llvmpipe: allocate regular texture memory upfront
authorRoland Scheidegger <sroland@vmware.com>
Tue, 1 Jul 2014 15:06:48 +0000 (17:06 +0200)
committerRoland Scheidegger <sroland@vmware.com>
Tue, 1 Jul 2014 23:55:59 +0000 (01:55 +0200)
commit90abdc15414f44862a82fe2e53e5419f8182b9ac
tree763d25e288b01a6254c90acfda3db2e23a86fff8
parent7e1521f191afe30fcf02a25b68fbf89278f014d3
llvmpipe: allocate regular texture memory upfront

The deferred allocation doesn't really make much sense anymore, since we no
longer allocate swizzled/linear memory in chunks and not per level / slice
neither.
This means we could fail resource creation a bit more (could already fail in
theory anyway) but should not fail maps later (right now, callers can't deal
with neither really).

Reviewed-by: Brian Paul <brianp@vmware.com>
src/gallium/drivers/llvmpipe/lp_texture.c