aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
...
* glx/dri2: Paper over errors in DRI2Connect when indirectChristopher James Halse Rogers2011-08-051-0/+9
| | | | | | | | | | | | DRI2 will throw BadRequest for this when the client is not local, but DRI2 is an implementation detail and not something callers should have to know about. Silently swallow errors in this case, and just propagate the failure through DRI2Connect's return code. Note: This is a candidate for the stable release branches. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=28125 Signed-off-by: Christopher James Halse Rogers <[email protected]>
* ir_to_mesa: Try to avoid emitting a MOV_SAT to saturate an expression tree.Eric Anholt2011-08-051-4/+24
| | | | | | Fixes a regression in codegen quality for ff_fragment_shader conversion to GLSL -- glean texCombine produces 7.5% fewer Mesa IR instructions.
* prog_optimize: Add support for saturates to _mesa_merge_mov_into_inst.Eric Anholt2011-08-051-3/+5
| | | | | This fixes the remaining regression from ff_fragment_shader in Mesa IR instruction count, to now being a 1.9% win overall.
* i965: Add dumping for gen6 WM constants too.Eric Anholt2011-08-053-1/+22
| | | | This looks just like the VS dump for now.
* mesa: Remove dead "MemPool" field of gl_shader_state.Eric Anholt2011-08-051-2/+0
|
* i965/fs: Don't upload unused uniform components.Eric Anholt2011-08-052-4/+95
| | | | | | | | | This saves both register space and upload bandwidth for unused values. Note that previously we were relying on the visitor not initially generating references to different sets of uniforms between the 8-wide and 16-wide code generation, and now we're relying on them dead-code eliminating the same stuff, too.
* i965/fs: Don't allocate the old backend's compile structs for our compile.Eric Anholt2011-08-051-4/+7
| | | | This saves some 35MB when the program only uses GLSL shaders.
* winsys/radeon: do the CS cleanup in the CS ioctl threadMarek Olšák2011-08-051-2/+4
|
* winsys/radeon: fix space checkingMarek Olšák2011-08-054-5/+36
| | | | | | | We should remove the relocations which caused a validation failure from the list, so that the kernel receives only the validated ones. NOTE: This is a candidate for the 7.11 branch.
* st/dri: remove an unused-but-set variableMarek Olšák2011-08-051-7/+0
|
* st/dri: remove a dummy function dri2_create_contextMarek Olšák2011-08-051-15/+1
| | | | It does nothing besides calling dri_create_context with the same parameters.
* st/mesa: remove unused-but-set variables in st_glsl_to_tgsi.cppMarek Olšák2011-08-051-3/+2
|
* r300g: handle new CAPsMarek Olšák2011-08-051-9/+4
|
* r300g: adapt to the resource_resolve interface changeMarek Olšák2011-08-051-21/+22
|
* vbo: do not call _mesa_max_buffer_index in debug buildsMarek Olšák2011-08-051-3/+1
| | | | | | | | | That code drops performance in Unigine Heaven and Tropics by a factor of 10. That's too crazy even for a debug build. NOTE: This is a candidate for the 7.11 branch. Reviewed-by: Brian Paul <[email protected]>
* glsl: empty declarations should be validChia-I Wu2011-08-052-12/+8
| | | | | | | | | | | | | | | Unlike C++, empty declarations such as float; should be valid. The spec is not explicit about this actually. Some apps that generate their shader sources may rely on this. This was noted when porting one of them to Linux from Windows. Reviewed-by: Chad Versace <[email protected]> Note: this is a candidate for the 7.11 branch.
* mesa: Remove MSVC stdint typedefs from compiler.h.Vinson Lee2011-08-041-23/+1
| | | | MSVC can now include the stdint.h at include/c99/stdint.h.
* mesa: pass correct constant type to _mesa_fetch_state()Brian Paul2011-08-041-1/+1
| | | | Fixes assorted warnings about float vs. gl_constant_value pointers.
* mesa: use gl_constant_value type in ARB program parserBrian Paul2011-08-042-29/+30
|
* mesa: use gl_constant_value type in _mesa_[Get]ProgramNamedParameter4fNV()Brian Paul2011-08-041-10/+10
|
* mesa: add st_glsl_to_tgsi.cpp to SconscriptBrian Paul2011-08-041-0/+1
|
* mesa: fix out of bounds array access in rtgc debug codeBrian Paul2011-08-041-1/+1
| | | | | Fixes https://bugs.freedesktop.org/show_bug.cgi?id=39841 This would only be hit if someone set RGTC_DEBUG=1.
