[Intel-gfx] [PATCH 0/3] More Haswell VBT code

[Paulo Zanoni]

From: Paulo Zanoni <paulo.r.zanoni at intel.com>

Hi

These 3 patches are on top of the 2 patches I recently sent to the list:
  [1/2] drm/i915: VBT's child_device_config changes over time
  [2/2] drm/i915: use the HDMI DDI buffer translations from VBT

Patch 1 is IMHO a "must have" since it allows us to discover if our current
assumptions are wrong. If we ever get a "black screen" bug report with one of
the WARNs added by this patch, then we'll have to write some additional code.

Patch 2 is just a bunch of WARNs that should really not ever appear in the log
files. If we ever see these WARNs on a machine, we should probably assume its
VBT is broken, and this may be a really good hint on why the user's eDP panel is
just black.

Patch 3 is just an "optimization" that will make our code not try to do DP AUX
transactions on HDMI-only ports.

I know some of us have really strong negative feelings about the VBT, but we
currently already rely on the VBT for some important display features (e.g.,
lots of SDVO, eDP and LVDS info, CRT DDC, etc.), and the WARNs added by this
series should actually help us decide if the bug reporter's VBT should be
trusted or not. Depending on the bug reports we get, we may consider adding a
"i915.trust_vbt" option or maybe even machine-specific quirks.

Thanks,
Paulo

Paulo Zanoni (3):
  drm/i915: check the DDC and AUX bits of the VBT on DDI machines
  drm/i915: add some assertions about VBT DDI port types
  drm/i915: don't init DP or HDMI when not supported by DDI port

 drivers/gpu/drm/i915/i915_drv.h   |  3 ++
 drivers/gpu/drm/i915/intel_bios.c | 58 ++++++++++++++++++++++++++++++++++++++-
 drivers/gpu/drm/i915/intel_ddi.c  | 30 ++++++++++++++------
 3 files changed, 82 insertions(+), 9 deletions(-)

-- 
1.8.1.2