OSDN Git Service
Kristian Høgsberg [Sun, 12 Oct 2008 00:41:14 +0000 (20:41 -0400)]
Update DRI2 implementation according to new specification.
Kristian Høgsberg [Wed, 15 Oct 2008 03:07:42 +0000 (23:07 -0400)]
Revert pointless reindents to avoid merge conflicts.
Why are we reindenting code that's work in progress...
RALOVICH, Kristóf [Mon, 13 Oct 2008 13:14:33 +0000 (15:14 +0200)]
glx: indent -br -i3 -npcs --no-tabs glxclient.h
RALOVICH, Kristóf [Mon, 13 Oct 2008 13:13:21 +0000 (15:13 +0200)]
glx: indent -br -i3 -npcs --no-tabs glcontextmodes.h
RALOVICH, Kristóf [Mon, 13 Oct 2008 13:10:59 +0000 (15:10 +0200)]
glx: indent -br -i3 -npcs --no-tabs glxextensions.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 13:07:23 +0000 (15:07 +0200)]
glx: indent -br -i3 -npcs --no-tabs glxextensions.h
RALOVICH, Kristóf [Mon, 13 Oct 2008 13:04:31 +0000 (15:04 +0200)]
glx: indent -br -i3 -npcs --no-tabs xfont.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 13:03:54 +0000 (15:03 +0200)]
glx: indent -br -i3 -npcs --no-tabs xf86dristr.h
RALOVICH, Kristóf [Mon, 13 Oct 2008 13:03:13 +0000 (15:03 +0200)]
glx: indent -br -i3 -npcs --no-tabs XF86dri.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 13:02:55 +0000 (15:02 +0200)]
glx: indent -br -i3 -npcs --no-tabs xf86dri.h
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:58:33 +0000 (14:58 +0200)]
glx: kill old K&R syntax in XF86dri.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:41:33 +0000 (14:41 +0200)]
glx: indent -br -i3 -npcs --no-tabs vertarr.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:37:06 +0000 (14:37 +0200)]
glx: indent -br -i3 -npcs --no-tabs singlepix.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:35:18 +0000 (14:35 +0200)]
glx: indent -br -i3 -npcs --no-tabs single2.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:34:43 +0000 (14:34 +0200)]
glx: indent -br -i3 -npcs --no-tabs renderpix.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:34:18 +0000 (14:34 +0200)]
glx: indent -br -i3 -npcs --no-tabs render2.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:33:28 +0000 (14:33 +0200)]
glx: indent -br -i3 -npcs --no-tabs pixelstore.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:32:52 +0000 (14:32 +0200)]
glx: indent -br -i3 -npcs --no-tabs pixel.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:31:49 +0000 (14:31 +0200)]
glx: indent -br -i3 -npcs --no-tabs packsingle.h
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:29:23 +0000 (14:29 +0200)]
glx: indent -br -i3 -npcs --no-tabs packrender.h
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:27:46 +0000 (14:27 +0200)]
glx: indent -br -i3 -npcs --no-tabs indirect_window_pos.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:27:07 +0000 (14:27 +0200)]
glx: indent -br -i3 -npcs --no-tabs indirect_vertex_program.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:25:55 +0000 (14:25 +0200)]
glx: indent -br -i3 -npcs --no-tabs indirect_vertex_array_priv.h
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:21:17 +0000 (14:21 +0200)]
glx: indent -br -i3 -npcs --no-tabs indirect_vertex_array.h
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:20:15 +0000 (14:20 +0200)]
glx: indent -br -i3 -npcs --no-tabs indirect_vertex_array.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:19:24 +0000 (14:19 +0200)]
glx: indent -br -i3 -npcs --no-tabs indirect_transpose_matrix.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:18:18 +0000 (14:18 +0200)]
glx: indent -br -i3 -npcs --no-tabs indirect_texture_compression.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:12:40 +0000 (14:12 +0200)]
glx: indent -br -i3 -npcs --no-tabs glx_query.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:12:02 +0000 (14:12 +0200)]
