[alsa-devel] [PATCH] ALSA: hda - Inform too slow responses

[Takashi Iwai]

For the sake of debugging, show each response that took too long time
and its corresponding last verb.  Since such an error happens in
series once when started, use the ratelimited variant for suppressing
the flood.

Signed-off-by: Takashi Iwai <tiwai at suse.de>
---
 sound/pci/hda/hda_controller.c | 13 +++++++++++--
 1 file changed, 11 insertions(+), 2 deletions(-)

diff --git a/sound/pci/hda/hda_controller.c b/sound/pci/hda/hda_controller.c
index 462559442f7a..6387c7e90918 100644
--- a/sound/pci/hda/hda_controller.c
+++ b/sound/pci/hda/hda_controller.c
@@ -794,6 +794,7 @@ static int azx_rirb_get_response(struct hdac_bus *bus, unsigned int addr,
 	unsigned long timeout;
 	unsigned long loopcounter;
 	int do_poll = 0;
+	bool warned = false;
 
  again:
 	timeout = jiffies + msecs_to_jiffies(1000);
@@ -813,9 +814,17 @@ static int azx_rirb_get_response(struct hdac_bus *bus, unsigned int addr,
 		spin_unlock_irq(&bus->reg_lock);
 		if (time_after(jiffies, timeout))
 			break;
-		if (hbus->needs_damn_long_delay || loopcounter > 3000)
+#define LOOP_COUNT_MAX	3000
+		if (hbus->needs_damn_long_delay ||
+		    loopcounter > LOOP_COUNT_MAX) {
+			if (loopcounter > LOOP_COUNT_MAX && !warned) {
+				dev_dbg_ratelimited(chip->card->dev,
+						    "too slow response, last cmd=%#08x\n",
+						    bus->last_cmd[addr]);
+				warned = true;
+			}
 			msleep(2); /* temporary workaround */
-		else {
+		} else {
 			udelay(10);
 			cond_resched();
 		}
-- 
2.16.4