Age | Commit message (Collapse) | Author | |
---|---|---|---|
2007-12-10 | nouveau: fix elusive dma bug | Ben Skeggs | |
In some situations WAIT_RING would get called while the GPU was processing data from outside the "master" ring, which caused dma.free to be updated incorrectly and much fun was had. WAIT_RING will now wait until it reads GET values from within the main ring buffer before calculating free space. | |||
2007-12-10 | nouveau: add callbacks for signalled fences | Ben Skeggs | |
2007-12-10 | nouveau: fix build | Ben Skeggs | |
2007-12-10 | Merge branch 'upstream-gallium-0.1' into darktama-gallium-0.1 | Ben Skeggs | |
Conflicts: src/mesa/drivers/dri/intel_winsys/intel_winsys_pipe.c src/mesa/pipe/Makefile src/mesa/pipe/draw/draw_vertex_fetch.c src/mesa/pipe/i915simple/i915_texture.c src/mesa/pipe/softpipe/sp_texture.c src/mesa/pipe/xlib/xm_winsys.c src/mesa/state_tracker/st_cb_fbo.c | |||
2007-12-09 | Adapt for winsys interface changes. | José Fonseca | |
2007-12-10 | nouveau: give resources a start property | Ben Skeggs | |
2007-12-09 | Merge branch 'nouveau-gallium-0.1' into darktama-gallium-0.1 | Ben Skeggs | |
2007-12-09 | nouveau: downgrade hacked syncs to hacked flushes | Ben Skeggs | |
Unfortunately we still need to keep the forced flushes as mesa's vbo path is still insane. But, at least this is a little less brutal. The bo code will detect when it needs to do a sync. | |||
2007-12-09 | nouveau: relocs are per-pushbuf, not per-channel | Ben Skeggs | |
2007-12-09 | nouveau: use "indirect" push buffers | Ben Skeggs | |
2007-12-09 | nouveau: adapt to gallium interface changes | Ben Skeggs | |
2007-12-09 | Move _mesa_init_glsl_driver_functions() into shader_api.c | Brian | |
This allows making a bunch of functions static, and removes a state tracker dependency on driverfuncs.c | |||
2007-12-09 | Define PIPE_FORMAT_ tokens as an enum set, rather than #defines. | Brian | |
This makes debugging a _lot_ easier. In gdb, "print format" used to display 613570600, now you see PIPE_FORMAT_A8R8G8B8_UNORM. | |||
2007-12-09 | Get rid of "duplicate" formats. | Brian | |
For example, replace PIPE_FORMAT_U_A8_R8_G8_B8 with PIPE_FORMAT_A8R8G8B8_UNORM | |||
2007-12-09 | Eliminate struct pipe_region. | Michel Dänzer | |
Directly use struct pipe_buffer_handle for storage and struct pipe_surface for (un)mapping. | |||
2007-12-09 | Hide texture layout details from the state tracker. | Michel Dänzer | |
pipe->get_tex_surface() has to be used for access to texture image data. | |||
2007-12-09 | Move dimensions from struct pipe_region to struct pipe_surface. | Michel Dänzer | |
2007-12-07 | Move _mesa_init_glsl_driver_functions() into shader_api.c | Brian | |
This allows making a bunch of functions static, and removes a state tracker dependency on driverfuncs.c | |||
2007-12-07 | Define PIPE_FORMAT_ tokens as an enum set, rather than #defines. | Brian | |
This makes debugging a _lot_ easier. In gdb, "print format" used to display 613570600, now you see PIPE_FORMAT_A8R8G8B8_UNORM. | |||
2007-12-07 | Get rid of "duplicate" formats. | Brian | |
For example, replace PIPE_FORMAT_U_A8_R8_G8_B8 with PIPE_FORMAT_A8R8G8B8_UNORM | |||
2007-12-07 | Eliminate struct pipe_region. | Michel Dänzer | |
Directly use struct pipe_buffer_handle for storage and struct pipe_surface for (un)mapping. | |||
2007-12-07 | Use write posting in the kickoff function too. | Stephane Marchesin | |
