This is done when constructing the fbconfigs, and the result is saved
for window system framebuffer creation.
Note: For dri2 the server needs to have an identical format selection
logic. Otherwise the dri state-tracker and the xorg driver (state-tracker)
will disagree on which format to use for the attachments. Some more work
is needed in this area.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
The default values true and false will expand to "1" and "0" when
gcc -std=c99, causing bool option defaults to generate runtime failures.
One solution is to specify bool option defaults quoted as "true" and "false".
Add a macro to assist this.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
We were packing according to the pitch, while the hardware appears to base
it on the base level width.
With this and the previous commit, fbo-cubemap now matches untiled behavior.
3D rendering to tiled textures was being done with non-tile-aligned offsets.
The G4X hardware has fields to let us support it easily and correctly, while
the pre-G4X hardware requires a path full of suffering, so we just fall back.
We flush the command buffer so we don't emit mixed
state (with new and previous buffer size) command
buffer, this is especialy affecting zbuffer states.
vap index offset is programmed to 0 by the kernel, it
would add work to kernel checker to allow userspace
programming of this so it's now disallowed with CS
on KMS.
Two parts to this:
One we don't keep pointers to possibly freed memory anymore once we unbind the
drawables from the context. Brian I need to figure out what the comment
you made there, can we get a glean/piglit test so we can fix it properly?
If the new gc is the same as the oldGC, we call the unbind even though
we just bound it in that function. doh.
glsl compiler will not generate OPCODE_SWZ, and as a first step it would
be translated away to a MOV anyway (why?), but later internally this opcode is
generated (for EXT_texture_swizzling).
Reverts part of commit d7ea9ddf58.
We were calling _mesa_reference_buffer_object() on some heap-allocated
memory that was uninitialized and could trigger an assertion.
We can actually go back to "looser" ref counting of the Null/default
buffer object in these cases.
We were failing to deal with:
- vsnprintf returns negative value on error.
- vsnprintf returns the number of chars that *would* have been
written on truncation.
Appart from separating r3xx/r5xx fragment shaders, a more consistent
naming scheme has been applied. From now on:
r300 = all chips
r3xx = R300/R400 only
r5xx = R500 only
This way r300_fragment_shader is the master struct, and the structs
r3xx_fragment_shader and r5xx_fragment_shader inherits it.
still always enable max, but the right values this time.
More work should probably be done for saner limits without mm, and/or
dri conf option allow_large_textures (which is ignored) removed.
3D limit on r100 is pretty arbitrary as still handled by swrast anyway.
Also fix r300 limits (except 3d I've no idea what the max is anyway so
keep using mesa default).
This driver does no transformation of the gallium calls
going to the real driver, like the identity matrix. It is
intended to be the basis for transforming and/or debug
drivers like trace and rbug.
Authors of this patch are:
Michal Krol, orignal heavy lifting.
José Fonesca, object wrapping code stolen from trace.
Jakob Bornecrantz, put it all toghether and renamed a stuff.
