Vendor import of llvm trunk r161861:

http://llvm.org/svn/llvm-project/llvm/trunk@161861

56 files changed, 772 insertions, 97 deletions

diff --git a/test/CodeGen/PowerPC/2005-09-02-LegalizeDuplicatesCalls.ll b/test/CodeGen/PowerPC/2005-09-02-LegalizeDuplicatesCalls.ll
index 5d1df46..4373660 100644
--- a/test/CodeGen/PowerPC/2005-09-02-LegalizeDuplicatesCalls.ll
+++ b/test/CodeGen/PowerPC/2005-09-02-LegalizeDuplicatesCalls.ll
@@ -1,7 +1,7 @@
 ; This function should have exactly one call to fixdfdi, no more!
 
 ; RUN: llc < %s -march=ppc32 -mattr=-64bit | \
-; RUN:    grep {bl .*fixdfdi} | count 1
+; RUN:    grep "bl .*fixdfdi" | count 1
 
 define double @test2(double %tmp.7705) {
         %mem_tmp.2.0.in = fptosi double %tmp.7705 to i64                ; <i64> [#uses=1]
diff --git a/test/CodeGen/PowerPC/2006-01-11-darwin-fp-argument.ll b/test/CodeGen/PowerPC/2006-01-11-darwin-fp-argument.ll
index 97bb48e..aeb28af 100644
--- a/test/CodeGen/PowerPC/2006-01-11-darwin-fp-argument.ll
+++ b/test/CodeGen/PowerPC/2006-01-11-darwin-fp-argument.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s | not grep {, f1}
+; RUN: llc < %s | not grep ", f1"
 
 target datalayout = "E-p:32:32"
 target triple = "powerpc-apple-darwin8.2.0"
diff --git a/test/CodeGen/PowerPC/2006-04-05-splat-ish.ll b/test/CodeGen/PowerPC/2006-04-05-splat-ish.ll
index 969772e..7e84538 100644
--- a/test/CodeGen/PowerPC/2006-04-05-splat-ish.ll
+++ b/test/CodeGen/PowerPC/2006-04-05-splat-ish.ll
@@ -1,5 +1,5 @@
 ; RUN: llc < %s -march=ppc32 -mtriple=powerpc-apple-darwin8 -mcpu=g5 | \
-; RUN:   grep {vspltish v.*, 10}
+; RUN:   grep "vspltish v.*, 10"
 
 define void @test(<8 x i16>* %P) {
         %tmp = load <8 x i16>* %P               ; <<8 x i16>> [#uses=1]
diff --git a/test/CodeGen/PowerPC/2007-04-24-InlineAsm-I-Modifier.ll b/test/CodeGen/PowerPC/2007-04-24-InlineAsm-I-Modifier.ll
index 86fd947..73736c5 100644
--- a/test/CodeGen/PowerPC/2007-04-24-InlineAsm-I-Modifier.ll
+++ b/test/CodeGen/PowerPC/2007-04-24-InlineAsm-I-Modifier.ll
@@ -1,5 +1,5 @@
-; RUN: llc < %s -march=ppc32 -mtriple=powerpc-apple-darwin8.8.0 | grep {foo r3, r4}
-; RUN: llc < %s -march=ppc32 -mtriple=powerpc-apple-darwin8.8.0 | grep {bari r3, 47}
+; RUN: llc < %s -march=ppc32 -mtriple=powerpc-apple-darwin8.8.0 | grep "foo r3, r4"
+; RUN: llc < %s -march=ppc32 -mtriple=powerpc-apple-darwin8.8.0 | grep "bari r3, 47"
 
 ; PR1351
 
diff --git a/test/CodeGen/PowerPC/2007-04-30-InlineAsmEarlyClobber.ll b/test/CodeGen/PowerPC/2007-04-30-InlineAsmEarlyClobber.ll
index 3489477..53231b4 100644
--- a/test/CodeGen/PowerPC/2007-04-30-InlineAsmEarlyClobber.ll
+++ b/test/CodeGen/PowerPC/2007-04-30-InlineAsmEarlyClobber.ll
@@ -1,5 +1,5 @@
 ; RUN: llc < %s | FileCheck %s
-; RUN: llc < %s -regalloc=fast | FileCheck %s
+; RUN: llc < %s -regalloc=fast -optimize-regalloc=0 | FileCheck %s
 ; The first argument of subfc must not be the same as any other register.
 
 ; CHECK: subfc [[REG:r.]],
diff --git a/test/CodeGen/PowerPC/2007-05-22-tailmerge-3.ll b/test/CodeGen/PowerPC/2007-05-22-tailmerge-3.ll
index c141551..382ba1f 100644
--- a/test/CodeGen/PowerPC/2007-05-22-tailmerge-3.ll
+++ b/test/CodeGen/PowerPC/2007-05-22-tailmerge-3.ll
@@ -1,8 +1,8 @@
-; RUN: llc < %s -march=ppc32 | grep bl.*baz | count 2
-; RUN: llc < %s -march=ppc32 | grep bl.*quux | count 2
-; RUN: llc < %s -march=ppc32 -enable-tail-merge | grep bl.*baz | count 1
-; RUN: llc < %s -march=ppc32 -enable-tail-merge=1 | grep bl.*quux | count 1
-; Check that tail merging is not the default on ppc, and that -enable-tail-merge works.
+; RUN: llc < %s -march=ppc32 -enable-tail-merge=0 | grep bl.*baz | count 2
+; RUN: llc < %s -march=ppc32 -enable-tail-merge=0 | grep bl.*quux | count 2
+; RUN: llc < %s -march=ppc32 | grep bl.*baz | count 1
+; RUN: llc < %s -march=ppc32 | grep bl.*quux | count 1
+; Check that tail merging is the default on ppc, and that -enable-tail-merge works.
 
 ; ModuleID = 'tail.c'
 target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64"
diff --git a/test/CodeGen/PowerPC/2007-05-30-dagcombine-miscomp.ll b/test/CodeGen/PowerPC/2007-05-30-dagcombine-miscomp.ll
index 72e93a9..b85792c 100644
--- a/test/CodeGen/PowerPC/2007-05-30-dagcombine-miscomp.ll
+++ b/test/CodeGen/PowerPC/2007-05-30-dagcombine-miscomp.ll
@@ -1,7 +1,7 @@
 target datalayout = "E-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64"
 target triple = "powerpc-apple-darwin8.8.0"
 
-; RUN: llc < %s -march=ppc32 | grep {rlwinm r3, r3, 23, 30, 30}
+; RUN: llc < %s -march=ppc32 | grep "rlwinm r3, r3, 23, 30, 30"
 ; PR1473
 
 define zeroext i8 @foo(i16 zeroext  %a)   {
diff --git a/test/CodeGen/PowerPC/2007-10-21-LocalRegAllocAssert.ll b/test/CodeGen/PowerPC/2007-10-21-LocalRegAllocAssert.ll
index 556a4a1c40..a60d11c 100644
--- a/test/CodeGen/PowerPC/2007-10-21-LocalRegAllocAssert.ll
+++ b/test/CodeGen/PowerPC/2007-10-21-LocalRegAllocAssert.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s -mtriple=powerpc64-apple-darwin9 -regalloc=fast -relocation-model=pic
+; RUN: llc < %s -mtriple=powerpc64-apple-darwin9 -regalloc=fast -optimize-regalloc=0 -relocation-model=pic
 
 	%struct.NSError = type opaque
 	%struct.NSManagedObjectContext = type opaque
diff --git a/test/CodeGen/PowerPC/2007-10-21-LocalRegAllocAssert2.ll b/test/CodeGen/PowerPC/2007-10-21-LocalRegAllocAssert2.ll
index b3b9280..3d1a328 100644
--- a/test/CodeGen/PowerPC/2007-10-21-LocalRegAllocAssert2.ll
+++ b/test/CodeGen/PowerPC/2007-10-21-LocalRegAllocAssert2.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s -mtriple=powerpc64-apple-darwin9 -regalloc=fast -relocation-model=pic
+; RUN: llc < %s -mtriple=powerpc64-apple-darwin9 -regalloc=fast -optimize-regalloc=0 -relocation-model=pic
 
 	%struct.NSError = type opaque
 	%struct.NSManagedObjectContext = type opaque
diff --git a/test/CodeGen/PowerPC/2008-02-09-LocalRegAllocAssert.ll b/test/CodeGen/PowerPC/2008-02-09-LocalRegAllocAssert.ll
index e03bd9e..e28a3e0 100644
--- a/test/CodeGen/PowerPC/2008-02-09-LocalRegAllocAssert.ll
+++ b/test/CodeGen/PowerPC/2008-02-09-LocalRegAllocAssert.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s -mtriple=powerpc-apple-darwin -regalloc=fast
+; RUN: llc < %s -mtriple=powerpc-apple-darwin -regalloc=fast -optimize-regalloc=0
 
 define i32 @bork(i64 %foo, i64 %bar) {
 entry:
diff --git a/test/CodeGen/PowerPC/2009-08-17-inline-asm-addr-mode-breakage.ll b/test/CodeGen/PowerPC/2009-08-17-inline-asm-addr-mode-breakage.ll
index 6a3c440..84aa40c 100644
--- a/test/CodeGen/PowerPC/2009-08-17-inline-asm-addr-mode-breakage.ll
+++ b/test/CodeGen/PowerPC/2009-08-17-inline-asm-addr-mode-breakage.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s -march=ppc32 -mtriple=powerpc-apple-darwin10 -mcpu=g5 | FileCheck %s
+; RUN: llc < %s -march=ppc32 -mtriple=powerpc-apple-darwin10 -mcpu=g5 -disable-ppc-ilp-pref | FileCheck %s
 ; ModuleID = '<stdin>'
 target datalayout = "E-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f128:64:128"
 target triple = "powerpc-apple-darwin10.0"
diff --git a/test/CodeGen/PowerPC/2010-03-09-indirect-call.ll b/test/CodeGen/PowerPC/2010-03-09-indirect-call.ll
index 6b31397..0003a17 100644
--- a/test/CodeGen/PowerPC/2010-03-09-indirect-call.ll
+++ b/test/CodeGen/PowerPC/2010-03-09-indirect-call.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s -march=ppc32 -mcpu=g5 -mtriple=powerpc-apple-darwin10.0 -join-physregs | FileCheck %s
+; RUN: llc < %s -march=ppc32 -mcpu=g5 -mtriple=powerpc-apple-darwin10.0 | FileCheck %s
 ; ModuleID = 'nn.c'
 target datalayout = "E-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f128:64:128"
 target triple = "powerpc-apple-darwin11.0"
@@ -9,7 +9,9 @@ target triple = "powerpc-apple-darwin11.0"
 
 define void @foo() nounwind ssp {
 entry:
-; CHECK: mtctr r12
+; Better: mtctr r12
+; CHECK: mr r12, [[REG:r[0-9]+]]
+; CHECK: mtctr [[REG]]
   %0 = load void (...)** @p, align 4              ; <void (...)*> [#uses=1]
   call void (...)* %0() nounwind
   br label %return
diff --git a/test/CodeGen/PowerPC/2011-12-05-NoSpillDupCR.ll b/test/CodeGen/PowerPC/2011-12-05-NoSpillDupCR.ll
index 6161b55..47d985c 100644
--- a/test/CodeGen/PowerPC/2011-12-05-NoSpillDupCR.ll
+++ b/test/CodeGen/PowerPC/2011-12-05-NoSpillDupCR.ll
@@ -1,5 +1,5 @@
-; RUN: llc < %s -mtriple=powerpc-apple-darwin -mcpu=g4 | FileCheck %s
-; RUN: llc < %s -mtriple=powerpc64-unknown-linux-gnu -mcpu=g4 | FileCheck %s
+; RUN: llc < %s -mtriple=powerpc-apple-darwin -mcpu=g4 -disable-ppc-ilp-pref | FileCheck %s
+; RUN: llc < %s -mtriple=powerpc64-unknown-linux-gnu -mcpu=g4 -disable-ppc-ilp-pref | FileCheck %s
 
