From: Paulo Zanoni <paulo.r.zanoni@xxxxxxxxx> Hi These patches save a lot of time on our eDP panel power sequencing times. We were doing a lot of needless waits, so now our load times should be much shorter. This series should improve a little bit our driver load time, and it should greatly improve our suspend time. The most relevant thing is that when we were waiting for the "panel off" timeout (50ms on my machine), we were not just waiting for it, but also waiting for the "backlight off" timeout (200ms) and the "panel power cycle" timeout (500ms). QA did some measurements of a branch containing these 6 patches + two patches removing the "wait for vblank timeout" problems we have on Haswell, and the results are reported here: https://bugs.freedesktop.org/show_bug.cgi?id=72211. One of my worries about this series is that, due to the extremely smaller wait times, we may get regression reports saying "my panel is now black!" due to broken VBT/BIOS/panel/whatever. If this happens, IMHO instead of just reverting the current patches, we should add some extra msleep() calls surrounded by FIXME comments explicitly saying why we are waiting more than we should, because the current code was silently hiding the fact that we were waiting more than we should. Thanks, Paulo Paulo Zanoni (6): drm/i915: init the DP panel power seq variables earlier drm/i915: save some time when waiting the eDP timings drm/i915: reset eDP timestamps on resume drm/i915: remove a column of zeros from the eDP wait definitions drm/i915: don't wait for power cycle when waiting for power off drm/i915: set the backlight panel delays registers to 1 drivers/gpu/drm/i915/i915_drv.c | 2 + drivers/gpu/drm/i915/i915_drv.h | 29 ++++++++++++ drivers/gpu/drm/i915/intel_ddi.c | 1 + drivers/gpu/drm/i915/intel_display.c | 11 +++++ drivers/gpu/drm/i915/intel_dp.c | 88 ++++++++++++++++++++++++++++-------- drivers/gpu/drm/i915/intel_drv.h | 6 +++ 6 files changed, 118 insertions(+), 19 deletions(-) -- 1.8.3.1 _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/intel-gfx
