Age | Commit message (Collapse) | Author |
|
The state tracker doesn't have to directly call winsys->wait_idle() anymore.
glFlush and glFinish both go through pipe->flush() now.
|
|
The winsys object is now per-screen and shared by multiple contexts.
The regionPool is now part of the i915 winsys layer.
The winsys wait_idle() and flush_frontbuffer() funcs will get more attention...
|
|
|
|
|
|
driFenceFinish() call
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
depth_value(). Special-case it.
|
|
|
|
It doesn't seem to be needed and disabling it fixes a big memory leak with
some programs like xdemos/wincopy.c
|
|
|
|
|
|
|
|
|
|
pipe_surface now has a pointer to the winsys which create/owns the surface.
This allows clean surface deallocation w/out a rendering context.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Fixes failed assertion in glxgears.
|
|
|
|
|
|
|
|
|
|
Use the "dummyContext" pointer (for now) instead of GET_CURRENT_CONTEXT().
|
|
Also, in st_init_limits(), clamp driver's values against Mesa's internal limits.
|
|
|