On Mon, 19 Dec 2016, Wang Elaine <elaine.wang@xxxxxxxxx> wrote:
> From: Elaine Wang <elaine.wang@xxxxxxxxx>
>
> when num_pipes is zero, it indicates display doesn't exist, so there
> is no need to initialize display hooks. And to avoid calling these
> uninitialized display hooks, respect num_pipes at the beginning of
> intel_modeset_init_hw and intel_init_clock_gating.
>
> intel_init_pm() calls FBC init function and then initializes
> water mark hooks. Both aren't needed when display doesn't exist. Add
> checking num_pipes at the beginning of intel_init_pm().
>
> v2: Move one check from caller to callee for consistency.
>
> Cc: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
> Cc: Joonas Lahtinen <joonas.lahtinen@xxxxxxxxxxxxxxx>
> Cc: Jani Nikula <jani.nikula@xxxxxxxxxxxxxxx>
> Signed-off-by: Elaine Wang <elaine.wang@xxxxxxxxx>
> ---
> drivers/gpu/drm/i915/intel_display.c | 6 ++++++
> drivers/gpu/drm/i915/intel_pm.c | 10 +++++++++-
> 2 files changed, 15 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
> index 9cc5dbf..e079ea7 100644
> --- a/drivers/gpu/drm/i915/intel_display.c
> +++ b/drivers/gpu/drm/i915/intel_display.c
> @@ -16030,6 +16030,9 @@ static void intel_atomic_state_free(struct drm_atomic_state *state)
> */
> void intel_init_display_hooks(struct drm_i915_private *dev_priv)
> {
> + if (INTEL_INFO(dev_priv)->num_pipes == 0)
> + return;
> +
> if (INTEL_INFO(dev_priv)->gen >= 9) {
> dev_priv->display.get_pipe_config = haswell_get_pipe_config;
> dev_priv->display.get_initial_plane_config =
> @@ -16412,6 +16415,9 @@ void intel_modeset_init_hw(struct drm_device *dev)
> {
> struct drm_i915_private *dev_priv = to_i915(dev);
>
> + if (INTEL_INFO(dev_priv)->num_pipes == 0)
> + return;
> +
> intel_update_cdclk(dev_priv);
>
> dev_priv->atomic_cdclk_freq = dev_priv->cdclk_freq;
> diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
> index d0834b3..6438ada 100644
> --- a/drivers/gpu/drm/i915/intel_pm.c
> +++ b/drivers/gpu/drm/i915/intel_pm.c
> @@ -7619,7 +7619,8 @@ static void i830_init_clock_gating(struct drm_i915_private *dev_priv)
>
> void intel_init_clock_gating(struct drm_i915_private *dev_priv)
> {
> - dev_priv->display.init_clock_gating(dev_priv);
> + if (INTEL_INFO(dev_priv)->num_pipes)
> + dev_priv->display.init_clock_gating(dev_priv);
> }
>
> void intel_suspend_hw(struct drm_i915_private *dev_priv)
> @@ -7644,6 +7645,10 @@ static void nop_init_clock_gating(struct drm_i915_private *dev_priv)
> */
> void intel_init_clock_gating_hooks(struct drm_i915_private *dev_priv)
> {
> +
> + if (INTEL_INFO(dev_priv)->num_pipes == 0)
> + return;
> +
Looks like this would be neater by doing
if (INTEL_INFO(dev_priv)->num_pipes == 0)
dev_priv->display.init_clock_gating = nop_init_clock_gating;
else if (IS_SKYLAKE(dev_priv))
...
and you can leave out the check in intel_init_clock_gating().
We really want to minimize the amount of these checks sprinkled around.
BR,
Jani.
> dev_priv->display.init_clock_gating = skylake_init_clock_gating;
> else if (IS_KABYLAKE(dev_priv))
> @@ -7685,6 +7690,9 @@ void intel_init_clock_gating_hooks(struct drm_i915_private *dev_priv)
> /* Set up chip specific power management-related functions */
> void intel_init_pm(struct drm_i915_private *dev_priv)
> {
> + if (INTEL_INFO(dev_priv)->num_pipes == 0)
> + return;
> +
> intel_fbc_init(dev_priv);
>
> /* For cxsr */
--
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
