summaryrefslogtreecommitdiff
path: root/src
AgeCommit message (Collapse)Author
2010-07-02ir_to_mesa: Fix up implementation of ir_unop_exp.Eric Anholt
OPCODE_EXP is not to ir_unop_exp what OPCODE_EX2 is to ir_unop_exp2. It's the weird VP approximation helper opcode. Just implement it with OPCODE_POW instead. Fixes glsl-fs-exp.
2010-07-02i965: Add support for the DP2 opcode, which we use for dot(vec2, vec2).Eric Anholt
The original glsl compiler would generate a.x * b.x + a.y * b.y, which we would do mul+mul+add for instead of this mul+mac. Fixes glsl-fs-dot-vec2.
2010-07-02ir_to_mesa: Add support for scalar * mat, vec * mat.Eric Anholt
This is not tested by piglit currently.
2010-07-02ir_to_mesa: Add support for shadow comparison to texture instructions.Eric Anholt
piglit lacks tests for this currently.
2010-07-02ir_to_mesa: Move projection handling out of ir_tex so txb and txl get it.Eric Anholt
Fixes: glsl-fs-texture2dproj-bias glsl-fs-texture2dproj-bias-2
2010-07-02glsl2: Print the linking info log in the stand-alone compilerIan Romanick
2010-07-02ir_to_mesa: Add support for projected non-shadow/bias/lod texturing.Eric Anholt
Fixes: glsl-fs-texture2dproj glsl-fs-texture2dproj-2
2010-07-02ir_to_mesa: Fix sparse swizzling of src regs when a writemask is present.Eric Anholt
Fixes glsl-fs-texture2d-masked.
2010-07-02ir_to_mesa: Don't forget to run the Mesa IR optimization passes.Eric Anholt
With how we generate assignments, the trivial copy propagation in it is really important, and some drivers will really want the register allocation, too.
2010-07-02glsl2: Add a pass to break ir_binop_div to _mul and _rcp.Eric Anholt
This results in constant folding of a constant divisor.
2010-07-01glsl2: Conditionally allow optional extensions to be enabledIan Romanick
The only optional extension currently supported by the compiler is GL_EXT_texture_array.
2010-07-01glsl2: Append _TOK to some parser tokensIan Romanick
This prevents conflicts with defines elsewhere in Mesa and allows including mtypes.h in the compiler.
2010-07-01glsl2: Conditionally define preprocessor tokens for optional extensionsIan Romanick
The only optional extension currently supported by the compiler is GL_EXT_texture_array.
2010-07-01glsl2: Define preprocessor tokens for extensionsIan Romanick
Currently only GL_ARB_draw_buffers and GL_ARB_texture_rectangle are defined because those extensions are always enabled. This make tex_rect-03.frag pass.
2010-07-01glsl2: Support AST-to-IR translation of invariant keywordIan Romanick
2010-07-01glsl2: Initialize ast_declarator_list::invariant in constructorIan Romanick
2010-07-01glsl2: Don't bounds check unsize array redeclarationsIan Romanick
This along with several previous commits fix test CorrectUnsizedArray.frag.
2010-07-01glsl2: Add gl_MaxTextureCoordsIan Romanick
2010-07-01glsl2: Default delcaration of gl_TexCoord is unsizedIan Romanick
2010-07-01glsl2: Change order of semaintic checks on variable declarationsIan Romanick
This will make it easier to support more (valid) kinds of redeclarations.
2010-07-01ir_to_mesa: Fill in remaining ops, remove default case for expression types.Eric Anholt
We should now have support for all the expression types we need for GLSL 1.20.
2010-07-01ir_to_mesa: Add support for ir_unop_rcp.Eric Anholt
This isn't used at the moment, but will be soon.
2010-07-01glsl2: Update README for what I've been thinking about with expr types work.Eric Anholt
2010-07-01glsl2: Add a pass to convert mod(a, b) to b * fract(a/b).Eric Anholt
This is used by the Mesa IR backend to implement mod, fixing glsl-fs-mod.
2010-07-01glsl2: Remove dead member from dead code visitor.Eric Anholt
2010-07-01glsl2: Add ir_unop_fract as an expression type.Eric Anholt
Most backends will prefer seeing this to seeing (a - floor(a)), so represent it explicitly.
2010-06-30ir_to_mesa: Add support for discard instructions.Eric Anholt
Fixes glsl-fs-discard-01.
2010-06-30glsl2: Don't break sign() down by vector components.Eric Anholt
2010-06-30ir_to_mesa: Add support for ir_unop_sign.Eric Anholt
Fixes glsl-fs-sign, glsl-vs-sign.
2010-06-30i965: Add support for OPCODE_SSG.Eric Anholt
The old compiler didn't use SSG, and instead emitted SGT/SGT/SUB. We can do a little better for SSG than we do for the SGT series.
2010-06-30ir_to_mesa: Add missing no-op type conversions.Eric Anholt
Fixes glsl-fs-step.
2010-06-30glsl2: Fix reversed value of step().Eric Anholt
It's 0.0 if x < edge, not 1.0. Partial fix for glsl-fs-step.
2010-06-30ir_to_mesa: Note which of our expr ops are unsupported 1.30 features.Eric Anholt
2010-06-30glsl2: Fix up the implementation of fract() for vector types.Eric Anholt
There's no need to split each vector component out, just do vector ops.
2010-06-30ir_to_mesa: Initialize the (we never use it) abs field of Mesa src regs.Eric Anholt
2010-06-30ir_to_mesa: Send the negate field on to Mesa IR.Eric Anholt
Fixes glsl-fs-neg.
2010-06-30ir_to_mesa: When generating a swizzle, respect the reg's current swizzle.Eric Anholt
Fixes depth-tex-modes-glsl.
2010-06-30glsl2: Implement AST->HIR support for the "discard" instruction.Kenneth Graunke
2010-06-30glsl2: Define new ir_discard instruction.Kenneth Graunke
2010-06-30ir_to_mesa: Support gl_FragDepth.Eric Anholt
Fixes glsl-bug-22603.
2010-06-30glsl2: Allow a fragment shader to not write a color.Eric Anholt
I can't find any text justifying this check, and it caused a reasonable-looking shader in glsl-bug-22603 (which writes only gl_FragDepth) to fail.
2010-06-30mesa: Don't look in unallocatd param slots for parameter values.Eric Anholt
glsl-derivs would add 40.0, 0.0, and 1.0 in order. When we went looking for 0.0, we'd find it in the second slot of the param, and use it, but param->Size would still be 1. When we went to add 1.0 and didn't find it, we'd put allocate it to that second slot and the 0.0 would actualy end up being 1.0. Fixes glsl-derivs, glsl-deriv-varyings.
2010-06-30ir_to_mesa: Fix the indexing of attributes in the program's Attributes.Eric Anholt
This fixes GetAttribLocation returning VERT_ATTRIB_GENERIC1 instead of 1, caught by glsl-dlist-getattriblocation.
2010-06-30glsl2: Fix storing of dead memory in the symbol table.Kenneth Graunke
decl->identifier is part of the AST, so it doesn't live very long. Instead, add var->name which is owned by var.
2010-06-30glsl2: Preprocessed source doesn't need to live past compile time.Kenneth Graunke
2010-06-30glsl2: Create new talloc contexts the "right" way.Kenneth Graunke
2010-06-30glsl2: Use _mesa_glsl_parse_state as the talloc parent, not glsl_shader.Kenneth Graunke
_mesa_glsl_parse_state should be the parent for all temporary allocation done while compiling a shader. glsl_shader should only be used as the parent for the shader's final IR---the _result_ of compilation. Since many IR instructions may be added or discarded during optimization passes, IR should not ever be allocated to glsl_shader directly. Done via sed -i s/talloc_parent(state)/state/g and s/talloc_parent(st)/st/g. This also removes a ton of talloc_parent calls, which may help performance.
2010-06-30glsl2: Steal the live IR and free the rest of the junk.Kenneth Graunke
2010-06-30linker: Don't dynamically allocate slots for linked shadersIan Romanick
The can be at most one shader per stage. There are currently only two stages. There is zero reason to dynamically size this array.
2010-06-30ir_to_mesa: Tell Mesa about our choices for vertex attribute locations.Eric Anholt