intel: Mark the FBO as incomplete if there's no intel_renderbuffer for it.
This happens to rendering with textures with a border, which had resulted
in a segfault on dereferencing the irb.
(cherry-picked from commit 8bba183b9e)
gallium debug: Add gcc printf hint to debug_printf
This causes gcc to issue warnings when format parameters do not match up
with the format string in calls to debug_printf.
Signed-off-by: Nicolai Hähnle <nhaehnle@gmail.com>
Some tests, eg small SubData are probably overwhelmed by the cost of
performing the draw after each upload. Add a varient which does a lot
of subdata uploads and then a single draw.
Also try to avoid cache-artifacts in the upload timings.
- MUL_LIT is ALU.Trans instruction
- some Trans instructions can take 3 arguments
- don't clobber dst.x, use dst.z as temp, it'll get written correct
value in last insn
- respect source swizzles
r300: Fix handling of NV_vertex_program parameters
The handling is a bit inefficient, unfortunately, but I don't want to make
any intrusive changes for Mesa 7.6.
Signed-off-by: Nicolai Hähnle <nhaehnle@gmail.com>
Since commit 2921a2555d ('intel: Deassociated
drawables from private context struct in intelUnbindContext'),
intel->driDrawable may be NULL in intel_flush().
radeon: Improve WARN_ONCE macro to appear as single statement.
Do-while makes macro safe to be used with if and for constructions.
Also remove __LINE__ macro from variable name because scope is local to macro anyway.
scons: Drop gprof support for profile builds; tweak optimization flags instead.
gprof is useful for shared libraries, hence our drivers. Nevertheless
profilers like oprofile can benefit from disabling some relatively
minor optimizations for more accurate / complete results.
mesa/st: Create front renderbuffer on the fly when supplied with a surface
Normally, the mesa/st would create a fake front buffer out of a
client-allocated surface.
In the DRI setting, however, st/dri provides a front buffer surface which is
created and maintained by the X server. Prefer to use this surface instead,
so that front buffer rendering and reading works correctly.
Signed-off-by: Nicolai Hähnle <nhaehnle@gmail.com>
r300/compiler: Fix R300 fragment program regression introduced by 0723cd1...
We obviously need to move the code addr register backwards because their may
be overlap.
This bug affected in particular the Compiz water plugin.
Signed-off-by: Nicolai Hähnle <nhaehnle@gmail.com>