From mboxrd@z Thu Jan 1 00:00:00 1970 From: Troy Kisky Subject: Re: [PATCH v4] ASoC: sgtl5000: Fix driver probe after reset Date: Thu, 09 May 2013 16:13:56 -0700 Message-ID: <518C2DB4.5000404@boundarydevices.com> References: <1368140116-15289-1-git-send-email-festevam@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; Format="flowed" Content-Transfer-Encoding: 7bit Return-path: Received: from mail-pb0-f49.google.com (mail-pb0-f49.google.com [209.85.160.49]) by alsa0.perex.cz (Postfix) with ESMTP id CD45226089A for ; Fri, 10 May 2013 01:13:59 +0200 (CEST) Received: by mail-pb0-f49.google.com with SMTP id rp8so2334396pbb.36 for ; Thu, 09 May 2013 16:13:58 -0700 (PDT) In-Reply-To: <1368140116-15289-1-git-send-email-festevam@gmail.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: Fabio Estevam Cc: Fabio Estevam , alsa-devel@alsa-project.org, lars@metafoo.de, matt@genesi-usa.com, eric.nelson@boundarydevices.com, broonie@kernel.org List-Id: alsa-devel@alsa-project.org On 5/9/2013 3:55 PM, Fabio Estevam wrote: > From: Fabio Estevam > > After a 'reboot' command in Linux or after pressing the system's reset button > the sgtl5000 driver fails to probe: > > sgtl5000 0-000a: Device with ID register ffff is not a sgtl5000 > sgtl5000 0-000a: ASoC: failed to probe CODEC -19 > imx-sgtl5000 sound.12: ASoC: failed to instantiate card -19 > imx-sgtl5000 sound.12: snd_soc_register_card failed (-19) > > sgtl5000 codec does not have a reset line, nor a reset command in software, so > after a system reset the codec does not contain the default register values > from sgtl5000_reg_defaults[] anymore, as these are only valid after a > power-on-reset cycle. > > Fix this issue by explicitly reading all the reset register values from > sgtl5000_reg_defaults[] and writing them back into sgtl5000 to ensure a sane > state. > > Signed-off-by: Fabio Estevam > --- > Changes since v3: > - Read sgtl5000_reg_defaults and write these values into sgtl5000 > Changes since v2: > - Do not use reg_defaults_raw as it is not the correct purpose > - Manually build sgtl5000_reg_default > - Improve commitlog > Changes since v1: > - Remove sgtl5000_reg_defaults array > - Do not use num_reg_defaults_raw > sound/soc/codecs/sgtl5000.c | 30 ++++++++++++++++++++++++++++++ > 1 file changed, 30 insertions(+) > > diff --git a/sound/soc/codecs/sgtl5000.c b/sound/soc/codecs/sgtl5000.c > index 327b443..be23fce 100644 > --- a/sound/soc/codecs/sgtl5000.c > +++ b/sound/soc/codecs/sgtl5000.c > @@ -1474,6 +1474,31 @@ static const struct regmap_config sgtl5000_regmap = { > .num_reg_defaults = ARRAY_SIZE(sgtl5000_reg_defaults), > }; > > +/* > + * Write all the default values from sgtl5000_reg_defaults[] array into the > + * sgtl5000 registers, to make sure we always start with the sane registers > + * values as stated in the datasheet. > + * > + * Since sgtl5000 does not have a reset line, nor a reset command in software, > + * we follow this approach to guarantee we always start from the default values > + * and avoid problems like, not being able to probe after an audio playback > + * followed by a system reset or a 'reboot' command in Linux > + */ > +static int sgtl5000_fill_defaults(struct sgtl5000_priv *sgtl5000) > +{ > + int i, ret, val, index; > + > + for (i = 0; i < ARRAY_SIZE(sgtl5000_reg_defaults); i++) { > + val = sgtl5000_reg_defaults[i].def; > + index = sgtl5000_reg_defaults[i].reg; > + ret = regmap_write(sgtl5000->regmap, index, val); > + if (ret) > + return ret; > + } > + > + return 0; > +} > + > static int sgtl5000_i2c_probe(struct i2c_client *client, > const struct i2c_device_id *id) > { > @@ -1494,6 +1519,11 @@ static int sgtl5000_i2c_probe(struct i2c_client *client, > > i2c_set_clientdata(client, sgtl5000); > > + /* Ensure sgtl5000 will start with sane register values */ > + ret = sgtl5000_fill_defaults(sgtl5000); > + if (ret) > + return ret; > + > ret = snd_soc_register_codec(&client->dev, > &sgtl5000_driver, &sgtl5000_dai, 1); > return ret; It seems strange to write all the registers before even verifying that yes this is an SGTL5000. Could you possibly write the 1 register that make SGTL5000_CHIP_ID read correctly, and write the others afterwards? Troy