Age | Commit message (Collapse) | Author |
|
Using radeon_platform_probe function breaks the Zaphod mode because
it attempts to call xf86AddEntityToScreen multiple times, but nobody
calls xf86SetEntityShared prior to that. Consequently, calls for all
but first device instance fail.
Prior to introduction of platform bus, the logic was that the Probe
function would make the entity sharable, which would cause Xserver
to later make it shared prior to adding it to screen. With the
platform bus loading, add to screen happens in the probe
function so we have to make it shared there.
v2: do not make the entity shared if it was not previously marked
sharable. Should fix:
https://bugs.freedesktop.org/show_bug.cgi?id=56680
https://bugs.freedesktop.org/show_bug.cgi?id=56663
Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
|
|
This reverts commit 39eac3104c2f08b4d78aab3f88fd104301eb4711.
This breaks non-zaphod configurations with xserver 1.13.
See:
https://bugs.freedesktop.org/show_bug.cgi?id=56680
https://bugs.freedesktop.org/show_bug.cgi?id=56663
|
|
Using radeon_platform_probe function breaks the Zaphod mode because
it attempts to call xf86AddEntityToScreen multiple times, but nobody
calls xf86SetEntityShared prior to that. Consequently, calls for all
but first device instance fail.
Prior to introduction of platform bus, the logic was that the Probe
function would make the entity sharable, which would cause Xserver
to later make it shared prior to adding it to screen. With the
platform bus loading, add to screen happens in the probe
function so we have to make it shared there.
Signed-off-by: Ilija Hadzic <ihadzic@research.bell-labs.com>
|
|
Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
|
|
Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
|
|
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Okay kwin did things different, and I really hope this time this is the
correct transformation operation. It works here with gears under kwin
and under mutter and under xcompmgr.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
when you use MPX or slave usb devices, the damage posting was incorrect,
and resulted in misrendered boxes on the screen.
Fix it by reporting damage correctly.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
Signed-off-by: Adam Jackson <ajax@redhat.com>
|
|
Improves performance of non-pageflipped 3D apps on SI.
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
|
|
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
|
|
Fixes https://bugs.freedesktop.org/show_bug.cgi?id=54711
Reported-by: wolput@onsneteindhoven.nl
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
This ensures radeon loads on output less GPUs as a GPU driver.
There are server bugs that also need to be fixed.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54519
Reported-By: scaty@dcinformatique.com
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
Another silly one.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
Older servers won't have DRI2UpdatePrime, so fix build.
bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=54433
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
This hooks into EXA and the dirty tracking to add sharing and output
offload support.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
This adds the new dri2 interfaces required for radeon to be offloaded to.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
this just adds the interface and shared support to the pixmap.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
This allows the radeon driver to be loaded via the platform bus mechanism.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
in order for offload to work we need to cause a full
screen damage to be registered after pageflip.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
Copied from fbdev, makes it so we can run without iopl.
Signed-off-by: Adam Jackson <ajax@redhat.com>
|
|
The only VBE code here went away when we dropped UMS.
Signed-off-by: Adam Jackson <ajax@redhat.com>
|
|
Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
|
|
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
|
|
Defaults to shadowfb. 3D acceleration is available with glamor. 2D
acceleration is disabled until the radeonsi driver can handle glamor's
shaders.
v2: add chip flags (Alex Deucher)
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Enable at build time with --enable-glamor and runtime with
Option "AccelMethod" "glamor"
The most notable lack of functionality is XVideo. Use something like VDPAU for
now.
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
|
|
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
|
|
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
This merges the removal of the UMS code and subsequent simplifications.
Hell yes.
* origin/kms-only: (24 commits)
radeon: fix radeonchipsets properly
radeon: fix RADEONChipsets
radeon: migrate remainder of radeon_driver.c to rest of driver
radeon: avoid including large static struct twice.
radeon: drop some more unused macros/inlines
radeon: drop another unused struct member.
radeon: more unused stuff
radeon: drop cardType shouldn't matter to userspace.
radeon: drop radeonGetPixmapOffset
radeon: move more functions into texture video
radeon: drop legacy memory, merge into radeon video
radeon: drop unused txoffset vars.
radeon: drop vb_mc_addr, not needed anymore
radeon: drop offset member of r600 accel object
radeon: move macros into radeon.h
radeon: drop more wrapper macros.
radeon: further macro cleanups.
radeon: make exa copy funcs static.
radeon: drop FUNC_NAME macro and ONCE_ONLY stuff
radeon: drop radeon_commonfuncs.
...
Conflicts:
configure.ac
|
|
It was using the pitch (displayWidth) for the virtual width (virtualX). This
prevented using page flipping in some cases, as displayWidth was already
overaligned for virtualX, so the DRI2 front and back buffers ended up having
a different pitch.
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
|
|
just fallback to the old behaviour on older servers.
should fix: https://bugs.freedesktop.org/show_bug.cgi?id=51388
Review-over-irc: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
|
|
reported on irc by okias.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
This helps make a few more things static and the driver generally
smaller.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
Just extern this in the one file that needs access.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
These are no longer used anywhere.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
Remove all CurrentLayout stuff.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
no longer needed.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
make more thing static.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
This merges these two functions into textured video code.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
These vars aren't being used anymore.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
This field is totally unused now, so drop it.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
This is always 0 now.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
Drop separate header file, there isn't many separate macros now.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|