 ; ModuleID = 'tsc.c'
 target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v128:128:128-n32:64"
diff --git a/test/CodeGen/PowerPC/Frames-leaf.ll b/test/CodeGen/PowerPC/Frames-leaf.ll
index c2e1d6b..7b1c464 100644
--- a/test/CodeGen/PowerPC/Frames-leaf.ll
+++ b/test/CodeGen/PowerPC/Frames-leaf.ll
@@ -1,35 +1,35 @@
 ; RUN: llc < %s -march=ppc32 | \
-; RUN:   not grep {stw r31, 20(r1)}
+; RUN:   not grep "stw r31, 20(r1)"
 ; RUN: llc < %s -march=ppc32 | \
-; RUN:   not grep {stwu r1, -.*(r1)}
+; RUN:   not grep "stwu r1, -.*(r1)"
 ; RUN: llc < %s -march=ppc32 | \
-; RUN:   not grep {addi r1, r1, }
+; RUN:   not grep "addi r1, r1, "
 ; RUN: llc < %s -march=ppc32 | \
-; RUN:   not grep {lwz r31, 20(r1)}
+; RUN:   not grep "lwz r31, 20(r1)"
 ; RUN: llc < %s -march=ppc32 -disable-fp-elim | \
-; RUN:   not grep {stw r31, 20(r1)}
+; RUN:   not grep "stw r31, 20(r1)"
 ; RUN: llc < %s -march=ppc32 -disable-fp-elim | \
-; RUN:   not grep {stwu r1, -.*(r1)}
+; RUN:   not grep "stwu r1, -.*(r1)"
 ; RUN: llc < %s -march=ppc32 -disable-fp-elim | \
-; RUN:   not grep {addi r1, r1, }
+; RUN:   not grep "addi r1, r1, "
 ; RUN: llc < %s -march=ppc32 -disable-fp-elim | \
-; RUN:   not grep {lwz r31, 20(r1)}
+; RUN:   not grep "lwz r31, 20(r1)"
 ; RUN: llc < %s -march=ppc64 | \
-; RUN:   not grep {std r31, 40(r1)}
+; RUN:   not grep "std r31, 40(r1)"
 ; RUN: llc < %s -march=ppc64 | \
-; RUN:   not grep {stdu r1, -.*(r1)}
+; RUN:   not grep "stdu r1, -.*(r1)"
 ; RUN: llc < %s -march=ppc64 | \
-; RUN:   not grep {addi r1, r1, }
+; RUN:   not grep "addi r1, r1, "
 ; RUN: llc < %s -march=ppc64 | \
-; RUN:   not grep {ld r31, 40(r1)}
+; RUN:   not grep "ld r31, 40(r1)"
 ; RUN: llc < %s -march=ppc64 -disable-fp-elim | \
-; RUN:   not grep {stw r31, 40(r1)}
+; RUN:   not grep "stw r31, 40(r1)"
 ; RUN: llc < %s -march=ppc64 -disable-fp-elim | \
-; RUN:   not grep {stdu r1, -.*(r1)}
+; RUN:   not grep "stdu r1, -.*(r1)"
 ; RUN: llc < %s -march=ppc64 -disable-fp-elim | \
-; RUN:   not grep {addi r1, r1, }
+; RUN:   not grep "addi r1, r1, "
 ; RUN: llc < %s -march=ppc64 -disable-fp-elim | \
-; RUN:   not grep {ld r31, 40(r1)}
+; RUN:   not grep "ld r31, 40(r1)"
 
 define i32* @f1() {
         %tmp = alloca i32, i32 2                ; <i32*> [#uses=1]
diff --git a/test/CodeGen/PowerPC/Frames-small.ll b/test/CodeGen/PowerPC/Frames-small.ll
index ecd5ecd..0f6bd10 100644
--- a/test/CodeGen/PowerPC/Frames-small.ll
+++ b/test/CodeGen/PowerPC/Frames-small.ll
@@ -1,26 +1,26 @@
 ; RUN: llc < %s -march=ppc32 -mtriple=powerpc-apple-darwin8 -o %t1
-; RUN: not grep {stw r31, -4(r1)} %t1
-; RUN: grep {stwu r1, -16448(r1)} %t1
-; RUN: grep {addi r1, r1, 16448} %t1
+; RUN: not grep "stw r31, -4(r1)" %t1
+; RUN: grep "stwu r1, -16448(r1)" %t1
+; RUN: grep "addi r1, r1, 16448" %t1
 ; RUN: llc < %s -march=ppc32 | \
-; RUN: not grep {lwz r31, -4(r1)}
+; RUN: not grep "lwz r31, -4(r1)"
 ; RUN: llc < %s -march=ppc32 -mtriple=powerpc-apple-darwin8 -disable-fp-elim \
 ; RUN:   -o %t2
-; RUN: grep {stw r31, -4(r1)} %t2
-; RUN: grep {stwu r1, -16448(r1)} %t2
-; RUN: grep {addi r1, r1, 16448} %t2
-; RUN: grep {lwz r31, -4(r1)} %t2
+; RUN: grep "stw r31, -4(r1)" %t2
+; RUN: grep "stwu r1, -16448(r1)" %t2
+; RUN: grep "addi r1, r1, 16448" %t2
+; RUN: grep "lwz r31, -4(r1)" %t2
 ; RUN: llc < %s -march=ppc64 -mtriple=powerpc-apple-darwin8 -o %t3
-; RUN: not grep {std r31, -8(r1)} %t3
-; RUN: grep {stdu r1, -16496(r1)} %t3
-; RUN: grep {addi r1, r1, 16496} %t3
-; RUN: not grep {ld r31, -8(r1)} %t3
+; RUN: not grep "std r31, -8(r1)" %t3
+; RUN: grep "stdu r1, -16496(r1)" %t3
+; RUN: grep "addi r1, r1, 16496" %t3
+; RUN: not grep "ld r31, -8(r1)" %t3
 ; RUN: llc < %s -march=ppc64 -mtriple=powerpc-apple-darwin8 -disable-fp-elim \
 ; RUN:   -o %t4
-; RUN: grep {std r31, -8(r1)} %t4
-; RUN: grep {stdu r1, -16512(r1)} %t4
-; RUN: grep {addi r1, r1, 16512} %t4
-; RUN: grep {ld r31, -8(r1)} %t4
+; RUN: grep "std r31, -8(r1)" %t4
+; RUN: grep "stdu r1, -16512(r1)" %t4
+; RUN: grep "addi r1, r1, 16512" %t4
+; RUN: grep "ld r31, -8(r1)" %t4
 
 define i32* @f1() {
         %tmp = alloca i32, i32 4095             ; <i32*> [#uses=1]
diff --git a/test/CodeGen/PowerPC/LargeAbsoluteAddr.ll b/test/CodeGen/PowerPC/LargeAbsoluteAddr.ll
index 7b0d69c..6f985c8 100644
--- a/test/CodeGen/PowerPC/LargeAbsoluteAddr.ll
+++ b/test/CodeGen/PowerPC/LargeAbsoluteAddr.ll
@@ -1,9 +1,9 @@
 ; RUN: llc < %s -march=ppc32 -mtriple=powerpc-apple-darwin | \
-; RUN:   grep {stw r4, 32751}
+; RUN:   grep "stw r4, 32751"
 ; RUN: llc < %s -march=ppc64 -mtriple=powerpc-apple-darwin | \
-; RUN:   grep {stw r4, 32751}
+; RUN:   grep "stw r4, 32751"
 ; RUN: llc < %s -march=ppc64 -mtriple=powerpc-apple-darwin | \
-; RUN:   grep {std r4, 9024}
+; RUN:   grep "std r4, 9024"
 
 define void @test() nounwind {
 	store i32 0, i32* inttoptr (i64 48725999 to i32*)
diff --git a/test/CodeGen/PowerPC/a2-fp-basic.ll b/test/CodeGen/PowerPC/a2-fp-basic.ll
index 932ad7a..de3aa7c 100644
--- a/test/CodeGen/PowerPC/a2-fp-basic.ll
+++ b/test/CodeGen/PowerPC/a2-fp-basic.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s -march=ppc64 -mcpu=a2 | FileCheck %s
+; RUN: llc < %s -march=ppc64 -mcpu=a2 -fp-contract=fast | FileCheck %s
 
 %0 = type { double, double }
 
diff --git a/test/CodeGen/PowerPC/and-imm.ll b/test/CodeGen/PowerPC/and-imm.ll
index 64a45e5..6fd484b 100644
--- a/test/CodeGen/PowerPC/and-imm.ll
+++ b/test/CodeGen/PowerPC/and-imm.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s -march=ppc32 | not grep {ori\\|lis}
+; RUN: llc < %s -march=ppc32 | not grep "ori\|lis"
 
