nv50/ir: rewrite the register allocator as GCRA, with spilling
authorChristoph Bumiller <e0425955@student.tuwien.ac.at>
Mon, 9 Apr 2012 18:58:39 +0000 (20:58 +0200)
committerChristoph Bumiller <e0425955@student.tuwien.ac.at>
Sat, 14 Apr 2012 19:54:03 +0000 (21:54 +0200)
commite43a3a66a9d8a99021d76ff4d07dec7b8cfd62ca
tree7bfd51391b8a2222c57e821cbcbe052187352939
parent99319328d4c249090d3e4a387f6bdc00f711b688
nv50/ir: rewrite the register allocator as GCRA, with spilling

This is more flexible than the linear scan, and we don't need the
separate allocation pass for constrained values anymore.
src/gallium/drivers/nv50/codegen/nv50_ir.cpp
src/gallium/drivers/nv50/codegen/nv50_ir.h
src/gallium/drivers/nv50/codegen/nv50_ir_bb.cpp
src/gallium/drivers/nv50/codegen/nv50_ir_from_tgsi.cpp
src/gallium/drivers/nv50/codegen/nv50_ir_inlines.h
src/gallium/drivers/nv50/codegen/nv50_ir_peephole.cpp
src/gallium/drivers/nv50/codegen/nv50_ir_print.cpp
src/gallium/drivers/nv50/codegen/nv50_ir_ra.cpp
src/gallium/drivers/nv50/codegen/nv50_ir_util.cpp
src/gallium/drivers/nv50/codegen/nv50_ir_util.h
src/gallium/drivers/nvc0/codegen/nv50_ir_emit_nvc0.cpp