| Age | Commit message (Collapse) | Author |
|---|
|
Two warnings and an error.
|
|
These registers are cleared after suspend/resume, so set them
up again in EnterVT.
|
|
The tiled surface registers were only setup in SAVAGEDRIScreenInit,
however, after suspend/resume these registers must be set up again.
So factor it out in a new function so it can be reused.
|
|
Removes the last bunch of -Wall build warnings.
savage_dri.c:490: warning: format ‘%08lx’ expects type ‘long unsigned int’,
but argument 5 has type ‘drm_handle_t’
...
|
|
It looks like "!" was meant for logic negation and the parentheses
were wrong.
I don't have a Savage 4 so I can't check if this fixes or breaks
anything.
|
|
Silence -Wall warnings.
The functions are just #if'ed out, to be left for inspiration.
|
|
savage_dga.c: In function ‘Savage_OpenFramebuffer’:
savage_dga.c:393: warning: cast to pointer from integer of different size
savage_dri.c: In function ‘SAVAGEDRIScreenInit’:
savage_dri.c:924: warning: cast to pointer from integer of different size
|
|
sed -i 's/xalloc/malloc/g; s/xcalloc/calloc/g; s/xfree/free/g'
Also replace xrealloc with realloc.
|
|
savage_exa.c: In function ‘SavageUploadToScreen’:
savage_exa.c:545: warning: passing argument 1 of ‘memcpy’ discards qualifiers from pointer target type
/usr/include/bits/string3.h:49: note: expected ‘void * restrict’ but argument is of type ‘volatile CARD32 *’
savage_video.c: In function ‘SavagePutImage’:
savage_video.c:2007: warning: format ‘%ld’ expects type ‘long int’, but argument 4 has type ‘drmSize’
savage_video.c:2007: warning: format ‘%ld’ expects type ‘long int’, but argument 5 has type ‘int’
|
|
Fixes bug 31614.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
|
|
Using common defaults will reduce errors and maintenance.
Only the very small or inexistent custom section need periodic maintenance
when the structure of the component changes. Do not edit defaults.
|
|
|
|
|
|
|
|
DPMS header was split into dpms.h (client) and dpmsconst.h (server). Drivers
need to include dpmsconst.h if xextproto 7.1 is available.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
|
|
|
|
The DRI code attempts to use XAAGetCopyROP without checking whether XAA or EXA is in effect. This results in the server crashing with an undefined-symbol error when enabling EXA, then starting glxgears under GNOME/Metacity and attempting to drag the glxgears window.
The EXA code happens to have a functional duplicate of the required code under the name SavageGetCopyROP(). This patch moves the code to savage_accel.c where it is now shared between EXA and the DRI code.
It is disturbing that this has not been caught before.
|
|
acceleration.
Under some conditions (documented in this patch), the UploadToScreen implementation can make use of the AGP scratch buffer used for XVideo as a convenient source for Mastered Image Transfer. The previous implementation is still available as a fallback for when AGP acceleration is not possible. This requires the AGP scratch buffer to be persistent, so the XVideo code is also made aware of an existing DRM mapping for the scratch buffer.
|
|
|
|
SavageGetPanelInfo will reset DisplayType if the panel is not
active, so check again before calling SavageAddPanelMode so that
bogus modes are not added.
This fixes startup failures in some cases, for instance when there is no
real panel and EDID is missing. The bogus sync ranges (typically 0 Hz)
would also block the default sync ranges from being applied.
Thanks to Peter Belew for testing.
https://bugs.launchpad.net/bugs/294899
|
|
use one memcpy per scanline instead of a conditional inside
a loop for every dword
|
|
Instructs the BCI Xv pixel formatter to use AGP memory as a scratch buffer
for Xv image upload
|
|
allocations
BCI-mediated planar conversion can now use arbitrary offset in framebuffer as planar
buffer, no longer restricted to space past packed buffer.
|
|
patch from Alex Villacís Lasso with minor updates from me
|
|
|
|
git+ssh://git.freedesktop.org/git/xorg/driver/xf86-video-savage
|
|
Red Hat bug #243589.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
arbitrary downscale
|
|
|
|
|
|
|
|
the end of the allocated video buffer.
|
|
for these (for now).
Small optimizations (use memcpy instead of copy loop).
|
|
Since some time, savage enforces video-mode and max. pixel clock
validation through EDID, which is great for autoconfiguration, but
prevents users from defining higher screen resolutions, even though the
monitor is capable of it. This patch implements the IgnoreEDID option so
that the EDID monitor information can be explictly disabled.
|
|
In the modified lines, the logical-not binds first to psav->videoFlags,
not to (psav->videoFlags & VF_STREAMS_ON). If more flags are ever
implemented, they could interfere with this check. More of a preventive
patch than anything else.
|
|
|
|
|
|
|
|
should fix bug 12260
|
|
|
|
|
|
|
|
|
