Here's an update of my SKL S3 series:
- Spotted a problem in the BXT resume path, we were missing a bxt_cdclk_init
in the S3 path (we only had it in the runtime resume path). Consolidated
the two paths to reduce the chances of this happening again.
- Updated the SKL display re-init on resume based on Ville's comments. The
major change is that we now poll the PCU for its ACK before changing CDCLK
and this does solve the problem I was seeing (it typically takes 3/4 cycles
separated by 10us to get the PCU to answer 'GO!')
--
Damien
Damien Lespiau (8):
drm/i915: Fix 'suspedn' typo
drm/i915/vlv/chv: Move resume_prepare() after uncore_early_sanitize()
drm/i915/bxt: Also add bxt_resume_prepare() to the S3/S4 path
drm/i915: Extract a common intel_resume_prepare() function
drm/i915: Add a space after ',' and don't capitalize mid-sentence
drm/i915: Kill the dev variable in intel_suspend_complete()
drm/i915: Store boot CDCLK in dev_priv on DDI platforms
drm/i915/skl: Deinit/init the display at suspend/resume
drivers/gpu/drm/i915/i915_drv.c | 63 ++++++-----
drivers/gpu/drm/i915/i915_reg.h | 3 +
drivers/gpu/drm/i915/intel_ddi.c | 6 +-
drivers/gpu/drm/i915/intel_display.c | 210 +++++++++++++++++++++++++++++++++++
drivers/gpu/drm/i915/intel_drv.h | 2 +
5 files changed, 253 insertions(+), 31 deletions(-)
--
2.1.0
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
