When the cube faces were stored in a compressed format, the img_stride
values were wrong and didn't match the per-face size computed in the
tex_image_face_size() function. This caused bad rendering or segfaults.
st/mesa: return success/fail from guess_and_alloc_texture()
Before we looked at stObj->pt to see if we may have run out of memory,
but that's not a good indicator. This fixes the spurious GL_OUT_OF_MEMORY
errors that could arise before.
In the lack of more fine grained capabilities in Gallium, assume that if
the pipe driver supports GLSL then native limits match Mesa software
limits.
(cherry picked from commit 40a90cd112)
glsl: change variable declared assertion into conditional
The slang_variable::declared field originated as a debug field but
can be promoted for use during sematic error checking.
Fixes fd.o bug 27921.
NOTE: this is a candidate for back-porting to the 7.8 stable branch.
This reverts commit 9446fd8f69.
It doesn't make sense to replace strcpy(a,b) with strncpy(a,b,strlen(b)).
The preceeding code effectively does bounds checking, btw.
Commit e648d4a1d1 changed the original
less-than test to a not-equal test. This was an effort to save some
memory by switching the texture layout to a non-mipmapped layout when
we mis-guessed about the original layout (thus saving some memory).
However, this causes us to hit a new (apparently broken) code path
when copying the old texture's data to the new texture. Simply
undo this change for the time being until the other/new bug is fixed.
Fixes fd.o bug 27933.
This gives a ~30% shader optimization time improvement on blender.
Tested by comparing the dumped LLVM modules.
Current ordering:
time ~/llvm-git/obj/Release-Asserts/bin/opt l.bc -constprop -instcombine
-mem2reg -gvn -simplifycfg
real 0m1.126s
user 0m1.108s
sys 0m0.012s
With this patch:
time ~/llvm-git/obj/Release-Asserts/bin/opt l.bc -mem2reg -constprop -instcombine -gvn -simplifycfg
real 0m0.885s
user 0m0.880s
sys 0m0.000s
The overall improvement in blender is ~15%.
Blender without the patch takes 1m13s:
edwin 5934 87.6 11.5 729440 458296 pts/5 SLl+ 17:35 1:13 blender
Blender with the patch takes 1m3s:
edwin 5726 94.2 11.2 716424 446168 pts/5 SLl+ 17:32 1:03 blender
It is still slow with the patch, but better (most of the optimization time is
taken up by GVN, see LLVM PR7023).
Signed-off-by: Török Edwin <edwintorok@gmail.com>
Signed-off-by: José Fonseca <jfonseca@vmware.com>