From: Takashi Iwai <tiwai@suse.de>
To: alsa-devel@alsa-project.org
Cc: Matwey Kornilov <matwey.kornilov@gmail.com>,
Dylan Robinson <dylan_robinson@motu.com>,
Keith Milner <kamilner@superlative.org>
Subject: [PATCH 28/41] ALSA: usb-audio: Use atomic_t for endpoint use_count
Date: Mon, 23 Nov 2020 09:53:34 +0100 [thread overview]
Message-ID: <20201123085347.19667-29-tiwai@suse.de> (raw)
In-Reply-To: <20201123085347.19667-1-tiwai@suse.de>
The endpoint objects may be started/stopped concurrently by different
substreams in the case of implicit feedback mode, while the current
code handles the reference counter without any protection.
This patch changes the refcount to atomic_t for avoiding the
inconsistency. We need no reference_t here as the refcount goes only
up to 2.
Also the name "use_count" is renamed to "running" since this is about
actually the running status, not the open refcount.
Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
---
sound/usb/card.h | 2 +-
sound/usb/endpoint.c | 26 ++++++++++++++------------
2 files changed, 15 insertions(+), 13 deletions(-)
diff --git a/sound/usb/card.h b/sound/usb/card.h
index 53f0ce61f858..f58c3769b058 100644
--- a/sound/usb/card.h
+++ b/sound/usb/card.h
@@ -60,7 +60,7 @@ struct snd_usb_endpoint {
struct snd_usb_audio *chip;
int opened; /* open refcount; protect with chip->mutex */
- int use_count;
+ atomic_t running; /* running status */
int ep_num; /* the referenced endpoint number */
int type; /* SND_USB_ENDPOINT_TYPE_* */
unsigned long flags;
diff --git a/sound/usb/endpoint.c b/sound/usb/endpoint.c
index 49fb934ee432..4d733b2d8287 100644
--- a/sound/usb/endpoint.c
+++ b/sound/usb/endpoint.c
@@ -1234,7 +1234,7 @@ int snd_usb_endpoint_configure(struct snd_usb_audio *chip,
*
* @ep: the endpoint to start
*
- * A call to this function will increment the use count of the endpoint.
+ * A call to this function will increment the running count of the endpoint.
* In case it is not already running, the URBs for this endpoint will be
* submitted. Otherwise, this function does nothing.
*
@@ -1253,11 +1253,12 @@ int snd_usb_endpoint_start(struct snd_usb_endpoint *ep)
if (ep->sync_master)
WRITE_ONCE(ep->sync_master->sync_slave, ep);
- usb_audio_dbg(ep->chip, "Starting %s EP 0x%x (count %d)\n",
- ep_type_name(ep->type), ep->ep_num, ep->use_count);
+ usb_audio_dbg(ep->chip, "Starting %s EP 0x%x (running %d)\n",
+ ep_type_name(ep->type), ep->ep_num,
+ atomic_read(&ep->running));
/* already running? */
- if (++ep->use_count != 1)
+ if (atomic_inc_return(&ep->running) != 1)
return 0;
/* just to be sure */
@@ -1319,7 +1320,7 @@ int snd_usb_endpoint_start(struct snd_usb_endpoint *ep)
if (ep->sync_master)
WRITE_ONCE(ep->sync_master->sync_slave, NULL);
clear_bit(EP_FLAG_RUNNING, &ep->flags);
- ep->use_count--;
+ atomic_dec(&ep->running);
deactivate_urbs(ep, false);
return -EPIPE;
}
@@ -1329,7 +1330,7 @@ int snd_usb_endpoint_start(struct snd_usb_endpoint *ep)
*
* @ep: the endpoint to stop (may be NULL)
*
- * A call to this function will decrement the use count of the endpoint.
+ * A call to this function will decrement the running count of the endpoint.
* In case the last user has requested the endpoint stop, the URBs will
* actually be deactivated.
