drm/i915: Fix ilk+ watermarks when disabling pipes
authorVille Syrjälä <ville.syrjala@linux.intel.com>
Thu, 25 Oct 2018 13:05:36 +0000 (16:05 +0300)
committerVille Syrjälä <ville.syrjala@linux.intel.com>
Mon, 5 Nov 2018 19:47:16 +0000 (21:47 +0200)
commita748faea3bfd7fd1d1485bc1c426c7d460cc6503
treef23bcef16289ce0b0189a0e9abaef86464ca8482
parent34b2f8da9d41b640a8331e4b8c3bcd9ba2c2b5e6
drm/i915: Fix ilk+ watermarks when disabling pipes

We're no longer programming any watermarks when we're disabling
a pipe. That means ilk_wm_merge() & co. will keep considering
the any pipe that is getting disabled as still enabled. Thus we
either get no LP1+ watermakrs (ilk-ivb), or we get suboptimal
ones (hsw-bdw).

This seems to have been broken by commit b6b178a77210 ("drm/i915:
Calculate ironlake intermediate watermarks correctly, v2."). Before
that we apparently had some difference between the intermediate
and optimal watermarks and so we would program the optiomal ones.
Now intermediate and optimal are identical for disabled pipes
and so we don't program either.

Fix this by programming the intermediate watermarks even for
disabled pipes. We were already doing that for skl+. We'll
leave out gmch platforms for now since those do the merging
in a different manner and should work as is. We'll want to
unify this eventually, but play it safe for now and just put
in a FIXME.

Cc: stable@vger.kernel.org
Cc: Matt Roper <matthew.d.roper@intel.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Fixes: b6b178a77210 ("drm/i915: Calculate ironlake intermediate watermarks correctly, v2.")
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20181025130536.29024-1-ville.syrjala@linux.intel.com
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> #irc
drivers/gpu/drm/i915/intel_display.c