| Age | Commit message (Collapse) | Author |
|---|
|
probably the last match-gallium-upstream merge for a bit, some cleanup+nv50
work coming RSN...
|
|
|
|
|
|
|
|
|
|
|
|
Conflicts:
src/gallium/include/pipe/p_util.h
|
|
Apparently gcc will omit to copy hidden padding bytes under some circumstances,
which means later on memcmp() will indicate a difference between structs
even though all the visible members are identical.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
GPU directly.
|
|
It is really the caller responsibility not to call pipebuffer with null
buffers, etc. But don't let the crash happen here, and still asserting
early.
|
|
In pipebuffer, fencing is done at on a level above sub-allocation, so no
matter how many times slab allocator retries no buffer will be freed. The
pipebuffer fencing implemention already retries allocating.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
also some minor cleanups
|
|
The caller can then free the token array immediately.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
