[pull] radeon drm-fixes-3.16

Dieter Nützel Dieter at nuetzel-hh.de
Thu Jul 17 17:50:06 PDT 2014


Am 17.07.2014 22:50, schrieb Alex Deucher:
> Hi Dave,
> 
> A few more fixes for 3.16.  The pageflipping fixes I dropped last week
> have finally shaped up so this is mostly fixes for fallout from the
> pageflipping code changes.  Also fix a memory leak and a black screen
> when restoring the backlight on console unblanking.
> 
> The following changes since commit 
> bf38b025d3f58f4c1273714ff1be5bfbf99574a4:
> 
>   Merge branch 'drm-fixes-3.16' of
> git://people.freedesktop.org/~agd5f/linux into drm-fixes (2014-07-11
> 11:24:13 +1000)
> 
> are available in the git repository at:
> 
> 
>   git://people.freedesktop.org/~agd5f/linux drm-fixes-3.16
> 
> for you to fetch changes up to 
> 5f87e090a7368adc2290ae17ffd82a070caadd20:
> 
>   drm/radeon: Make classic pageflip completion path less racy.
> (2014-07-17 09:04:03 -0400)
> 
> ----------------------------------------------------------------
> Alex Deucher (2):
>       drm/radeon: avoid leaking edid data
>       drm/radeon: set default bl level to something reasonable
> 
> Mario Kleiner (4):
>       drm/radeon: Prevent too early kms-pageflips triggered by vblank.
>       drm/radeon: Remove redundant fence unref in pageflip path.
>       drm/radeon: Add missing vblank_put in pageflip ioctl error path.
>       drm/radeon: Make classic pageflip completion path less racy.
> 
> Michel Dänzer (2):
>       drm/radeon: Move pinning the BO back to radeon_crtc_page_flip()
>       drm/radeon: Complete page flip even if waiting on the BO fence 
> fails

Hello Alex and Michel,

isn't this needed anylonger?
drm-radeon-Disable-pflip-interrupts.patch

[-]
 From 462febcc7b5f6e54648f2fd941b9f90de16e54f1 Mon Sep 17 00:00:00 2001
 From: Michel Dänzer <michel.daenzer at amd.com>
Date: Mon, 14 Jul 2014 06:42:06 +0000
Subject: drm/radeon: Disable pflip interrupts

With Mario's previous fix, there are three possible scenarios for the
pflip interrupts:

1) If a pflip interrupt can occur before the corresponding vblank
    interrupt, the sequence number of the userspace event will be too
    small by one.
2) If a pflip interrupt can occur after the vblank interrupt and after
    the next flip is programmed to the hardware, 
radeon_crtc_handle_flip()
    will complete the next flip earlier than expected by userspace.
3) Otherwise, radeon_crtc_handle_flip() doesn't perform any actual work
    when called from the pflip interrupt handler, i.e. the pflip 
interrupt
    is useless.

In summary, Mario's fix made the pflip interrupts useless in the best
case and harmful in the worst case, so let's disable them.

Signed-off-by: Michel Dänzer <michel.daenzer at amd.com>
Signed-off-by: Alex Deucher <alexander.deucher at amd.com>
[-]

Thanks,
   Dieter


More information about the dri-devel mailing list