drm/i915: Drop redundant watermark programming
authorMatt Roper <matthew.d.roper@intel.com>
Thu, 24 Sep 2015 22:53:06 +0000 (15:53 -0700)
committerDaniel Vetter <daniel.vetter@ffwll.ch>
Wed, 30 Sep 2015 15:15:13 +0000 (17:15 +0200)
commit37d9078b9b395185947784ba8d71c0980e5aacad
treee288fc9a458887b200a2874b46fdf857cc86dc4b
parentd8135109e45d4cf352387b7de66c216a176e37f0
drm/i915: Drop redundant watermark programming

In commit

        commit e4ca061275ec6a48b66c6edebe08644e666994c0
        Author: Patrik Jakobsson <patrik.jakobsson@linux.intel.com>
        Date:   Wed Jul 8 15:31:52 2015 +0200

            drm/i915: Don't forget to mark crtc as inactive after disable

we added extra watermark updates to all of the .crtc_disable()
entrypoints to avoid problems problems with system resume on SKL.  Those
disable entrypoints are currently called in just two places in the
driver: intel_atomic_commit (i.e., during a modeset) and
intel_crtc_disable_noatomic (which is called during hardware readout).
It seems that this extra watermark recalculation should only be
important in the latter case (which happens during a resume operation);
the former case should always have appropriate watermark programming
happening at other points in the modeset sequence.

Let's move the watermark update out of the .crtc_disable() entrypoints
and place it directly in intel_crtc_disable_noatomic() so that it only
happens on S3 resume and not during a regular modeset (since the
existing watermark handling should properly update watermarks during
normal atomic commits).

Cc: Patrik Jakobsson <patrik.jakobsson@linux.intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
drivers/gpu/drm/i915/intel_display.c