RegAlloc superpass: includes phi elimination, coalescing, and scheduling.
[oota-llvm.git] / test / CodeGen / X86 / inline-asm-tied.ll
index 79b688551eb9fb29809587483618c3195ec38d5b..91576fb09ec29f4a91d32d0eba318d9a895dcf00 100644 (file)
@@ -1,6 +1,8 @@
-; RUN: llc < %s -mtriple=i386-apple-darwin9 -O0 -regalloc=linearscan | grep {movl      %edx, 4(%esp)} | count 2
+; RUN: llc < %s -mtriple=i386-apple-darwin9 -O0 -optimize-regalloc -regalloc=basic | FileCheck %s
 ; rdar://6992609
 
+; CHECK: movl [[EDX:%e..]], 4(%esp)
+; CHECK: movl [[EDX]], 4(%esp)
 target triple = "i386-apple-darwin9.0"
 @llvm.used = appending global [1 x i8*] [i8* bitcast (i64 (i64)* @_OSSwapInt64 to i8*)], section "llvm.metadata"               ; <[1 x i8*]*> [#uses=0]