Re: [PATCH v2 17/27] drm/tegra: Add Tegra114 HDMI support

[Thierry Reding <thierry.reding-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>]

[-- Attachment #1: Type: text/plain, Size: 2068 bytes --]

On Fri, Oct 11, 2013 at 04:19:19PM -0600, Stephen Warren wrote:
> On 10/07/2013 02:34 AM, Thierry Reding wrote:
> > From: Mikko Perttunen <mperttunen-DDmLM1+adcrQT0dZR+AlfA@public.gmane.org>
> > 
> > Tegra114 TMDS configuration requires a new peak_current field and the
> > driver current override bit has changed position.
> 
> > diff --git a/drivers/gpu/drm/tegra/hdmi.c b/drivers/gpu/drm/tegra/hdmi.c
> 
> >  static const struct tmds_config tegra2_tmds_config[] = {
> > @@ -223,6 +224,85 @@ static const struct tmds_config tegra3_tmds_config[] = {
> 
> Not related to this patch, but those should have been named
> tegra20_tmds_config[] and tegra30_tmds_config[].
> 
> >  static void tegra_hdmi_setup_tmds(struct tegra_hdmi *hdmi,
> 
> > -	value = tmds->drive_current | DRIVE_CURRENT_FUSE_OVERRIDE;
> > -	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_SOR_LANE_DRIVE_CURRENT);
> > +	if (of_device_is_compatible(np, "nvidia,tegra114-hdmi")) {
> 
> Let's not check this at run-time. Instead, host1x_drm_subdevs[]'s .data
> field should be used to contain either flags or a pointer to a
> configuration structure, either of which can be directly consulted to
> determine the properties of the HW in a feature-oriented/semantic way.
> 
> drivers/gpio/gpio-tegra.c's
> tegra20_gpio_config/tegra30_gpio_config/tegra_gpio_of_match provide a
> good example of this.
> 
> This means that if Tegra124 is identical to Tegra114, yet a hypothetical
> Tegra999 is different, you don't have to keep adjusting these if
> conditions throughout the code; they can simply refer to the same
> feature bit forever.

Okay, I'll see what I can come up with. It's unfortunately not as simple
as the GPIO driver's parameterization, and who knows what other
differences will be introduced in some later versions of this block.

What I mean is that at some point it becomes questionable whether it
makes sense to parameterize at all if you have to encode the register
offset and bit position within that register for a large number of bits.

Thierry

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]