Age | Commit message (Collapse) | Author | |
---|---|---|---|
2007-01-06 | i965: Avoid branch instructions while in single program flow mode. | Eric Anholt | |
There is an errata for Broadwater that threads don't have the instruction/loop mask stacks initialized on thread spawn. In single program flow mode, those stacks are not writable, so we can't initialize them. However, they do get read during ELSE and ENDIF instructions. So, instead, replace branch instructions in single program flow mode with predicated jumps (ADD to the ip register), avoiding use of the more complicated branch instructions that may fail. This is also a minor optimization as no ENDIF equivalent is necessary. Signed-off-by: Keith Packard <keithp@neko.keithp.com> | |||
2007-01-06 | i965: Connect INTEL_DEBUG=sync up to cmd/batch ioctls. | Eric Anholt | |
Signed-off-by: Keith Packard <keithp@neko.keithp.com> | |||
2006-12-30 | Use the tiled flag in the sarea to determine region tiling. | Haihao Xiang | |
This fixes mis-rendering if back/depth fail to get set up as tiled. While it probably won't ever be the case now that the pitch limits are loosened, this is still the right thing to do. | |||
2006-12-26 | Support linear format in i965. | Haihao Xiang | |
Fix bug #117 #118 | |||
2006-12-14 | Share code to lay out >= 945 style 2D mipmaps between i915tex and i965 drivers. | Michel Dänzer | |
Use the i965 version as it has some fixes over the i915tex version. | |||
2006-12-13 | Merge git://proxy01.pd.intel.com:9419/git/mesa/mesa into crestline | Nian Wu | |
2006-12-13 | Fix bug #93, i965 driver not thread safe. | Zou Nan hai | |
I am not confident of it is 100% thread safe now. bufmgr_fake.c need a total rewrite later (cherry picked from 606632ca27558ee1335be2f4a5906f2baa240a6a commit) | |||
2006-12-13 | fix bug #99. | Zou Nan hai | |
prim_count overflow when there is more than 1 cliprect (cherry picked from 84b958d66fe7d3fe03ed12b493e3f3197f656531 commit) | |||
2006-12-13 | ARB_occlusion_query support | Zou Nan hai | |
2006-12-13 | if (tex width < 4), mipmap calculation will be out of range | Zou Nan hai | |
2006-12-10 | Avoid branch instructions while in single program flow mode. | Eric Anholt | |
There is an errata for Broadwater that threads don't have the instruction/loop mask stacks initialized on thread spawn. In single program flow mode, those stacks are not writable, so we can't initialize them. However, they do get read during ELSE and ENDIF instructions. So, instead, replace branch instructions in single program flow mode with predicated jumps (ADD to the ip register), avoiding use of the more complicated branch instructions that may fail. This is also a minor optimization as no ENDIF equivalent is necessary. | |||
2006-12-10 | Connect INTEL_DEBUG=sync up to cmd/batch ioctls. | Eric Anholt | |
2006-12-10 | adding pci id of Crestline | Wang Zhenyu | |
2006-12-09 | i965: Fix a crash with wine by not allocating >1MB on the stack. | Eric Anholt | |
2006-12-08 | fix bug#9237 | Xiang, Haihao | |
2006-12-08 | fix bug#9045 | Xiang, Haihao | |
2006-12-02 | Structure CopyPixels similarly to i915 do_texture_copypixels, to ease future ↵ | Gary Wong | |
unification. | |||
2006-11-29 | Add accelerated CopyPixels for non-overlapping, 1:1 blits. | Eric Anholt | |
Submitted by Gary Wong <gtw@gnu.org> | |||
2006-11-23 | Haihao Xiang's edgeflag patch. | Keith Whitwell | |
2006-11-23 | Haihao Xiang's pointsize initialization patch. | Keith Whitwell | |
2006-11-21 | Gary Wong's patches for CopyPixels Logiop (enable) and Blend | Keith Whitwell | |
(disallow). Slightly cleaned to disallow on all blend states for code consiseness and turn a table lookup into a function to match other code in the driver. | |||
