[PATCH] ALSA: hda: Clear RIRB status before reading WP

[PATCH] ALSA: hda: Clear RIRB status before reading WP

[Viswanath L]

From: Mohan Kumar <mkumard@nvidia.com>

RIRB interrupt status getting cleared after the write pointer is read
causes a race condition, where last response(s) into RIRB may remain
unserviced by IRQ, eventually causing azx_rirb_get_response to fall
back to polling mode. Clearing the RIRB interrupt status ahead of
write pointer access ensures that this condition is avoided.

Signed-off-by: Mohan Kumar <mkumard@nvidia.com>
Signed-off-by: Viswanath L <viswanathl@nvidia.com>
---
 sound/pci/hda/hda_controller.c | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

diff --git a/sound/pci/hda/hda_controller.c b/sound/pci/hda/hda_controller.c
index 9757667..2609e39 100644
--- a/sound/pci/hda/hda_controller.c
+++ b/sound/pci/hda/hda_controller.c
@@ -1110,16 +1110,23 @@ irqreturn_t azx_interrupt(int irq, void *dev_id)
 		if (snd_hdac_bus_handle_stream_irq(bus, status, stream_update))
 			active = true;
 
-		/* clear rirb int */
 		status = azx_readb(chip, RIRBSTS);
 		if (status & RIRB_INT_MASK) {
+			/*
+			 * Clearing the interrupt status here ensures that no
+			 * interrupt gets masked after the RIRB wp is read in
+			 * snd_hdac_bus_update_rirb. This avoids a possible
+			 * race condition where codec response in RIRB may
+			 * remain unserviced by IRQ, eventually falling back
+			 * to polling mode in azx_rirb_get_response.
+			 */
+			azx_writeb(chip, RIRBSTS, RIRB_INT_MASK);
 			active = true;
 			if (status & RIRB_INT_RESPONSE) {
 				if (chip->driver_caps & AZX_DCAPS_CTX_WORKAROUND)
 					udelay(80);
 				snd_hdac_bus_update_rirb(bus);
 			}
-			azx_writeb(chip, RIRBSTS, RIRB_INT_MASK);
 		}
 	} while (active && ++repeat < 10);
 
-- 
2.7.4

Re: [PATCH] ALSA: hda: Clear RIRB status before reading WP

[Takashi Iwai]

On Thu, 06 Feb 2020 11:10:53 +0100,
Viswanath L wrote:
> 
> From: Mohan Kumar <mkumard@nvidia.com>
> 
> RIRB interrupt status getting cleared after the write pointer is read
> causes a race condition, where last response(s) into RIRB may remain
> unserviced by IRQ, eventually causing azx_rirb_get_response to fall
> back to polling mode. Clearing the RIRB interrupt status ahead of
> write pointer access ensures that this condition is avoided.
> 
> Signed-off-by: Mohan Kumar <mkumard@nvidia.com>
> Signed-off-by: Viswanath L <viswanathl@nvidia.com>

Applied, thanks.


Takashi