fdo-mirrors/mesa
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/mesa/mesa.git synced 2026-05-06 13:48:06 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions 1
34935 commits 66 branches 933 tags 1,019 MiB
Commit graph

34935 commits

This branch
This branch
All branches
Author SHA1 Message Date
José Fonseca
b57e18c35c util: Add inline function for approximate floating point comparison. 2010-05-30 16:38:41 +01:00
José Fonseca
f503b3dd9d llvmpipe: Use pointer_to_func() instead of custom wrappers. 2010-05-30 16:38:23 +01:00
José Fonseca
4b29da1226 draw: Use pointer_to_func() instead of custom wrappers. 2010-05-30 16:38:03 +01:00
José Fonseca
6f67a71aad util: Generic pointer to func conversion helper. 
That observes strict-aliasing rules.
 2010-05-30 16:36:08 +01:00
José Fonseca
18fb9ff6d8 llvmpipe: Pass the fs variant to the rasterizer. 2010-05-30 10:30:01 +01:00
José Fonseca
dc10943038 python/retrace: Interpret resource_copy_region. 2010-05-30 10:30:01 +01:00
José Fonseca
311227abbb st/python: Expose region_copy/fill_region instead of deprecated surface_copy/fill 2010-05-30 10:30:01 +01:00
Chia-I Wu
0f37f242dd st/vega: Use FREE. 
Match MALLOC with FREE.  Otherwise, it causes weird segfaults when built
with memory debugging.
 2010-05-30 11:11:27 +08:00
Chia-I Wu
a5183a38c2 st/egl: Hook eglCreatePbufferFromClientBuffer. 
This is some refactoring works.  Creating a pbuffer from an
EGL_OPENVG_IMAGE is still not supported.
 2010-05-30 11:09:30 +08:00
Chia-I Wu
9d7219c560 st/egl: Fix eglCopyBuffers. 
It was broken since surface_copy -> resource_copy_region transition.
 2010-05-30 11:05:15 +08:00
Chia-I Wu
20c3d7190e st/egl: Initialize KMS backend from an fd. 
Check if the native native is EGL_DEFAULT_DISPLAY.  If not, treat it as
an fd.
 2010-05-30 10:52:50 +08:00
Marek Olšák
b202c78586 r300: fix blits for textures of width/height greater than 2048 on r5xx 
Yes I am fixing r300c ... who knew?
 2010-05-29 17:44:20 +02:00
Marek Olšák
b5c147caba r600g: remove unused variable 2010-05-29 15:41:22 +02:00
Marek Olšák
7d9396d7cf r600g: fix gallium function parameters 2010-05-29 15:41:22 +02:00
Marek Olšák
1d57b6ebd1 r600g: make a local copy of viewport state 2010-05-29 15:41:22 +02:00
Jakob Bornecrantz
e7f0f6bb72 i915g: Rename winsys c file 2010-05-29 13:31:30 +02:00
Jakob Bornecrantz
835f5a65a1 i915g: Move pci id to winsys struct 2010-05-29 13:30:25 +02:00
Jakob Bornecrantz
bbaaf823fa targets/egl: Give egl targets a good build enviroment 2010-05-29 13:19:13 +02:00
Kristian Høgsberg
0defc25977 egl_dri2: Split driver loading out into its own function 2010-05-28 21:31:22 -04:00
Kristian Høgsberg
c5ddaa2cd7 egl_dri2: Check for xcb_connect() failure the right way 
It always returns a valid xcb_connection_t, but if connection failed, it's in the
error state.
 2010-05-28 21:31:22 -04:00
Marek Olšák
ba10b2732b r600g: fix build 2010-05-29 02:31:44 +02:00
Roland Scheidegger
1e17178fc4 llvmpipe: adapt to clear interface changes 
with some newfangled code, should support separate depth/stencil clears.
Needs some testing.
 2010-05-29 01:31:01 +02:00
Roland Scheidegger
837d32062e i965g: adapt to clear interface changes 
just like i915g, should in theory support separate depth/stencil clears
(untested), with the same 2d blitter hacks.
 2010-05-29 01:29:31 +02:00
Roland Scheidegger
4bc7b14224 i915g: adapt to clear interface changes 
should in theory support separate depth/stencil clears (untested).
Also fix some format omissions?
 2010-05-29 01:28:21 +02:00
Roland Scheidegger
10baf7ec1d r300g: adapt to clear interface changes 
should support separate depth/stencil clears just fine with changed u_blitter.
 2010-05-29 01:26:22 +02:00
Roland Scheidegger
e5b82c8222 nv50: adapt to clear interface changes 
should support separate depth/stencil clears just fine.
 2010-05-29 01:25:09 +02:00
Roland Scheidegger
c5cccf8a49 nvfx: adapt to clear interface changes 2010-05-29 01:24:25 +02:00
Roland Scheidegger
d1998962ef svga: adapt to clear interface changes 
this should support separate stencil/depth clears just fine.
 2010-05-29 01:22:23 +02:00
Roland Scheidegger
72c77d472b softpipe: adapt to clear interface changes 2010-05-29 01:21:15 +02:00
Roland Scheidegger
9d596ea492 rbug: adapt to clear interface changes 2010-05-29 01:20:11 +02:00
Roland Scheidegger
e4de4b0755 trace: adapt to clear interface changes 2010-05-29 01:19:59 +02:00
Roland Scheidegger
25296a3ed3 identity: adapt to clear interface changes 2010-05-29 01:19:48 +02:00
Roland Scheidegger
9eae3fd992 fo: adapt to clear interface changes 2010-05-29 01:19:24 +02:00
Roland Scheidegger
821498bd6f cell: adapt to clear interface changes 2010-05-29 01:18:47 +02:00
Roland Scheidegger
c350068720 st/python: adapt to clear interface changes 2010-05-29 01:18:17 +02:00
Roland Scheidegger
783083c3b8 st/mesa: use new ability to clear only depth or stencil 2010-05-29 00:04:14 +02:00
Roland Scheidegger
a1d5131d2e util: adapt to clear interface changes 2010-05-29 00:02:05 +02:00
Roland Scheidegger
0cd70b554c gallium: clear interface changes 
clears were a bit limited in gallium:
- no scissoring (OGL only) nor explicit rectangle list (d3d9)
- no color/stencil masks (OGL only)
- no separate depth/stencil clears (d3d9/d3d10/OGL)
- cannot really clear single color buffer (only with resource_fill_region)

