[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20150610180912.GQ14071@sirena.org.uk>
Date: Wed, 10 Jun 2015 19:09:12 +0100
From: Mark Brown <broonie@...nel.org>
To: Zidan Wang <zidan.wang@...escale.com>
Cc: nicoleotsuka@...il.com, perex@...ex.cz, tiwai@...e.de,
timur@...i.org, Xiubo.Lee@...il.com, devicetree@...r.kernel.org,
linux-kernel@...r.kernel.org, alsa-devel@...a-project.org,
linuxppc-dev@...ts.ozlabs.org
Subject: Re: [alsa-devel][PATCH] ASoC: fsl: add imx-wm8960 machine driver
On Wed, Jun 10, 2015 at 04:26:27PM +0800, Zidan Wang wrote:
> + - hp-det : ADCLRC/GPIO1, LINPUT3/JD2 and RINPUT3/JD3 pins can
> + be selected as headphone jack detect inputs to
> + automatically disable the speaker output and enable
> + the headphone output.
> + hp-det = <hp-det-pin hp-det-polarity>;
> + hp-det-pin = 1: ADCLRC/GPIO1 used as detect pin
> + hp-det-pin = 2: LINPUT3/JD2 used as detect pin
> + hp-det-pin = 3: RINPUT3/JD3 used as detect pin
> + hp-det-polarity = 0: hp detect high for headphone
> + hp-det-polarity = 1: hp detect high for speaker
This looks like something that should be in the DT binding for the
CODEC, not the machine driver.
> +static int hp_set_status_check(void *data)
> +{
Why is this in the driver?
> + hp_status = gpio_get_value(priv->hp_set_gpio) ? 1 : 0;
gpio_get_value() already returns a boolean.
> + if (hp_status != priv->hp_active_low) {
> + snprintf(buf, 32, "STATE=%d", 2);
> + snd_soc_dapm_disable_pin(&priv->codec->dapm, "Ext Spk");
> + snd_soc_dapm_disable_pin(&priv->codec->dapm, "Main MIC");
> + ret = imx_hp_set_gpio.report;
The generic jack code already has support for disabling and enabling
pins, though disabling the speaker looks like a policy decision which
probably doesn't belong here.
> + /*
> + * As the hp MIC only connect the input for left channel, we
> + * need to route it for right channel.
> + */
> + snd_soc_update_bits(priv->codec, WM8960_ADDCTL1, 3<<2, 1<<2);
This looks like routing which we'd expect userspace to be doing.
> + snd_kctl_jack_report(priv->snd_card, priv->headset_kctl, 1);
Use the ASoC level helpers.
> + envp[0] = "NAME=headset";
> + envp[1] = buf;
> + envp[2] = NULL;
> + kobject_uevent_env(&pdev->dev.kobj, KOBJ_CHANGE, envp);
> + kfree(buf);
Let the core deal with notifying userspace - it looks like you want to
implement extcon integration there.
> + /* set cpu DAI configuration */
> + ret = snd_soc_dai_set_fmt(cpu_dai, fmt);
> + if (ret) {
> + dev_err(dev, "failed to set cpu dai fmt: %d\n", ret);
> + return ret;
> + }
> + /* set codec DAI configuration */
> + ret = snd_soc_dai_set_fmt(codec_dai, fmt);
> + if (ret) {
> + dev_err(dev, "failed to set codec dai fmt: %d\n", ret);
> + return ret;
> + }
Initialise these in the dai_link structure.
> + for (i = 0; i < ARRAY_SIZE(sysclk_divs); ++i) {
> + if (sysclk_divs[i] == -1)
> + continue;
> + sysclk /= sysclk_divs[i];
> + for (j = 0; j < ARRAY_SIZE(dac_divs); ++j) {
> + if (sysclk == sample_rate * dac_divs[j]) {
> + for (k = 0; k < ARRAY_SIZE(bclk_divs); ++k)
> + if (sysclk == bclk * bclk_divs[k] / 10)
> + break;
> + if (k != ARRAY_SIZE(bclk_divs))
> + break;
> + }
> + }
> + if (j != ARRAY_SIZE(dac_divs))
> + break;
> + }
> + if (i != ARRAY_SIZE(sysclk_divs)) {
> + /* Set codec sysclk */
> + snd_soc_dai_set_sysclk(codec_dai,
> + WM8960_SYSCLK_MCLK, sysclk, 0);
> + snd_soc_dai_set_clkdiv(codec_dai, WM8960_SYSCLKDIV, i << 1);
> + return 0;
> + }
Better, upgrade the CODEC driver to do this.
> + /* codec mclk should be enabled early to avoid jack detect error */
> + ret = clk_prepare_enable(data->codec_clk);
> + if (ret) {
> + dev_err(card->dev, "Failed to enable MCLK: %d\n", ret);
> + return ret;
> + }
Similarly integrating clock API support into the CODEC is better - it
makes all the DT stuff work a lot more smoothly and avoids things having
to get duplicated.
Download attachment "signature.asc" of type "application/pgp-signature" (474 bytes)
Powered by blists - more mailing lists