Age | Commit message (Collapse) | Author |
|
|
|
Based on initial patch from Stephan Schmid <stephan_2303@gmx.de>.
Basic idea is to dump the zpass count at the start and end of the query
and subtract to get the total number of visible fragments. HW writes
alternating qwords for up to 4 DBs. On the first pass, we start at
buffer address + 0; on the second pass, we start at buffer address + 8
(bytes). The resulting buffer at the end of the query looks like:
qw[0]: db0 start
qw[1]: db0 end
...
qw[6]: db3 start
qw[7]: db3 end
The MSB of each qword is the valid bit and the lower 63 bits are
the zpass count for that DB.
OQ on RV740 is disabled at the moment as it only seems to report
results for half of its DBs. This needs further investigation.
Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
|
|
Adapted from osiris' version on his tree.
|
|
|
|
|
|
Still need to actually get reference info from winsys somehow. Doing added
buffers is easy, but knowing whether a flush has happened is a bit tricky.
|
|
Nothing strange here.
|
|
Signed-off-by: Corbin Simpson <MostAwesomeDude@gmail.com>
DSA really needs its head examined someday. ~ C.
|
|
Signed-off-by: Corbin Simpson <MostAwesomeDude@gmail.com>
|
|
Signed-off-by: Corbin Simpson <MostAwesomeDude@gmail.com>
|
|
Signed-off-by: Corbin Simpson <MostAwesomeDude@gmail.com>
|
|
Conflicts:
src/mesa/shader/lex.yy.c
src/mesa/shader/program_lexer.l
|
|
|
|
|
|
Signed-off-by: Brian Paul <brianp@vmware.com>
|
|
Basically mimic the llvm 2.6 way of linking execution engines and
targets.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Anything that matched IDENTIFIER was strdup'ed and returned to the
parser. However, almost every case of IDENTIFIER in the parser just
dropped the returned string on the floor. Every swizzle string, every
option string, every use of a variable, etc. leaked memory.
Create a temporary buffer in the parser state (string_dumpster and
dumpster_size). Return strings from the lexer to the parser in the
buffer. Grow the buffer as needed. When the parser needs to keep a
string (i.e., delcaring a new variable), let it make a copy then.
The only leak that valgrind now detects is /occasionally/ the copy of
the program string in gl_program::String is leaked. I'm not seeing
how. :(
|
|
The program string is kept in the program object. On the second call
into glProgramStringARB the previous kept string would be leaked.
|
|
|
|
Bug #24435
(cherry picked from commit d56125a298106d81e10674f1c4b3b43b51a5139d)
|
|
This fixes the second part of bug 23552.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Plus add code for verbose/debugging.
|
|
|
|
Use src->draw_offset intead of zero. Zero usually worked, except when
the src renderbuffer is actually a texture mipmap level higher than zero.
Fixes progs/test/blitfb.c test.
|
|
A slightly modified version of a patch from Vinson Lee.
|
|
Bug #24734.
|
|
|
|
This reverts commit 8810b8f67135185d1044746bb861fe2ff997626c.
It turns out the i965 driver uses the intel->Fallback field as a boolean,
not as a bitmask. The intelFallback() function is a no-op in the i965
driver. It would have been nice if there were some comments about this.
I'll fix that next...
|
|
This would only be hit if we got and invalid index_size.
|
|
Need to push texture state and polygon state too.
Fixes rendering glitches seen in progs/demos/engine when changing
the rendering mode (wireframe, texture modes).
This makes bitmap rendering a little slower, unfortunately.
|
|
Some cards have crippling defaults set and use only 4
of 32 lanes. This should activate 16 on these.
Those that allow 32 by default should still do so.
Found out by Marcin Kościelnicki.
|
|
Use src->draw_offset intead of zero. Zero usually worked, except when
the src renderbuffer is actually a texture mipmap level higher than zero.
Fixes progs/test/blitfb.c test.
|
|
A slightly modified version of a patch from Vinson Lee.
|