Subzero: Use the linear-scan register allocator for Om1 as well.
This removes the need for Om1's postLower() code which did its own ad-hoc register allocation. And it actually speeds up Om1 translation significantly.
This mode of register allocation only allocates for infinite-weight Variables, while respecting live ranges of pre-colored Variables.
BUG= none
R=jvoung@chromium.org
Review URL: https://codereview.chromium.org/733643005
diff --git a/src/IceDefs.h b/src/IceDefs.h
index 9812e3e..a28da4b 100644
--- a/src/IceDefs.h
+++ b/src/IceDefs.h
@@ -102,6 +102,11 @@
Liveness_Intervals
};
+enum RegAllocKind {
+ RAK_Global, // full, global register allocation
+ RAK_InfOnly // allocation only for infinite-weight Variables
+};
+
enum VerboseItem {
IceV_None = 0,
IceV_Instructions = 1 << 0,