*
@@ -1343,16 +1344,17 @@ void snd_usb_endpoint_stop(struct snd_usb_endpoint *ep)
if (!ep)
return;
- usb_audio_dbg(ep->chip, "Stopping %s EP 0x%x (count %d)\n",
- ep_type_name(ep->type), ep->ep_num, ep->use_count);
+ usb_audio_dbg(ep->chip, "Stopping %s EP 0x%x (running %d)\n",
+ ep_type_name(ep->type), ep->ep_num,
+ atomic_read(&ep->running));
- if (snd_BUG_ON(ep->use_count == 0))
+ if (snd_BUG_ON(!atomic_read(&ep->running)))
return;
if (ep->sync_master)
WRITE_ONCE(ep->sync_master->sync_slave, NULL);
- if (--ep->use_count == 0) {
+ if (!atomic_dec_return(&ep->running)) {
deactivate_urbs(ep, false);
set_bit(EP_FLAG_STOPPING, &ep->flags);
}
@@ -1363,7 +1365,7 @@ void snd_usb_endpoint_stop(struct snd_usb_endpoint *ep)
*
* @ep: the endpoint to release
*
- * This function does not care for the endpoint's use count but will tear
+ * This function does not care for the endpoint's running count but will tear
* down all the streaming URBs immediately.
*/
void snd_usb_endpoint_release(struct snd_usb_endpoint *ep)
@@ -1410,7 +1412,7 @@ static void snd_usb_handle_sync_urb(struct snd_usb_endpoint *ep,
* will take care of them later.
*/
if (snd_usb_endpoint_implicit_feedback_sink(ep) &&
- ep->use_count != 0) {
+ atomic_read(&ep->running)) {
/* implicit feedback case */
int i, bytes = 0;
--
2.16.4
next prev parent reply other threads:[~2020-11-23 9:06 UTC|newest]
Thread overview: 52+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-11-23 8:53 [PATCH 00/41] USB audio refactoring for better implicit feedback support Takashi Iwai
2020-11-23 8:53 ` [PATCH 01/41] ALSA: usb-audio: Handle discrete rates properly in hw constraints Takashi Iwai
2020-11-23 8:53 ` [PATCH 02/41] ALSA: usb-audio: Don't call usb_set_interface() at trigger callback Takashi Iwai
2020-11-23 8:53 ` [PATCH 03/41] ALSA: usb-audio: Check valid altsetting at parsing rates for UAC2/3 Takashi Iwai
2020-11-23 8:53 ` [PATCH 04/41] ALSA: usb-audio: Check implicit feedback EP generically for UAC2 Takashi Iwai
2020-11-23 8:53 ` [PATCH 05/41] ALSA: usb-audio: Add snd_usb_get_endpoint() helper Takashi Iwai
2020-11-23 8:53 ` [PATCH 06/41] ALSA: usb-audio: Set and clear sync EP link properly Takashi Iwai
2020-11-23 8:53 ` [PATCH 07/41] ALSA: usb-audio: Improve some debug prints Takashi Iwai
2020-11-23 8:53 ` [PATCH 08/41] ALSA: usb-audio: Track implicit fb sync endpoint in audioformat list Takashi Iwai
2020-11-23 8:53 ` [PATCH 09/41] ALSA: usb-audio: Move snd_usb_autoresume() call out of setup_hw_info() Takashi Iwai
2020-11-23 8:53 ` [PATCH 10/41] ALSA: usb-audio: Add hw constraint for implicit fb sync Takashi Iwai
2020-11-23 8:53 ` [PATCH 11/41] ALSA: usb-audio: Simplify hw_params rules Takashi Iwai
2020-11-23 8:53 ` [PATCH 12/41] ALSA: usb-audio: Drop debug.h Takashi Iwai
2020-11-23 8:53 ` [PATCH 13/41] ALSA: usb-audio: Avoid doubly initialization for implicit fb Takashi Iwai
2020-11-23 8:53 ` [PATCH 14/41] ALSA: usb-audio: Create endpoint objects at parsing phase Takashi Iwai
2021-01-03 17:09 ` [PATCH 14/41] ALSA: usb-audio: Create endpoint objects at parsing phase - Pioneer DJ DJM-250MK2 stopped working František Kučera
2021-01-03 17:19 ` Takashi Iwai
2021-01-03 18:15 ` František Kučera
