| Age | Commit message (Collapse) | Author |
|---|
|
I suspect this might break TFP in some way but it makes firecube run here
|
|
1. the data type of <src1> (JMPI offset) must be D
2. execution size must be 1
3. NoMask
4. instruction compression isn't allowed.
|
|
Also just noticed that demos/copypix walks around the overlapping blit rules.
Bad, bad Mesa. :3
|
|
This makes demos/copypix better-looking. Horizontal dimensions are right now.
|
|
Note that trace debugging is temporarily gone. I'll rework it later.
|
|
Bug #22000.
|
|
This improves the performance of my GLSL demo by 30%. It also fixes the
VS deadlock that ut2004 had, for reasons I can't explain. Bug #21330.
|
|
|
|
|
|
|
|
|
|
Remove the drm_api from the functions in the softpipe and
i915simple drivers. Create wrapper functions in the
backends instead.
|
|
If we can't fit all the VS outputs into the MRF, we need to overflow into
temporary GRF registers, then use some MOVs and a second brw_urb_WRITE()
instruction to place the overflow vertex results into the URB.
This is hit when a vertex/fragment shader pair has a large number of varying
variables (12 or more).
There's still something broken here, but it seems close...
|
|
|
|
|
|
|
|
|
|
|
|
Failure to set the obj->Pointer back to null tripped up the assertion.
Bug #22428.
|
|
Conflicts:
src/mesa/vbo/vbo_exec_draw.c
|
|
This was only present for the sake of GL_ARB_shadow_ambient which we
never implemented in Gallium. If we someday want GL_ARB_shadow_ambient
we can implement it in the state tracker by adding a MAD after the
relevant TEX instructions.
|
|
Signed-off-by: Arthur HUILLET <arthur.huillet@free.fr>
|
|
Swrast was missing a free for the culmination of driConcatConfigs.
Use free(), not _mesa_free() since we shouldn't be calling any Mesa
functions from the GLX code. driConcatConfigs() should probably use
regular malloc/free to be consistant but the Mesa functions just wrap
the libc functions anyway.
|
|
|
|
Call Init() from CompileShaderFile, was previously only called for the
Text version of this function.
|
|
As prescribed by ARB_map_buffer_range.
|
|
buffer_flush_mapped_range.
|
|
buffer_flush_mapped_range.
When a buffer was mapped for write and no explicit flush range was provided
the existing semantics were that the whole buffer would be flushed, mostly
for backwards compatability with non map-buffer-range aware code.
However if the buffer was mapped/unmapped with nothing really written --
something that often happens with the vbo -- we were unnecessarily assuming
that the whole buffer was written.
The new PIPE_BUFFER_USAGE_FLUSH_EXPLICIT flag (based from ARB_map_buffer_range
's GL_MAP_FLUSH_EXPLICIT_BIT flag) allows to clearly distinguish the
legacy usage from the nothing written usage.
|
|
|
|
|
|
|
|
|
|
Never crashed on x86, ptx is in the same stack slot in either case.
Thanks to Bob Gleitsmann for catching this.
|
|
|
|
There doesn't appear to be any driver impact for enabling this, and
tests/prog_parameter passes.
|
|
|
|
The gl_PointCoord attribute is currently expected to be in the fog coord
register's z/w components. This was never totally fleshed out though.
This is just some placeholder code.
|
|
All gallium drivers should be able to support mixed-size color/depth/stencil
buffers. If not, we'll need a new PIPE_CAP_ query.
|
|
|
|
|
|
This fixes a segfault seen with piglit's fdo20701 test.
|
|
This reverts commit de447afff26706e3bf8bdcd5cfb8b1daf49b4b21 but
puts the lock under DRI1-only.
From keithw:
> It's there because the DRI1 code doesn't actually achieve the mutexing
> which it looks as if it should. For multi-threaded applications it was
> always possible to get two threads inside locked regions -- I have no
> idea how, but it certainly was and presumably still is possible.
|
|
|
|
This would cause LOCK_HARDWARE to mutex all contexts in this process on
both DRI1 and DRI2. On DRI1, LOCK_HARDWARE already does it for all
processes on the system. On DRI2, LOCK_HARDWARE doesn't, but there shouldn't
be any state outside the context that needs any additional protection.
Notably, the bufmgr is protected by its own mutex and not
LOCK_HARDWARE.
This code was originally introduced with the i915tex code dump, so it's not
clear what it was there for.
|
|
|
|
|
|
|
|
Get most of these working with scons.
|
|
Unfortunately scons does not check if a tool exists before it invokes
its generate function.
|
|
gcc-4.2's optimizer has a strange bug where it looses code from inner
loops in certain situations. For example, if the appearently innocent
looking code below is compiled with gcc-4.2 -S -O1, the inner loop's
code is missing from the outputed assembly.
struct Size {
unsigned width;
};
struct Command {
unsigned length;
struct Size sizes[32];
};
extern void emit_command(void *command, unsigned length);
void
create_surface( struct Size size, unsigned faces, unsigned levels)
{
struct Command cmd;
unsigned face;
unsigned level;
cmd.length = faces*levels*sizeof(cmd.sizes[0]);
for(face = 0; face < faces; ++face) {
for(level = 0; level < levels; ++level) {
cmd.sizes[face*levels + level] = size;
// This should generate a shrl statement, but the whole for body
// disappears in gcc-4.2 -O1/-O2/-O3!
size.width >>= 1;
}
}
emit(&cmd, sizeof cmd.length + cmd.length);
}
Note that this is not specific to MinGW's gcc-4.2 crosscompiler (the
version typically found in debian/ubuntu's mingw32 packages). gcc-4.2 on
Linux also displays the same error. gcc-4.3 and above gets this
correctly though.
Updated MinGW debian packages with gcc-4.3 are available from
http://people.freedesktop.org/~jrfonseca/debian/pool/main/m/
|
