| Age | Commit message (Collapse) | Author |
|---|
|
|
|
If lod or bias can be non-constant across a quad of fragments,
we need to execute TEX separately for each value.
Don't ask why.
|
|
(Warsow creates a 1x1 front buffer)"
This reverts commit ec7844537ecdb0b598447e37bf0b7120acd029f3.
|
|
creates a 1x1 front buffer)
|
|
... and don't set the 'live' flag for TEX anymore, we'd
have to know if results affect the inputs for another TEX,
and I'm not going to do that kind of analysis now.
|
|
|
|
Unfortunately it seems that if depth comparison is
active and we read a 2D texture, i.e. provide only
2 inputs, the second is used for comparison ...
|
|
|
|
Keeping this dynamically allocated for texture arrays.
Since we don't use it to store zslice offsets anymore
it's either 1 or 6 integers (cube) ...
|
|
So we can use the 'mod' member without concern
if a source is used multiple times in 1 insn.
|
|
|
|
Conflicts:
src/gallium/state_trackers/xorg/xorg_xv.c
src/mesa/drivers/dri/intel/intel_span.c
|
|
Thanks for Keith to finding this.
Fixes Jedi Knight 2 menus.
|
|
Since we're reusing buffers we're effectively transforming all
of them into dynamic buffers.
It would be nice to not cache long lived static buffers. But there
is no way to detect the long lived from short lived ones yet. A
good heuristic would be buffer size.
|
|
The texture owns the surface handle. All derivatives need to keep
a reference to texture.
This fixes several assertions failures starting up Jedi Knight 2.
Should cause no change for DRM surface sharing -- reference count still
done as before there.
|
|
|
|
Generating mipmaps finally works, among other things. Yay!
|
|
|
|
Unfortunately we can't fix this easily in the R300 fragment shader,
and it's probably not worth the effort.
|
|
|
|
This reverts commit eb926ddf9eee1095c7fc12013f0b8375bbaeca6f.
|
|
We really just need to know whether the format is compressed or not.
For more detailed information format colorspace should suffice.
|
|
|
|
|
|
|
|
|
|
In particular, gcc man page warns that
union a_union {
int i;
double d;
};
int f() {
double d = 3.0;
return ((union a_union *) &d)->i;
}
"might" not be ok (why not?), even though it doesn't seem to generate
any warnings. Hence don't use this and do the extra step to actually use
assignment to get the values in/out of the union.
This changes parts of 3456f9149b3009fcfce80054759d05883d3c4ee5.
|
|
Depricate pf_type(), pf_size_*(), pf_layout() and pf_exp2().
Map depricated PIPE_FORMAT_TYPE to new UTIL_FORMAT_ values:
UNKNOWN = TYPE_VOID
UNORM = TYPE_UNSIGNED + LAYOUT_ARITH
SNORM = TYPE_SIGNED + LAYOUT_ARITH
FIXED = TYPE_FIXED
FLOAT = TYPE_FLOAT
USCALED = TYPE_UNSIGNED + LAYOUT_ARRAY
SSCALED = TYPE_SIGNED + LAYOUT_ARRAY
SRGB = TYPE_COLORSPACE_SRGB
|
|
|
|
Conflicts:
src/mesa/drivers/dri/r600/r700_assembler.c
src/mesa/main/version.h
|
|
|
|
|
|
Conflicts:
progs/demos/projtex.c
progs/xdemos/glxinfo.c
src/mesa/main/version.h
To fix the confilicts in projtex.c and glxinfo.c I just took the code from
mesa_7_6_branch. The conflicts seem to have occured from cherry-picks from
mesa_7_7_branch to mesa_7_6_branch followed by commmits just to
mesa_7_6_branch.
|
|
use pointer to union instead of void pointer.
gcc complained a lot, depending what the pointer originally actually was.
Looks like it's in fact maybe legal to cast for instance uint pointers to
union pointers as long as union contains a uint type, hence use this with some
callers, other just use union util_color in the first place.
|
|
|
|
If *I* can't read it, there's a strong possibility others can't,
either.
|
|
|
|
Thanks to Bob Gleitsmann for the patch.
I'll clean this up in a better way later if noone else beats me to it.
|
|
Otherwise a CS is refused by kernel 2.6.31 (and maybe all later
versions, not sure).
|
|
|
|
Just make it cover the whole framebuffer in that case. Otherwise the kernel CS
checker may complain, e.g. running progs/demos/gearbox. That runs fast now
here, but doesn't look right yet.
|
|
|
|
The dump calls should be wrapped in #ifdef DEBUG .. #endif.
|
|
Helps verifying udis86 output.
|
|
Otherwise the terminal gets full of garbage.
|
|
Just enough boilerplate code to avoid segfaulting.
|
|
This shouldn't happen but it does by some misterious reason. Fail the
assertion but at least do not segfault on release builds.
|
|
|
|
|
|
Also, overlapping occlusion queries seems to work now.
|
