From: Takashi Iwai Subject: Re: snd_pcm_set_params faild on set_buffer_time_near and set_period_time_near Date: Thu, 23 Jul 2015 10:11:48 +0200 Message-ID: References: <55AE0831.2010206@streamunlimited.com> Mime-Version: 1.0 (generated by SEMI 1.14.6 - "Maruoka") Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Received: from mx2.suse.de (cantor2.suse.de [195.135.220.15]) by alsa0.perex.cz (Postfix) with ESMTP id 572E726007E for ; Thu, 23 Jul 2015 10:11:49 +0200 (CEST) In-Reply-To: <55AE0831.2010206@streamunlimited.com> 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: Martin Geier Cc: alsa-devel@alsa-project.org List-Id: alsa-devel@alsa-project.org On Tue, 21 Jul 2015 10:52:01 +0200, Martin Geier wrote: > > Hallo, > > I tried to configure alsa device with snd_pcm_set_params function, > requested > parameters was: > channel 2, rate: 22050, soft_resample: 1, latency: 50000 > > Function snd_pcm_hw_params_set_buffer_time_near and > snd_pcm_hw_param_set_near > is called with: > min: 50000, max: 50000, mindir: 0, maxdir: -1 > > and input params was: > > ACCESS: RW_INTERLEAVED > FORMAT: S16_LE > SUBFORMAT: STD > SAMPLE_BITS: 16 > FRAME_BITS: 32 > CHANNELS: 2 > RATE: 22050 > PERIOD_TIME: (362 743039) > PERIOD_SIZE: [8 16384] > PERIOD_BYTES: [32 65536] > PERIODS: [2 19] > BUFFER_TIME: (725 1486078) > BUFFER_SIZE: [16 32768] > BUFFER_BYTES: [64 131072] > TICK_TIME: ALL > > unfortunately, this function fails, on function snd_pcm_hw_param_set_first: > > ALSA ERROR hw_params: set_near (BUFFER_TIME) > value = 50000 : Invalid argument > ACCESS: RW_INTERLEAVED > FORMAT: S16_LE > SUBFORMAT: STD > SAMPLE_BITS: 16 > FRAME_BITS: 32 > CHANNELS: 2 > RATE: 22050 > PERIOD_TIME: (4580 5533) > PERIOD_SIZE: NONE > PERIOD_BYTES: [404 488] > PERIODS: 10 > BUFFER_TIME: (50022 50023) > BUFFER_SIZE: 1103 > BUFFER_BYTES: 4412 > TICK_TIME: ALL > > This shouldn't be problem because in snd_pcm_set_params is > snd_pcm_hw_params_set_period_time_near called. > This function should be called with same params as > snd_pcm_hw_params_set_buffer_time_near > but the input params are: > > snd_pcm_hw_param_set_near: min: 12500, max: 12500, mindir: 0, maxdir: -1 > hw_params: snd_pcm_hw_param_set_near (PERIOD_TIME) > ACCESS: RW_INTERLEAVED > FORMAT: S16_LE > SUBFORMAT: STD > SAMPLE_BITS: 16 > FRAME_BITS: 32 > CHANNELS: 2 > RATE: 22050 > PERIOD_TIME: (4580 5533) > PERIOD_SIZE: NONE > PERIOD_BYTES: [404 488] > PERIODS: 10 > BUFFER_TIME: (50022 50023) > BUFFER_SIZE: 1103 > BUFFER_BYTES: 4412 > TICK_TIME: ALL > > because in snd_pcm_hw_param_set_near is > > if (last) > err = snd_pcm_hw_param_set_last(pcm, params, var, val, dir); > else > err = snd_pcm_hw_param_set_first(pcm, params, var, val, dir); > if (err < 0) > dump_hw_params(params, "set_near", var, *val, err); > return err; > > and in error part is only debug output and not restore original parameters. > > Environment: > alsa-lib 1.0.27.1 > TI AM335x soc > > Simple fix is attached, but I am not sure if it is correct. Thanks for analysis and patch. It's however rather a bug in snd_pcm_set_params(). There is no guarantee to keep the old value in snd_pcm_set_*(), thus the caller needs to take care of it instead. Takashi