2021-01-05 9:29 ` Takashi Iwai
2021-01-05 13:20 ` Takashi Iwai
2021-01-05 22:27 ` František Kučera
2021-01-06 9:03 ` Takashi Iwai
2021-01-06 19:01 ` František Kučera
2021-01-07 13:30 ` Takashi Iwai
2020-11-23 8:53 ` [PATCH 15/41] ALSA: usb-audio: Drop keep_interface flag again Takashi Iwai
2020-11-23 8:53 ` [PATCH 16/41] ALSA: usb-audio: Add snd_usb_get_host_interface() helper Takashi Iwai
2020-11-23 8:53 ` [PATCH 17/41] ALSA: usb-audio: Don't set altsetting before initializing sample rate Takashi Iwai
2020-11-23 8:53 ` [PATCH 18/41] ALSA: usb-audio: Pass snd_usb_audio object to quirk functions Takashi Iwai
2020-11-23 8:53 ` [PATCH 19/41] ALSA: usb-audio: Simplify snd_usb_init_sample_rate() arguments Takashi Iwai
2020-11-23 8:53 ` [PATCH 20/41] ALSA: usb-audio: Simplify snd_usb_init_pitch() arguments Takashi Iwai
2020-11-23 8:53 ` [PATCH 21/41] ALSA: usb-audio: Stop both endpoints properly at error Takashi Iwai
2020-11-23 8:53 ` [PATCH 22/41] ALSA: usb-audio: Set callbacks via snd_usb_endpoint_set_callback() Takashi Iwai
2020-11-23 8:53 ` [PATCH 23/41] ALSA: usb-audio: Always set up the parameters after resume Takashi Iwai
2020-11-23 8:53 ` [PATCH 24/41] ALSA: usb-audio: Fix EP matching for continuous rates Takashi Iwai
2020-11-23 8:53 ` [PATCH 25/41] ALSA: usb-audio: Refactor endpoint management Takashi Iwai
2020-11-23 8:53 ` [PATCH 26/41] ALSA: usb-audio: Fix possible stall of implicit fb packet ring-buffer Takashi Iwai
2020-11-23 8:53 ` [PATCH 27/41] ALSA: usb-audio: Constify audioformat pointer references Takashi Iwai
2020-11-23 8:53 ` Takashi Iwai [this message]
2020-11-23 8:53 ` [PATCH 29/41] ALSA: usb-audio: Refactoring endpoint URB deactivation Takashi Iwai
2020-11-23 8:53 ` [PATCH 30/41] ALSA: usb-audio: Drop unneeded snd_usb_substream fields Takashi Iwai
2020-11-23 8:53 ` [PATCH 31/41] ALSA: usb-audio: Unify the code for the next packet size calculation Takashi Iwai
2020-11-23 8:53 ` [PATCH 32/41] ALSA: usb-audio: Simplify rate_min/max and rates set up Takashi Iwai
2020-11-23 8:53 ` [PATCH 33/41] ALSA: usb-audio: Replace slave/master terms Takashi Iwai
2020-11-23 8:53 ` [PATCH 34/41] ALSA: usb-audio: Use unsigned char for iface and altsettings fields Takashi Iwai
2020-11-23 8:53 ` [PATCH 35/41] ALSA: usb-audio: Show sync endpoint information in proc outputs Takashi Iwai
2020-11-23 8:53 ` [PATCH 36/41] ALSA: usb-audio: Quirk for BOSS GT-001 Takashi Iwai
2020-11-23 8:53 ` [PATCH 37/41] ALSA: usb-audio: Factor out the implicit feedback quirk code Takashi Iwai
2020-11-23 8:53 ` [PATCH 38/41] ALSA: usb-audio: Add generic implicit fb parsing Takashi Iwai
2020-11-23 8:53 ` [PATCH 39/41] ALSA: usb-audio: Add implicit_fb module option Takashi Iwai
2020-11-23 9:46 ` Pavel Hofman
2020-11-23 8:53 ` [PATCH 40/41] ALSA: usb-audio: Fix quirks for other BOSS devices Takashi Iwai
2020-11-23 8:53 ` [PATCH 41/41] ALSA: usb-audio: Fix MOTU M-Series quirks 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=20201123085347.19667-29-tiwai@suse.de \
--to=tiwai@suse.de \
--cc=alsa-devel@alsa-project.org \
--cc=dylan_robinson@motu.com \
--cc=kamilner@superlative.org \
--cc=matwey.kornilov@gmail.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).