Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | cell: Generate blend / depth test code when state atom is created | Ian Romanick | 2008-03-21 | 1 | -21/+9 |
| | | | | | Code generation should be performed when the device-specific state atom is created, not when it is bound. | ||||
* | cell: Change code-gen for CONST_COLOR blend factor | Ian Romanick | 2008-03-21 | 5 | -49/+53 |
| | | | | | | | | | | | | Previously the constant color blend factor was compiled into the generated code. This meant that the code had to be regenerated each time the constant color was changed. This doesn't fit with the model used in Gallium. As-is, the code could be better. The constant color is loaded for every quad processed, even if it is not used. Also, if a lot of (1-x) blend factors are used, 1.0 will be loaded and reloaded into registers many times. | ||||
* | Tabs to spaces | Ian Romanick | 2008-03-21 | 1 | -27/+27 |
| | |||||
* | cell: Remove unnecessary default_blend work-around | Ian Romanick | 2008-03-21 | 1 | -21/+0 |
| | | | | | I suspect that there was some other bug in the blend code-gen that made this work-around necessary. | ||||
* | cell: Fix several bugs in blend code-gen | Ian Romanick | 2008-03-20 | 1 | -20/+92 |
| | | | | | | | | | | - Alpha factor set to a _COLOR mode was mishandled - Cases when either dst factor or src factor was ZERO were mishandled - MIN and MAX cases were backwards - Case when blend was disabled was mishandled - Incorrect comments about number of instructions generated The tests blendminmax and blendsquare run correctly. | ||||
* | cell: Call the correct function to generate blending code | Ian Romanick | 2008-03-20 | 1 | -1/+1 |
| | | | | Cut-and-paste for the lose. :( | ||||
* | cell: Fix bus error when there is no depth buffer | Ian Romanick | 2008-03-20 | 1 | -0/+3 |
| | |||||
* | cell: Use code-gen for alpha blend | Ian Romanick | 2008-03-20 | 9 | -132/+129 |
| | | | | So far this is only tested when GL_BLEND is disabled. | ||||
* | cell: use pipe_framebuffer_state.width, height | Brian | 2008-03-20 | 2 | -13/+4 |
| | |||||
* | gallium: add face, dirtyLevels params to pipe->texture_update() | Brian Paul | 2008-03-19 | 1 | -1/+2 |
| | | | | | This provides better information about which images in texture object have changed. Also, call texture_update() from more places previously missed. | ||||
* | cell: Fix various stencil test code-gen bugs | Ian Romanick | 2008-03-18 | 1 | -27/+43 |
| | |||||
* | cell: Correctly load stencil for PIPE_FORMAT_S8Z24_UNORM | Ian Romanick | 2008-03-18 | 1 | -3/+2 |
| | |||||
* | cell: Fix depth read / write for s8z24. | Ian Romanick | 2008-03-18 | 2 | -2/+24 |
| | | | | Stencil is still broken. | ||||
* | cell: Fix added whitespace errors. | Ian Romanick | 2008-03-18 | 1 | -2/+2 |
| | |||||
* | cell: Minor changes to make stencil not crash | Ian Romanick | 2008-03-18 | 2 | -5/+11 |
| | | | | | | I'm not sure these are quite correct. The reflect demo doesn't assert anymore, but it doesn't produce correct results either. SPE-based vertex shader code needs to be disabled for relfect to run. | ||||
* | cell: Move and (conditionally) silence debug code | Ian Romanick | 2008-03-18 | 2 | -10/+39 |
| | |||||
* | cell: Don't segfault when unbinding alpha / stencil / depth test state | Ian Romanick | 2008-03-17 | 3 | -13/+29 |
| | |||||
* | cell: Initial code-gen for alpha / stencil / depth testing | Ian Romanick | 2008-03-17 | 15 | -170/+1409 |
| | | | | | | | | Alpha test is currently broken because all per-fragment testing occurs before alpha is calculated. Stencil test is currently broken because the Z-clear code asserts if there is a stencil buffer. | ||||
* | cell: Fix to work with commit fa9e7e9a8debb68611909ac2ffab527c6c39a3e5 | Ian Romanick | 2008-03-13 | 3 | -32/+15 |
| | |||||
* | Replicate TXP changes in the SPU version of TGSI exec | Ian Romanick | 2008-03-13 | 1 | -31/+14 |
| | | | | | Replicate changes from commit ba75e82b6ebaf88dd2e4a8f764b2d296d715bf8a in spu_exec.c | ||||
* | cell: check for NULL shader pointer in cell_bind_vs_state() | Brian | 2008-03-11 | 1 | -1/+2 |
| | |||||
* | Move SPE register allocator to rtasm code | Ian Romanick | 2008-03-10 | 1 | -72/+29 |
