);
+		if (bus->polling_mode)
 			snd_hdac_bus_update_rirb(bus);
 		if (!bus->rirb.cmds[addr]) {
-			if (!do_poll)
-				bus->poll_count = 0;
 			if (res)
 				*res = bus->rirb.res[addr]; /* the last value */
+			if (!bus->polling_mode)
+				finish_wait(&bus->rirb_wq, &wait);
 			spin_unlock_irq(&bus->reg_lock);
 			return 0;
 		}
@@ -814,7 +818,9 @@ static int azx_rirb_get_response(struct hdac_bus *bus, unsigned int addr,
 		if (time_after(jiffies, timeout))
 			break;
 #define LOOP_COUNT_MAX	3000
-		if (hbus->needs_damn_long_delay ||
+		if (!bus->polling_mode) {
+			schedule_timeout(msecs_to_jiffies(2));
+		} else if (hbus->needs_damn_long_delay ||
 		    loopcounter > LOOP_COUNT_MAX) {
 			if (loopcounter > LOOP_COUNT_MAX && !warned) {
 				dev_dbg_ratelimited(chip->card->dev,
@@ -829,19 +835,12 @@ static int azx_rirb_get_response(struct hdac_bus *bus, unsigned int addr,
 		}
 	}
 
+	if (!bus->polling_mode)
+		finish_wait(&bus->rirb_wq, &wait);
+
 	if (hbus->no_response_fallback)
 		return -EIO;
 
-	if (!bus->polling_mode && bus->poll_count < 2) {
-		dev_dbg(chip->card->dev,
-			"azx_get_response timeout, polling the codec once: last cmd=0x%08x\n",
-			bus->last_cmd[addr]);
-		do_poll = 1;
-		bus->poll_count++;
-		goto again;
-	}
-
-
 	if (!bus->polling_mode) {
 		dev_warn(chip->card->dev,
 			 "azx_get_response timeout, switching to polling mode: last cmd=0x%08x\n",