 ; andi. r3, r3, 32769	
 define i32 @test(i32 %X) {
diff --git a/test/CodeGen/PowerPC/big-endian-actual-args.ll b/test/CodeGen/PowerPC/big-endian-actual-args.ll
index 009f468..898ad7c 100644
--- a/test/CodeGen/PowerPC/big-endian-actual-args.ll
+++ b/test/CodeGen/PowerPC/big-endian-actual-args.ll
@@ -1,7 +1,7 @@
 ; RUN: llc < %s -march=ppc32 -mtriple=powerpc-unknown-linux-gnu | \
-; RUN:   grep {addc 4, 4, 6}
+; RUN:   grep "addc 4, 4, 6"
 ; RUN: llc < %s -march=ppc32 -mtriple=powerpc-unknown-linux-gnu | \
-; RUN:   grep {adde 3, 3, 5}
+; RUN:   grep "adde 3, 3, 5"
 
 define i64 @foo(i64 %x, i64 %y) {
   %z = add i64 %x, %y
diff --git a/test/CodeGen/PowerPC/big-endian-call-result.ll b/test/CodeGen/PowerPC/big-endian-call-result.ll
index fe85404..760833c 100644
--- a/test/CodeGen/PowerPC/big-endian-call-result.ll
+++ b/test/CodeGen/PowerPC/big-endian-call-result.ll
@@ -1,7 +1,7 @@
 ; RUN: llc < %s -march=ppc32 -mtriple=powerpc-unknown-linux-gnu | \
-; RUN:   grep {addic 4, 4, 1}
+; RUN:   grep "addic 4, 4, 1"
 ; RUN: llc < %s -march=ppc32 -mtriple=powerpc-unknown-linux-gnu | \
-; RUN:   grep {addze 3, 3}
+; RUN:   grep "addze 3, 3"
 
 declare i64 @foo()
 
diff --git a/test/CodeGen/PowerPC/branch-opt.ll b/test/CodeGen/PowerPC/branch-opt.ll
index cc02e40..dda1538 100644
--- a/test/CodeGen/PowerPC/branch-opt.ll
+++ b/test/CodeGen/PowerPC/branch-opt.ll
@@ -1,5 +1,5 @@
 ; RUN: llc < %s -march=ppc32 | \
-; RUN:   grep {b LBB.*} | count 4
+; RUN:   grep "b LBB.*" | count 4
 
 target datalayout = "E-p:32:32"
 target triple = "powerpc-apple-darwin8.7.0"
diff --git a/test/CodeGen/PowerPC/calls.ll b/test/CodeGen/PowerPC/calls.ll
index 29bcb20..dcdda57 100644
--- a/test/CodeGen/PowerPC/calls.ll
+++ b/test/CodeGen/PowerPC/calls.ll
@@ -1,11 +1,11 @@
 ; Test various forms of calls.
 
 ; RUN: llc < %s -march=ppc32 | \
-; RUN:   grep {bl } | count 1
+; RUN:   grep "bl " | count 1
 ; RUN: llc < %s -march=ppc32 | \
-; RUN:   grep {bctrl} | count 1
+; RUN:   grep "bctrl" | count 1
 ; RUN: llc < %s -march=ppc32 | \
-; RUN:   grep {bla } | count 1
+; RUN:   grep "bla " | count 1
 
 declare void @foo()
 
diff --git a/test/CodeGen/PowerPC/coalesce-ext.ll b/test/CodeGen/PowerPC/coalesce-ext.ll
new file mode 100644
index 0000000..cc80f83
--- /dev/null
+++ b/test/CodeGen/PowerPC/coalesce-ext.ll
@@ -0,0 +1,17 @@
+; RUN: llc -march=ppc64 -mtriple=powerpc64-apple-darwin < %s | FileCheck %s
+; Check that the peephole optimizer knows about sext and zext instructions.
+; CHECK: test1sext
+define i32 @test1sext(i64 %A, i64 %B, i32* %P, i64 *%P2) nounwind {
+  %C = add i64 %A, %B
+  ; CHECK: add [[SUM:r[0-9]+]], r3, r4
+  %D = trunc i64 %C to i32
+  %E = shl i64 %C, 32
+  %F = ashr i64 %E, 32
+  ; CHECK: extsw [[EXT:r[0-9]+]], [[SUM]]
+  store volatile i64 %F, i64 *%P2
+  ; CHECK: std [[EXT]]
+  store volatile i32 %D, i32* %P
+  ; Reuse low bits of extended register, don't extend live range of SUM.
+  ; CHECK: stw [[EXT]]
+  ret i32 %D
+}
diff --git a/test/CodeGen/PowerPC/compare-simm.ll b/test/CodeGen/PowerPC/compare-simm.ll
index 92d1dbe..94c5c02 100644
--- a/test/CodeGen/PowerPC/compare-simm.ll
+++ b/test/CodeGen/PowerPC/compare-simm.ll
@@ -1,5 +1,5 @@
 ; RUN: llc < %s -march=ppc32 -mtriple=powerpc-apple-darwin8 | \
-; RUN:   grep {cmpwi cr0, r3, -1}
+; RUN:   grep "cmpwi cr0, r3, -1"
 
 define i32 @test(i32 %x) nounwind {
         %c = icmp eq i32 %x, -1
diff --git a/test/CodeGen/PowerPC/constants.ll b/test/CodeGen/PowerPC/constants.ll
index 8901e02..9efca91 100644
--- a/test/CodeGen/PowerPC/constants.ll
+++ b/test/CodeGen/PowerPC/constants.ll
@@ -4,7 +4,7 @@
 ; RUN: llc < %s -march=ppc32 | \
 ; RUN:   grep ori | count 3
 ; RUN: llc < %s -march=ppc32 | \
-; RUN:   grep {li } | count 4
+; RUN:   grep "li " | count 4
 
