summaryrefslogtreecommitdiffstats
path: root/src/gallium/auxiliary/tgsi
Commit message (Collapse)AuthorAgeFilesLines
* gallium: in drivers, make copy of tokens passed to pipe->create_vs/fs_state()Brian Paul2008-05-172-1/+45
| | | | The caller can then free the token array immediately.
* gallium: remove ^M (CR) charsBrian Paul2008-05-021-23/+23
|
* tgsi: Enable fast high precision rsqrt.Michal Krol2008-05-021-22/+0
|
* tgsi: Implement fast rsqrtf. Not tested, inactive.Keith Whitwell2008-05-023-19/+40
|
* tgsi: Do not assume IN and OUT registers are declared sequentially.Michal Krol2008-05-021-8/+7
|
* tgsi: Fix build on Win32.Michal Krol2008-05-021-2/+2
|
* tgsi: Dump destination register modulate modifier.Michal Krol2008-05-021-0/+25
|
* gallium: temporarily disable broken SSE2 code for ARL opcodeBrian2008-05-011-0/+5
|
* gallium: implement TGSI_OPCODE_EXPBrian2008-05-011-14/+15
|
* gallium: implement TGSI_OPCODE_LOGBrian2008-05-011-14/+16
|
* tgsi: remove some bogus win vs. linux crudKeith Whitwell2008-05-011-72/+46
| | | | | Pass arguments properly in linux now. Still need to change this to use a single calling convention on both platforms.
* rtasm: fix labels after (not so) recent change to allow dynamic fn growthKeith Whitwell2008-05-011-2/+2
| | | | | Using char * for labels doesn't work if you realloc the function during assembly and free the old storage...
* tgsi: restore HIGH_PRECISION settingKeith Whitwell2008-05-011-1/+1
|
* tgsi: add some const qualifiers to immediate pointersKeith Whitwell2008-05-011-2/+2
|
* tgsi: use EBX everywhere, be sure to push/pop itKeith Whitwell2008-05-011-8/+10
|
* tgsi: use x86_fn_arg instead of get_argument() -- it knows about push/pops ↵Keith Whitwell2008-05-011-29/+34
| | | | to the stack
* tgsi: use ESI instead of EBX on non-win32 platformsKeith Whitwell2008-05-011-7/+7
|
* tgsi: make loop structure clearer, use x86_lea for incrementsKeith Whitwell2008-04-281-66/+68
|
* tsgi: add a makefileKeith Whitwell2008-04-281-0/+2
|
* gallium: Generate SSE code to swizzle and unswizzle vs inputs and outputs.Michal Krol2008-04-282-4/+142
| | | | Change SSE_SWIZZLES #define to 0 to disable it.
* gallium: fix broken SGT, SLEBrian2008-04-251-2/+16
|
* use PIPE_ARCH_X86Brian2008-04-251-2/+2
|
* tgsi: fix compile when HIGH_PRECISION not definedKeith Whitwell2008-04-251-1/+1
|
* gallium: Include dependent header.José Fonseca2008-04-221-0/+2
|
* tgsi: use new float math funcs, drop local disassembly codeKeith Whitwell2008-04-212-501/+178
|
* gallium: a few commentsBrian Paul2008-04-181-0/+5
|
* gallium: implement recip sqrt() with C code for now.Brian Paul2008-04-181-0/+30
| | | | Some conformance lighting tests fail with the SSE rsqrt instruction.
* tgsi: add const qualifier to tokens on sse emitKeith Whitwell2008-04-182-2/+2
|
* tsgi: make Consts constKeith Whitwell2008-04-171-1/+1
|
* gallium: Eliminate stdio file usage. Remove unused stuff.José Fonseca2008-04-152-310/+34
|
* gallium: Don't assume snprintf are always available.José Fonseca2008-04-151-4/+5
|
* tgsi: Fix source register short dump code.Michal Krol2008-04-121-30/+32
|
* gallium: merge the tgsi_emit_sse2() and tgsi_emit_sse2_fs() functions.Brian Paul2008-04-112-130/+78
| | | | | The two functions were mostly the same. We can look at the shader header info to determine if it's a vertex or fragment shader.
* gallium: handle TGSI immediates in SSE code for vertex shadersBrian Paul2008-04-112-5/+22
|
* gallium: implement immediates (aka literals) for SSE fragment shadersBrian Paul2008-04-112-5/+81
|
* gallium: commentsBrian Paul2008-04-111-0/+31
|
* gallium: fix SCS codegen (sin scalar src comes from X, not Y)Brian Paul2008-04-111-1/+1
|
* tgsi: Dump processor type and version as a single token.Michal Krol2008-04-111-1/+0
|
* tgsi: Dump semantics before interpolator.Michal Krol2008-04-111-7/+9
|
* gallium: added tgsi_num_tokens() function to return number of tokens in ↵Brian2008-03-242-0/+17
| | | | | | token array. Maybe move to a different file someday.
* tgsi: Use debug_printf().Michal Krol2008-03-141-2/+2
|
* gallium: print warning rather than assert(0) for LOG/EXP opcodesBrian Paul2008-03-141-2/+32
| | | | Glean vertProg1 runs all the way through, rather than aborting.
* tgsi: add debug_printf version of tgsi_dumpKeith Whitwell2008-03-142-0/+26
|
* tgsi: Drop pre-ps_2_0 opcodes.Michal Krol2008-03-131-40/+6
|
* tgsi: replace erroneous use of FETCH with emit_tempfKeith Whitwell2008-03-131-3/+15
|
* tgsi: bump MAX_SRC_REGS to 4, for TXDKeith Whitwell2008-03-131-1/+1
|
* tgsi: Remove ExtDivide field from existence. Implement OPCODE_TXP.Michal Krol2008-03-124-46/+15
|
* tgsi: Dump TXP opcode.Michal Krol2008-03-121-2/+4
|
* tgsi: Dump source register divide component.Michal Krol2008-03-121-0/+5
|
* gallium: Fix MSVC warnings.José Fonseca2008-02-281-4/+4
|