Re: [PATCH 2/6] ASoC: tegra: Add Tegra186 based ASRC driver

From: Mark Brown <broonie@kernel.org>
To: Sameer Pujar <spujar@nvidia.com>
Cc: robh+dt@kernel.org, krzk+dt@kernel.org, thierry.reding@gmail.com,
	lgirdwood@gmail.com, perex@perex.cz, tiwai@suse.com,
	jonathanh@nvidia.com, catalin.marinas@arm.com, will@kernel.org,
	alsa-devel@alsa-project.org, devicetree@vger.kernel.org,
	linux-tegra@vger.kernel.org, linux-kernel@vger.kernel.org,
	linux-arm-kernel@lists.infradead.org
Subject: Re: [PATCH 2/6] ASoC: tegra: Add Tegra186 based ASRC driver
Date: Mon, 28 Mar 2022 16:44:31 +0100	[thread overview]
Message-ID: <YkHX3/8BbXo4obWI@sirena.org.uk> (raw)
In-Reply-To: <1648447526-14523-3-git-send-email-spujar@nvidia.com>

[-- Attachment #1: Type: text/plain, Size: 2478 bytes --]

On Mon, Mar 28, 2022 at 11:35:22AM +0530, Sameer Pujar wrote:

> +	regcache_cache_only(asrc->regmap, false);
> +	regcache_sync(asrc->regmap);
> +
> +	/* Setup global registers */
> +	regmap_write(asrc->regmap, TEGRA186_ASRC_GLOBAL_SOFT_RESET, 0x1);
> +	regmap_write(asrc->regmap, TEGRA186_ASRC_GLOBAL_SCRATCH_ADDR,
> +		     TEGRA186_ASRC_ARAM_START_ADDR);
> +	regmap_write(asrc->regmap, TEGRA186_ASRC_GLOBAL_INT_MASK, 0x01);
> +	regmap_write(asrc->regmap, TEGRA186_ASRC_GLOBAL_ENB,
> +		     TEGRA186_ASRC_GLOBAL_EN);
> +	regmap_write(asrc->regmap, TEGRA186_ASRC_GLOBAL_INT_CLEAR, 0x01);

This seems weird - we resync the cache, then do a soft reset (which
presumably desyncs the cache) and then explicitly restore a bunch of
things (hopefully everything that was in the cached state?).  This is
certainly very much not idiomatic and looks worrying.  Are you sure that
the device is getting anything out of the register cache?

> +static int tegra186_asrc_put_ratio_source(struct snd_kcontrol *kcontrol,
> +					  struct snd_ctl_elem_value *ucontrol)
> +{
> +	struct soc_enum *asrc_private =
> +		(struct soc_enum  *)kcontrol->private_value;
> +	struct snd_soc_component *cmpnt = snd_soc_kcontrol_component(kcontrol);
> +	struct tegra186_asrc *asrc = snd_soc_component_get_drvdata(cmpnt);
> +	unsigned int id = asrc_private->reg / TEGRA186_ASRC_STREAM_STRIDE;
> +
> +	asrc->lane[id].ratio_source = ucontrol->value.enumerated.item[0];
> +
> +	regmap_update_bits(asrc->regmap, asrc_private->reg,
> +			   TEGRA186_ASRC_STREAM_RATIO_TYPE_MASK,
> +			   asrc->lane[id].ratio_source);
> +
> +	return 1;
> +}

This should only return 1 if the value actually changed, you can use
regmap_update_bits_check() to detect the change.  Current mixer-test
ought to spot this.

> +static const struct snd_kcontrol_new tegra186_asrc_controls[] = {
> +	/* Controls for integer part of ratio */
> +	SOC_SINGLE_EXT("Ratio1 Integer Part",
> +		       ASRC_STREAM_REG(TEGRA186_ASRC_RATIO_INT_PART, 0),
> +		       0, TEGRA186_ASRC_STREAM_RATIO_INT_PART_MASK, 0,
> +		       tegra186_asrc_get_ratio_int,
> +		       tegra186_asrc_put_ratio_int),

Can't the driver work out the ratios based on...

> +	/* Source of ratio provider */
> +	SOC_ENUM_EXT("Ratio1 Source", src_select1,
> +		     tegra186_asrc_get_ratio_source,
> +		     tegra186_asrc_put_ratio_source),

...the sources?  Or does it need to be configured before either side is
ready in which case this might be the best we can do for now.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]