[Intel-gfx] [PATCH 00/19] ddi: respin of runtime PM for DPMS

[Daniel Vetter]

On Tue, Jul 01, 2014 at 06:33:50PM -0300, Paulo Zanoni wrote:
> 2014-06-25 16:01 GMT-03:00 Imre Deak <imre.deak at intel.com>:
> > This is a respin of the unmerged part of Daniel's runtime PM for DPMS
> > patchset [1]. The original one also included a refactoring of the DDI
> > PCH/CRT encoder modesetting path, I left the corresponding patches out
> > from this series. This is because there hasn't been yet an agreement on
> > those parts, but people would like to see the RPM DPMS support already
> > applied.
> >
> > Some patches needed to be updated/rebased because of the above omission,
> > but these weren't anywhere significant so I just marked the fact
> > with my s-o-b line. I also added two minor change to keep the
> > modeset sequence at its current order and collected all the reviewed-by
> > lines.
> >
> > Tested on HSW DP/VGA, with basic DPMS on/off and igt/pm_rpm.
> 
> For patches 2, 4, 5, 6, 7, 19: Reviewed-by: Paulo Zanoni
> <paulo.r.zanoni at intel.com>
> 
> However, I tested these patches on a HSW Machine with eDP+HDMI, and
> there are WARNs on dmesg for the dpms-non-lpsp subtest. I found at
> least two problems:
> 1 - Function hsw_ddi_pll_get_hw_state() reads registers while the
> device is suspended.
> 2 - When _intel_set_mode() calls intel_crtc_disable(), it calls
> assert_plane() which reads register 0x71180, which triggers an
> "Unclaimed register" error. I didn't investigate this deeply, so I
> don't have a suggestion for a solution.
> 
> I can reproduce these errors 100% of the time.

Ok, I've pulled in the series (hopefully all of it with all the right
fixup and in the right order), except for the last patch. I'll do that one
once your unclaimed register write fixes are in to avoid regressions.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch