summaryrefslogtreecommitdiffstats
path: root/src/mesa
Commit message (Collapse)AuthorAgeFilesLines
* ir_to_mesa: Don't do function inlining until linking.Eric Anholt2010-08-051-1/+0
| | | | | | | | | | | | Optimizations at compile time should generally be done with the goal of reducing instruction count so that other work, particularly linking, is less time-consuming if the shader is used multiple times. However, function inlining increases instruction count for the inlined function bodies without removing the original function body, since we don't know if it will be used at link time or not. Reduces the runtime of linking and executing a Yo Frankie fragment shader from 0.9 seconds to 0.5 seconds (-45.9%, +/- 2.2%, n=5).
* glsl2: Don't try to dump GLSL IR for a shader that didn't compile.Eric Anholt2010-08-051-3/+5
|
* ir_to_mesa: Remove debug force-enablement of EmitNoIfs.Eric Anholt2010-08-051-1/+1
|
* i965: Settle on printing our program debug to stdout.Eric Anholt2010-08-045-26/+34
| | | | | | Mixing stderr (_mesa_print_program, _mesa_print_instruction, _mesa_print_alu) with stdout means that when writing both to a file, there isn't a consistent ordering between the two.
* ir_to_mesa: Print shader source and compiled IR under MESA_GLSL=dump.Eric Anholt2010-08-041-1/+17
| | | | | | | While the Mesa IR dumping includes some corresponding GLSL IR for correlating Mesa IR to GLSL IR, it doesn't completely express it. This printing includes things like variable declarations and control flow structure that is hard to read otherwise.
* glsl2: Add ir_assignment::write_mask and associated methodsIan Romanick2010-08-041-42/+25
| | | | | | | | | | | | | | | | | | | | | | | Replace swizzles on the LHS with additional swizzles on the RHS and a write mask in the assignment instruction. As part of this add ir_assignment::set_lhs. Ideally we'd make ir_assignment::lhs private to prevent erroneous writes, but that would require a lot of code butchery at this point. Add ir_assignment constructor that takes an explicit write mask. This is required for ir_assignment::clone, but it can also be used in other places. Without this, ir_assignment clones lose their write masks, and incorrect IR is generated in optimization passes. Add ir_assignment::whole_variable_written method. This method gets the variable on the LHS if the whole variable is written or NULL otherwise. This is different from ir->lhs->whole_variable_referenced() because the latter has no knowledge of the write mask stored in the ir_assignment. Gut all code from ir_to_mesa that handled swizzles on the LHS of assignments. There is probably some other refactoring that could be done here, but that can be left for another day.
* ir_to_mesa: Clean up the mapping of samplers to Mesa's sampler uniforms.Eric Anholt2010-08-041-20/+22
| | | | | | Instead of using a linker-assigned location (since samplers don't actually take up uniform space, being a link-time choice), use the sampler's varaible pointer as a hash key.
* mesa: Don't null deref looking for Mesa IR code at compile time.Eric Anholt2010-08-041-5/+3
| | | | | | | The new compiler doesn't generate Mesa IR at compile time, and that compile time code previously wouldn't have reflected the link time code that actually got used. But do dump the info log of the compile regardless.
* glsl2: Skip talloc_parent in constant_expression of non-constant arrays.Eric Anholt2010-08-041-1/+1
|
* ir_to_mesa: Add support for 1.20 uniform initializers.Eric Anholt2010-08-021-30/+97
| | | | | | | | | Fixes: glsl-uniform-initializer-1 glsl-uniform-initializer-2 glsl-uniform-initializer-3 glsl-uniform-initializer-4 glsl1-GLSL 1.20 uniform array constructor
* Initialize a couple of HasIndex2 fields on Mesa IR src regs.Eric Anholt2010-08-022-0/+2
|
* ir_to_mesa: Support for struct uniforms.Eric Anholt2010-08-021-0/+78
| | | | Fixes glsl-uniform-struct.
* ir_to_mesa: Add a constructor for ir_to_mesa_src_reg.Eric Anholt2010-08-021-48/+34
| | | | | This helps makes sure we don't miss any new fields, and makes totally uninitialized src_regs be PROGRAM_UNDEFINED.
* Keep a local copy of the symbol name in the symbol tableIan Romanick2010-08-021-3/+4
| | | | | | | | | | | | | | | | | | | | | The symbol_header structure that tracks symbols with a particular name may have a different (longer) life time than the symbols it tracks. Not keeping a local copy of the name can lead to use-after-free errors. For example, the following sequence would trigger such an error: char *copy = strdup(name); _mesa_symbol_table_push_scope(st); _mesa_symbol_table_add_symbol(st, 0, name, NULL); _mesa_symbol_table_pop_scope(st); free(name); _mesa_symbol_table_find_symbol(st, 0, copy); With this change, the symbol table keeps a local copy of the name that has the same life time as the symbol_header for that name. This resolves some use-after-free errors with built-in functions in the GLSL compiler.
* ir_to_mesa: Add support for MESA_GLSL=log.Eric Anholt2010-08-011-0/+4
| | | | | This is the option that dumps shader source to files in the current directory.
* glsl2: Add new tree grafting optimization pass.Eric Anholt2010-07-311-0/+1
|
* ir_to_mesa: Add the function name as a comment to BGNSUB and ENDSUB.Eric Anholt2010-07-301-4/+18
|
* ir_to_mesa: Don't emit a duplicate return at the end of a function.Eric Anholt2010-07-291-2/+6
| | | | It was harmless, but ugly.
* ir_to_mesa: Respect the driver if it rejects a shader.Eric Anholt2010-07-282-9/+11
|
* glsl2: Add support for redeclaring layout of gl_FragCoord for ARB_fcc.Eric Anholt2010-07-281-1/+6
| | | | | Fixes: glsl-arb-fragment-coord-conventions
* ir_to_mesa: Add remaining state variable (builtin uniforms) support.Eric Anholt2010-07-281-0/+201
| | | | | | | | | | Fixes: glsl1-GL state variable reference (diffuse product) glsl1-GL state variable reference (gl_FrontMaterial.ambient) glsl1-GL state variable reference (gl_LightSource[0].diffuse) glsl1-GL state variable reference (point attenuation) glsl1-GL state variable reference (point size) glsl1-linear fog
* ir_to_mesa: Provide a restricted type size to _mesa_add_uniform.Eric Anholt2010-07-271-1/+10
| | | | | Fixes: glsl-uniform-out-of-bounds.
* ir_to_mesa: Add support for array constants.Eric Anholt2010-07-271-17/+31
| | | | | | | | | Fixes: glsl1-GLSL 1.20 array constructor 1 glsl1-GLSL 1.20 array constructor 2 glsl1-GLSL 1.20 array.length() glsl1-GLSL 1.20 const array constructor 1 glsl1-GLSL 1.20 const array constructor 2
* glsl2: Make the dead code handler make its own talloc context.Eric Anholt2010-07-271-1/+1
| | | | | This way, we don't need to pass in a parse state, and the context doesn't grow with the number of passes through optimization.
* mesa: Allow large temporary indices coming into the temporary reg allocator.Eric Anholt2010-07-271-21/+27
| | | | | This gets glsl-vs-raytrace, glsl-fs-raytrace running on the new compiler.
* ir_to_mesa: Set the swizzle on constant struct src regs.Eric Anholt2010-07-271-0/+1
| | | | MESA_GLSL=nopt now produces believable output for glsl-fs-raytrace.
* ir_to_mesa: Fix stray "break" that broke functions of >1 argument.Eric Anholt2010-07-271-1/+0
|
* glsl2: Add optimization pass for algebraic simplifications.Eric Anholt2010-07-271-0/+1
| | | | | | This cleans up the assembly output of almost all the non-logic tests glsl-algebraic-*. glsl-algebraic-pow-two needs love (basically, flattening to a temporary and squaring it).
* ir_to_mesa: Fix up handling of void function returns.Eric Anholt2010-07-261-1/+1
| | | | void functions have a type of glsl_type::void_type, not a null type.
* ir_to_mesa: Actually allocate the right size for constant matrix temps.Eric Anholt2010-07-261-1/+1
|
* ir_to_mesa: Add support for structure constants.Eric Anholt2010-07-261-5/+30
| | | | | Fixes: TPPStreamCompiler::assignOperands
* Merge remote branch 'origin/master' into glsl2Eric Anholt2010-07-26296-7800/+12406
|\ | | | | | | | | | | | | | | | | | | | | This pulls in multiple i965 driver fixes which will help ensure better testing coverage during development, and also gets past the conflicts of the src/mesa/shader -> src/mesa/program move. Conflicts: src/mesa/Makefile src/mesa/main/shaderapi.c src/mesa/main/shaderobj.h
| * i965: Fix reversed naming of the operations in compute-to-mrf optimization.Eric Anholt2010-07-263-6/+11
| | | | | | | | | | Also fix up comments, so that the difference between the two passes is clarified.
| * i965: Clean up a few magic numbers to use brw_defines.h defs.Eric Anholt2010-07-263-18/+20
| |
| * i965: Use MIN2, MAX2 instead of rolling our own.Eric Anholt2010-07-261-15/+12
| |
| * i965: Fold the "is arithmetic" bit of 965 opcodes into the opcode list.Eric Anholt2010-07-261-50/+26
| |
| * i965: Remove some duped register size/count definitionsEric Anholt2010-07-262-34/+26
| |
| * i965: Move the GRF-to-MRF optimizations to brw_optimize.c.Eric Anholt2010-07-263-619/+618
| |
| * i965: Improve (i.e. remove) some grf-to-mrf unnecessary movesBenjamin Segovia2010-07-261-2/+626
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Several routines directly analyze the grf-to-mrf moves from the Gen binary code. When it is possible, the mov is removed and the message register is directly written in the arithmetic instruction Also redundant mrf-to-grf moves are removed (frequently for example, when sampling many textures with the same uv) Code was tested with piglit, warsow and nexuiz on an Ironlake machine. No regression was found there Note that the optimizations are *deactivated* on Gen4 and Gen6 since I did test them properly yet. No reason there are bugs but who knows The optimizations are currently done in branch free programs *only*. Considering branches is more complicated and there are actually two paths: one for branch free programs and one for programs with branches Also some other optimizations should be done during the emission itself but considering that some code is shader between vertex shaders (AOS) and pixel shaders (SOA) and that we may have branches or not, it is pretty hard to both factorize the code and have one good set of strategies
| * i965: Allow VS MOVs to use immediate constants.Eric Anholt2010-07-261-0/+1
| | | | | | | | | | Clarifies program assembly, and with a little tweak to always use constant_map, we could cut down on constant buffer payload.
| * mesa: Fix Cygwin build with llvm enabled.Vinson Lee2010-07-231-1/+1
| | | | | | | | | | On Cygwin locale_t in not available but 'llvm-config --cppflags' adds the compiler flag -D_GNU_SOURCE to the build.
| * gallium: Fix build with llvm installed in non-standard locationChristopher James Halse Rogers2010-07-231-0/+7
| | | | | | | | | | | | | | | | | | | | The es1, es2 and gl state trackers include draw_pipe.h, which includes the llvm headers if MESA_LLVM is true, so we also need to add the llvm seachpaths. Similarly, gallivm and other gallium drivers need LLVM_CFLAGS to build when enabled. Also fix xorg drivers, they didn't include LDFLAGS.
| * st/mesa: get rid of unneeded ureg_writemask()Brian Paul2010-07-231-3/+2
| |
| * st/mesa: fix bug in emit_adjusted_wpos()Brian Paul2010-07-231-2/+5
| | | | | | | | | | | | | | | | | | If we bias x,y we still need to pass through z,w in case the shader reads gl_FragCoord.z or .w. Fixes fd.o bug 29183 (piglit glsl-bug-22603). NOTE: This is a candidate for the 7.8 branch.
| * i965: Cleanly fail programs with unsupported array access.Eric Anholt2010-07-231-1/+28
| | | | | | | | | | This should be more useful for developers and for bug triaging than just generating wrong code.
| * i965: Add support for VS relative addressing of temporary arrays.Eric Anholt2010-07-231-2/+49
| | | | | | | | Fixes glsl-vs-arrays. Bug #27388.
| * glsl: remove invalid _mesa_problem() callBrian Paul2010-07-221-0/+2
| | | | | | | | Fixes fd.o bug 29206.
| * i965: Respect VS/VP point size result when enabled.Eric Anholt2010-07-221-3/+4
| | | | | | | | Fixes glsl-vs-point-size.
| * i965: Fix the disasm output for da16 src widths.Eric Anholt2010-07-221-1/+1
| | | | | | | | | | | | This has confused me twice now. It's a fixed width of 4 (usually a region description of <4,4,1>), not 1. If it was 1, we'd have been skipping all over register space.
| * i965: Avoid extra MOV in VS indirect register reads.Eric Anholt2010-07-221-15/+16
| |
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-25 21:34:22 +0000