[Mesa-dev] [PATCH 00/11] radeonsi: rasterization/clipping fixes and cleanups

[Nicolai Hähnle]

Hi all,

the situation of wide points and lines in ES is a bit confused since
EXT_primitive_bounding_box (and ES 3.2, which promotes that to core).

The traditional language surrounding clipping says that points are
discarded when their coordinates are outside the viewport, and similar
for lines. This means that wide points will "pop out" as soon as they
leave the viewport.

On the other hand, the language for the primitive bounding box says
that the window clip coordinates will be expanded to account for point
size and line width, and there are dEQP tests that rely on it. The
spec language is pretty badly written in my opinion, and I'm pushing
for a clarification (https://gitlab.khronos.org/opengl/API/issues/48),
but in the expectation that it'll be resolved in favor of *not*
discarding wide points and lines that are only partially visible,
this is what the present series does.

This series also contains a whole bunch of cleanups following the
recent fork from r600, moving the viewport code to the radeonsi
directory.

Please review!
Thanks,
Nicolai
--
 src/gallium/drivers/radeon/Makefile.sources  |   1 -
 .../drivers/radeon/r600_pipe_common.c        |   8 +-
 .../drivers/radeon/r600_pipe_common.h        |  37 --
 .../drivers/radeonsi/Makefile.sources        |   3 +-
 src/gallium/drivers/radeonsi/si_blit.c       |   4 +-
 src/gallium/drivers/radeonsi/si_hw_context.c |  10 +-
 src/gallium/drivers/radeonsi/si_pipe.c       |   3 +-
 src/gallium/drivers/radeonsi/si_pipe.h       |  32 ++
 src/gallium/drivers/radeonsi/si_state.c      |  27 +-
 src/gallium/drivers/radeonsi/si_state.h      |   2 +
 src/gallium/drivers/radeonsi/si_state_draw.c |  18 +-
 .../drivers/radeonsi/si_state_shaders.c      |   6 +-
 .../si_viewport.c}                           | 300 +++++++++--------
 13 files changed, 235 insertions(+), 216 deletions(-)