glx: indent -br -i3 -npcs --no-tabs glx_pbuffer.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:10:16 +0000 (14:10 +0200)]
glx: indent -br -i3 -npcs --no-tabs glxhash.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:10:06 +0000 (14:10 +0200)]
glx: indent -br -i3 -npcs --no-tabs glxhash.h
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:08:45 +0000 (14:08 +0200)]
glx: indent -br -i3 -npcs --no-tabs glxext.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:07:07 +0000 (14:07 +0200)]
glx: indent -br -i3 -npcs --no-tabs glxcurrent.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 12:06:30 +0000 (14:06 +0200)]
glx: indent -br -i3 -npcs --no-tabs glxcmds.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 11:49:49 +0000 (13:49 +0200)]
glx: indent -br -i3 -npcs --no-tabs glcontextmodes.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 11:48:07 +0000 (13:48 +0200)]
glx: indent -br -i3 -npcs --no-tabs eval.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 11:47:37 +0000 (13:47 +0200)]
glx: indent -br -i3 -npcs --no-tabs drisw_glx.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 11:40:09 +0000 (13:40 +0200)]
glx: indent -br -i3 -npcs --no-tabs dri_glx.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 11:26:09 +0000 (13:26 +0200)]
glx: indent -br -i3 -npcs --no-tabs dri_common.h
RALOVICH, Kristóf [Mon, 13 Oct 2008 11:25:28 +0000 (13:25 +0200)]
glx: indent -br -i3 -npcs --no-tabs dri_common.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 11:18:41 +0000 (13:18 +0200)]
glx: indent -br -i3 -npcs --no-tabs dri2.h
RALOVICH, Kristóf [Mon, 13 Oct 2008 11:17:32 +0000 (13:17 +0200)]
glx: indent -br -i3 -npcs --no-tabs dri2_glx.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 11:16:28 +0000 (13:16 +0200)]
glx: indent -br -i3 -npcs --no-tabs dri2.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 11:12:37 +0000 (13:12 +0200)]
glx: indent -br -i3 -npcs --no-tabs compsize.c
RALOVICH, Kristóf [Mon, 13 Oct 2008 11:11:30 +0000 (13:11 +0200)]
glx: indent -br -i3 -npcs --no-tabs clientattrib.c
RALOVICH, Kristóf [Sat, 11 Oct 2008 12:27:07 +0000 (14:27 +0200)]
glx: add a line of Emacs helping variables
RALOVICH, Kristóf [Sat, 11 Oct 2008 11:09:50 +0000 (13:09 +0200)]
glx: No need to zero a local variable.
My previous commit
e2060348630b59a446bac7f734fdde40033093ab introduced this.
RALOVICH, Kristóf [Sat, 11 Oct 2008 10:56:32 +0000 (12:56 +0200)]
glx: make INIT_MESA_SPARC more robust
Embed the macro body into do { ... } while(0) .
Xiang, Haihao [Mon, 13 Oct 2008 05:04:04 +0000 (13:04 +0800)]
i915: Texture instructions use r/t/oC/oD register as texture coordinate.
Fix http://bugs.freedesktop.org/show_bug.cgi?id=16287.
Eric Anholt [Fri, 26 Sep 2008 02:11:03 +0000 (19:11 -0700)]
intel: Add acceleration for glDrawPixels(GL_STENCIL_INDEX).
This is nasty because there's no way in GL to output data to the stencil
buffer directly, so we have to do a dance to wrap the depth/stencil buffer
in an ARGB renderbuffer.
Improves performance of several oglconform testcases by better than a factor
of 2.
Brian Paul [Tue, 14 Oct 2008 23:22:40 +0000 (17:22 -0600)]
cell: initial bits for 3D texture support
Brian Paul [Tue, 14 Oct 2008 23:22:06 +0000 (17:22 -0600)]
cell: replace FREE() with align_free()
Brian Paul [Tue, 14 Oct 2008 23:19:57 +0000 (17:19 -0600)]
cell: fix incorrect parameter type
Brian Paul [Tue, 14 Oct 2008 23:11:29 +0000 (17:11 -0600)]
cell: support for cubemaps
Though, progs/demos/cubemap.c doesn't quite work right...
Brian Paul [Tue, 14 Oct 2008 23:09:56 +0000 (17:09 -0600)]
cell: fall-through case for TGSI_OPCODE_TXB
Brian Paul [Tue, 14 Oct 2008 21:19:01 +0000 (15:19 -0600)]
cell: support NPOT textures, clamp/repeat mode, normalized/unorm texcoords
glDrawPixels works now.