 define i32 @f1() {
 entry:
diff --git a/test/CodeGen/PowerPC/ctrloop-reg.ll b/test/CodeGen/PowerPC/ctrloop-reg.ll
new file mode 100644
index 0000000..874e571
--- /dev/null
+++ b/test/CodeGen/PowerPC/ctrloop-reg.ll
@@ -0,0 +1,87 @@
+; ModuleID = 'bugpoint-reduced-simplified.bc'
+target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v128:128:128-n32:64"
+target triple = "powerpc64-unknown-linux-gnu"
+; RUN: llc < %s -march=ppc64 | FileCheck %s
+
+%struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211 = type { %union.v.0.48.90.114.120.138.144.150.156.162.168.174.180.210, i16, i16 }
+%union.v.0.48.90.114.120.138.144.150.156.162.168.174.180.210 = type { i64 }
+%struct.stream_s.5.53.95.119.125.143.149.155.161.167.173.179.185.215 = type { i8*, i8*, i8*, i32, i8, i8, i64, %struct.stream_procs.2.50.92.116.122.140.146.152.158.164.170.176.182.212, i32, %struct._IO_FILE.4.52.94.118.124.142.148.154.160.166.172.178.184.214*, %struct.stream_s.5.53.95.119.125.143.149.155.161.167.173.179.185.215*, i16, i32 }
+%struct.stream_procs.2.50.92.116.122.140.146.152.158.164.170.176.182.212 = type { i32 (%struct.stream_s.5.53.95.119.125.143.149.155.161.167.173.179.185.215*)*, i32 (%struct.stream_s.5.53.95.119.125.143.149.155.161.167.173.179.185.215*, i8)*, i32 (%struct.stream_s.5.53.95.119.125.143.149.155.161.167.173.179.185.215*, i64*)*, i32 (%struct.stream_s.5.53.95.119.125.143.149.155.161.167.173.179.185.215*, i64)*, i32 (%struct.stream_s.5.53.95.119.125.143.149.155.161.167.173.179.185.215*)*, i32 (%struct.stream_s.5.53.95.119.125.143.149.155.161.167.173.179.185.215*)* }
+%struct._IO_FILE.4.52.94.118.124.142.148.154.160.166.172.178.184.214 = type { i32, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, i8*, %struct._IO_marker.3.51.93.117.123.141.147.153.159.165.171.177.183.213*, %struct._IO_FILE.4.52.94.118.124.142.148.154.160.166.172.178.184.214*, i32, i32, i64, i16, i8, [1 x i8], i8*, i64, i8*, i8*, i8*, i8*, i64, i32, [20 x i8] }
+%struct._IO_marker.3.51.93.117.123.141.147.153.159.165.171.177.183.213 = type { %struct._IO_marker.3.51.93.117.123.141.147.153.159.165.171.177.183.213*, %struct._IO_FILE.4.52.94.118.124.142.148.154.160.166.172.178.184.214*, i32 }
+
+@special_ops = external global [7 x i32 (%struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211*)*], align 8
+@ostack = external global [520 x %struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211], align 8
+@osbot = external global %struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211*, align 8
+@osp = external global %struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211*, align 8
+@ostop = external global %struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211*, align 8
+@osp_nargs = external global [6 x %struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211*], align 8
+@estack = external global [150 x %struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211], align 8
+@esp = external global %struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211*, align 8
+@estop = external global %struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211*, align 8
+@dstack = external global [20 x %struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211], align 8
+@dsp = external global %struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211*, align 8
+@dstop = external global %struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211*, align 8
+@name_errordict = external global %struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211
+@name_ErrorNames = external global %struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211
+@error_object = external global %struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211, align 8
+
+declare i32 @zadd(%struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211*)
+
+declare i32 @zdup(%struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211*)
+
+declare i32 @zexch(%struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211*)
+
+declare i32 @zifelse(%struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211*)
+
+declare i32 @zle(%struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211*)
+
+declare i32 @zpop(%struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211*)
+
+declare i32 @zsub(%struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211*)
+
+declare void @interp_init(i32) nounwind
+
+declare void @interp_fix_op(%struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211* nocapture) nounwind
+
+define i32 @interpret(%struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211* %pref, i32 %user_errors) nounwind {
+entry:
+  %erref = alloca %struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211, align 8
+  br i1 undef, label %retry.us, label %retry
+
+retry.us:                                         ; preds = %if.end18, %retry, %retry, %retry, %retry, %entry
+  ret i32 undef
+
+retry:                                            ; preds = %if.end18, %entry
+  %0 = phi %struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211* [ null, %entry ], [ %erref, %if.end18 ]
+  %call = call i32 @interp(%struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211* %0)
+  switch i32 %call, label %if.end18 [
+    i32 -3, label %retry.us
+    i32 -5, label %retry.us
+    i32 -16, label %retry.us
+    i32 -25, label %retry.us
+  ]
+
+if.end18:                                         ; preds = %retry
+  br i1 false, label %retry.us, label %retry
+}
+
+; CHECK: @interpret
+
+declare i32 @interp_exit(%struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211* nocapture) nounwind readnone
+
+declare i32 @interp(%struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211*) nounwind
+
+declare i32 @dict_lookup(%struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211*, %struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211*, %struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211*, %struct.ref_s.1.49.91.115.121.139.145.151.157.163.169.175.181.211**)
+
+declare void @llvm.memcpy.p0i8.p0i8.i64(i8* nocapture, i8* nocapture, i64, i32, i1) nounwind
+
+declare i32 @obj_compare(...)
+
+declare i32 @file_check_read(...)
+
+declare i32 @scan_token(...)
+
+declare i32 @file_close(...)
+
+declare void @sread_string(%struct.stream_s.5.53.95.119.125.143.149.155.161.167.173.179.185.215*, i8*, i32)
diff --git a/test/CodeGen/PowerPC/ctrloop-s000.ll b/test/CodeGen/PowerPC/ctrloop-s000.ll
new file mode 100644
index 0000000..dcea06f
--- /dev/null
+++ b/test/CodeGen/PowerPC/ctrloop-s000.ll
@@ -0,0 +1,156 @@
+; ModuleID = 'tsc_s000.c'
+target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v128:128:128-n32:64"
+target triple = "powerpc64-unknown-linux-gnu"
+; RUN: llc < %s -march=ppc64 | FileCheck %s
+
+@Y = common global [16000 x double] zeroinitializer, align 32
+@X = common global [16000 x double] zeroinitializer, align 32
+@Z = common global [16000 x double] zeroinitializer, align 32
+@U = common global [16000 x double] zeroinitializer, align 32
+@V = common global [16000 x double] zeroinitializer, align 32
+@aa = common global [256 x [256 x double]] zeroinitializer, align 32
+@bb = common global [256 x [256 x double]] zeroinitializer, align 32
+@cc = common global [256 x [256 x double]] zeroinitializer, align 32
+@array = common global [65536 x double] zeroinitializer, align 32
+@x = common global [16000 x double] zeroinitializer, align 32
+@temp = common global double 0.000000e+00, align 8
+@temp_int = common global i32 0, align 4
+@a = common global [16000 x double] zeroinitializer, align 32
+@b = common global [16000 x double] zeroinitializer, align 32
+@c = common global [16000 x double] zeroinitializer, align 32
+@d = common global [16000 x double] zeroinitializer, align 32
+@e = common global [16000 x double] zeroinitializer, align 32
+@tt = common global [256 x [256 x double]] zeroinitializer, align 32
+@indx = common global [16000 x i32] zeroinitializer, align 32
+@xx = common global double* null, align 8
+@yy = common global double* null, align 8
+
+define i32 @s000() nounwind {
+entry:
+  br label %for.cond1.preheader
+
+for.cond1.preheader:                              ; preds = %for.end, %entry
+  %nl.010 = phi i32 [ 0, %entry ], [ %inc7, %for.end ]
+  br label %for.body3
+
+for.body3:                                        ; preds = %for.body3, %for.cond1.preheader
+  %indvars.iv = phi i64 [ 0, %for.cond1.preheader ], [ %indvars.iv.next.15, %for.body3 ]
+  %arrayidx = getelementptr inbounds [16000 x double]* @Y, i64 0, i64 %indvars.iv
+  %0 = load double* %arrayidx, align 32, !tbaa !0
+  %add = fadd double %0, 1.000000e+00
+  %arrayidx5 = getelementptr inbounds [16000 x double]* @X, i64 0, i64 %indvars.iv
+  store double %add, double* %arrayidx5, align 32, !tbaa !0
+  %indvars.iv.next11 = or i64 %indvars.iv, 1
+  %arrayidx.1 = getelementptr inbounds [16000 x double]* @Y, i64 0, i64 %indvars.iv.next11
+  %1 = load double* %arrayidx.1, align 8, !tbaa !0
+  %add.1 = fadd double %1, 1.000000e+00
+  %arrayidx5.1 = getelementptr inbounds [16000 x double]* @X, i64 0, i64 %indvars.iv.next11
+  store double %add.1, double* %arrayidx5.1, align 8, !tbaa !0
+  %indvars.iv.next.112 = or i64 %indvars.iv, 2
+  %arrayidx.2 = getelementptr inbounds [16000 x double]* @Y, i64 0, i64 %indvars.iv.next.112
+  %2 = load double* %arrayidx.2, align 16, !tbaa !0
+  %add.2 = fadd double %2, 1.000000e+00
+  %arrayidx5.2 = getelementptr inbounds [16000 x double]* @X, i64 0, i64 %indvars.iv.next.112
+  store double %add.2, double* %arrayidx5.2, align 16, !tbaa !0
+  %indvars.iv.next.213 = or i64 %indvars.iv, 3
+  %arrayidx.3 = getelementptr inbounds [16000 x double]* @Y, i64 0, i64 %indvars.iv.next.213
+  %3 = load double* %arrayidx.3, align 8, !tbaa !0
+  %add.3 = fadd double %3, 1.000000e+00
+  %arrayidx5.3 = getelementptr inbounds [16000 x double]* @X, i64 0, i64 %indvars.iv.next.213
+  store double %add.3, double* %arrayidx5.3, align 8, !tbaa !0
+  %indvars.iv.next.314 = or i64 %indvars.iv, 4
+  %arrayidx.4 = getelementptr inbounds [16000 x double]* @Y, i64 0, i64 %indvars.iv.next.314
+  %4 = load double* %arrayidx.4, align 32, !tbaa !0
+  %add.4 = fadd double %4, 1.000000e+00
+  %arrayidx5.4 = getelementptr inbounds [16000 x double]* @X, i64 0, i64 %indvars.iv.next.314
+  store double %add.4, double* %arrayidx5.4, align 32, !tbaa !0
+  %indvars.iv.next.415 = or i64 %indvars.iv, 5
+  %arrayidx.5 = getelementptr inbounds [16000 x double]* @Y, i64 0, i64 %indvars.iv.next.415
+  %5 = load double* %arrayidx.5, align 8, !tbaa !0
+  %add.5 = fadd double %5, 1.000000e+00
+  %arrayidx5.5 = getelementptr inbounds [16000 x double]* @X, i64 0, i64 %indvars.iv.next.415
+  store double %add.5, double* %arrayidx5.5, align 8, !tbaa !0
+  %indvars.iv.next.516 = or i64 %indvars.iv, 6
+  %arrayidx.6 = getelementptr inbounds [16000 x double]* @Y, i64 0, i64 %indvars.iv.next.516
+  %6 = load double* %arrayidx.6, align 16, !tbaa !0
+  %add.6 = fadd double %6, 1.000000e+00
+  %arrayidx5.6 = getelementptr inbounds [16000 x double]* @X, i64 0, i64 %indvars.iv.next.516
+  store double %add.6, double* %arrayidx5.6, align 16, !tbaa !0
+  %indvars.iv.next.617 = or i64 %indvars.iv, 7
+  %arrayidx.7 = getelementptr inbounds [16000 x double]* @Y, i64 0, i64 %indvars.iv.next.617
+  %7 = load double* %arrayidx.7, align 8, !tbaa !0
+  %add.7 = fadd double %7, 1.000000e+00
+  %arrayidx5.7 = getelementptr inbounds [16000 x double]* @X, i64 0, i64 %indvars.iv.next.617
+  store double %add.7, double* %arrayidx5.7, align 8, !tbaa !0
+  %indvars.iv.next.718 = or i64 %indvars.iv, 8
+  %arrayidx.8 = getelementptr inbounds [16000 x double]* @Y, i64 0, i64 %indvars.iv.next.718
+  %8 = load double* %arrayidx.8, align 32, !tbaa !0
+  %add.8 = fadd double %8, 1.000000e+00
+  %arrayidx5.8 = getelementptr inbounds [16000 x double]* @X, i64 0, i64 %indvars.iv.next.718
+  store double %add.8, double* %arrayidx5.8, align 32, !tbaa !0
+  %indvars.iv.next.819 = or i64 %indvars.iv, 9
+  %arrayidx.9 = getelementptr inbounds [16000 x double]* @Y, i64 0, i64 %indvars.iv.next.819
+  %9 = load double* %arrayidx.9, align 8, !tbaa !0
+  %add.9 = fadd double %9, 1.000000e+00
+  %arrayidx5.9 = getelementptr inbounds [16000 x double]* @X, i64 0, i64 %indvars.iv.next.819
+  store double %add.9, double* %arrayidx5.9, align 8, !tbaa !0
+  %indvars.iv.next.920 = or i64 %indvars.iv, 10
+  %arrayidx.10 = getelementptr inbounds [16000 x double]* @Y, i64 0, i64 %indvars.iv.next.920
+  %10 = load double* %arrayidx.10, align 16, !tbaa !0
+  %add.10 = fadd double %10, 1.000000e+00
+  %arrayidx5.10 = getelementptr inbounds [16000 x double]* @X, i64 0, i64 %indvars.iv.next.920
+  store double %add.10, double* %arrayidx5.10, align 16, !tbaa !0
+  %indvars.iv.next.1021 = or i64 %indvars.iv, 11
+  %arrayidx.11 = getelementptr inbounds [16000 x double]* @Y, i64 0, i64 %indvars.iv.next.1021
+  %11 = load double* %arrayidx.11, align 8, !tbaa !0
+  %add.11 = fadd double %11, 1.000000e+00
+  %arrayidx5.11 = getelementptr inbounds [16000 x double]* @X, i64 0, i64 %indvars.iv.next.1021