* mesa: don't use K&R style function pointer callsBrian Paul2011-08-041-3/+3
|
* Merge branch 'glsl-to-tgsi'Bryan Cain2011-08-0443-323/+5610
|\ | | | | | | | | | | Conflicts: src/mesa/state_tracker/st_atom_pixeltransfer.c src/mesa/state_tracker/st_program.c
| * st/mesa: replace duplicated create_color_map_texture() function with shared ↵Bryan Cain2011-08-044-43/+26
| | | | | | | | function
| * glsl_to_tgsi: remove debugging printfBryan Cain2011-08-041-1/+0
| |
| * glsl_to_tgsi: copy reladdr in st_src_reg(st_dst_reg) constructorBryan Cain2011-08-011-1/+1
| | | | | | | | This is a glsl_to_tgsi port of commit f7cd9a858c04.
| * glsl_to_tgsi: add each relative address to the previousBryan Cain2011-08-011-0/+12
| | | | | | | | This is a glsl_to_tgsi port of commit d6e1a8f71437.
| * glsl_to_tgsi: lower all ir_quadop_vector expressionsBryan Cain2011-08-011-1/+1
| | | | | | | | Unlike Mesa IR, TGSI doesn't have a SWZ opcode.
| * glsl_to_tgsi: rework immediate tracking to not use gl_program_parameter_listBryan Cain2011-08-011-47/+88
| |
| * glsl_to_tgsi: update commentsBryan Cain2011-08-011-12/+11
| |
| * glsl_to_tgsi: make coding style more consistentBryan Cain2011-08-011-137/+126
| |
| * glsl_to_tgsi: make assignment hack saferBryan Cain2011-08-011-3/+4
| | | | | | | | Fixes an assertion failure in piglit test glsl-texcoord-array.
| * glsl_to_tgsi: separate immediates from array constants during IR translationBryan Cain2011-08-011-48/+95
| | | | | | | | | | | | | | | | Before, if any uniform or constant array was accessed with indirect addressing, st_translate_program() would emit uniform constants in the place of immediates. This behavior was unavoidable with ir_to_mesa/mesa_to_tgsi, but glsl_to_tgsi can work around it since the GLSL IR backend and the TGSI emission are both inside the state tracker.
| * glsl_to_tgsi: fix mistakes in get_pixel_transfer_visitor()Bryan Cain2011-08-011-3/+3
| | | | | | | | I noticed these issues while working on get_bitmap_visitor().
| * st/mesa, glsl_to_tgsi: support glBitmap with a GLSL fragment shader activeBryan Cain2011-08-013-6/+105
| |
| * st/mesa, glsl_to_tgsi: support glDrawPixels/glCopyPixels with a GLSL ↵Bryan Cain2011-08-013-0/+194
| | | | | | | | | | | | | | fragment shader active Since this was previously implemented using Mesa IR and _mesa_combine_programs, this commit adds a new code path that works with glsl_to_tgsi.
| * glsl_to_tgsi: replace MAX_PROGRAM_TEMPS (256) with MAX_TEMPS (4096)Bryan Cain2011-08-011-6/+8
| |
| * r200, r600c, i965: fix buildBryan Cain2011-08-017-30/+30
| |
| * glsl_to_tgsi: always run copy_propagate() and eliminate_dead_code_advanced()Bryan Cain2011-08-011-5/+4
| | | | | | | | These two passes are written to handle indirect addressing properly.
| * glsl_to_tgsi: fix mistake in new dead code elimination passBryan Cain2011-08-011-4/+4
| | | | | | | | | | The conditions of IF opcodes were not being counted as reads, which sometimes led to the condition register being wrong or undefined.
| * glsl_to_tgsi: improve assignment handlingBryan Cain2011-08-011-3/+14
| | | | | | | | | | This is a hack, but it's better than emitting an unnecessary MOV instruction and hoping the optimization passes clean it up.
| * glsl_to_tgsi: improve eliminate_dead_code_advanced()Bryan Cain2011-08-011-0/+9
| |
| * glsl_to_tgsi: use the correct writemask in try_emit_mad() and try_emit_sat()Bryan Cain2011-08-011-2/+7
| |
| * glsl_to_tgsi: use a more specific condition for gl_FragDepth hack in ↵Bryan Cain2011-08-011-2/+2
| | | | | | | | | | | | | | generating assignments This reduces the number of instructions in the fragment shader of glsl-fs-atan-2 from 174 to 146 with EmitNoIfs enabled.
| * glsl_to_tgsi: add a better, more advanced dead code elimination passBryan Cain2011-08-011-0/+140
| |
| * glsl_to_tgsi: remove handling of XPD opcode in compile_tgsi_instruction()Bryan Cain2011-08-011-8/+0
| | | | | | | | | | The opcode is never emitted by the glsl_to_tgsi_visitor, so its special case in compile_tgsi_instruction() was dead code.
| * glsl_to_tgsi: use swizzle_for_size for src reg in conditional movesBryan Cain2011-08-011-2/+5
| | | | | | | | | | This prevents the copy propagation pass from being confused by undefined channels and thus missing optimization opportunities.
| * glsl_to_tgsi: execute merge_registers() after eliminate_dead_code()Emil Velikov2011-08-011-1/+1
| | | | | | | | | | | | | | | | | | Fixes a regression unintentionally introduced by "glsl_to_tgsi: fix shaders with indirect addressing of temps" that caused missing leaves in 3dmark01 test 4 (Nature) and missing/displaced textures on human models in Counter-Strike: Source. Signed-off-by: Emil Velikov <[email protected]> Signed-off-by: Bryan Cain <[email protected]>
| * glsl_to_tgsi: implement simplify_cmp pass needed by r300gBryan Cain2011-08-011-0/+95
| |