Brian Paul [Tue, 14 Oct 2008 20:02:07 +0000 (14:02 -0600)]
cell: fix npot texture tiling bugs
Brian Paul [Tue, 14 Oct 2008 18:52:16 +0000 (12:52 -0600)]
cell: fix broken negative texcoord conversion
Brian Paul [Tue, 14 Oct 2008 18:42:21 +0000 (12:42 -0600)]
cell: use minify vs magnify filters
Brian Paul [Tue, 14 Oct 2008 18:41:46 +0000 (12:41 -0600)]
cell: fix assertions
Brian Paul [Tue, 14 Oct 2008 18:11:52 +0000 (12:11 -0600)]
cell: basic mipmap filtering works now
Though, only GL_MIPMAP_NEAREST / GL_LINEAR works right now.
Brian Paul [Tue, 14 Oct 2008 18:10:27 +0000 (12:10 -0600)]
cell: fix tex image stride bugs
Brian Paul [Tue, 14 Oct 2008 16:55:38 +0000 (10:55 -0600)]
cell: now do texture twiddling in the right way, at the right time.
Also handles images smaller than 32x32 now.
Brian Paul [Tue, 14 Oct 2008 16:05:23 +0000 (10:05 -0600)]
cell: remove old code, clean-ups, etc.
Alan Hourihane [Tue, 14 Oct 2008 10:54:20 +0000 (11:54 +0100)]
dri: don't check the number of cliprects before swap, let
the swap handle the requirements.
Brian Paul [Tue, 14 Oct 2008 02:19:51 +0000 (20:19 -0600)]
cell: more clean-up in spu_tri.c
Brian Paul [Tue, 14 Oct 2008 01:50:20 +0000 (19:50 -0600)]
cell: remove dead code, clean-up, reformatting
Brian Paul [Tue, 14 Oct 2008 01:41:26 +0000 (19:41 -0600)]
cell: finish-up perspective-corrected interpolation
Brian Paul [Mon, 13 Oct 2008 22:43:11 +0000 (16:43 -0600)]
cell: initial work for mipmap texture filtering
Brian Paul [Mon, 13 Oct 2008 21:17:01 +0000 (15:17 -0600)]
cell: remove old texture-related fields
Brian Paul [Mon, 13 Oct 2008 20:31:11 +0000 (14:31 -0600)]
cell: use fewer memory references in sample_texture4_bilinear_2()
Brian Paul [Mon, 13 Oct 2008 20:10:36 +0000 (14:10 -0600)]
cell: remove more old texture code
Brian Paul [Mon, 13 Oct 2008 20:09:54 +0000 (14:09 -0600)]
cell: bilinear texture filtering using integer arithmetic
Fewer float/int conversions involved.
Brian Paul [Mon, 13 Oct 2008 17:16:04 +0000 (11:16 -0600)]
cell: remove old texture code
Brian Paul [Mon, 13 Oct 2008 16:55:08 +0000 (10:55 -0600)]
cell: do texture sampling/filtering for four pixels at a time.
Brian Paul [Mon, 13 Oct 2008 16:54:11 +0000 (10:54 -0600)]
cell: added spu_unpack_A8R8G8B8_transpose4()
Plus, clearer shuffle masks in other funcs.
Alan Hourihane [Mon, 13 Oct 2008 11:30:40 +0000 (12:30 +0100)]
mesa: when emitting an address load instruction, for
indexed elements ensure we write to a single register.
Brian Paul [Fri, 10 Oct 2008 23:48:16 +0000 (17:48 -0600)]
cell: more instruction scheduling optimizations (MIN/MAX/LERP/etc)
Also, optimize register->memory stores.
Brian Paul [Fri, 10 Oct 2008 22:40:16 +0000 (16:40 -0600)]
Merge commit 'origin/master' into gallium-0.2
Brian Paul [Fri, 10 Oct 2008 22:36:40 +0000 (16:36 -0600)]
cell: pass texture unit (sampler number) to txp() function
The glsl/multitex demo runs now.
Ian Romanick [Fri, 10 Oct 2008 18:47:43 +0000 (11:47 -0700)]
intel: GLSL 1.20 is broken in Mesa, so disable it in the i965 driver
Eric Anholt [Fri, 10 Oct 2008 16:43:15 +0000 (09:43 -0700)]
i965: Add missing intel_pixel_draw.c symlink to fix build.
