[Intel-gfx] [PATCH] drm/vblank: Document drm_crtc_vblank_restore constraints

[Ville Syrjälä]

On Tue, Feb 09, 2021 at 11:15:23AM +0100, Daniel Vetter wrote:
> I got real badly confused when trying to review a fix from Ville for
> this. Let's try to document better what's required for this, and check
> the minimal settings at runtime - we can't check ofc that there's
> indeed no races in the driver callback.
> 
> Also noticed that the drm_vblank_restore version is unused, so lets
> unexport that while at it.
> 
> Cc: Ville Syrjala <ville.syrjala at linux.intel.com>
> Signed-off-by: Daniel Vetter <daniel.vetter at intel.com>
> ---
>  drivers/gpu/drm/drm_vblank.c | 25 ++++++++++---------------
>  include/drm/drm_vblank.h     |  1 -
>  2 files changed, 10 insertions(+), 16 deletions(-)
> 
> diff --git a/drivers/gpu/drm/drm_vblank.c b/drivers/gpu/drm/drm_vblank.c
> index c914b14cfb43..05f4d4c078fd 100644
> --- a/drivers/gpu/drm/drm_vblank.c
> +++ b/drivers/gpu/drm/drm_vblank.c
> @@ -1471,20 +1471,7 @@ void drm_crtc_vblank_on(struct drm_crtc *crtc)
>  }
>  EXPORT_SYMBOL(drm_crtc_vblank_on);
>  
> -/**
> - * drm_vblank_restore - estimate missed vblanks and update vblank count.
> - * @dev: DRM device
> - * @pipe: CRTC index
> - *
> - * Power manamement features can cause frame counter resets between vblank
> - * disable and enable. Drivers can use this function in their
> - * &drm_crtc_funcs.enable_vblank implementation to estimate missed vblanks since
> - * the last &drm_crtc_funcs.disable_vblank using timestamps and update the
> - * vblank counter.
> - *
> - * This function is the legacy version of drm_crtc_vblank_restore().
> - */
> -void drm_vblank_restore(struct drm_device *dev, unsigned int pipe)
> +static void drm_vblank_restore(struct drm_device *dev, unsigned int pipe)
>  {
>  	ktime_t t_vblank;
>  	struct drm_vblank_crtc *vblank;
> @@ -1520,7 +1507,6 @@ void drm_vblank_restore(struct drm_device *dev, unsigned int pipe)
>  		    diff, diff_ns, framedur_ns, cur_vblank - vblank->last);
>  	store_vblank(dev, pipe, diff, t_vblank, cur_vblank);
>  }
> -EXPORT_SYMBOL(drm_vblank_restore);
>  
>  /**
>   * drm_crtc_vblank_restore - estimate missed vblanks and update vblank count.
> @@ -1531,9 +1517,18 @@ EXPORT_SYMBOL(drm_vblank_restore);
>   * &drm_crtc_funcs.enable_vblank implementation to estimate missed vblanks since
>   * the last &drm_crtc_funcs.disable_vblank using timestamps and update the
>   * vblank counter.
> + *
> + * Note that drivers must have race-free high-precision timestamping support,
> + * i.e.  &drm_crtc_funcs.get_vblank_timestamp must be hooked up and
> + * &drm_driver.vblank_disable_immediate must be set to indicate the
> + * time-stamping functions are race-free against vblank hardware counter
> + * increments.

Looks good. Might prevent someone from shooting themselves in
the foot.

Reviewed-by: Ville Syrjälä <ville.syrjala at linux.intel.com>

>   */
>  void drm_crtc_vblank_restore(struct drm_crtc *crtc)
>  {
> +	WARN_ON_ONCE(!crtc->funcs->get_vblank_timestamp);
> +	WARN_ON_ONCE(!crtc->dev->vblank_disable_immediate);
> +
>  	drm_vblank_restore(crtc->dev, drm_crtc_index(crtc));
>  }
>  EXPORT_SYMBOL(drm_crtc_vblank_restore);
> diff --git a/include/drm/drm_vblank.h b/include/drm/drm_vblank.h
> index dd125f8c766c..733a3e2d1d10 100644
> --- a/include/drm/drm_vblank.h
> +++ b/include/drm/drm_vblank.h
> @@ -247,7 +247,6 @@ void drm_crtc_vblank_off(struct drm_crtc *crtc);
>  void drm_crtc_vblank_reset(struct drm_crtc *crtc);
>  void drm_crtc_vblank_on(struct drm_crtc *crtc);
>  u64 drm_crtc_accurate_vblank_count(struct drm_crtc *crtc);
> -void drm_vblank_restore(struct drm_device *dev, unsigned int pipe);
>  void drm_crtc_vblank_restore(struct drm_crtc *crtc);
>  
>  void drm_calc_timestamping_constants(struct drm_crtc *crtc,
> -- 
> 2.30.0

-- 
Ville Syrjälä
Intel