[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <87tszz6mmk.wl-tiwai@suse.de>
Date: Thu, 16 Oct 2025 14:45:39 +0200
From: Takashi Iwai <tiwai@...e.de>
To: Mark Brown <broonie@...nel.org>
Cc: Richard Fitzgerald <rf@...nsource.cirrus.com>,
Takashi Iwai <tiwai@...e.de>,
tiwai@...e.com,
linux-sound@...r.kernel.org,
linux-kernel@...r.kernel.org,
patches@...nsource.cirrus.com
Subject: Re: [PATCH 05/11] ALSA: hda/cs35l56: Create sysfs files for factory calibration
On Thu, 16 Oct 2025 14:13:16 +0200,
Mark Brown wrote:
>
> On Thu, Oct 16, 2025 at 12:58:15PM +0100, Richard Fitzgerald wrote:
> > On 16/10/2025 12:27 pm, Mark Brown wrote:
>
> > > This feels like it might be a better fit for debugfs or possibly
> > > configfs? It's not really within the sysfs rules, and especially
> > > debugfs is a lot more relaxed about everything.
>
> > debugfs is an easy change in the driver but more complex for the
> > tooling. Unlike sysfs there's no standard layout or naming convention
> > for the ASoC debugfs tree, so it's more troublesome to locate the amp
> > entries.
>
> > ASoC creates a debugfs root by default if DEBUGFS is enabled. But HDA
> > doesn't have its own debugfs. There is an ALSA debugfs enabled by
> > CONFIG_SND_DEBUG, which is normally disabled. But enabling
> > CONFIG_SND_DEBUG has other side-effects, it enables more than debugfs.
>
> > Or we could create a new debugfs node specifically for Cirrus Logic
> > directly in /sys/kernel/debug.
>
> Dunno about Takashi but for this application I'd be happy for you to
> just put something in the root of debugfs.
I'm fine with debugfs root. The current approach with sysfs looks too
generic for the need of this access, so I find debugfs would be a
better fit, too.
thanks,
Takashi
Powered by blists - more mailing lists