[Bug 91831] New: Corrupted video output with 32bit kernel
bugzilla-daemon at freedesktop.org
bugzilla-daemon at freedesktop.org
Mon Aug 31 16:39:40 PDT 2015
https://bugs.freedesktop.org/show_bug.cgi?id=91831
Bug ID: 91831
Summary: Corrupted video output with 32bit kernel
Product: xorg
Version: git
Hardware: x86 (IA32)
OS: Linux (All)
Status: NEW
Severity: normal
Priority: medium
Component: Driver/AMDgpu
Assignee: xorg-driver-ati at lists.x.org
Reporter: nholland at tisys.org
QA Contact: xorg-team at lists.x.org
I'm not entirely sure if this is the right place to mention this issue, but I'm
fairly confident that the right people will see it here:
I'm using a Kabini GPU on a 32-bit (kernel + userland) system using linux 4.2
and have just decided to give the amdgpu kernel drm driver and
xf86-video-amdgpu a try (previously I was using radeon). What I noticed is that
screen output in X11 is garbled using these drivers - i.e. pixels are messed up
/ missing / corrupt in various places all over the screen.
I've seen this issue before in the radeon driver, there's actually a (resolved)
bug report about it at https://bugs.freedesktop.org/show_bug.cgi?id=84627.
Back then, the bug was fixed in the kernel in
linux/drivers/gpu/drm/radeon/radeon_object.c. Here's the relevant section (from
radeon_bo_create()):
#ifdef CONFIG_X86_32
/* XXX: Write-combined CPU mappings of GTT seem broken on 32-bit
* See https://bugs.freedesktop.org/show_bug.cgi?id=84627
*/
bo->flags &= ~RADEON_GEM_GTT_WC;
I'm not into this stuff enough, but I believe that in contrast to the radeon
driver, the amdgpu driver does not disable write-combined CPU mappings on 32
bit kernels. The question is: Would it be possible and reasonable to change
amdgpu to follow radeon in this regard and thus - hopefully - fix the garbled
screen problem on 32 bit?
As a side note, the problem goes away when disabling hardware acceleration:
Option "NoAccel" "true"
Screen output looks good then. Still, this is probably not the best solution.
--
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.x.org/archives/xorg-driver-ati/attachments/20150831/35914c69/attachment.html>
More information about the xorg-driver-ati
mailing list