2006-11-21 | Gary Wong's fix for 64 bit cleanness of vertex program inputs bitmask. | Keith Whitwell | |
2006-11-21 | Commit Gary Wong & Keith Packard's changes for bug 8867 -- adjust code | Keith Whitwell | |
after repositioning of INDEX value in BRW_ATTRIB enum. | |||
2006-11-20 | Don't perform dangling attribute check on POS attribute. It can't by | Keith Whitwell | |
definition dangle (every vertex has a position). However save->currentsz isn't properly maintained for this attribute, as there is no current position value to track. Reported by Haihao Xiang. | |||
2006-11-16 | Make sure vbo's are mapped before accessing their contents in | Keith Whitwell | |
api_arrayelt.c. Reported by Haihao Xiang. | |||
2006-11-15 | Add a size parameter to _mesa_add_unnamed_constant() and | Brian Paul | |
_mesa_add_named_constant() to indicate vector size (1, 2, 3 or 4). Always 4 for now... | |||
2006-11-02 | merge current trunk into vbo branch | Alan Hourihane | |
2006-11-01 | Remove x/y/width/height parameters from Clear functions. | Brian Paul | |
2006-10-31 | cleanup code, compiles with vbo changes | Keith Whitwell | |
2006-10-30 | switch several dri drivers over | Keith Whitwell | |
2006-10-30 | checkpoint - remove dead files, otherwise untested | Keith Whitwell | |
2006-10-23 | Emit cliprects in the userspace driver as required, rather than | Keith Whitwell | |
passing them to the kernel. This works because all drawing commands in the 965 driver are emitted with the lock held and the batchbuffer is always flushed prior to releasing the lock. This allows multiple cliprects to be dealt with, without replaying entire batchbuffers and redundantly re-emitting state. | |||
2006-10-15 | Remove a bunch of "ctx->Driver.function = _swrast_Function" lines since | Brian Paul | |
default/fallback functions are already plugged in by the call to _mesa_init_driver_functions(). | |||
2006-10-15 | Remove Driver.ResizeBuffers = _mesa_resize_framebuffer lines. | Brian Paul | |
This is already done by the preceeding call to _mesa_init_driver_functions() which plugs in default functions like that. | |||
2006-10-13 | Upload of interleaved arrays currently assumes that position is the | Keith Whitwell | |
first element in the interleaved group. Add a test to catch cases where this isn't true and use per-array uploads instead. Fixes compiz glitches on x64. | |||
2006-10-13 | Make sure instructions are zero'd out before use. | Keith Whitwell | |
2006-10-13 | Use unsigned long batchbuffer offset, fixes x64 warnings. | Keith Whitwell | |
2006-10-13 | Fix printf warning. | Keith Whitwell | |
2006-10-13 | Use unsigned long cast for checking pointer alignment, fixes x64 warnings. | Keith Whitwell | |
2006-10-13 | Use mesa function to get opcode name rather than an internal table. | Keith Whitwell | |
2006-10-05 | missing file | Keith Whitwell | |
2006-10-05 | Accelerate glBitmap with a color expand blit. Nice speedup for demos | Keith Whitwell | |
like 'fire' that display a help message or fps number this way. | |||
2006-10-05 | Quieten debug message. | Keith Whitwell | |
2006-10-05 | eliminate rhw divide under some circumstances | Keith Whitwell | |
2006-10-05 | Quieten debug message. | Keith Whitwell | |
2006-09-25 | Fixup one place where attrsz was being used instead of active_sz | Keith Whitwell | |
2006-09-22 | Explicit test for +ve pitches in intelEmitCopyBlit() | Keith Whitwell | |
2006-09-21 | Do a better job of choosing texture formats to avoid image conversions. | Keith Whitwell | |
2006-09-21 | use the requested internal texture format where possible | Keith Whitwell | |
2006-09-21 | Catch a couple of cases where we would have to delete and recreate the | Keith Whitwell | |
miptree object. |