-; RUN: llc -mcpu=pwr7 -O0 -disable-fp-elim < %s | FileCheck %s
-
-; FIXME: The code generation for packed structs is very poor because the
-; PowerPC target wrongly rejects all unaligned loads. This test case will
-; need to be revised when that is fixed.
+; RUN: llc -mcpu=ppc64 -O0 -disable-fp-elim -fast-isel=false < %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"
%p6 = alloca %struct.s6, align 4
%p7 = alloca %struct.s7, align 4
%0 = bitcast %struct.s1* %p1 to i8*
- call void @llvm.memcpy.p0i8.p0i8.i64(i8* %0, i8* getelementptr inbounds (%struct.s1* @caller1.p1, i32 0, i32 0), i64 1, i32 1, i1 false)
+ call void @llvm.memcpy.p0i8.p0i8.i64(i8* %0, i8* getelementptr inbounds (%struct.s1, %struct.s1* @caller1.p1, i32 0, i32 0), i64 1, i32 1, i1 false)
%1 = bitcast %struct.s2* %p2 to i8*
call void @llvm.memcpy.p0i8.p0i8.i64(i8* %1, i8* bitcast (%struct.s2* @caller1.p2 to i8*), i64 2, i32 2, i1 false)
%2 = bitcast %struct.s3* %p3 to i8*
store i32 %z6, i32* %z6.addr, align 4
store i32 %z7, i32* %z7.addr, align 4
store i32 %z8, i32* %z8.addr, align 4
- %a = getelementptr inbounds %struct.s1* %v1, i32 0, i32 0
- %0 = load i8* %a, align 1
+ %a = getelementptr inbounds %struct.s1, %struct.s1* %v1, i32 0, i32 0
+ %0 = load i8, i8* %a, align 1
%conv = zext i8 %0 to i32
- %a1 = getelementptr inbounds %struct.s2* %v2, i32 0, i32 0
- %1 = load i16* %a1, align 2
+ %a1 = getelementptr inbounds %struct.s2, %struct.s2* %v2, i32 0, i32 0
+ %1 = load i16, i16* %a1, align 2
%conv2 = sext i16 %1 to i32
%add = add nsw i32 %conv, %conv2
- %a3 = getelementptr inbounds %struct.s3* %v3, i32 0, i32 0
- %2 = load i16* %a3, align 2
+ %a3 = getelementptr inbounds %struct.s3, %struct.s3* %v3, i32 0, i32 0
+ %2 = load i16, i16* %a3, align 2
%conv4 = sext i16 %2 to i32
%add5 = add nsw i32 %add, %conv4
- %a6 = getelementptr inbounds %struct.s4* %v4, i32 0, i32 0
- %3 = load i32* %a6, align 4
+ %a6 = getelementptr inbounds %struct.s4, %struct.s4* %v4, i32 0, i32 0
+ %3 = load i32, i32* %a6, align 4
%add7 = add nsw i32 %add5, %3
- %a8 = getelementptr inbounds %struct.s5* %v5, i32 0, i32 0
- %4 = load i32* %a8, align 4
+ %a8 = getelementptr inbounds %struct.s5, %struct.s5* %v5, i32 0, i32 0
+ %4 = load i32, i32* %a8, align 4
%add9 = add nsw i32 %add7, %4
- %a10 = getelementptr inbounds %struct.s6* %v6, i32 0, i32 0
- %5 = load i32* %a10, align 4
+ %a10 = getelementptr inbounds %struct.s6, %struct.s6* %v6, i32 0, i32 0
+ %5 = load i32, i32* %a10, align 4
%add11 = add nsw i32 %add9, %5
- %a12 = getelementptr inbounds %struct.s7* %v7, i32 0, i32 0
- %6 = load i32* %a12, align 4
+ %a12 = getelementptr inbounds %struct.s7, %struct.s7* %v7, i32 0, i32 0
+ %6 = load i32, i32* %a12, align 4
%add13 = add nsw i32 %add11, %6
ret i32 %add13
; CHECK: lha {{[0-9]+}}, 126(1)
-; CHECK: lbz {{[0-9]+}}, 119(1)
; CHECK: lha {{[0-9]+}}, 132(1)
+; CHECK: lbz {{[0-9]+}}, 119(1)
; CHECK: lwz {{[0-9]+}}, 140(1)
