Marc-André Lureau
|
e9ca2d4ad3
|
blitter: fix invalid texelFetch call
Triggered by piglit/bin/ext_framebuffer_multisample-no-color 2 depth single -auto -fbo
|
10 years ago |
|
Marc-André Lureau
|
043320c4f1
|
shader: better type promotion for comparison op
This is an improvement from ee7852525: the right source type is
used.
Fixes piglit spec/glsl-1.50/execution/built-in-functions/gs-equal-ivec2-ivec2
|
10 years ago |
|
Marc-André Lureau
|
444f661736
|
formats: fix byte order of BGRA
I am a bit confused by the docs, but it fixes piglit spec/!opengl
1.1/draw-pixels.
|
10 years ago |
|
Marc-André Lureau
|
400c7df8b5
|
shader: use uint comparison for UCMP
Fixes a few piglit tests, such as:
spec/glsl-1.30/execution/switch/fs-default_last
|
10 years ago |
|
Marc-André Lureau
|
7f050c7e5b
|
shader: restore the original CMP
e4880f6b broke the original CMP translation that looks correct
|
10 years ago |
|
Marc-André Lureau
|
002354dd91
|
shader: use writemask for op1
Fixes piglit glsl-vec-array test
|
10 years ago |
Dave Airlie
|
26aa5bf620
|
renderer: remove unneeded printf
|
10 years ago |
|
Dave Airlie
|
803e90e76a
|
renderer: fix frontfacing conversion from TGSI->GLSL
TGSI expects -1.0, 1.0, GLSL is 0.0, 1.0 so convert.
Alternate fix to Marc-Andre's fix.
|
10 years ago |
|
Marc-André Lureau
|
fc7e0ef0fb
|
shader: add a check for number of imm and ubo
I think UBO limit is 32, but IMM is not, so we better had some checks to
avoid overflow
|
10 years ago |
|
Marc-André Lureau
|
5f3f9459d6
|
Fix segfault if vrend_shader_select() failed on GS
Signed-off-by: Dave Airlie <airlied@gmail.com>
|
10 years ago |
Dave Airlie
|
9e5bd0dbf0
|
renderer: handle illegal so values better
This stops the guest from crashing the parser here,
if it gives us so entries that are invalid.
|
10 years ago |
|
Dave Airlie
|
ac972775ef
|
renderer: fix encoding on stream output
|
10 years ago |
|
Dave Airlie
|
44a7c8cf99
|
renderer: some minor cleanups
don't call same functions twice in 2 places, don't
bother hand cleaning up stuff
|
10 years ago |
|
Dave Airlie
|
431c818c6a
|
renderer: move some more stuff into ctx state
with contexts we don't need this in the global state.
|
10 years ago |
|
Dave Airlie
|
b1109ee664
|
renderer: move program id into context
no point in having this in global state anymore
|
10 years ago |
|
Dave Airlie
|
c2d68e9995
|
renderer: move some global state into contexts.
This stuff isn't required since the renderer uses contexts now
|
10 years ago |
|
Dave Airlie
|
c13ac6fcd5
|
renderer: limit immediates to a higher number
varying-packing-simple float separate was overflowing
the immediate pool, so add a limit, then increase
the limit so test passes.
|
10 years ago |
|
Dave Airlie
|
dac4578a37
|
renderer: make sure to get glsl 1.50 for interface blocks
we can't emit an interface block in glsl 1.40
|
10 years ago |
|
Dave Airlie
|
f6a786c6ea
|
renderer: allow binding 0 vbos
This was broken calculation, we should allow binding 0 vbos.
|
10 years ago |
|
Dave Airlie
|
869a6aa5f5
|
renderer: add ARB_vertex_type_10f_11f_11f_rev support
|
10 years ago |
|
Dave Airlie
|
cba784dfaf
|
renderer: add arb_texture_gather support
This just adds support for the basic TG4 opcode,
the GS5 variant is a bit more work.
