draw: use float version of LLVM Mul/Add instructions
authorBrian Paul <brianp@vmware.com>
Wed, 20 Oct 2010 20:54:38 +0000 (14:54 -0600)
committerBrian Paul <brianp@vmware.com>
Wed, 20 Oct 2010 20:56:42 +0000 (14:56 -0600)
commitc492066071c17e55740f3eed69b3344e6f1793ff
treeeb840d7e1b0472d3aee0e73c69d5493e4814e5b6
parentf36346c11662901cc2f2b2239d9adb38bbdc54b6
draw: use float version of LLVM Mul/Add instructions

LLVM 2.8 is pickier about int vs float instructions and operands.
src/gallium/auxiliary/draw/draw_llvm.c