happen before any state had been set, causing a hang later on. Fix this by
calling r200Flush instead of FIREVERTICES (which checks if any state has been
emitted but not flushed, before calling Flush) in r200Clear. While here, add
some more debugging info which was useful, and remove an unnecessary
save/restore in BackUpAndEmit.
it's not that big of a deal in more normal apps, and axes a good bit of code.
And I assume that t_vertex will only get faster. Removes ~43k from compiled
binary.
Tested with: quake3, ut, ipers, texcyl, chromium, tuxracer, neverball (kinda)
scratch. There were just too many cut-and-paste errors in the code,
and it was too hard to follow.
This fixes Mesa bugzilla #1508. #1509 is probably also fixed, but
I'd like to do some more testing before I close that particular bug.
Additionally, this cuts almost 4,600 lines of code from the driver.
state with a function. This function sets the bits correctly. Did some
trivial refactoring on some of the GL_COMBINE code.
First pass at replacing classic texture environments with GL_COMBINE.
This is controlled by the EXPERIMENTAL_COMBINE_MODE define.
for GL_ARB_texture_mirrored_repeat. Enabled GL_NV_blend_square. It has
always actually been supported. Removed redundant EXT versions of ARB
extension strings.
state in a ready-to-emit cmdbuf, which avoids the issue Nicolai Haehnle reported
where the check() could return differently during backup-and-emit than it should
have if it were called at the right time. Move the lit emission before most of
the TCL state emission on r200, which fixes neverball issues.
Tested with: r100/r200 with neverball, tuxracer, chromium, quake3, ipers
linux-dri. Turn on -O by default (-O2 provides a slight improvement in
performance, at a large cost to debuggability). Turn off -fPIC on x86 by
default. Turn on -Wall, to catch more stupid mistakes. This could be both done
cleaner, and done for more config files, but it works for me for now.
a new cmdbuf, to ensure that state wasn't lost across UNLOCK/LOCK pairs (in the
case of context switching). This was rather inefficient. Instead, after
flushing a cmdbuf, mark the state as needing to be saved on unlock. Then, at
the beginning of flushing a cmdbuf, if we actually have lost the context, go
back and emit a new cmdbuf with the full set of state, before continuing with
the cmdbuf flush. Provides a 10-15% improvement in ipers performance in my
tests, along with other apps.
Tested with: ipers, glxgears, quake3
Brian Paul
Eric Anholt
Ian Romanick
Karl Schultz
Dave Airlie
Nicolai Haehnle
Daniel Borca