From mboxrd@z Thu Jan 1 00:00:00 1970 From: Vinod Koul Subject: Re: [PATCH v5] ALSA: compress: fix drain calls blocking other compress functions Date: Thu, 7 Nov 2013 12:56:47 +0530 Message-ID: <20131107072647.GA2990@intel.com> References: <1383803409-2269-1-git-send-email-vinod.koul@intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Received: from mga09.intel.com (mga09.intel.com [134.134.136.24]) by alsa0.perex.cz (Postfix) with ESMTP id C4485260376 for ; Thu, 7 Nov 2013 09:21:54 +0100 (CET) Content-Disposition: inline In-Reply-To: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: alsa-devel-bounces@alsa-project.org Sender: alsa-devel-bounces@alsa-project.org To: Takashi Iwai Cc: alsa-devel@alsa-project.org, broonie@kernel.org, lgirdwood@gmail.com List-Id: alsa-devel@alsa-project.org On Thu, Nov 07, 2013 at 08:02:14AM +0100, Takashi Iwai wrote: > At Thu, 7 Nov 2013 11:20:09 +0530, > Vinod Koul wrote: > > > > The drain and drain_notify callback were blocked by low level driver untill the > > draining was complete. Due to this being invoked with big fat mutex held, others > > ops like reading timestamp, calling pause, drop were blocked. > > > > So to fix this we add a new snd_compr_drain_notify() API. This would be required > > to be invoked by low level driver when drain or partial drain has been completed > > by the DSP. Thus we make the drain and partial_drain callback as non blocking > > and driver returns immediately after notifying DSP. > > The waiting is done while relasing the lock so that other ops can go ahead. > > > > Signed-off-by: Vinod Koul > > CC: stable@vger.kernel.org > > --- > > v5: > > remove the drain_wake and wait queue and use the existing wait queue > > make wait in draining state as interruptible > > > > v4: > > move pr_err -> pr_debug to avoid spamming kernel log > > make wait in drain interruptible > > v3: > > call snd_compr_drain_notify from compress_stop() > > rename draining -> drain_wake > > add some comments on state transistion after drain > > v2: > > fix the 80 line warn > > move the state change to compress_drain() > > > > include/sound/compress_driver.h | 9 ++++++ > > sound/core/compress_offload.c | 52 ++++++++++++++++++++++++++++++++++++--- > > 2 files changed, 57 insertions(+), 4 deletions(-) > > > > diff --git a/include/sound/compress_driver.h b/include/sound/compress_driver.h > > index 9031a26..ae6c3b8 100644 > > --- a/include/sound/compress_driver.h > > +++ b/include/sound/compress_driver.h > > @@ -171,4 +171,13 @@ static inline void snd_compr_fragment_elapsed(struct snd_compr_stream *stream) > > wake_up(&stream->runtime->sleep); > > } > > > > +static inline void snd_compr_drain_notify(struct snd_compr_stream *stream) > > +{ > > + if (snd_BUG_ON(!stream)) > > + return; > > + > > + stream->runtime->state = SNDRV_PCM_STATE_SETUP; > > + wake_up(&stream->runtime->sleep); > > +} > > + > > #endif > > diff --git a/sound/core/compress_offload.c b/sound/core/compress_offload.c > > index bea523a..e60e0c9 100644 > > --- a/sound/core/compress_offload.c > > +++ b/sound/core/compress_offload.c > > @@ -682,12 +682,48 @@ static int snd_compr_stop(struct snd_compr_stream *stream) > > if (!retval) { > > stream->runtime->state = SNDRV_PCM_STATE_SETUP; > > wake_up(&stream->runtime->sleep); > > + snd_compr_drain_notify(stream); > > The two lines above this do the exactly same as > snd_compr_drain_notify(), thus it can be dropped here. ah yes will send updated one.. -- ~Vinod