Eric Anholt
503eb57a00
i965: Avoid calloc/free in the CURBE upload process.
...
In exchange we end up with an extra memcpy, but that seems better than
calloc/free. Each buffer is 4k maximum, and on the i965-streaming
branch this allocation was showing up as the top entry in
brw_validate_state profiling for cairo-gl.
2010-06-09 14:17:52 -07:00
Marek Olšák
45fb47d50c
r300g: try harder to create a transfer object
...
I was told this wouldn't help to fix the FDO bug #28443 , but still,
it's a harmless last resort.
Also, linear textures safely fallback to an unpipelined transfer here.
2010-06-09 21:43:31 +02:00
Zack Rusin
1d11eac93f
draw: make sure geometry shader correctly iterates the output buffer
2010-06-09 15:02:08 -04:00
Marek Olšák
f571a569a0
r300g: debug output texture size
2010-06-09 20:22:52 +02:00
Marek Olšák
950be3fa3c
r300g: return NULL texture early if buffer_create fails
2010-06-09 20:17:27 +02:00
Zack Rusin
da4185ca77
draw: make sure the buffer is big enough to fit everything emitted by the gs
2010-06-09 14:11:43 -04:00
Roland Scheidegger
26d5d4d995
llvmpipe: fix copy & paste bug in clear logic
...
fixes bug 28450.
2010-06-09 18:46:52 +02:00
Zack Rusin
04490ad31d
draw: fix geometry shader which emit multiplie primitives
2010-06-09 11:59:53 -04:00
Zack Rusin
d27d45a2df
graw: add a multiple primitive geometry shader test
2010-06-09 11:44:19 -04:00
Zack Rusin
db7ed20730
nv: fix compile after the latest gs changes
2010-06-09 11:29:45 -04:00
Zack Rusin
d4ef0f6c67
geometry shaders: make gs work with changable primitives and variable number of vertices
...
lots and lots of fixes for geometry shaders. in particular now we work when the gs
emits a different primitive than the one the pipeline was started with and also
we work when gs emits more vertices than would fit in the original buffer.
2010-06-09 11:13:34 -04:00
Thomas Hellstrom
cec9955acc
xorg/vmwgfx: Disable 3D by default.
...
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
2010-06-09 07:40:07 +02:00
Thomas Hellstrom
2e27bf8b7c
st/xorg: Add a customizable option to disable 3D.
...
If no customizer is present, 3D will be enabled by default.
Otherwise the option will default to the customizer value.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
2010-06-09 07:40:07 +02:00
Ben Skeggs
ebd98b3496
nvfx: fix typo from gallium-msaa merge
2010-06-09 11:59:49 +10:00
Ben Skeggs
851ca2f7f9
nv50: fix typo from gallium-msaa merge
2010-06-09 11:59:01 +10:00
Ben Skeggs
7bf678f7ec
nv50: use MARK_RING correctly
2010-06-09 11:59:01 +10:00
Vinson Lee
c217d94cca
draw: Remove unnecessary headers.
2010-06-08 17:40:23 -07:00
Marek Olšák
eee570a0fe
r600g: save vertex buffers in blitter
2010-06-09 02:26:44 +02:00
Marek Olšák
26f3f13fa1
r300g: save vertex buffers in blitter
2010-06-09 02:26:44 +02:00
Marek Olšák
8f17e011c3
util: reference vertex buffers in blitter
2010-06-09 02:26:24 +02:00
Marek Olšák
89dc925150
r300g: report that transfers are always supported
2010-06-09 00:26:28 +02:00
Marek Olšák
25470dcbb1
r300g: fix loading compressed textures
2010-06-09 00:18:28 +02:00
Marek Olšák
fccfb7b165
r300g: skip rendering for unsupported vertex formats instead of aborting
...
The next step is to replace skipping by an actual fallback.
2010-06-08 23:50:05 +02:00
Eric Anholt
bcf63dbbd8
intel: Flag NEW_BUFFERS when changing draw buffers.
...
There were entries to this function (most imporantly, prepare_render
-> update_renderbuffers) that wouldn't have had NEW_BUFFERS set, but
brw_wm_surface_state (the i965 state tracking the drawing regions)
expected this to change.
2010-06-08 13:42:02 -07:00
Eric Anholt
df3c1a563f
intel: Convert remaining dri_bo_emit_reloc to drm_intel_bo_emit_reloc.
