AArch64/PowerPC/SystemZ/X86: This patch fixes the interface, usage, and all

[oota-llvm.git] / test / CodeGen / PowerPC / bswap-load-store.ll

diff --git a/test/CodeGen/PowerPC/bswap-load-store.ll b/test/CodeGen/PowerPC/bswap-load-store.ll
index 664a2aa4bc914674ee3b18ccbc0994ee723f0d68..53bbc52167c4c144144b9521aa3c4176bf006490 100644 (file)
--- a/test/CodeGen/PowerPC/bswap-load-store.ll
+++ b/test/CodeGen/PowerPC/bswap-load-store.ll
@@ -1,42 +1,82 @@
-; RUN: llvm-upgrade < %s | llvm-as | llc -march=ppc32 | grep 'stwbrx\|lwbrx\|sthbrx\|lhbrx' | wc -l | grep 4 &&
-; RUN: llvm-upgrade < %s | llvm-as | llc -march=ppc32 | not grep rlwinm &&
-; RUN: llvm-upgrade < %s | llvm-as | llc -march=ppc32 | not grep rlwimi &&
-; RUN: llvm-upgrade < %s | llvm-as | llc -march=ppc64 | grep 'stwbrx\|lwbrx\|sthbrx\|lhbrx' | wc -l | grep 4 &&
-; RUN: llvm-upgrade < %s | llvm-as | llc -march=ppc64 | not grep rlwinm &&
-; RUN: llvm-upgrade < %s | llvm-as | llc -march=ppc64 | not grep rlwimi
-
-void %STWBRX(uint %i, sbyte* %ptr, int %off) {
-       %tmp1 = getelementptr sbyte* %ptr, int %off
-       %tmp1 = cast sbyte* %tmp1 to uint*
-       %tmp13 = tail call uint %llvm.bswap.i32.i32(uint %i)
-       store uint %tmp13, uint* %tmp1
-       ret void
+; RUN: llc < %s -march=ppc32 -mcpu=ppc32 | FileCheck %s -check-prefix=X32
+; RUN: llc < %s -march=ppc64 -mcpu=ppc64 | FileCheck %s -check-prefix=X64
+; RUN: llc < %s -march=ppc64 -mcpu=pwr7 | FileCheck %s -check-prefix=PWR7
+; RUN: llc < %s -march=ppc32 -mcpu=pwr7 | FileCheck %s -check-prefix=X32
+
+
+define void @STWBRX(i32 %i, i8* %ptr, i32 %off) {
+        %tmp1 = getelementptr i8* %ptr, i32 %off                ; <i8*> [#uses=1]
+        %tmp1.upgrd.1 = bitcast i8* %tmp1 to i32*               ; <i32*> [#uses=1]
+        %tmp13 = tail call i32 @llvm.bswap.i32( i32 %i )                ; <i32> [#uses=1]
+        store i32 %tmp13, i32* %tmp1.upgrd.1
+        ret void
+}
+
+define i32 @LWBRX(i8* %ptr, i32 %off) {
+        %tmp1 = getelementptr i8* %ptr, i32 %off                ; <i8*> [#uses=1]
+        %tmp1.upgrd.2 = bitcast i8* %tmp1 to i32*               ; <i32*> [#uses=1]
+        %tmp = load i32* %tmp1.upgrd.2          ; <i32> [#uses=1]
+        %tmp14 = tail call i32 @llvm.bswap.i32( i32 %tmp )              ; <i32> [#uses=1]
+        ret i32 %tmp14
+}
+
+define void @STHBRX(i16 %s, i8* %ptr, i32 %off) {
+        %tmp1 = getelementptr i8* %ptr, i32 %off                ; <i8*> [#uses=1]
+        %tmp1.upgrd.3 = bitcast i8* %tmp1 to i16*               ; <i16*> [#uses=1]
+        %tmp5 = call i16 @llvm.bswap.i16( i16 %s )              ; <i16> [#uses=1]
+        store i16 %tmp5, i16* %tmp1.upgrd.3
+        ret void
 }
 
-uint %LWBRX(sbyte* %ptr, int %off) {
-       %tmp1 = getelementptr sbyte* %ptr, int %off
-       %tmp1 = cast sbyte* %tmp1 to uint*              
-       %tmp = load uint* %tmp1         
-       %tmp14 = tail call uint %llvm.bswap.i32.i32( uint %tmp )
-       ret uint %tmp14
+define i16 @LHBRX(i8* %ptr, i32 %off) {
+        %tmp1 = getelementptr i8* %ptr, i32 %off                ; <i8*> [#uses=1]
+        %tmp1.upgrd.4 = bitcast i8* %tmp1 to i16*               ; <i16*> [#uses=1]
+        %tmp = load i16* %tmp1.upgrd.4          ; <i16> [#uses=1]
+        %tmp6 = call i16 @llvm.bswap.i16( i16 %tmp )            ; <i16> [#uses=1]
+        ret i16 %tmp6
 }
 
-void %STHBRX(ushort %s, sbyte* %ptr, int %off) {
-       %tmp1 = getelementptr sbyte* %ptr, int %off
-       %tmp1 = cast sbyte* %tmp1 to ushort*
-       %tmp5 = call ushort %llvm.bswap.i16.i16( ushort %s )
-       store ushort %tmp5, ushort* %tmp1
-       ret void
+define void @STDBRX(i64 %i, i8* %ptr, i64 %off) {
+        %tmp1 = getelementptr i8* %ptr, i64 %off                ; <i8*> [#uses=1]
+        %tmp1.upgrd.1 = bitcast i8* %tmp1 to i64*               ; <i64*> [#uses=1]
+        %tmp13 = tail call i64 @llvm.bswap.i64( i64 %i )                ; <i64> [#uses=1]
+        store i64 %tmp13, i64* %tmp1.upgrd.1
+        ret void
 }
 
-ushort %LHBRX(sbyte* %ptr, int %off) {
-       %tmp1 = getelementptr sbyte* %ptr, int %off
-       %tmp1 = cast sbyte* %tmp1 to ushort*
-       %tmp = load ushort* %tmp1
-       %tmp6 = call ushort %llvm.bswap.i16.i16(ushort %tmp)
-       ret ushort %tmp6
+define i64 @LDBRX(i8* %ptr, i64 %off) {
+        %tmp1 = getelementptr i8* %ptr, i64 %off                ; <i8*> [#uses=1]
+        %tmp1.upgrd.2 = bitcast i8* %tmp1 to i64*               ; <i64*> [#uses=1]
+        %tmp = load i64* %tmp1.upgrd.2          ; <i64> [#uses=1]
+        %tmp14 = tail call i64 @llvm.bswap.i64( i64 %tmp )              ; <i64> [#uses=1]
+        ret i64 %tmp14
 }
 
-declare uint %llvm.bswap.i32.i32(uint)
+declare i32 @llvm.bswap.i32(i32)
+
+declare i16 @llvm.bswap.i16(i16)
+
+declare i64 @llvm.bswap.i64(i64)
+
+
+; X32: stwbrx
+; X32: lwbrx
+; X32: sthbrx
+; X32: lhbrx
+; X32-NOT: ldbrx
+; X32-NOT: stdbrx
+
+; X64: stwbrx
+; X64: lwbrx
+; X64: sthbrx
+; X64: lhbrx
+; X64-NOT: ldbrx
+; X64-NOT: stdbrx
+
+; PWR7: stwbrx
+; PWR7: lwbrx
+; PWR7: sthbrx
+; PWR7: lhbrx
+; PWR7: stdbrx
+; PWR7: ldbrx
 
-declare ushort %llvm.bswap.i16.i16(ushort)