Since texture images are now stored in miptrees, we cannot usually
access them directly via the Data pointer.
So we wrap Mesa's implementation by map/unmap calls.
This crash was triggered by Sauerbraten, Piglit now contains a
regression test.
Signed-off-by: Nicolai Haehnle <nhaehnle@gmail.com>
The mip tree creation would crash if the first baselevel image to be uploaded
was not the positive-x image.
Found with Sauerbraten, also added a regression test to Piglit.
Signed-off-by: Nicolai Haehnle <nhaehnle@gmail.com>
R3xx/R5xx fragment program texture constants must come from a hardware
register instead of the constant file, so we redirect if necessary during
the native rewrite phase.
The symptoms of this bug started appearing when the Mesa fixed function
texenvprogram code started using STATE_CURRENT_ATTRIB constants for
texture coordinates when the corresponding attributes were constant across
all vertices.
Signed-off-by: Nicolai Haehnle <nhaehnle@gmail.com>
While running conform with render-to-texture:
conform -d 33 -v 2 -t -direct
the i965 driver failed this assertion:
intel_clear.c:77: intel_clear_tris: Assertion `(mask & ~((1 << BUFFER_BACK_LEFT) | (1 << BUFFER_FRONT_LEFT) | (1 << BUFFER_DEPTH) | (1 << BUFFER_STENCIL))) == 0' failed.
The problem is that intel_clear_tris() is called by intelClear() to
clear any and all of the available color buffers, but intel_clear_tris()
actually only handles the back left and front left color buffers; so
the assertion fails as soon as you try to clear a non-standard color
buffer.
The fix is to have intelClear() only call intel_clear_tris() with
buffers that intel_clear_tris() can support. intelClear() already backs
down to _swrast_Clear() for all buffers that aren't handled explicitly.
Omit math/m_xform.c from gallium builds since it's not used and it's the
one place we were pulling in the Mesa x86 codegen which collides with
gallium's x86 codegen.
Can now omit ASM_C_SOURCES from gallium build too.
Only VBO uses the evaluator code so call _math_init_eval() there.
Only TNL uses the transform/translate code so call _math_init_transformation()
and _math_init_translate9) there.
This is a step toward resolving some symbol collisions between Mesa's and
gallium's x86 codegen.
Have VBO and TNL modules call _math_init_transformation()
Commit 90b2beb661 changed the assembly
variables. Without this change, the glapi assembly wasn't being built
when it was supposed to, resulting in missing symbols in libGL.
Signed-off-by: Dan Nicholson <dbn.lists@gmail.com>
This is not the proper approach to disabling assembly. If there are bugs,
they should be fixed or the assembly can default to off.
Signed-off-by: Dan Nicholson <dbn.lists@gmail.com>
With the pixman CFLAGS at the beginning of the command line, my build
was picking up the glcore.h from glproto, which doesn't have any of the
TFP fields in __GLcontextmodes.
Signed-off-by: Dan Nicholson <dbn.lists@gmail.com>
