[Intel-gfx] [PATCH 4/9] drm/i915: Add infrastructure for choosing DPLLs before disabling crtcs

[Daniel Vetter]

On Wed, Oct 29, 2014 at 11:32:33AM +0200, Ander Conselvan de Oliveira wrote:
> It is possible for a mode set to fail if there aren't shared DPLLS that
> match the new configuration requirement or other errors in clock
> computation. If that step is executed after disabling crtcs, in the
> failure case the hardware configuration is changed and needs to be
> restored. Doing those things early will allow the mode set to fail
> before actually touching the hardware.
> 
> Follow up patches will convert different platforms to use the new
> infrastructure.
> 
> v2: Keep pll->new_config valid only during mode set (Ville)
>     Use kmemdup() in i915_shared_dpll_start_config() (Ville)
>     Restore old pll config if something fails before commit (Ville)
>     Don't set compute_clock hooks since dev_priv is kzalloc()'d (Ville)
> 
> Signed-off-by: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira at intel.com>

Ran into a blocking question with this one, merged thus far.

> @@ -7395,6 +7456,9 @@ static int ironlake_crtc_mode_set(struct intel_crtc *crtc,
>  		else
>  			crtc->new_config->dpll_hw_state.fp1 = fp;
>  
> +		if (intel_crtc_to_shared_dpll(crtc))
> +			intel_put_shared_dpll(crtc);

Don't we need the same fixup in intel_ddi_pll_select?
> +
>  		pll = intel_get_shared_dpll(crtc);
>  		if (pll == NULL) {
>  			DRM_DEBUG_DRIVER("failed to find PLL for pipe %c\n",
> @@ -10739,6 +10803,22 @@ static int __intel_set_mode(struct drm_crtc *crtc,
>  		prepare_pipes &= ~disable_pipes;
>  	}
>  
> +	if (dev_priv->display.crtc_compute_clock) {
> +		unsigned clear_pipes = modeset_pipes | disable_pipes;
> +
> +		ret = intel_shared_dpll_start_config(dev_priv, clear_pipes);
> +		if (ret)
> +			goto done;
> +
> +		for_each_intel_crtc_masked(dev, modeset_pipes, intel_crtc) {
> +			ret = dev_priv->display.crtc_compute_clock(intel_crtc);
> +			if (ret) {
> +				intel_shared_dpll_abort_config(dev_priv);
> +				goto done;
> +			}
> +		}
> +	}

Might be useful to shuffle this and the vlv-specific code above into a new
intel_compute_global_config kind of helper function. But that can be done
later on I think.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch