- Jun 17, 2008
-
-
Brice Goglin authored
-
Brice Goglin authored
-
-
Brice Goglin authored
with 04_cleanup-osmesa-configs.patch renamed into 04_debian-configs.patch
-
Wilfried Holzke authored
-
Brian Paul authored
-
Brian Paul authored
-
- Jun 16, 2008
-
-
Brian Paul authored
-
Brian Paul authored
-
Brian Paul authored
Fixes stack overflow on Windows. cherry-picked from master
-
Brian Paul authored
This was introduced by the "i965 GLSL merge" from master (ce7a9efb)
-
Brian Paul authored
-
- Jun 13, 2008
-
-
Haihao Xiang authored
-
Brice Goglin authored
-
- Jun 11, 2008
-
-
Brice Goglin authored
-
-
because really we should be able to just fix the driver. (cherry picked from commit 7013eecf) There are actually even better reasons for this, the bottom line being that enabling these fallbacks makes a lot of apps unusable mostly for no gain whatsoever.
-
Haihao Xiang authored
-
Haihao Xiang authored
Cherry picked from commit 3369cd9a
-
- Jun 10, 2008
-
-
Brice Goglin authored
-
-
Haihao Xiang authored
(cherry picked from commit a742bed9)
-
(cherry picked from commit 7f747204)
-
While OPCODE_SWZ has usually been optimized away in pass0, it may still exist if a SWZ with dst saturate was emitted in pass_fp. Fixes an error in oglconform fpalu.c. (cherry picked from commit 13a6f73a)
-
The pass1 optimization stage clears out writemasks and registers, but the instructions themselves are still being processed at this stage, and could have resulted in them still being emitted. (cherry picked from commit c60b5dfd)
-
(cherry picked from commit bb419970)
-
(cherry picked from commit 6c1a98e9)
-
This fixes the arbvptorus demo. (cherry picked from commit 5f10438f)
-
(cherry picked from commit 31fe7cf5)
-
Also, use wrapped ffs() instead of native. (cherry picked from commit 3105bc1d)
-
(cherry picked from commit e9809a36)
-
The 32-bit immediate value in the i965 instruction word must contain two copies of any 16-bit constants. brw_imm_uw and brw_imm_w just needed to copy the value into both halves of the immediate value instruction field. (cherry picked from commit ca73488f)
-
This patch is a variant of a submission by Michal Wajdeczko to fix oglconform fpalu failures. (cherry picked from commit b4cbf698)
-
-
(cherry picked from commit b0f681b4)
-
(cherry picked from commit 9c8f27ba)
-
The previous COS function failed badly outside of [-pi/2, pi/2].
-
This is described in the link in the comment, and is the same technique that r300 uses.
-
The Taylor series notably fails at producing sin(pi) == 0, which leads to discontinuity every 2*pi. The quadratic gets us sin(pi) == 0 behavior, at the expense of going from 2.4% THD with working Taylor series to 3.8% THD (easily seen on comparative graphs of the two). However, our previous implementation was producing sin(pi) < -1 and worse, so any reasonable approximation is an improvement. This also fixes the repeating behavior, where the previous implementation would repeat sin(x) for x>pi as sin(x % pi) and the opposite for x < -pi.
-