| Age | Commit message (Collapse) | Author |
|---|
|
|
|
Revert "radeon: Default to 1x again with non-v3 AGP cards."
This reverts commit b653e5a628bfa4dfb168e96f93f41eb910f409fb.
|
|
git+ssh://agd5f@git.freedesktop.org/git/xorg/driver/xf86-video-ati into atombios-support
merge master and fix conflicts
|
|
Seems more reliable in general than what was set up by firmware - fingers
crossed...
|
|
git+ssh://agd5f@git.freedesktop.org/git/xorg/driver/xf86-video-ati into atombios-support
merge and fix conflicts
|
|
should fix bug 13610
|
|
|
|
(DE)ALLOCATE_LOCAL are gone.
|
|
bit is wrong on at least X700 cards with rialto pcie-agp bridge chip,
should be safe to use just the bit from the bridge hopefully to make
agp setup work on these cards and not adversely affect others.
|
|
Surprisingly easy, thanks to George's pci-rework changes.
|
|
Exclude the DRI window(s) though to avoid scribbling over 3D rendering.
|
|
|
|
|
|
Conflicts:
src/radeon.h
src/radeon_bios.c
src/radeon_display.c
src/radeon_dri.c
src/radeon_driver.c
src/radeon_modes.c
src/radeon_probe.h
src/radeon_video.c
|
|
Actively warn the user in the log about the effects of AGPFastWrite and
sanitise AGPFastWrite handling while we're here.
|
|
|
|
|
|
|
|
Also disable vbl when 3D isn't running
|
|
|
|
It can be enabled at runtime by increasing the log verbosity level.
Also change the prefix from (**) to (II) to make grepping the log file for
defaults overridden by xorg.conf more useful again.
Turn some MC related debugging output into normal informational output as it's
useful for recognizing corner cases that can cause stability issues.
|
|
We don't automatically require bumped minor versions.
|
|
|
|
This requires a drm > 1.26 to work
|
|
My 8500 in i845 doesn't work with fastwrites even setup by the firmware.
|
|
Based on the assumption that firmware should have set up the card and host
bridge appropriately for these settings, this may actually be safer, at least
for the transfer rate; leaving fast writes enabled is hopefully safe as well,
it certainly is on my sytem.
See https://bugs.freedesktop.org/show_bug.cgi?id=9284 .
|
|
Instead of calling the DRM CP idle ioctl, just emit the cache flush commands
into the CP stream.
|
|
In theory the driver should be able to handle the front buffer not at VRAM 0
In practice it didn't.. this is cleanup for at least XAA parts of the driver
to allow for the frontbuffer to move. It has to re-organise a large part of ScreenInit so things happen in the correct order otherwise some things get setup in-correctly. (not sure EXA with fb not at 0 works yet...)
|
|
We still have to force-sync the pages on enabling page flipping with XAA as the
second page may have been clobbered by the offscreen pixmap cache since they
were last synced.
|
|
This way we don't ever have to forcedly synchronize the pages so we hopefully
never clobber one page's contents when we shouldn't.
|
|
Also move the DamageEmpty() call into RADEONDRIRefreshArea().
|
|
Also use the damage layer directly instead of via shadowfb and blit dirty
rectangles to the second page in LeaveServer in order to try and improve the
tradeoff between performance and correctness.
|
|
|
|
Rather than publish a non-existant driver name for RN50, just don't bother
initializing it in the first place.
|
|
OTOH, flush everything when the previous engine mode is unknown, and mark the
engine mode as unknown in a couple more cases.
|
|
|
|
|
|
Also make the formatting of the if statements consistent.
|
|
This allows getting higher depth buffer precision and stencil hardware
acceleration in depth 16 or saving some video RAM at the cost of these features
in depth 24.
|
|
The radeon driver builds warning-free with gcc -Wall here now.
|
|
|
|
|
|
Sauerbeck).
|
|
various problems up to system hangs on X server startup have been
reported with that and DRI enabled.
|
|
switching to console with 3D clients running.
|
|
is double-buffered by default, which is desirable in a number of
situations, in particular when running a GLX compositing manager with
AIGLX.
|
|
aperture size. Otherwise it tries to overmap things if they are up near
the end of memory.
|
|
suspend/resume work a lot better on PCIE cards.
|
|
symbols directly. The xf86* versions aren't supposed to be used
directly.
Drop libc wrapper; don't include xf86_ansic.h and add includes now missing.
|
|
right for requesting drm versions, replace the use of some numbers with
the respective macro defines in radeon_video.c, and add some more macro
defines. None of that really matters.
|
