ACK/Cmnt: [SRU][B/C/D][PATCH v2 0/4] Audio doesn't work after s3 or s4 on cnl and glk machines
stefan.bader at canonical.com
Fri Jun 28 11:46:03 UTC 2019
On 10.06.19 09:14, Hui Wang wrote:
> BugLink: https://bugs.launchpad.net/bugs/1826868
> The patch is already in the v5.2-rc1, so don't send the patch
> to unstable.
> In the V2, I added 3 more patches. The 0001-xxx.patch is for fixing
> the issue of "audio doesn't work after s3/s4 on cln and glk machines",
> but it introduces a new issue of "screen flickering when CDCLK changes",
> So I added 0002/3/4-xxx.patch to fix the new issue.
> After s3 or s4, the hdmi audio driver will re-probe the hw, if the hdmi
> monitor is not plugged before s3 or s4, the re-probe will fail on cnl
> and glk machines.
> Keep the cdclk alive when re-probing
> [Test Case]
> I applied the patches to bionic kernel and disco kernel, and tested them
> on Whiskeylake, cannonlake, kabylake, skylake and geminilake
> machines, before s3, I verified analogue audio, hdmi audio and screen
> dispaly all worked well, then let the system enter s3 and resume, recheck
> the analogue audio, hdmi audio and screen display, all worked well too.
> And repeated the test 5 times.
> [Regression Risk]
> Low. I have tested the patched kernel on different Dell and Lenovo machines,
> they include Whiskeylake, cannonlake, kabylake, skylake and geminilake.
> Imre Deak (2):
> drm/i915: Save the old CDCLK atomic state
> drm/i915: Remove redundant store of logical CDCLK state
> Ville Syrjälä (2):
> drm/i915: Force 2*96 MHz cdclk on glk/cnl when audio power is enabled
> drm/i915: Skip modeset for cdclk changes if possible
> drivers/gpu/drm/i915/i915_drv.h | 6 +-
> drivers/gpu/drm/i915/intel_audio.c | 57 +++++++++
> drivers/gpu/drm/i915/intel_cdclk.c | 171 ++++++++++++++++++++-------
> drivers/gpu/drm/i915/intel_display.c | 56 +++++++--
> drivers/gpu/drm/i915/intel_drv.h | 22 +++-
> 5 files changed, 257 insertions(+), 55 deletions(-)
Rather hesitant for the Bionic generic kernel because, like I said many times
before, I would limit the changes to released LTS kernels to the bug fixing
changes. For hardware enablement we have either the OEM kernel or the HWE
kernels. So Cosmic and Disco, yes. But not so much for Bionic.
Especially not the i915 driver which handles everything and is notorious to
enable one thing and breaking old stuff in the process... But likely it will be
done anyway, so...
Acked-by: Stefan Bader <stefan.bader at canonical.com>
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 833 bytes
Desc: OpenPGP digital signature
More information about the kernel-team