+  store double %add.11, double* %arrayidx5.11, align 8, !tbaa !0
+  %indvars.iv.next.1122 = or i64 %indvars.iv, 12
+  %arrayidx.12 = getelementptr inbounds [16000 x double]* @Y, i64 0, i64 %indvars.iv.next.1122
+  %12 = load double* %arrayidx.12, align 32, !tbaa !0
+  %add.12 = fadd double %12, 1.000000e+00
+  %arrayidx5.12 = getelementptr inbounds [16000 x double]* @X, i64 0, i64 %indvars.iv.next.1122
+  store double %add.12, double* %arrayidx5.12, align 32, !tbaa !0
+  %indvars.iv.next.1223 = or i64 %indvars.iv, 13
+  %arrayidx.13 = getelementptr inbounds [16000 x double]* @Y, i64 0, i64 %indvars.iv.next.1223
+  %13 = load double* %arrayidx.13, align 8, !tbaa !0
+  %add.13 = fadd double %13, 1.000000e+00
+  %arrayidx5.13 = getelementptr inbounds [16000 x double]* @X, i64 0, i64 %indvars.iv.next.1223
+  store double %add.13, double* %arrayidx5.13, align 8, !tbaa !0
+  %indvars.iv.next.1324 = or i64 %indvars.iv, 14
+  %arrayidx.14 = getelementptr inbounds [16000 x double]* @Y, i64 0, i64 %indvars.iv.next.1324
+  %14 = load double* %arrayidx.14, align 16, !tbaa !0
+  %add.14 = fadd double %14, 1.000000e+00
+  %arrayidx5.14 = getelementptr inbounds [16000 x double]* @X, i64 0, i64 %indvars.iv.next.1324
+  store double %add.14, double* %arrayidx5.14, align 16, !tbaa !0
+  %indvars.iv.next.1425 = or i64 %indvars.iv, 15
+  %arrayidx.15 = getelementptr inbounds [16000 x double]* @Y, i64 0, i64 %indvars.iv.next.1425
+  %15 = load double* %arrayidx.15, align 8, !tbaa !0
+  %add.15 = fadd double %15, 1.000000e+00
+  %arrayidx5.15 = getelementptr inbounds [16000 x double]* @X, i64 0, i64 %indvars.iv.next.1425
+  store double %add.15, double* %arrayidx5.15, align 8, !tbaa !0
+  %indvars.iv.next.15 = add i64 %indvars.iv, 16
+  %lftr.wideiv.15 = trunc i64 %indvars.iv.next.15 to i32
+  %exitcond.15 = icmp eq i32 %lftr.wideiv.15, 16000
+  br i1 %exitcond.15, label %for.end, label %for.body3
+
+for.end:                                          ; preds = %for.body3
+  %call = tail call i32 @dummy(double* getelementptr inbounds ([16000 x double]* @X, i64 0, i64 0), double* getelementptr inbounds ([16000 x double]* @Y, i64 0, i64 0), double* getelementptr inbounds ([16000 x double]* @Z, i64 0, i64 0), double* getelementptr inbounds ([16000 x double]* @U, i64 0, i64 0), double* getelementptr inbounds ([16000 x double]* @V, i64 0, i64 0), [256 x double]* getelementptr inbounds ([256 x [256 x double]]* @aa, i64 0, i64 0), [256 x double]* getelementptr inbounds ([256 x [256 x double]]* @bb, i64 0, i64 0), [256 x double]* getelementptr inbounds ([256 x [256 x double]]* @cc, i64 0, i64 0), double 0.000000e+00) nounwind
+  %inc7 = add nsw i32 %nl.010, 1
+  %exitcond = icmp eq i32 %inc7, 400000
+  br i1 %exitcond, label %for.end8, label %for.cond1.preheader
+
+for.end8:                                         ; preds = %for.end
+  ret i32 0
+
+; CHECK: @s000
+; CHECK: mtctr
+; CHECK: bdnz
+}
+
+declare i32 @dummy(double*, double*, double*, double*, double*, [256 x double]*, [256 x double]*, [256 x double]*, double)
+
+!0 = metadata !{metadata !"double", metadata !1}
+!1 = metadata !{metadata !"omnipotent char", metadata !2}
+!2 = metadata !{metadata !"Simple C/C++ TBAA"}
diff --git a/test/CodeGen/PowerPC/ctrloop-sums.ll b/test/CodeGen/PowerPC/ctrloop-sums.ll
new file mode 100644
index 0000000..eae8c38
--- /dev/null
+++ b/test/CodeGen/PowerPC/ctrloop-sums.ll
@@ -0,0 +1,134 @@
+; ModuleID = 'SingleSource/Regression/C/sumarray2d.c'
+target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v128:128:128-n32:64"
+target triple = "powerpc64-unknown-linux-gnu"
+; RUN: llc < %s -march=ppc64 | FileCheck %s
+
+@.str = private unnamed_addr constant [23 x i8] c"Sum(Array[%d,%d] = %d\0A\00", align 1
+
+define i32 @SumArray([100 x i32]* nocapture %Array, i32 %NumI, i32 %NumJ) nounwind readonly {
+entry:
+  %cmp12 = icmp eq i32 %NumI, 0
+  br i1 %cmp12, label %for.end8, label %for.cond1.preheader.lr.ph
+
+for.cond1.preheader.lr.ph:                        ; preds = %entry
+  %cmp29 = icmp eq i32 %NumJ, 0
+  br i1 %cmp29, label %for.inc6, label %for.body3.lr.ph.us
+
+for.inc6.us:                                      ; preds = %for.body3.us
+  %indvars.iv.next17 = add i64 %indvars.iv16, 1
+  %lftr.wideiv18 = trunc i64 %indvars.iv.next17 to i32
+  %exitcond19 = icmp eq i32 %lftr.wideiv18, %NumI
+  br i1 %exitcond19, label %for.end8, label %for.body3.lr.ph.us
+
+for.body3.us:                                     ; preds = %for.body3.us, %for.body3.lr.ph.us
+  %indvars.iv = phi i64 [ 0, %for.body3.lr.ph.us ], [ %indvars.iv.next, %for.body3.us ]
+  %Result.111.us = phi i32 [ %Result.014.us, %for.body3.lr.ph.us ], [ %add.us, %for.body3.us ]
+  %arrayidx5.us = getelementptr inbounds [100 x i32]* %Array, i64 %indvars.iv16, i64 %indvars.iv
+  %0 = load i32* %arrayidx5.us, align 4, !tbaa !0
+  %add.us = add nsw i32 %0, %Result.111.us
+  %indvars.iv.next = add i64 %indvars.iv, 1
+  %lftr.wideiv = trunc i64 %indvars.iv.next to i32
+  %exitcond = icmp eq i32 %lftr.wideiv, %NumJ
+  br i1 %exitcond, label %for.inc6.us, label %for.body3.us
+
+for.body3.lr.ph.us:                               ; preds = %for.inc6.us, %for.cond1.preheader.lr.ph
+  %indvars.iv16 = phi i64 [ %indvars.iv.next17, %for.inc6.us ], [ 0, %for.cond1.preheader.lr.ph ]
+  %Result.014.us = phi i32 [ %add.us, %for.inc6.us ], [ 0, %for.cond1.preheader.lr.ph ]
+  br label %for.body3.us
+
+for.inc6:                                         ; preds = %for.inc6, %for.cond1.preheader.lr.ph
+  %i.013 = phi i32 [ %inc7, %for.inc6 ], [ 0, %for.cond1.preheader.lr.ph ]
+  %inc7 = add i32 %i.013, 1
+  %exitcond20 = icmp eq i32 %inc7, %NumI
+  br i1 %exitcond20, label %for.end8, label %for.inc6
+
+for.end8:                                         ; preds = %for.inc6.us, %for.inc6, %entry
+  %Result.0.lcssa = phi i32 [ 0, %entry ], [ %add.us, %for.inc6.us ], [ 0, %for.inc6 ]
+  ret i32 %Result.0.lcssa
+; CHECK: @SumArray
+; CHECK: mtctr
+; CHECK: bdnz
+}
+
+define i32 @main() nounwind {
+entry:
+  %Array = alloca [100 x [100 x i32]], align 4
+  br label %for.body
+
+for.body:                                         ; preds = %for.body, %entry
+  %indvars.iv33 = phi i64 [ 0, %entry ], [ %indvars.iv.next34, %for.body ]
+  %0 = trunc i64 %indvars.iv33 to i32
+  %sub = sub i32 0, %0
+  %arrayidx2 = getelementptr inbounds [100 x [100 x i32]]* %Array, i64 0, i64 %indvars.iv33, i64 %indvars.iv33
+  store i32 %sub, i32* %arrayidx2, align 4, !tbaa !0
+  %indvars.iv.next34 = add i64 %indvars.iv33, 1
+  %lftr.wideiv35 = trunc i64 %indvars.iv.next34 to i32
+  %exitcond36 = icmp eq i32 %lftr.wideiv35, 100
+  br i1 %exitcond36, label %for.cond6.preheader, label %for.body
+
+for.cond6.preheader:                              ; preds = %for.body, %for.inc17
+  %indvars.iv29 = phi i64 [ %indvars.iv.next30, %for.inc17 ], [ 0, %for.body ]
+  br label %for.body8
+
+for.body8:                                        ; preds = %for.inc14, %for.cond6.preheader
+  %indvars.iv = phi i64 [ 0, %for.cond6.preheader ], [ %indvars.iv.next, %for.inc14 ]
+  %1 = trunc i64 %indvars.iv to i32
+  %2 = trunc i64 %indvars.iv29 to i32
+  %cmp9 = icmp eq i32 %1, %2
+  br i1 %cmp9, label %for.inc14, label %if.then
+
+if.then:                                          ; preds = %for.body8
+  %3 = add i64 %indvars.iv, %indvars.iv29
+  %arrayidx13 = getelementptr inbounds [100 x [100 x i32]]* %Array, i64 0, i64 %indvars.iv29, i64 %indvars.iv
+  %4 = trunc i64 %3 to i32
+  store i32 %4, i32* %arrayidx13, align 4, !tbaa !0
+  br label %for.inc14
+
+for.inc14:                                        ; preds = %for.body8, %if.then
+  %indvars.iv.next = add i64 %indvars.iv, 1
+  %lftr.wideiv27 = trunc i64 %indvars.iv.next to i32
+  %exitcond28 = icmp eq i32 %lftr.wideiv27, 100
+  br i1 %exitcond28, label %for.inc17, label %for.body8
+
+for.inc17:                                        ; preds = %for.inc14
+  %indvars.iv.next30 = add i64 %indvars.iv29, 1
+  %lftr.wideiv31 = trunc i64 %indvars.iv.next30 to i32
+  %exitcond32 = icmp eq i32 %lftr.wideiv31, 100
+  br i1 %exitcond32, label %for.body3.lr.ph.us.i, label %for.cond6.preheader
+
+for.inc6.us.i:                                    ; preds = %for.body3.us.i
+  %indvars.iv.next17.i = add i64 %indvars.iv16.i, 1
+  %lftr.wideiv24 = trunc i64 %indvars.iv.next17.i to i32
+  %exitcond25 = icmp eq i32 %lftr.wideiv24, 100
+  br i1 %exitcond25, label %SumArray.exit, label %for.body3.lr.ph.us.i
+
+for.body3.us.i:                                   ; preds = %for.body3.lr.ph.us.i, %for.body3.us.i
+  %indvars.iv.i = phi i64 [ 0, %for.body3.lr.ph.us.i ], [ %indvars.iv.next.i, %for.body3.us.i ]
+  %Result.111.us.i = phi i32 [ %Result.014.us.i, %for.body3.lr.ph.us.i ], [ %add.us.i, %for.body3.us.i ]
+  %arrayidx5.us.i = getelementptr inbounds [100 x [100 x i32]]* %Array, i64 0, i64 %indvars.iv16.i, i64 %indvars.iv.i
+  %5 = load i32* %arrayidx5.us.i, align 4, !tbaa !0
+  %add.us.i = add nsw i32 %5, %Result.111.us.i
+  %indvars.iv.next.i = add i64 %indvars.iv.i, 1
+  %lftr.wideiv = trunc i64 %indvars.iv.next.i to i32
+  %exitcond = icmp eq i32 %lftr.wideiv, 100
+  br i1 %exitcond, label %for.inc6.us.i, label %for.body3.us.i
+
+for.body3.lr.ph.us.i:                             ; preds = %for.inc17, %for.inc6.us.i
+  %indvars.iv16.i = phi i64 [ %indvars.iv.next17.i, %for.inc6.us.i ], [ 0, %for.inc17 ]
+  %Result.014.us.i = phi i32 [ %add.us.i, %for.inc6.us.i ], [ 0, %for.inc17 ]
+  br label %for.body3.us.i
+
+SumArray.exit:                                    ; preds = %for.inc6.us.i
+  %call20 = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([23 x i8]* @.str, i64 0, i64 0), i32 100, i32 100, i32 %add.us.i) nounwind
+  ret i32 0
+
+; CHECK: @main
+; CHECK: mtctr
+; CHECK: bdnz
+}
+
+declare i32 @printf(i8* nocapture, ...) nounwind
+
+!0 = metadata !{metadata !"int", metadata !1}
+!1 = metadata !{metadata !"omnipotent char", metadata !2}
+!2 = metadata !{metadata !"Simple C/C++ TBAA"}
diff --git a/test/CodeGen/PowerPC/ctrloops.ll b/test/CodeGen/PowerPC/ctrloops.ll
new file mode 100644
index 0000000..4b6f7b9
--- /dev/null
+++ b/test/CodeGen/PowerPC/ctrloops.ll
@@ -0,0 +1,79 @@
+target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v128:128:128-n32:64"
+target triple = "powerpc64-unknown-freebsd10.0"
+; RUN: llc < %s -march=ppc64 | FileCheck %s
+
+@a = common global i32 0, align 4
+
+define void @test1(i32 %c) nounwind {
+entry:
+  br label %for.body
+
+for.body:                                         ; preds = %for.body, %entry
+  %i.01 = phi i32 [ 0, %entry ], [ %inc, %for.body ]
+  %0 = load volatile i32* @a, align 4, !tbaa !0
+  %add = add nsw i32 %0, %c
+  store volatile i32 %add, i32* @a, align 4, !tbaa !0
+  %inc = add nsw i32 %i.01, 1
+  %exitcond = icmp eq i32 %inc, 2048
+  br i1 %exitcond, label %for.end, label %for.body
+
+for.end:                                          ; preds = %for.body
+  ret void
+; CHECK: @test1
+; CHECK-NOT: or 3, 3, 3
+; CHECK: mtctr
+; CHECK-NOT: addi
+; CHECK-NOT: cmplwi
+; CHECK: bdnz
+}
+
+define void @test2(i32 %c, i32 %d) nounwind {
+entry:
+  %cmp1 = icmp sgt i32 %d, 0
+  br i1 %cmp1, label %for.body, label %for.end
+
+for.body:                                         ; preds = %entry, %for.body
+  %i.02 = phi i32 [ %inc, %for.body ], [ 0, %entry ]
+  %0 = load volatile i32* @a, align 4, !tbaa !0
+  %add = add nsw i32 %0, %c
+  store volatile i32 %add, i32* @a, align 4, !tbaa !0
+  %inc = add nsw i32 %i.02, 1
+  %exitcond = icmp eq i32 %inc, %d
+  br i1 %exitcond, label %for.end, label %for.body
+
+for.end:                                          ; preds = %for.body, %entry
+  ret void
+; CHECK: @test2
+; CHECK: mtctr
+; CHECK-NOT: addi
+; CHECK-NOT: cmplwi
+; CHECK: bdnz
+}
+
+define void @test3(i32 %c, i32 %d) nounwind {
+entry:
+  %cmp1 = icmp sgt i32 %d, 0
+  br i1 %cmp1, label %for.body, label %for.end
+
+for.body:                                         ; preds = %entry, %for.body
+  %i.02 = phi i32 [ %inc, %for.body ], [ 0, %entry ]
+  %mul = mul nsw i32 %i.02, %c
+  %0 = load volatile i32* @a, align 4, !tbaa !0
+  %add = add nsw i32 %0, %mul
+  store volatile i32 %add, i32* @a, align 4, !tbaa !0
+  %inc = add nsw i32 %i.02, 1
+  %exitcond = icmp eq i32 %inc, %d
+  br i1 %exitcond, label %for.end, label %for.body
+
+for.end:                                          ; preds = %for.body, %entry
+  ret void
+; CHECK: @test3
+; CHECK: mtctr
+; CHECK-NOT: addi
+; CHECK-NOT: cmplwi
+; CHECK: bdnz
+}
+
+!0 = metadata !{metadata !"int", metadata !1}
+!1 = metadata !{metadata !"omnipotent char", metadata !2}
+!2 = metadata !{metadata !"Simple C/C++ TBAA"}
diff --git a/test/CodeGen/PowerPC/darwin-labels.ll b/test/CodeGen/PowerPC/darwin-labels.ll
index af23369..56f7782 100644
--- a/test/CodeGen/PowerPC/darwin-labels.ll
+++ b/test/CodeGen/PowerPC/darwin-labels.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s | grep {foo bar":}
+; RUN: llc < %s | grep 'foo bar":'
 
