[PATCH v5 4/7] drm/bridge: dw-hdmi: Add API dw_hdmi_set_sample_iec958() for iec958 format

[Maxime Ripard]

Hi,

On Wed, Sep 03, 2025 at 06:41:05PM +0800, Shengjiu Wang wrote:
> On Tue, Sep 2, 2025 at 12:52 AM Luca Ceresoli <luca.ceresoli at bootlin.com> wrote:
> >
> > Hello Shengjiu,
> >
> > On Thu, 21 Aug 2025 15:31:28 +0800
> > Shengjiu Wang <shengjiu.wang at nxp.com> wrote:
> >
> > > Add API dw_hdmi_set_sample_iec958() for IEC958 format because audio device
> > > driver needs IEC958 information to configure this specific setting.
> > >
> > > Signed-off-by: Shengjiu Wang <shengjiu.wang at nxp.com>
> > > Acked-by: Liu Ying <victor.liu at nxp.com>
> >
> > [...]
> >
> > > +void dw_hdmi_set_sample_iec958(struct dw_hdmi *hdmi, unsigned int iec958)
> > > +{
> > > +     mutex_lock(&hdmi->audio_mutex);
> > > +     hdmi->sample_iec958 = iec958;
> > > +     mutex_unlock(&hdmi->audio_mutex);
> > > +}
> >
> > Apologies for jumping in the discussion as late as in v5, but I noticed
> > this patch and I was wondering whether this mutex_lock/unlock() is
> > really needed, as you're copying an int.
> 
> Thanks for your comments.
> 
> Seems it is not necessary to add mutex here. I just follow the code as
> other similar functions.  I will send a new version to update it.

Let's not be smart about it. Next thing you know, someone will add
another field in there that would absolutely require a mutex and now
you're not race free anymore.

Unless there's a real concern, the mutex must stay.

Maxime
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 273 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20250909/1f8165e7/attachment.sig>