Age | Commit message (Collapse) | Author |
|
Add a GLbitfield64 type and several macros to operate on 64-bit
fields. The OutputsWritten field of gl_program is changed to use that
type. This results in a fair amount of fallout in drivers that use
programs.
No changes are strictly necessary at this point as all bits used are
below the 32-bit boundary. Fairly soon several bits will be added for
clip distances written by a vertex shader. This will cause several
bits used for varyings to be pushed above the 32-bit boundary. This
will affect any drivers that support GLSL.
At this point, only the i965 driver has been modified to support this
eventuality.
I did this as a "squash" merge. There were several places through the
outputswritten64 branch where things were broken. I foresee this
causing difficulties later for bisecting. The history is still
available in the branch.
Conflicts:
src/mesa/drivers/dri/i965/brw_wm.h
|
|
processing"
This reverts commit f058b25881e08c9d89a33345e5c84e1357396932.
This change is completely wrong in so many ways. When clip distances
are generated as part of vertex processing, they must be interpolated
to perform clipping. Geometric clipping goes right out the window.
|
|
Once the clipping distance is calculated and stored per vertex, the
distances can be re-used when clipping is actually performed. This
doesn't have any immediate benefit, but it paves the way for
implementing gl_ClipDistance in vertex shaders and result.clip[] in
vertex programs.
This has not produces any oglconform regressions on my G31 system
which uses software TNL.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
Fixed gallium build breakage.
|
|
This currently doesn't include fixing up the cliptests in the assembly
paths to support ARB_depth_clamp, so enabling depth_clamp forces the C path.
|
|
|
|
|
|
Use MAX_VERTEX_GENERIC_ATTRIBS instead. No need for two #defines for
the same quantity.
|
|
The code's cleaner and a step toward supporting float-valued texture sampling.
Some optimizations for common cases can be added and re-enabled...
|
|
|
|
t_vb_program.c
|
|
|
|
|
|
do the same math as for fixed function pipe, including
user clip planes.
(mostly resurrected from the dead t_vb_arbprogram.c code)
|
|
of -I flags.
|
|
|
|
This is a bit of a hack for now because the tnl module is using the swrast
module to fetch texels. The texture fetch/filter code should probably be
moved into the main/ module since it doesn't really depend upon other
swrast code.
|
|
use VertexProgram._Current instead of VertexProgram.Current in a few more places.
Only fixup fogc and psiz in case this is really a nv program (others are fine
if undefined), and fix this case up so the values actually get written.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
handle branch instructions, etc.
|
|
|
|
|
|
|
|
Among changes:
Remove ctx->FragmentProgram._Active
Remove _UseTexEnvProgram
Move _MaintainTnlProgram, _MaintainTexEnvProgram, _TexEnvProgram and
_TnlProgram fields.
Remove/disable old GLSL interpreter code.
|
|
Massage nvvertexec.c code to work more like s_nvfragprog.c - another step
toward unifying vertex/fragment program execution.
|
|
|
|
"gl_" to match other structs.
|
|
just aliases for members of the VB->AttribPtr[] array.
Begin replacing FogCoordPtr with VB->AttribPtr[_TNL_ATTRIB_FOG], and similarly
for NormalPtr, TexCoordPtr, PointSizePtr, etc.
|
|
|
|
in various places.
Note that ctx->Texture.CurrentUnit needs to be tested against Coord/Image
limits when referenced, not just in glActiveTexture().
|
|
only includes the 6 frustum bits, not the user-clip plane bit, nor the vertex
cull bit.
|
|
- generate error on NULL pointers in glShaderSourceARB;
- reinstall program object, if current, in glLinkProgramARB;
- vertex and fragment shaders are optional in program object;
- floor asm was wrongly computed for x86 back-end;
- allow for (void) idiom in function prototypes;
- all fixed-state uniforms are updated;
- local variable initializers are working;
- implement texture* and shadow* functions for vertex processor;
- generate error if too many arguments in general constructor;
- trim unused data in general constructor;
- struct r-value field select was badly relocated;
Changes:
- add derived state gl_fog_attrib::_Scale;
- add derived state gl_light::_CosCutoffNeg;
|
|
If enabled, all other stages, except render, are disabled.
|
|
Instructions, InputsRead, OutputsWritten, and Parameters.
Also, added debug functions: _mesa_print_instruction(),
_mesa_print_program_parameters() and revamp _mesa_print_program().
|
|
Based on patch by Ian (#4967) but also unify instruction opcodes.
|
|
vertex_program struct.
Allow switching between regular and vertex_program implementations
of fixed function TNL with the MESA_TNL_PROG environment var
(previously this required recompilation).
Ensure program compilation only references program data, not the
wider context. This means that compiled programs only need to be
invalidated when the program string changes, not on other state
changes.
|
|
- remove input/output fields, input tracking removed.
- remove state fields, the validate function now called
on every statechange.
- add an explicit 'create' function.
Add in code to build vertex program to implement current t&l state. Still
disabled, but turn on with a #define in t_vp_build.h.
|
|
|
|
|
|
program is enabled AND the currently bound program is valid.
Check _Enable instead of Enable to prevent things from blowing up
when someone calls glEnable(GL_VERTEX_PROGRAM_ARB) without actually
defining a program.
|
|
Some need to be set per-vertex, other per-primitive. Cleared that up.
Only need to init temp/result registers if executing an NV vertex program.
|
|
- cope with input vectors with size less than that of the emitted
attribute.
- cope with vertices with 'holes' inside and between vertices.
Fix calculation of tnl->render_inputs to work with fp programs.
Mirror VB->PointSizePtr in VB->AttribPtr so that it can work with t_vertex.c.
Transition swrast_setup/ to use t_vertex.c to build swrast vertices.
|
|
|
|
|