All of lore.kernel.org
 help / color / mirror / Atom feed
From: Takashi Iwai <tiwai@suse.de>
To: alsa-devel@alsa-project.org
Cc: John Keeping <john@metanate.com>
Subject: [PATCH] ALSA: usb-audio: Fix recursive locking at XRUN during syncing
Date: Mon, 20 Mar 2023 15:28:38 +0100	[thread overview]
Message-ID: <20230320142838.494-1-tiwai@suse.de> (raw)

The recent support of low latency playback in USB-audio driver made
the snd_usb_queue_pending_output_urbs() function to be called via PCM
ack ops.  In the new code path, the function is performed alread in
the PCM stream lock.  The problem is that, when an XRUN is detected,
the  function calls snd_pcm_xrun() to notify, but snd_pcm_xrun() is
supposed to be called only outside the stream lock.  As a result, it
leads to a deadlock of PCM stream locking.

For avoiding such a recursive locking, this patch adds an additional
check to the code paths in PCM core that call the ack callback; now it
checks the error code from the callback, and if it's -EPIPE, the XRUN
is handled in the PCM core side gracefully.  Along with it, the
USB-audio driver code is changed to follow that, i.e. -EPIPE is
returned instead of the explicit snd_pcm_xrun() call when the function
is performed already in the stream lock.

Fixes: d5f871f89e21 ("ALSA: usb-audio: Improved lowlatency playback support")
Reported-and-tested-by: John Keeping <john@metanate.com>
Link: https://lore.kernel.org/r/20230317195128.3911155-1-john@metanate.com
Cc: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
---
 sound/core/pcm_lib.c |  2 ++
 sound/usb/endpoint.c | 22 ++++++++++++++--------
 sound/usb/endpoint.h |  4 ++--
 sound/usb/pcm.c      |  2 +-
 4 files changed, 19 insertions(+), 11 deletions(-)

diff --git a/sound/core/pcm_lib.c b/sound/core/pcm_lib.c
index 8b6aeb8a78f7..02fd65993e7e 100644
--- a/sound/core/pcm_lib.c
+++ b/sound/core/pcm_lib.c
@@ -2155,6 +2155,8 @@ int pcm_lib_apply_appl_ptr(struct snd_pcm_substream *substream,
 		ret = substream->ops->ack(substream);
 		if (ret < 0) {
 			runtime->control->appl_ptr = old_appl_ptr;
+			if (ret == -EPIPE)
+				__snd_pcm_xrun(substream);
 			return ret;
 		}
 	}
diff --git a/sound/usb/endpoint.c b/sound/usb/endpoint.c
index 419302e2057e..647fa054d8b1 100644
--- a/sound/usb/endpoint.c
+++ b/sound/usb/endpoint.c
@@ -455,8 +455,8 @@ static void push_back_to_ready_list(struct snd_usb_endpoint *ep,
  * This function is used both for implicit feedback endpoints and in low-
  * latency playback mode.
  */
-void snd_usb_queue_pending_output_urbs(struct snd_usb_endpoint *ep,
-				       bool in_stream_lock)
+int snd_usb_queue_pending_output_urbs(struct snd_usb_endpoint *ep,
+				      bool in_stream_lock)
 {
 	bool implicit_fb = snd_usb_endpoint_implicit_feedback_sink(ep);
 
@@ -480,7 +480,7 @@ void snd_usb_queue_pending_output_urbs(struct snd_usb_endpoint *ep,
 		spin_unlock_irqrestore(&ep->lock, flags);
 
 		if (ctx == NULL)
-			return;
+			break;
 
 		/* copy over the length information */
 		if (implicit_fb) {
@@ -495,11 +495,14 @@ void snd_usb_queue_pending_output_urbs(struct snd_usb_endpoint *ep,
 			break;
 		if (err < 0) {
 			/* push back to ready list again for -EAGAIN */
-			if (err == -EAGAIN)
+			if (err == -EAGAIN) {
 				push_back_to_ready_list(ep, ctx);
-			else
+				break;
+			}
+
+			if (!in_stream_lock)
 				notify_xrun(ep);
-			return;
+			return -EPIPE;
 		}
 
 		err = usb_submit_urb(ctx->urb, GFP_ATOMIC);
@@ -507,13 +510,16 @@ void snd_usb_queue_pending_output_urbs(struct snd_usb_endpoint *ep,
 			usb_audio_err(ep->chip,
 				      "Unable to submit urb #%d: %d at %s\n",
 				      ctx->index, err, __func__);
-			notify_xrun(ep);
-			return;
+			if (!in_stream_lock)
+				notify_xrun(ep);
+			return -EPIPE;
 		}
 
 		set_bit(ctx->index, &ep->active_mask);
 		atomic_inc(&ep->submitted_urbs);
 	}
+
+	return 0;
 }
 
 /*
diff --git a/sound/usb/endpoint.h b/sound/usb/endpoint.h
index 924f4351588c..c09f68ce08b1 100644
--- a/sound/usb/endpoint.h
+++ b/sound/usb/endpoint.h
@@ -52,7 +52,7 @@ int snd_usb_endpoint_implicit_feedback_sink(struct snd_usb_endpoint *ep);
 int snd_usb_endpoint_next_packet_size(struct snd_usb_endpoint *ep,
 				      struct snd_urb_ctx *ctx, int idx,
 				      unsigned int avail);
-void snd_usb_queue_pending_output_urbs(struct snd_usb_endpoint *ep,
-				       bool in_stream_lock);
+int snd_usb_queue_pending_output_urbs(struct snd_usb_endpoint *ep,
+				      bool in_stream_lock);
 
 #endif /* __USBAUDIO_ENDPOINT_H */
diff --git a/sound/usb/pcm.c b/sound/usb/pcm.c
index d959da7a1afb..eec5232f9fb2 100644
--- a/sound/usb/pcm.c
+++ b/sound/usb/pcm.c
@@ -1639,7 +1639,7 @@ static int snd_usb_pcm_playback_ack(struct snd_pcm_substream *substream)
 	 * outputs here
 	 */
 	if (!ep->active_mask)
-		snd_usb_queue_pending_output_urbs(ep, true);
+		return snd_usb_queue_pending_output_urbs(ep, true);
 	return 0;
 }
 
-- 
2.35.3


             reply	other threads:[~2023-03-20 14:29 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-03-20 14:28 Takashi Iwai [this message]
2023-03-20 14:42 ` [PATCH] ALSA: usb-audio: Fix recursive locking at XRUN during syncing Jaroslav Kysela
2023-03-21  4:02 ` Takashi Sakamoto
2023-03-21  4:23   ` Takashi Sakamoto
2023-03-23  6:19   ` Takashi Iwai
2023-03-23 13:29     ` Takashi Sakamoto
2023-03-23 13:37       ` Takashi Iwai

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20230320142838.494-1-tiwai@suse.de \
    --to=tiwai@suse.de \
    --cc=alsa-devel@alsa-project.org \
    --cc=john@metanate.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.