 target datalayout = "E-p:32:32"
 target triple = "powerpc-apple-darwin8.2.0"
diff --git a/test/CodeGen/PowerPC/fabs.ll b/test/CodeGen/PowerPC/fabs.ll
index 6ef740f..ddcce745 100644
--- a/test/CodeGen/PowerPC/fabs.ll
+++ b/test/CodeGen/PowerPC/fabs.ll
@@ -1,7 +1,7 @@
-; RUN: llc < %s -march=ppc32 -mtriple=powerpc-apple-darwin | grep {fabs f1, f1}
+; RUN: llc < %s -march=ppc32 -mtriple=powerpc-apple-darwin | grep "fabs f1, f1"
 
 define double @fabs(double %f) {
 entry:
-	%tmp2 = tail call double @fabs( double %f )		; <double> [#uses=1]
+	%tmp2 = tail call double @fabs( double %f ) readnone	; <double> [#uses=1]
 	ret double %tmp2
 }
diff --git a/test/CodeGen/PowerPC/fma.ll b/test/CodeGen/PowerPC/fma.ll
index 815c72c..27496f7 100644
--- a/test/CodeGen/PowerPC/fma.ll
+++ b/test/CodeGen/PowerPC/fma.ll
@@ -1,5 +1,5 @@
-; RUN: llc < %s -march=ppc32 | \
-; RUN:   egrep {fn?madd|fn?msub} | count 8
+; RUN: llc < %s -march=ppc32 -fp-contract=fast | \
+; RUN:   egrep "fn?madd|fn?msub" | count 8
 
 define double @test_FMADD1(double %A, double %B, double %C) {
 	%D = fmul double %A, %B		; <double> [#uses=1]
diff --git a/test/CodeGen/PowerPC/fnabs.ll b/test/CodeGen/PowerPC/fnabs.ll
index bbd5c71..9fa2dcb 100644
--- a/test/CodeGen/PowerPC/fnabs.ll
+++ b/test/CodeGen/PowerPC/fnabs.ll
@@ -3,7 +3,7 @@
 declare double @fabs(double)
 
 define double @test(double %X) {
-        %Y = call double @fabs( double %X )             ; <double> [#uses=1]
+        %Y = call double @fabs( double %X ) readnone     ; <double> [#uses=1]
         %Z = fsub double -0.000000e+00, %Y               ; <double> [#uses=1]
         ret double %Z
 }
diff --git a/test/CodeGen/PowerPC/fsqrt.ll b/test/CodeGen/PowerPC/fsqrt.ll
index 74a8725..bf8c4a2 100644
--- a/test/CodeGen/PowerPC/fsqrt.ll
+++ b/test/CodeGen/PowerPC/fsqrt.ll
@@ -2,13 +2,13 @@
 ; otherwise.
 
 ; RUN: llc < %s -march=ppc32 -mtriple=powerpc-apple-darwin8 -mattr=+fsqrt | \
-; RUN:   grep {fsqrt f1, f1}
+; RUN:   grep "fsqrt f1, f1"
 ; RUN: llc < %s -march=ppc32 -mtriple=powerpc-apple-darwin8 -mcpu=g5 | \
-; RUN:   grep {fsqrt f1, f1}
+; RUN:   grep "fsqrt f1, f1"
 ; RUN: llc < %s -march=ppc32 -mtriple=powerpc-apple-darwin8 -mattr=-fsqrt | \
-; RUN:   not grep {fsqrt f1, f1}
+; RUN:   not grep "fsqrt f1, f1"
 ; RUN: llc < %s -march=ppc32 -mtriple=powerpc-apple-darwin8 -mcpu=g4 | \
-; RUN:   not grep {fsqrt f1, f1}
+; RUN:   not grep "fsqrt f1, f1"
 
 declare double @llvm.sqrt.f64(double)
 
diff --git a/test/CodeGen/PowerPC/iabs.ll b/test/CodeGen/PowerPC/iabs.ll
index a43f09c..7d089bb 100644
--- a/test/CodeGen/PowerPC/iabs.ll
+++ b/test/CodeGen/PowerPC/iabs.ll
@@ -1,5 +1,5 @@
-; RUN: llc < %s -march=ppc32 -stats |& \
-; RUN:   grep {4 .*Number of machine instrs printed}
+; RUN: llc < %s -march=ppc32 -stats 2>&1 | \
+; RUN:   grep "4 .*Number of machine instrs printed"
 
