If a vertex shader uses gl_Vertex, gl_Normal, etc, we need to include them
when the user queries the list of active attributes. Before this we were
just including the user-defined attributes.
This saves mapping the index buffer to get a bounds on the indices that
drivers just drop on the floor in the VBO case (cache win), saves a bonus
walk of the indices in the CheckArrayBounds case, and other miscellaneous
validation. On intel it's a particularly a large win (50-100% in my app)
because even though we let the indices stay in both CPU and GPU caches, we
still end up waiting for the GPU to be done with the buffer before reading
from it.
Drivers that want the min/max_index fields must now check index_bounds_valid
and use vbo_get_minmax_index before using them.
Reference counting protects DRM lock call from recursive locking that would
cause hang. Code also adds optional debugging output for recursive call that
is compiled only if NDEBUG is not defined.
This code is not 100% thread safe because mesa doesn't include increment and
test atomic operation. There is built-in gcc functions but they are only
available from gcc 4.2.
The old GLX dispatch table stuff isn't needed (same story for the Mesa/Xlib
driver). The intention of that code was being able to switch on the fly
between the real GLX library and the fake/Xlib-based emulation. That hasn't
been used in a long time.
Next up: some file renaming.
Thread infos and displays are usually not freed by applications. This
commit add atexit calls to free them.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Current thread management was initialized in _eglInitGlobals, which is
called only in eglGetDisplay. Since EGL does not require eglGetDisplay
to be called first, the initialization is better to be done on demand.
_eglFiniCurrent is removed, as it is not called at all.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
This is done through pthread TSD destructor. It destroys all thread
infos except for main thread's. The thread info of the main thread is
destroyed by _eglFiniCurrent.
TLS case is not supported yet.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
We need aditional meta data about the usage of the surface
in softpipe because we need to be able tell the diffrence
between PRIMARY and DISPLAY_TARGET surfaces.
This reverts part of commit 2c9812e3d3.
The calls to _mesa_set_vp_override() were causing extra state validation
and caused the gallium state tracker's bitmap cache to get flushed on
every call.
