linux-kernel - Re: [PATCH v2] ALSA: hda/conexant: Fix pop noise on conexant codec

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <d57548e1-3f23-4b9e-84c2-cfdc31e92e1e@intel.com>
Date: Wed, 5 Nov 2025 16:45:00 +0100
From: Cezary Rojewski <cezary.rojewski@...el.com>
To: <wangdich9700@....com>
CC: <linux-kernel@...r.kernel.org>, <alsa-devel@...a-project.org>,
	<linux-arm-kernel@...ts.infradead.org>, wangdicheng <wangdicheng@...inos.cn>,
	<lgirdwood@...il.com>, <broonie@...nel.org>, <perex@...ex.cz>,
	<tiwai@...e.com>
Subject: Re: [PATCH v2] ALSA: hda/conexant: Fix pop noise on conexant codec

On 2025-11-05 3:25 AM, wangdich9700@....com wrote:
> From: wangdicheng <wangdich9700@....com>
> 
> Pop noise mitigation: When headphones are unplugged during playback,mute
> speaker DAC(0x17)immediately and restore after 20ms delay to avoid
> audible popping.

What's the device ID for the "conexant codec"? I doubt all of them share 
the problem. Once obtained, mention the ID within the title.

Now, the problem you're aiming to address seem generic - addressing pop 
noises during plug/unplug is a common subject in sound/ domain. I'd 
check if sound/hda/codecs doesn't have something for such occasion 
already. Takashi is more fluent in sound/hda, perhaps he could help here.

If such code exists, you could get rid of ugly:
	codec->spec <-> spec->codec

relation which you're adding with:
	struct hda_codec *conexant_codec;

> 
> Signed-off-by: wangdicheng <wangdicheng@...inos.cn>
> ---
>   sound/hda/codecs/conexant.c | 65 +++++++++++++++++++++++++++++++++++++
>   1 file changed, 65 insertions(+)
> 
> diff --git a/sound/hda/codecs/conexant.c b/sound/hda/codecs/conexant.c
> index 5fcbc1312c69..2c16d1930a75 100644
> --- a/sound/hda/codecs/conexant.c
> +++ b/sound/hda/codecs/conexant.c
> @@ -43,6 +43,9 @@ struct conexant_spec {
>   	unsigned int gpio_mute_led_mask;
>   	unsigned int gpio_mic_led_mask;
>   	bool is_cx11880_sn6140;
> +
> +	struct hda_codec *conexant_codec;
> +	struct delayed_work change_pinctl_work;

Not fan of adding 'struct hda_codec *' as part of the spec. As stated 
above, I'd first check if something pop noise -related doesn't exist in 
sound/hda already. If it does, hook into it.

>   };
>   
>   
> @@ -216,6 +219,63 @@ static void cx_remove(struct hda_codec *codec)
>   	snd_hda_gen_remove(codec);
>   }
>   
> +static void mute_unmute_speaker(struct hda_codec *codec, hda_nid_t nid, bool mute)
> +{
> +	unsigned int conn_sel, dac, conn_list, gain_left, gain_right;
> +
> +	conn_sel = snd_hda_codec_read(codec, nid, 0, 0xf01, 0x0);
> +	conn_list = snd_hda_codec_read(codec, nid, 0, 0xf02, 0x0);
> +
> +	dac = ((conn_list >> (conn_sel * 8)) & 0xff);
> +	if (dac == 0)
> +		return;
> +
> +	gain_left = snd_hda_codec_read(codec, dac, 0, 0xba0, 0x0);
> +	gain_right = snd_hda_codec_read(codec, dac, 0, 0xb80, 0x0);
> +
> +	if (mute) {
> +		gain_left |= 0x80;
> +		gain_right |= 0x80;
> +	} else {
> +		gain_left &= (~(0x80));
> +		gain_right &= (~(0x80));
> +	}
> +
> +	snd_hda_codec_write(codec, dac, 0, 0x3a0, gain_left);
> +	snd_hda_codec_write(codec, dac, 0, 0x390, gain_right);
> +
> +	if (mute) {
> +		snd_hda_codec_write(codec, nid, 0, 0x707, 0);
> +		codec_dbg(codec, "mute_speaker, set 0x%x  PinCtrl to 0.\n", nid);
> +	} else {
> +		snd_hda_codec_write(codec, nid, 0, 0x707, 0x40);
> +		codec_dbg(codec, "unmute_speaker, set 0x%x  PinCtrl to 0x40.\n", nid);
> +	}
> +}
> +
> +static void change_pinctl_worker(struct work_struct *work)
> +{
> +	struct hda_codec *codec;
> +	struct conexant_spec *spec;
> +
> +	spec = container_of(work, struct conexant_spec, change_pinctl_work.work);
> +	codec = spec->conexant_codec;
> +
> +	return mute_unmute_speaker(codec, 0x17, false);

return-statement in void? The code won't compile. Was the patch tested 
at all?

> +}