 ;; Integer absolute value, should produce something as good as:
 ;;      srawi r2, r3, 31
diff --git a/test/CodeGen/PowerPC/isel.ll b/test/CodeGen/PowerPC/isel.ll
new file mode 100644
index 0000000..ed494c5
--- /dev/null
+++ b/test/CodeGen/PowerPC/isel.ll
@@ -0,0 +1,23 @@
+target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v128:128:128-n32:64"
+target triple = "powerpc64-unknown-linux-gnu"
+; RUN: llc -mcpu=a2 < %s | FileCheck %s
+; RUN: llc -mcpu=pwr7 < %s | FileCheck %s
+
+define i64 @test1(i64 %a, i64 %b, i64 %c, i64 %d) {
+entry:
+	%p = icmp uge i64 %a, %b
+	%x = select i1 %p, i64 %c, i64 %d
+	ret i64 %x
+; CHECK: @test1
+; CHECK: isel
+}
+
+define i32 @test2(i32 %a, i32 %b, i32 %c, i32 %d) {
+entry:
+	%p = icmp uge i32 %a, %b
+	%x = select i1 %p, i32 %c, i32 %d
+	ret i32 %x
+; CHECK: @test2
+; CHECK: isel
+}
+
diff --git a/test/CodeGen/PowerPC/ispositive.ll b/test/CodeGen/PowerPC/ispositive.ll
index 4161e34..78cdf4a 100644
--- a/test/CodeGen/PowerPC/ispositive.ll
+++ b/test/CodeGen/PowerPC/ispositive.ll
@@ -1,5 +1,5 @@
 ; RUN: llc < %s -march=ppc32 -mtriple=powerpc-apple-darwin8 | \
-; RUN:   grep {srwi r3, r3, 31}
+; RUN:   grep "srwi r3, r3, 31"
 
 define i32 @test1(i32 %X) {
 entry:
diff --git a/test/CodeGen/PowerPC/lbzux.ll b/test/CodeGen/PowerPC/lbzux.ll
new file mode 100644
index 0000000..12f1d1f
--- /dev/null
+++ b/test/CodeGen/PowerPC/lbzux.ll
@@ -0,0 +1,49 @@
+target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v128:128:128-n32:64"
+target triple = "powerpc64-unknown-linux-gnu"
+; RUN: llc < %s | FileCheck %s
+
+define fastcc void @allocateSpace(i1 %cond1, i1 %cond2) nounwind {
+entry:
+  %0 = load i8** undef, align 8, !tbaa !0
+  br i1 undef, label %return, label %lor.lhs.false
+
+lor.lhs.false:                                    ; preds = %entry
+  br i1 undef, label %if.end7, label %return
+
+if.end7:                                          ; preds = %lor.lhs.false
+  br i1 undef, label %if.then15, label %if.end71
+
+if.then15:                                        ; preds = %if.end7
+  br label %while.cond
+
+while.cond:                                       ; preds = %while.body, %if.then15
+  %idxprom17 = sext i32 0 to i64
+  %arrayidx18 = getelementptr inbounds i8* %0, i64 %idxprom17
+  %or = or i32 undef, undef
+  br i1 %cond1, label %if.end71, label %while.body
+
+while.body:                                       ; preds = %while.cond
+  br i1 %cond2, label %while.cond, label %if.then45
+
+if.then45:                                        ; preds = %while.body
+  %idxprom48139 = zext i32 %or to i64
+  %arrayidx49 = getelementptr inbounds i8* %0, i64 %idxprom48139
+  %1 = bitcast i8* %arrayidx49 to i16*
+  %2 = bitcast i8* %arrayidx18 to i16*
+  %3 = load i16* %1, align 1
+  store i16 %3, i16* %2, align 1
+  br label %return
+
+if.end71:                                         ; preds = %while.cond, %if.end7
+  unreachable
+
+return:                                           ; preds = %if.then45, %lor.lhs.false, %entry
+  ret void
+
+; CHECK: @allocateSpace
+; CHECK: lbzux
+}
+
+!0 = metadata !{metadata !"any pointer", metadata !1}
+!1 = metadata !{metadata !"omnipotent char", metadata !2}
+!2 = metadata !{metadata !"Simple C/C++ TBAA"}
diff --git a/test/CodeGen/PowerPC/long-compare.ll b/test/CodeGen/PowerPC/long-compare.ll
index 94c2526..915595f 100644
--- a/test/CodeGen/PowerPC/long-compare.ll
+++ b/test/CodeGen/PowerPC/long-compare.ll
@@ -1,7 +1,7 @@
 ; RUN: llc < %s -march=ppc32 | grep cntlzw 
 ; RUN: llc < %s -march=ppc32 | not grep xori 
-; RUN: llc < %s -march=ppc32 | not grep {li }
-; RUN: llc < %s -march=ppc32 | not grep {mr }
+; RUN: llc < %s -march=ppc32 | not grep "li "
+; RUN: llc < %s -march=ppc32 | not grep "mr "
 
 define i1 @test(i64 %x) {
   %tmp = icmp ult i64 %x, 4294967296
diff --git a/test/CodeGen/PowerPC/lsr-postinc-pos.ll b/test/CodeGen/PowerPC/lsr-postinc-pos.ll
index f441e42..42472c5 100644
--- a/test/CodeGen/PowerPC/lsr-postinc-pos.ll
+++ b/test/CodeGen/PowerPC/lsr-postinc-pos.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s -print-lsr-output |& FileCheck %s
+; RUN: llc < %s -print-lsr-output 2>&1 | FileCheck %s
 
 ; The icmp is a post-inc use, and the increment is in %bb11, but the
 ; scevgep needs to be inserted in %bb so that it is dominated by %t.
diff --git a/test/CodeGen/PowerPC/mem_update.ll b/test/CodeGen/PowerPC/mem_update.ll
index 17e7e28..39af11a 100644
--- a/test/CodeGen/PowerPC/mem_update.ll
+++ b/test/CodeGen/PowerPC/mem_update.ll
@@ -1,6 +1,6 @@
-; RUN: llc < %s -march=ppc32 -enable-ppc-preinc | \
+; RUN: llc < %s -march=ppc32 | \
 ; RUN:   not grep addi
-; RUN: llc < %s -march=ppc64 -enable-ppc-preinc | \
+; RUN: llc < %s -march=ppc64 | \
 ; RUN:   not grep addi
 
 @Glob = global i64 4
diff --git a/test/CodeGen/PowerPC/no-dead-strip.ll b/test/CodeGen/PowerPC/no-dead-strip.ll
index 3459413..6320e28 100644
--- a/test/CodeGen/PowerPC/no-dead-strip.ll
+++ b/test/CodeGen/PowerPC/no-dead-strip.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s | grep {no_dead_strip.*_X}
+; RUN: llc < %s | grep "no_dead_strip.*_X"
 
 target datalayout = "E-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64"
 target triple = "powerpc-apple-darwin8.8.0"
diff --git a/test/CodeGen/PowerPC/ppc440-fp-basic.ll b/test/CodeGen/PowerPC/ppc440-fp-basic.ll
index 1fad2fa..77b726c 100644
--- a/test/CodeGen/PowerPC/ppc440-fp-basic.ll
+++ b/test/CodeGen/PowerPC/ppc440-fp-basic.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s -march=ppc32 -mcpu=440 | FileCheck %s
+; RUN: llc < %s -march=ppc32 -mcpu=440 -fp-contract=fast | FileCheck %s
 
 %0 = type { double, double }
 
diff --git a/test/CodeGen/PowerPC/ppc64-cyclecounter.ll b/test/CodeGen/PowerPC/ppc64-cyclecounter.ll
new file mode 100644
index 0000000..38406ca
--- /dev/null
+++ b/test/CodeGen/PowerPC/ppc64-cyclecounter.ll
@@ -0,0 +1,15 @@
+target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v128:128:128-n32:64"
+target triple = "powerpc64-unknown-linux-gnu"
+; RUN: llc < %s | FileCheck %s
+
+define i64 @test1() nounwind {
+entry:
+  %r = call i64 @llvm.readcyclecounter()
+  ret i64 %r
+}
+
+; CHECK: @test1
+; CHECK: mfspr 3, 268
+
+declare i64 @llvm.readcyclecounter()
+
diff --git a/test/CodeGen/PowerPC/retaddr.ll b/test/CodeGen/PowerPC/retaddr.ll
index cf16b4c..c931dfe 100644
--- a/test/CodeGen/PowerPC/retaddr.ll
+++ b/test/CodeGen/PowerPC/retaddr.ll
@@ -1,6 +1,6 @@
 ; RUN: llc < %s -march=ppc32 | grep mflr
 ; RUN: llc < %s -march=ppc32 | grep lwz
-; RUN: llc < %s -march=ppc64 | grep {ld r., 16(r1)}
+; RUN: llc < %s -march=ppc64 | grep "ld r., 16(r1)"
 
 target triple = "powerpc-apple-darwin8"
 
diff --git a/test/CodeGen/PowerPC/rlwimi-commute.ll b/test/CodeGen/PowerPC/rlwimi-commute.ll
index 6410c63..3f90008 100644
--- a/test/CodeGen/PowerPC/rlwimi-commute.ll
+++ b/test/CodeGen/PowerPC/rlwimi-commute.ll
@@ -1,5 +1,5 @@
 ; RUN: llc < %s -march=ppc32 | grep rlwimi
-; RUN: llc < %s -march=ppc32 | not grep {or }
+; RUN: llc < %s -march=ppc32 | not grep "or "
 
 ; Make sure there is no register-register copies here.
 
