Add support for emitting both signed- and zero-extend loads.  Fix 
SimplifyAddress to handle either a 12-bit unsigned offset or the ARM +/-imm8
offsets (addressing mode 3).  This enables a load followed by an integer 
extend to be folded into a single load.

For example:
ldrb r1, [r0]       ldrb r1, [r0]
uxtb r2, r1     =>
mov  r3, r2         mov  r3, r1


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@144488 91177308-0d34-0410-b5e6-96231b3b80d8
diff --git a/test/CodeGen/ARM/fast-isel-fold.ll b/test/CodeGen/ARM/fast-isel-fold.ll
new file mode 100644
index 0000000..61bd185
--- /dev/null
+++ b/test/CodeGen/ARM/fast-isel-fold.ll
@@ -0,0 +1,80 @@
+; RUN: llc < %s -O0 -fast-isel-abort -relocation-model=dynamic-no-pic -mtriple=armv7-apple-darwin | FileCheck %s --check-prefix=ARM
+; RUN: llc < %s -O0 -fast-isel-abort -relocation-model=dynamic-no-pic -mtriple=thumbv7-apple-darwin | FileCheck %s --check-prefix=THUMB
+
+@a = global i8 1, align 1
+@b = global i16 2, align 2
+
+define void @t1() nounwind uwtable ssp {
+; ARM: t1
+; ARM: ldrb
+; ARM-NOT: uxtb
+; THUMB: t1
+; THUMB: ldrb
+; THUMB-NOT: uxtb
+  %1 = load i8* @a, align 1
+  call void @foo1(i8 zeroext %1)
+  ret void
+}
+
+define void @t2() nounwind uwtable ssp {
+; ARM: t2
+; ARM: ldrh
+; ARM-NOT: uxth
+; THUMB: t2
+; THUMB: ldrh
+; THUMB-NOT: uxth
+  %1 = load i16* @b, align 2
+  call void @foo2(i16 zeroext %1)
+  ret void
+}
+
+declare void @foo1(i8 zeroext)
+declare void @foo2(i16 zeroext)
+
+define i32 @t3() nounwind uwtable ssp {
+; ARM: t3
+; ARM: ldrb
+; ARM-NOT: uxtb
+; THUMB: t3
+; THUMB: ldrb
+; THUMB-NOT: uxtb
+  %1 = load i8* @a, align 1
+  %2 = zext i8 %1 to i32
+  ret i32 %2
+}
+
+define i32 @t4() nounwind uwtable ssp {
+; ARM: t4
+; ARM: ldrh
+; ARM-NOT: uxth
+; THUMB: t4
+; THUMB: ldrh
+; THUMB-NOT: uxth
+  %1 = load i16* @b, align 2
+  %2 = zext i16 %1 to i32
+  ret i32 %2
+}
+
+define i32 @t5() nounwind uwtable ssp {
+; ARM: t5
+; ARM: ldrsh
+; ARM-NOT: sxth
+; THUMB: t5
+; THUMB: ldrsh
+; THUMB-NOT: sxth
+  %1 = load i16* @b, align 2
+  %2 = sext i16 %1 to i32
+  ret i32 %2
+}
+
+define i32 @t6() nounwind uwtable ssp {
+; ARM: t6
+; ARM: ldrsb
+; ARM-NOT: sxtb
+; THUMB: t6
+; THUMB: ldrsb
+; THUMB-NOT: sxtb
+  %1 = load i8* @a, align 2
+  %2 = sext i8 %1 to i32
+  ret i32 %2
+}