2007-12-06 | Hide texture layout details from the state tracker. | Michel Dänzer | |
pipe->get_tex_surface() has to be used for access to texture image data. | |||
2007-12-05 | nv50: 3d driver skeleton | Ben Skeggs | |
2007-12-05 | nouveau: g8x winsys support, yay softpipe! | Ben Skeggs | |
2007-11-29 | Move dimensions from struct pipe_region to struct pipe_surface. | Michel Dänzer | |
2007-11-25 | Merge branch 'upstream-gallium-0.1' into darktama-gallium-0.1 | Ben Skeggs | |
2007-11-25 | nv40: point sprites | Ben Skeggs | |
2007-11-23 | nv40: support for keeping multiple vtxprogs on the hw at the same time. | Ben Skeggs | |
2007-11-22 | Remove fences from the i915simple winsys interface. | José Fonseca | |
Fences will be part of the pipe winsys interface, so remove this to avoid merge conflicts later on. This reverts commit ca7f68a7cf25a51f382bba8c42d8c6ab7db57b5d. This reverts commit dec60d33b2570cf2bdce72a00a1539ee93133f91. This reverts commit 90dd0cb822f2fe14258c786e5c37da69472b7d17. | |||
2007-11-21 | nouveau: wait for sync after m2mf also, to be thourough. | Ben Skeggs | |
These will all go away at some point I promise, want to rule out sync issues while I bash nv40 into shape. | |||
2007-11-21 | Merge branch 'upstream-gallium-0.1' into darktama-gallium-0.1 | Ben Skeggs | |
2007-11-20 | nouveau: temporary workaround for dodgy buffer code | Ben Skeggs | |
2007-11-20 | nouveau: update to latest header | Ben Skeggs | |
2007-11-18 | nouveau: m2mf fallback path for region copies. | Ben Skeggs | |
2007-11-18 | nouveau: oops, when'd that disappear.. | Ben Skeggs | |
2007-11-18 | nouveau: Very rough cut at gallium winsys + nv40 pipe driver. | Ben Skeggs | |
2007-11-17 | Fix typo | José Fonseca | |
2007-11-17 | Proper fence reference counting. | José Fonseca | |
2007-11-15 | Include fences in the i915simple winsys interface. | José Fonseca | |
2007-11-08 | check for fence==NULL in intel_batchbuffer_finish(), fixes glxinfo crash on exit | Brian | |
2007-11-08 | Rename struct field 'private' to 'priv'. | Michel Dänzer | |
This broke the LLVM build because 'private' is a C++ keyword. | |||
2007-11-07 | Remove winsys->wait_idle(), subsumed by pipe->flush(). | Brian | |
Connect intel_i915_batch_finish() into i915_winsys, just like intel_i915_batch_flush(). Call i915_winsys->batch_finish() in response to pipe->flush(PIPE_FLUSH_WAIT). Now all the batchbuffer/fence code is in one place and a little cleaner. | |||
2007-11-07 | New PIPE_FLUSH_WAIT flag for pipe->flush(). | Brian | |
The state tracker doesn't have to directly call winsys->wait_idle() anymore. glFlush and glFinish both go through pipe->flush() now. | |||
2007-11-07 | Remove context dependencies in winsys layer. | Brian | |
The winsys object is now per-screen and shared by multiple contexts. The regionPool is now part of the i915 winsys layer. The winsys wait_idle() and flush_frontbuffer() funcs will get more attention... | |||
2007-11-07 | cosmetic fix to driFenceFinish() call | Brian | |
2007-11-07 | replace magic number 3 with DRM_FENCE_TYPE_EXE | DRM_I915_FENCE_TYPE_RW in ↵ | Brian | |
driFenceFinish() call | |||
2007-11-07 | remove DriverCtx=intel assigment/hack | Brian | |
2007-11-07 | Pass winsys as a parameter to intel_create_i915simple(), intel_create_softpipe() | Brian | |
2007-11-07 | var renaming: s/sws/winsys/ | Brian | |