|
10 years ago |
|
Dave Airlie
|
c66f8b23d9
|
renderer: add support for ARB_texture_query_levels
|
10 years ago |
|
Marc-André Lureau
|
19fae107e9
|
Fix VIRGL_SET_SCISSOR_MAXX_MAXY
|
10 years ago |
|
Dave Airlie
|
3f4e3e1cad
|
renderer: add support for ARB_viewport_array
|
10 years ago |
|
Dave Airlie
|
78370e879f
|
renderer: add initial texture_query_lod support
|
10 years ago |
|
Dave Airlie
|
3dcc980055
|
renderer: add some more caps info
this is jsut for 3 trivial extensions
|
10 years ago |
|
Dave Airlie
|
0ff22a06b7
|
renderer: ABI break: overhaul viewport/scissor state
add support for multiple viewports, and reduce
viewport size.
|
10 years ago |
|
Dave Airlie
|
e13ebc57e1
|
renderer: move away from pipe bind flag definitions.
|
10 years ago |
|
Dave Airlie
|
ebcc110592
|
tests: make buffer size larger for bad stride test
|
10 years ago |
|
Dave Airlie
|
621629e32f
|
tests: add 2d array bad layer stride
pass a layer stride too small for other dimensions
|
10 years ago |
|
Dave Airlie
|
53ab094df9
|
tests: use a box with height to test stride validation
|
10 years ago |
|
Dave Airlie
|
dfd7749ddb
|
renderer: refactor iov bounds checking
take the user supplied stride and layer stride into a/c
when calculating the amount of memory required to read/write
from the user supplied iovec.
|
10 years ago |
|
Dave Airlie
|
a423879e87
|
tests: don't pass a bad stride in validation tests
|
10 years ago |
|
Dave Airlie
|
93072467c4
|
tests: add bad stride test
stride shouldn't be less than resource width
|
10 years ago |
|
Dave Airlie
|
5fd93680a7
|
tests: don't pass a stride to bad level test
|
10 years ago |
|
Dave Airlie
|
ee10a18ef5
|
tests: add transfer test with a bad level.
|
10 years ago |
|
Dave Airlie
|
22f2944e52
|
tests: don't pass a layer stride on the transfer tests
|
10 years ago |
|
Dave Airlie
|
1f9ab04501
|
tests: do some tests for layer stride with bad texture types
|
10 years ago |
|
Dave Airlie
|
20e86f1771
|
renderer: validate layer stride isn't set for some texture types
Some texture types have no layers, so a non-zero layer stride makes
no sense.
|
10 years ago |
|
Dave Airlie
|
1fe0822ff4
|
renderer: fix bug transfer inline write decoding
cleanup a little, but also read layer stride from correct place
found writing validation tests.
|
10 years ago |
|
Dave Airlie
|
a6a0b09d5e
|
tests: drop unused test
this is subsumed into other tests
|
10 years ago |
|
Dave Airlie
|
5f7031e117
|
tests: add mipmap tests to resource creation
|
10 years ago |
|
Dave Airlie
|
6495f870b4
|
renderer: block resource creation for mipmaps
Some scenarios for mipmap creation are illegal, so block them.
|
10 years ago |
|
Dave Airlie
|
9006de318e
|
tests: more inline transfer tests
This adds code to the encoder that needs to be backported
to the mesa virgl driver to support inline writes for
non-buffer objects.
|
10 years ago |
|
Dave Airlie
|
44d4cf4fd3
|
tests: add option to test harness for large buffer
|
10 years ago |
|
Dave Airlie
|
60ee9c4457
|
tests: allocate cmd buf to correct size
the value is in dwords, so calloc the correct size.
|
10 years ago |
|
Dave Airlie
|
c6fd7d55ac
|
tests: add flush callback to test large transfers
|
10 years ago |
|
Dave Airlie
|
052835544f
|
tests: add overlapping id test
|
10 years ago |
|
Dave Airlie
|
5e975c6a2e
|
renderer: switch hw contexts outside the transfer iov
this fixes a regression since the transfer inline rewrite
|
10 years ago |
|
Dave Airlie
|
5109a1ff26
|
tests: add new transfer tests
this does a transfer per resource type, also start
adding inline transfer tests, these need a bit more work
|
10 years ago |