Age | Commit message (Collapse) | Author |
|
|
|
|
|
|
|
|
|
Comments about the deleted stuff:
- openaren hang: likely caused by the vertex corruptions, fixed by Jakob.
- tiling: Y-tiling works with my hw-clear branch. X-tiling works as
merged to master a while ago (execbuf2 version).
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
|
|
|
|
ARB_instanced_arrays is a subset of D3D9.
ARB_draw_instanced is a subset of D3D10.
The point of this change is to allow D3D9-level drivers to enable
ARB_instanced_arrays without ARB_draw_instanced.
|
|
|
|
|
|
Use MULHI_UINT instead of the more complex
INT_TO_FLT->MUL->TRUNC->FLT_TO_INT
|
|
|
|
|
|
bc.ndw is altered in r600_bc_build, respect that
in fragment shader size calculation.
|
|
If an array redeclaration includes an initializer, the initializer
would previously be dropped on the floor. Instead, directly apply the
initializer to the correct ir_variable instance and append the
generated instructions.
Fixes bugzilla #34374 and piglit tests glsl-{vs,fs}-array-redeclaration.
NOTE: This is a candidate for stable release branches. 0292ffb8 and
8e6cb9fe are also necessary.
|
|
|
|
|
|
Fixes piglit/tex-srgb.
|
|
|
|
Removed sampler view support for USCALED/SSCALED, the texture unit
refuses to convert to non-normalized float. The enums are treated
like UNORM.
Removed duplicate format related headers.
|
|
Which is always the case, but we'll keep the 2D engine blitter
nonetheless.
|
|
|
|
|
|
The hw doesn't like it - demos/shadowtex is broken. The emitted shader
isn't totally empty though, the depth write fixup gets emitted instead.
Maybe that one is somewhat fishy, too?
Idea for this patch from Jakob Bornecrantz.
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
|
|
|
|
This hopefully fixes wayland-egl's dependency
resolution for autogenerated wayland-drm headers.
|
|
This is an awful hack and will hurt performance on Ironlake, but we're
at a loss as to what's going wrong otherwise. This is the only common
variable we've found that avoids the problem on 4 applications
(CelShading, gnome-shell, Pill Popper, and my GLSL demo), while other
variables we've tried appear to only be confounding. Neither the
specifications nor the hardware team have been able to provide any
enlightenment, despite much searching.
https://bugs.freedesktop.org/show_bug.cgi?id=29172
Tested by: Chris Lord <chris@linux.intel.com> (Pill Popper)
Tested by: Ryan Lortie <desrt@desrt.ca> (gnome-shell)
|
|
|
|
|
|
Blending is now fully supported with:
- R8_UNORM
- R8G8_UNORM
- B8G8R8A8_UNORM
- R16G16B16A16_FLOAT (r500-only)
Blending is partially supported (DST_ALPHA not working) with:
- L8A8_UNORM
- I8_UNORM
- B5G5R5A1_UNORM
- B10G10R10A2_UNORM
The other formats can't do blending.
|
|
|
|
It's broken now that tgsi_exec_machine::Inputs/Ouputs are pointers.
Temporary if anybody still cares about tgsi_sse2.c. Permanent otherwise.
|
|
|
|
Because the format can be changed to UNORM in a surface.
This fixes:
state_tracker/st_atom_framebuffer.c:163:update_framebuffer_state:
Assertion `framebuffer->cbufs[i]->texture->bind & (1 << 1)' failed.
|
|
|
|
|
|
|
|
|
|
|
|
Computation of the delta of this array from the last had a silly little
bug and ignored any initial delta==0 causing grief in Nexuiz and
friends.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
There is a silicon bug which causes unpredictable behaviour if the
URB_FENCE command should cross a cache-line boundary. Pad before the
command to avoid such occurrences. As this command only applies to
gen4/5, do the fixup unconditionally as the specs do not actually state
for which chip it was fixed (and the cost is negligible)...
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
|
|
we still have a lot of corner cases that aren't working.
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
Elements(mach->Inputs) is wrong now that mach->Inputs is dynamically
allocated.
|
|
These are now unnecessary.
|
|
Previously, the rule deleted by this commit was matched every single
time (being the longest match). If not skipping, it used REJECT to
continue on to the actual correct rule.
The flex manual advises against using REJECT where possible, as it is
one of the most expensive lexer features. So using it on every match
seems undesirable. Perhaps more importantly, it made it necessary for
the #if directive rules to contain a look-ahead pattern to make them
as long as the (now deleted) "skip the whole line" rule.
This patch introduces an exclusive start state, SKIP, to avoid REJECTs.
Each time the lexer is called, the code at the top of the rules section
will run, implicitly switching the state to the correct one.
Fixes piglit tests 16384-consecutive-chars.frag and
16385-consecutive-chars.frag.
|
|
The expected result has been out of sync with what glcpp produces for
some time; glcpp's actual result seems to be correct and is very close to
GCC's cpp. Updating this will make it easier to catch regressions in
upcoming commits.
|
|
|