Additionally, d3d can clear surfaces not currently bound to the framebuffer.

It is, however, not easy to find some common ground what a clear should be able
to do, due to both API requirements and also hw differences (a case which might
be able to use a special clear path on one hw might need a "normal" quad render
on another).

Hence several clear methods are provided, and a driver should implement all of
them.
- clear: slightly modified to also be able to clear only depth or stencil in a
  combined depth/stencil surface. This is however optional based on driver
  capability though ideally it wouldn't be optional. AFAIK this is in fact
  something used by applications quite a bit.
  Otherwise, for now still doesn't allow clearing with scissors/mask (or single
  color buffers)
- clearRT: clears a single (potentially unbound) color surface. This was formerly
  roughly known as resource_fill_region. mesa st will not currently use this,
  though potentially would be useful for GL ClearBuffer.
- clearDS: similar to above except for depth stencil surfaces.

Note that clearDS/clearRT currently handle can handle partial clear. This might
change however.
 2010-05-28 23:57:47 +02:00
Eric Anholt
828f648ac3 i965: Add cache unit -> bo name mapping for more gen6 state objects. 
This will help in bufmgr debugging and aub dumping.
 2010-05-28 12:51:48 -07:00
Zou Nan hai
9390af0d96 i965: fix PIPE_CONTROL command for gen6. 
Signed-off-by: Zou Nan hai <nanhai.zou@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
 2010-05-28 12:51:48 -07:00
Brian Paul
e6d8b4090f llvmpipe: add PIPE_BUFFER to assertion 
Fixes regression from commit f2e05acac9.
 2010-05-28 13:31:23 -06:00
Brian Paul
f57edc58b5 softpipe: add PIPE_BUFFER to assertion 
Fixes regression from commit f2e05acac9.
 2010-05-28 13:30:24 -06:00
Brian Paul
151b3bb3af llvmpipe: add out of memory checking to triangle setup path 2010-05-28 12:52:41 -06:00
Brian Paul
5d0e136eff glsl: change uniform location/offset encoding 
This lets Mesa work like other OpenGL implementations with regard
to indexing uniform arrays.  See comments for details.

Note: this is a candidate for the 7.8 branch.
 2010-05-28 12:52:34 -06:00
Marek Olšák
f4bcd0cab0 r300g: remove unnecessary breaks 2010-05-28 03:30:11 +02:00
Marek Olšák
ebe2b54663 r300g: report vertex format support in is_format_supported 2010-05-28 03:30:11 +02:00
Marek Olšák
f2e05acac9 st/mesa: advertise ARB_half_float_vertex when available 
Both softpipe and llvmpipe pass the piglit half float test.
 2010-05-28 03:30:11 +02:00
Marek Olšák
3262554bb3 r300g/swtcl: fix WPOS 
There you are! r300->viewport_state is undefined when using SW TCL.

Piglit score (r300g.tests):
HW TCL: 231/275
SW TCL: 233/275
 2010-05-28 03:30:11 +02:00
Marek Olšák
49244df5a8 r300g/swtcl: avoid an infinite loop if an occl. query doesn't return anything 2010-05-28 03:30:11 +02:00
Jerome Glisse
6a143b8424 r600g: disable rendering 
Disable rendering to avoid GPU lockup.
Use radeondb to debug shader compiler :
radeondb -c gallium.bof
radeondb -s gallium.json
Will print shader generated, best is to use fp demos to test
the compiler.

Signed-off-by: Jerome Glisse <jglisse@redhat.com>
 2010-05-27 23:24:26 +02:00