diff --git a/test/CodeGen/PowerPC/rlwimi3.ll b/test/CodeGen/PowerPC/rlwimi3.ll
index 05d37bf..7efdbe9 100644
--- a/test/CodeGen/PowerPC/rlwimi3.ll
+++ b/test/CodeGen/PowerPC/rlwimi3.ll
@@ -1,5 +1,5 @@
-; RUN: llc < %s -march=ppc32 -stats |& \
-; RUN:   grep {Number of machine instrs printed} | grep 12
+; RUN: llc < %s -march=ppc32 -stats 2>&1 | \
+; RUN:   grep "Number of machine instrs printed" | grep 12
 
 define i16 @Trans16Bit(i32 %srcA, i32 %srcB, i32 %alpha) {
 	%tmp1 = shl i32 %srcA, 15		; <i32> [#uses=1]
diff --git a/test/CodeGen/PowerPC/seteq-0.ll b/test/CodeGen/PowerPC/seteq-0.ll
index 688b29a..7319583 100644
--- a/test/CodeGen/PowerPC/seteq-0.ll
+++ b/test/CodeGen/PowerPC/seteq-0.ll
@@ -1,5 +1,5 @@
 ; RUN: llc < %s -march=ppc32 -mtriple=powerpc-apple-darwin8 | \
-; RUN:   grep {srwi r., r., 5}
+; RUN:   grep "srwi r., r., 5"
 
 define i32 @eq0(i32 %a) {
         %tmp.1 = icmp eq i32 %a, 0              ; <i1> [#uses=1]
diff --git a/test/CodeGen/PowerPC/small-arguments.ll b/test/CodeGen/PowerPC/small-arguments.ll
index b4767b0..19ca098 100644
--- a/test/CodeGen/PowerPC/small-arguments.ll
+++ b/test/CodeGen/PowerPC/small-arguments.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s -march=ppc32 | not grep {extsh\\|rlwinm}
+; RUN: llc < %s -march=ppc32 | not grep "extsh\|rlwinm"
 
 declare signext i16 @foo()  
 
diff --git a/test/CodeGen/PowerPC/stack-protector.ll b/test/CodeGen/PowerPC/stack-protector.ll
index 2020361..810630f6 100644
--- a/test/CodeGen/PowerPC/stack-protector.ll
+++ b/test/CodeGen/PowerPC/stack-protector.ll
@@ -1,5 +1,5 @@
-; RUN: llc -march=ppc32 < %s -o - | grep {__stack_chk_guard}
-; RUN: llc -march=ppc32 < %s -o - | grep {__stack_chk_fail}
+; RUN: llc -march=ppc32 < %s -o - | grep "__stack_chk_guard"
+; RUN: llc -march=ppc32 < %s -o - | grep "__stack_chk_fail"
 
 @"\01LC" = internal constant [11 x i8] c"buf == %s\0A\00"		; <[11 x i8]*> [#uses=1]
 
diff --git a/test/CodeGen/PowerPC/stwu-gta.ll b/test/CodeGen/PowerPC/stwu-gta.ll
new file mode 100644
index 0000000..4febe7e
--- /dev/null
+++ b/test/CodeGen/PowerPC/stwu-gta.ll
@@ -0,0 +1,22 @@
+target datalayout = "E-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v128:128:128-n32"
+target triple = "powerpc-unknown-linux"
+; RUN: llc < %s | FileCheck %s
+
+%class.Two.0.5 = type { i32, i32, i32 }
+
+@foo = external global %class.Two.0.5, align 4
+
+define void @_GLOBAL__I_a() nounwind section ".text.startup" {
+entry:
+  store i32 5, i32* getelementptr inbounds (%class.Two.0.5* @foo, i32 0, i32 0), align 4, !tbaa !0
+  store i32 6, i32* getelementptr inbounds (%class.Two.0.5* @foo, i32 0, i32 1), align 4, !tbaa !0
+  ret void
+}
+
+; CHECK: @_GLOBAL__I_a
+; CHECK-NOT: stwux
+; CHECK: stwu
+
+!0 = metadata !{metadata !"int", metadata !1}
+!1 = metadata !{metadata !"omnipotent char", metadata !2}
+!2 = metadata !{metadata !"Simple C/C++ TBAA"}
diff --git a/test/CodeGen/PowerPC/stwu8.ll b/test/CodeGen/PowerPC/stwu8.ll
new file mode 100644
index 0000000..897bfc6
--- /dev/null
+++ b/test/CodeGen/PowerPC/stwu8.ll
@@ -0,0 +1,28 @@
+; RUN: llc < %s | FileCheck %s
+target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v128:128:128-n32:64"
+target triple = "powerpc64-unknown-linux-gnu"
+
+%class.spell_checker.21.103.513.538 = type { %"class.std::map.20.102.512.537" }
+%"class.std::map.20.102.512.537" = type { %"class.std::_Rb_tree.19.101.511.536" }
+%"class.std::_Rb_tree.19.101.511.536" = type { %"struct.std::_Rb_tree<std::pair<const char *, const char *>, std::pair<const std::pair<const char *, const char *>, int>, std::_Select1st<std::pair<const std::pair<const char *, const char *>, int>>, std::less<std::pair<const char *, const char *>>, std::allocator<std::pair<const std::pair<const char *, const char *>, int>> >::_Rb_tree_impl.18.100.510.535" }
+%"struct.std::_Rb_tree<std::pair<const char *, const char *>, std::pair<const std::pair<const char *, const char *>, int>, std::_Select1st<std::pair<const std::pair<const char *, const char *>, int>>, std::less<std::pair<const char *, const char *>>, std::allocator<std::pair<const std::pair<const char *, const char *>, int>> >::_Rb_tree_impl.18.100.510.535" = type { %"struct.std::less.16.98.508.533", %"struct.std::_Rb_tree_node_base.17.99.509.534", i64 }
+%"struct.std::less.16.98.508.533" = type { i8 }
+%"struct.std::_Rb_tree_node_base.17.99.509.534" = type { i32, %"struct.std::_Rb_tree_node_base.17.99.509.534"*, %"struct.std::_Rb_tree_node_base.17.99.509.534"*, %"struct.std::_Rb_tree_node_base.17.99.509.534"* }
+
+define void @test1(%class.spell_checker.21.103.513.538* %this) unnamed_addr align 2 {
+entry:
+  %_M_header.i.i.i.i.i.i = getelementptr inbounds %class.spell_checker.21.103.513.538* %this, i64 0, i32 0, i32 0, i32 0, i32 1
+  %0 = bitcast %"struct.std::_Rb_tree_node_base.17.99.509.534"* %_M_header.i.i.i.i.i.i to i8*
+  call void @llvm.memset.p0i8.i64(i8* %0, i8 0, i64 40, i32 4, i1 false) nounwind
+  store %"struct.std::_Rb_tree_node_base.17.99.509.534"* %_M_header.i.i.i.i.i.i, %"struct.std::_Rb_tree_node_base.17.99.509.534"** undef, align 8, !tbaa !0
+  unreachable
+}
+
+; CHECK: @test1
+; CHECK: stwu
+
+declare void @llvm.memset.p0i8.i64(i8* nocapture, i8, i64, i32, i1) nounwind
+
+!0 = metadata !{metadata !"any pointer", metadata !1}
+!1 = metadata !{metadata !"omnipotent char", metadata !2}
+!2 = metadata !{metadata !"Simple C/C++ TBAA"}
diff --git a/test/CodeGen/PowerPC/stwux.ll b/test/CodeGen/PowerPC/stwux.ll
new file mode 100644
index 0000000..737e9d9
--- /dev/null
+++ b/test/CodeGen/PowerPC/stwux.ll
@@ -0,0 +1,47 @@
+target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v128:128:128-n32:64"
+target triple = "powerpc64-unknown-linux-gnu"
+; RUN: llc < %s | FileCheck %s
+
+@multvec_i = external unnamed_addr global [100 x i32], align 4
+
+define fastcc void @subs_STMultiExceptIntern() nounwind {
+entry:
+  br i1 undef, label %while.body.lr.ph, label %return
+
+while.body.lr.ph:                                 ; preds = %entry
+  br label %while.body
+
+while.body:                                       ; preds = %if.end12, %while.body.lr.ph
+  %i.0240 = phi i32 [ -1, %while.body.lr.ph ], [ %i.1, %if.end12 ]
+  br i1 undef, label %if.end12, label %if.then
+
+if.then:                                          ; preds = %while.body
+  br label %if.end12
+
+if.end12:                                         ; preds = %if.then, %while.body
+  %i.1 = phi i32 [ %i.0240, %while.body ], [ undef, %if.then ]
+  br i1 undef, label %while.body, label %while.end
+
+while.end:                                        ; preds = %if.end12
+  br i1 undef, label %return, label %if.end15
+
+if.end15:                                         ; preds = %while.end
+  %idxprom.i.i230 = sext i32 %i.1 to i64
+  %arrayidx18 = getelementptr inbounds [100 x i32]* @multvec_i, i64 0, i64 %idxprom.i.i230
+  store i32 0, i32* %arrayidx18, align 4
+  br i1 undef, label %while.body21, label %while.end90
+
+while.body21:                                     ; preds = %if.end15
+  unreachable
+
+while.end90:                                      ; preds = %if.end15
+  store i32 0, i32* %arrayidx18, align 4
+  br label %return
+
+return:                                           ; preds = %while.end90, %while.end, %entry
+  ret void
+
+; CHECK: @subs_STMultiExceptIntern
+; CHECK: stwux
+}
+
diff --git a/test/CodeGen/PowerPC/tls.ll b/test/CodeGen/PowerPC/tls.ll
new file mode 100644
index 0000000..713893b
--- /dev/null
+++ b/test/CodeGen/PowerPC/tls.ll
@@ -0,0 +1,16 @@
+target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v128:128:128-n32:64"
+target triple = "powerpc64-unknown-freebsd10.0"
+; RUN: llc < %s -march=ppc64 | FileCheck %s
+
+@a = thread_local global i32 0, align 4
+
+;CHECK:          localexec:
+define i32 @localexec() nounwind {
+entry:
+;CHECK:          addis [[REG1:[0-9]+]], 13, a@tprel@ha
+;CHECK-NEXT:     li [[REG2:[0-9]+]], 42
+;CHECK-NEXT:     addi [[REG1]], [[REG1]], a@tprel@l
+;CHECK-NEXT:     stw [[REG2]], 0([[REG1]])
+  store i32 42, i32* @a, align 4
+  ret i32 0
+}
diff --git a/test/CodeGen/PowerPC/trampoline.ll b/test/CodeGen/PowerPC/trampoline.ll
index 91b2011..3ea46f5 100644
--- a/test/CodeGen/PowerPC/trampoline.ll
+++ b/test/CodeGen/PowerPC/trampoline.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s -march=ppc32 | grep {__trampoline_setup}
+; RUN: llc < %s -march=ppc32 | grep "__trampoline_setup"
 
 module asm "\09.lazy_reference .objc_class_name_NSImageRep"
 module asm "\09.objc_class_name_NSBitmapImageRep=0"
diff --git a/test/CodeGen/PowerPC/vec_buildvector_loadstore.ll b/test/CodeGen/PowerPC/vec_buildvector_loadstore.ll
index 015c086..7e58ec0 100644
--- a/test/CodeGen/PowerPC/vec_buildvector_loadstore.ll
+++ b/test/CodeGen/PowerPC/vec_buildvector_loadstore.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s -march=ppc32 -mtriple=powerpc-apple-darwin -mattr=+altivec  | FileCheck %s
+; RUN: llc < %s -march=ppc32 -mtriple=powerpc-apple-darwin -mattr=+altivec -disable-ppc-ilp-pref  | FileCheck %s
 ; Formerly this did byte loads and word stores.
 @a = external global <16 x i8>
 @b = external global <16 x i8>