| | | | | | Move the register allocator to a common location. There is more code on the way that will make use of this interface. | ||||
* | cell: sync up with sampler/texture state-setting changes | Brian | 2008-03-10 | 3 | -12/+28 |
| | |||||
* | cell: convert all points/lines to tris for the time being | Brian | 2008-02-28 | 1 | -0/+4 |
| | | | | Allows more programs to look correct. We'll want native points/lines someday. | ||||
* | cell: remove obsolete texture field | Brian | 2008-02-28 | 1 | -1/+0 |
| | |||||
* | cell: fix minor get_tex_surface() breakage | Brian | 2008-02-27 | 1 | -1/+2 |
| | |||||
* | cell: Use unified data cache for textures too | Ian Romanick | 2008-02-27 | 3 | -117/+72 |
| | |||||
* | gallium: remove pipe_context->texture_create/release/get_tex_surface() | Brian | 2008-02-27 | 1 | -30/+0 |
| | | | | These functions are now per-screen, not per-context. | ||||
* | gallium: move is_format_supported() to pipe_screen struct | Brian | 2008-02-27 | 2 | -23/+19 |
| | |||||
* | gallium: start removing pipe_context->get_name/vendor/param/paramf | Brian | 2008-02-27 | 1 | -29/+0 |
| | | | | | These are now per-screen functions, not per-context. State tracker updated, code in drivers and p_context.h disabled. | ||||
* | Cell: implement pipe_screen for cell driver | Brian | 2008-02-27 | 7 | -63/+249 |
| | |||||
* | cell: insert a (disabled) call to spe_cpu_info_get() | Brian | 2008-02-26 | 1 | -0/+3 |
| | | | | Found on the Cell devel forum, but doesn't appear to be available in SDK 2.1. | ||||
* | gallium: remove pipe parameter from pipe_texture_reference() | Brian | 2008-02-26 | 1 | -2/+1 |
| | | | | | Added pipe field to pipe_texture (temporary, see comments). First step toward context-less texture creation... | ||||
* | cell: Multiple rendering contexts don't work yet | Ian Romanick | 2008-02-26 | 1 | -0/+10 |
| | | | | | Log a message and forcibly exit. This prevents silly fools from thinking there's a bug...instead of just an unimplemented feature. :) | ||||
* | cell: Additional changes to match changes in draw/draw_vertex_shader.c | Ian Romanick | 2008-02-25 | 1 | -1/+2 |
| | |||||
* | cell: Fix off-by-one error in spu_dcache_fetch_unaligned | Ian Romanick | 2008-02-25 | 1 | -18/+32 |
| | | | | | This time the off-by-one error caused an extra qword to be fetched under certain circumstances when the source ea was not qword aligned. | ||||
* | cell: Trivial compiler warning clean-ups. | Ian Romanick | 2008-02-25 | 4 | -5/+8 |
| | |||||
* | Cell: Remove unnecessary include files | Ian Romanick | 2008-02-25 | 2 | -4/+0 |
| | |||||
* | Cell: Use multiple DMA tags for the dcache. | Ian Romanick | 2008-02-25 | 2 | -2/+5 |
| | |||||
* | cell: added function prototypes to silence warnings | Brian | 2008-02-22 | 2 | -0/+6 |
| | |||||
* | cell: fix build: s/dest/vertex/ | Brian | 2008-02-22 | 1 | -1/+1 |
| | |||||
* | Cell: Remove erroneous ALIGN16_ATTRIB attributes | Ian Romanick | 2008-02-21 | 1 | -3/+3 |
| | | | | | | | If a structure is marked as being aligned the SPE compiler performs extra optimizations (sadly, only -O2 is used) when reading the structure. Since most of the structures sent in batch buffers are only 8-byte aligned, this resulted in mysterous bugs with -O2. | ||||
* | Cell: Add spu_dcache.c to Makefile. | Ian Romanick | 2008-02-21 | 1 | -0/+1 |
| | | | | This was erroneously missing in previous commits. | ||||
* | Cell: Initial scalar implementation of spu_dcache_mark_dirty | Ian Romanick | 2008-02-21 | 1 | -6/+16 |
| | |||||
* | Cell: Fix off-by-one error in spu_dcache_fetch_unaligned | Ian Romanick | 2008-02-21 | 2 | -6/+6 |
| | | | | | An off-by-one error caused an extra qword to be fetched under certain alignment / size combinations. | ||||
* | Cell: Initial pass at unified data cache | Ian Romanick | 2008-02-21 | 8 | -95/+194 |
| | |||||
* | cell: use pipe_texture_reference() | Brian | 2008-02-20 | 1 | -1/+3 |
| | |||||
* | gallium: replace some ordinary assignments with pipe_reference_texture() | Brian | 2008-02-20 | 1 | -0/+1 |
| | | | | This fixes at least one instance of dereferencing an invalid texture pointer. | ||||
* | cell: rename cell_state_fs.c -> cell_state_shader.c | Brian | 2008-02-20 | 2 | -1/+1 |
| | |||||
* | cell: init shader-related functions in cell_init_shader_functions() | Brian | 2008-02-20 | 4 | -36/+55 |
| |