...
The new API makes so much more sense, I'd like to forget how the old
one worked.
2010-06-08 13:42:02 -07:00
Eric Anholt
34474fa411
intel: Change dri_bo_* to drm_intel_bo* to consistently use new API.
...
The slightly less mechanical change of converting the emit_reloc calls
will follow.
2010-06-08 13:42:02 -07:00
Eric Anholt
22409756d4
intel: Clean up stale comments in intel_batchbuffer.c.
2010-06-08 13:42:02 -07:00
Eric Anholt
754fcd605b
intel: Remove the non-gem paths for batchbuffer upload.
2010-06-08 13:42:02 -07:00
Eric Anholt
4527faa344
intel: Update comment in intel_tex_copy from before miptree x/y rework.
2010-06-08 13:42:02 -07:00
Zack Rusin
551467d4ef
softpipe: make sure we don't fallback to assert on so stats query
...
spotted by Vinson
2010-06-08 16:20:25 -04:00
Zack Rusin
38b13459be
softpipe: make sure that invalid calls to sosettargets don't crash us
2010-06-08 16:18:38 -04:00
Henri Verbeet
8744c36ea4
r600: Make next_inst() static.
2010-06-08 15:40:25 -04:00
Henri Verbeet
1bf75a921b
r600: Assert output registers have a valid export index.
2010-06-08 15:40:25 -04:00
Henri Verbeet
1ec492a366
r600: Process exports for all written fragment outputs.
2010-06-08 15:40:25 -04:00
Henri Verbeet
1f7bc87391
r600: Fill uiFP_OutputMap for all written fragment outputs.
2010-06-08 15:40:25 -04:00
Keith Whitwell
e3c6d4108c
SConscript: don't build mesa components if mesa not in state trackers
2010-06-08 17:38:56 +01:00
Brian Paul
10379e4edf
softpipe: added sp_state_so.c to file list
2010-06-08 08:21:27 -06:00
Zack Rusin
b5d2fc2b48
gallium: scream when one of two unimplemented features in draw is used
...
plus used the correctly specified stride
2010-06-08 06:28:11 -04:00
Zack Rusin
be7d8ddf0c
gallium: add basic support for stream output queries
2010-06-08 06:28:11 -04:00
Zack Rusin
34f7681669
gallium: adjust the query interface to support custom types
...
we need to change it to support composite types
2010-06-08 06:28:11 -04:00
Zack Rusin
3a7fc2e4f3
gallium: rename draw_auto to draw_stream_output, plus fix a comment
...
Brian spotted those
2010-06-08 06:28:11 -04:00
Zack Rusin
2c22b8e61d
gallium: make draw auto work and add relevant caps and docs
2010-06-08 06:28:11 -04:00
Zack Rusin
ba5975dd32
gallium: add interface for DrawAuto and implement it in softpipe
2010-06-08 06:28:11 -04:00
Zack Rusin
c9db97c822
gallium: a lot more complete implementation of stream output
...
interface wise we have everything needed by d3d10 and gl transform feedback.
the draw module misses implementation of some corner cases (e.g. when stream
output wants different number of components per output than normal rendering
paths)
2010-06-08 06:28:10 -04:00
Zack Rusin
a45b7f47ee
gallium: basic and initial implementation of the stream output interface
...
aka transform feedback
2010-06-08 06:28:10 -04:00
Dave Airlie
a2817f6ae5
st/mesa: attempt to fix TFP by using sampler views (v1)
...
Okay I think this is good enough for now, I can't see any other reason
for mesa to want to use a sampler view so lets just leave it at all the A->X conversions for now.
I've been running gnome-shell under r300g with this for day or so and it seems fine.
Signed-off-by: Dave Airlie <airlied@redhat.com>
2010-06-08 19:08:36 +10:00
Jakob Bornecrantz
3a876e8473
llvmpipe: Add geom shader file to make build
2010-06-08 04:46:52 +02:00
Vinson Lee
380f08f4ec
graw: Fix file handle leaks.
2010-06-07 18:34:13 -07:00
Vinson Lee
d09271d6ef
graw: Remove unnecessary headers.
2010-06-07 18:14:34 -07:00
Vinson Lee
b0acaaac53
sw: Remove unnecessary header.
2010-06-07 18:08:33 -07:00