Brian Paul [Fri, 10 Oct 2008 22:30:43 +0000 (16:30 -0600)]
mesa: fix asst. issues in _mesa_texstore_argb8888()
If we shift bytes into the texel word (or use the PACK_COLOR_8888 macro),
we don't have to worry about big vs. little endian. See comments about
texel formats in texformat.h.
Remove an unneeded/incorrect else-if clause that produced wrong results
on big-endian systems.
Brian Paul [Fri, 10 Oct 2008 22:30:18 +0000 (16:30 -0600)]
mesa: include needed header
Brian Paul [Fri, 10 Oct 2008 22:29:54 +0000 (16:29 -0600)]
mesa: remove unneeded includes
Brian Paul [Fri, 10 Oct 2008 22:26:15 +0000 (16:26 -0600)]
mesa: fix error codes in _mesa_GetObjectParameterivARB(), bug 17861
Brian Paul [Fri, 10 Oct 2008 22:23:31 +0000 (16:23 -0600)]
mesa: rename macro params to emphasize that there's no particular color ordering
Brian Paul [Fri, 10 Oct 2008 22:21:52 +0000 (16:21 -0600)]
mesa: remove unneeded includes
Brian Paul [Fri, 10 Oct 2008 20:57:57 +0000 (14:57 -0600)]
cell: add emit_RI10s() which does range checking on the 10-bit signed immediate field
This type of checking should be expanded to cover more instructions...
Brian Paul [Fri, 10 Oct 2008 20:44:52 +0000 (14:44 -0600)]
cell: additional 'offset' checking in spe_lqd(), spe_stqd()
Brian Paul [Fri, 10 Oct 2008 20:35:56 +0000 (14:35 -0600)]
cell: fix assertions in spe_lqd(), spe_stqd()
Brian Paul [Fri, 10 Oct 2008 20:34:43 +0000 (14:34 -0600)]
cell: fix function prologue/epilogue code for large stack frames
The ai instruction is limited to a 10-bit signed immediate value.
Robert Ellison [Fri, 10 Oct 2008 20:13:13 +0000 (14:13 -0600)]
CELL: fixing stencil bugs
These are the defects found and fixed so far. Several more have
been observed; I'm working on them.
- Fixed an error in spe_load_uint() that caused incorrect values to be
loaded if the given unsigned value had the low 18 bits as 0,
and that caused inefficient code to be emitted if the given value
had the high 14 bits as 0.
- Fixed a problem in stencil code generation where optional registers
weren't tracked correctly.
- Fixed a problem that the stencil function NEVER was acting as ALWAYS.
- Fixed several problems that could occur if stenciling were enabled but
depth was disabled.
- Fixed a problem with two-sided stencil writemask handling that could
cause a stencil writemask to not be applied.
- Fixed several state permutations that were incorrectly flagged as
not requiring stencil values to be calculated.
Ian Romanick [Fri, 10 Oct 2008 18:47:43 +0000 (11:47 -0700)]
intel: GLSL 1.20 is broken in Mesa, so disable it in the i965 driver
Brian Paul [Fri, 10 Oct 2008 18:38:27 +0000 (12:38 -0600)]
cell: call cell_flush_int() at end of cell_create_context()
Ensures that SPUs are initialized/ready before proceeding.
This fixes a spurious assertion failure when the SPU-side shader function info
hasn't been returned to the PPU before shader codegen.
Brian Paul [Fri, 10 Oct 2008 18:24:39 +0000 (12:24 -0600)]
cell: fix LERP when dst reg is a src reg
Also, bump up frame size and fix some assertions.
Brian Paul [Fri, 10 Oct 2008 18:04:49 +0000 (12:04 -0600)]
replace 1.0/sqrt() with inversesqrt()
Brian Paul [Fri, 10 Oct 2008 18:04:19 +0000 (12:04 -0600)]
cell: fix fm/fs copy & paste bug from a few commits ago
Brian Paul [Fri, 10 Oct 2008 17:56:03 +0000 (11:56 -0600)]
cell: fix bug in emit_FLR() when src reg == dst reg
Brian Paul [Fri, 10 Oct 2008 17:52:55 +0000 (11:52 -0600)]
cell: fix bug in emit_FRC() when src register == dst register.
With this fix, the glsl/brick demo runs.