; CHECK: lwz {{[0-9]+}}, 144(1)
; CHECK: lwz {{[0-9]+}}, 152(1)
%p6 = alloca %struct.t6, align 1
%p7 = alloca %struct.t7, align 1
%0 = bitcast %struct.t1* %p1 to i8*
- call void @llvm.memcpy.p0i8.p0i8.i64(i8* %0, i8* getelementptr inbounds (%struct.t1* @caller2.p1, i32 0, i32 0), i64 1, i32 1, i1 false)
+ call void @llvm.memcpy.p0i8.p0i8.i64(i8* %0, i8* getelementptr inbounds (%struct.t1, %struct.t1* @caller2.p1, i32 0, i32 0), i64 1, i32 1, i1 false)
%1 = bitcast %struct.t2* %p2 to i8*
call void @llvm.memcpy.p0i8.p0i8.i64(i8* %1, i8* bitcast ({ i16 }* @caller2.p2 to i8*), i64 2, i32 1, i1 false)
%2 = bitcast %struct.t3* %p3 to i8*
store i32 %z6, i32* %z6.addr, align 4
store i32 %z7, i32* %z7.addr, align 4
store i32 %z8, i32* %z8.addr, align 4
- %a = getelementptr inbounds %struct.t1* %v1, i32 0, i32 0
- %0 = load i8* %a, align 1
+ %a = getelementptr inbounds %struct.t1, %struct.t1* %v1, i32 0, i32 0
+ %0 = load i8, i8* %a, align 1
%conv = zext i8 %0 to i32
- %a1 = getelementptr inbounds %struct.t2* %v2, i32 0, i32 0
- %1 = load i16* %a1, align 1
+ %a1 = getelementptr inbounds %struct.t2, %struct.t2* %v2, i32 0, i32 0
+ %1 = load i16, i16* %a1, align 1
%conv2 = sext i16 %1 to i32
%add = add nsw i32 %conv, %conv2
- %a3 = getelementptr inbounds %struct.t3* %v3, i32 0, i32 0
- %2 = load i16* %a3, align 1
+ %a3 = getelementptr inbounds %struct.t3, %struct.t3* %v3, i32 0, i32 0
+ %2 = load i16, i16* %a3, align 1
%conv4 = sext i16 %2 to i32
%add5 = add nsw i32 %add, %conv4
- %a6 = getelementptr inbounds %struct.t4* %v4, i32 0, i32 0
- %3 = load i32* %a6, align 1
+ %a6 = getelementptr inbounds %struct.t4, %struct.t4* %v4, i32 0, i32 0
+ %3 = load i32, i32* %a6, align 1
%add7 = add nsw i32 %add5, %3
- %a8 = getelementptr inbounds %struct.t5* %v5, i32 0, i32 0
- %4 = load i32* %a8, align 1
+ %a8 = getelementptr inbounds %struct.t5, %struct.t5* %v5, i32 0, i32 0
+ %4 = load i32, i32* %a8, align 1
%add9 = add nsw i32 %add7, %4
- %a10 = getelementptr inbounds %struct.t6* %v6, i32 0, i32 0
- %5 = load i32* %a10, align 1
+ %a10 = getelementptr inbounds %struct.t6, %struct.t6* %v6, i32 0, i32 0
+ %5 = load i32, i32* %a10, align 1
%add11 = add nsw i32 %add9, %5
- %a12 = getelementptr inbounds %struct.t7* %v7, i32 0, i32 0
- %6 = load i32* %a12, align 1
+ %a12 = getelementptr inbounds %struct.t7, %struct.t7* %v7, i32 0, i32 0
+ %6 = load i32, i32* %a12, align 1
%add13 = add nsw i32 %add11, %6
ret i32 %add13
-; CHECK: lbz {{[0-9]+}}, 149(1)
-; CHECK: lbz {{[0-9]+}}, 150(1)
-; CHECK: lbz {{[0-9]+}}, 147(1)
-; CHECK: lbz {{[0-9]+}}, 148(1)
-; CHECK: lbz {{[0-9]+}}, 133(1)
-; CHECK: lbz {{[0-9]+}}, 134(1)
; CHECK: lha {{[0-9]+}}, 126(1)
+; CHECK: lha {{[0-9]+}}, 133(1)
; CHECK: lbz {{[0-9]+}}, 119(1)
; CHECK: lwz {{[0-9]+}}, 140(1)
-; CHECK: lhz {{[0-9]+}}, 154(1)
-; CHECK: lhz {{[0-9]+}}, 156(1)
-; CHECK: lbz {{[0-9]+}}, 163(1)
-; CHECK: lbz {{[0-9]+}}, 164(1)
-; CHECK: lbz {{[0-9]+}}, 161(1)
-; CHECK: lbz {{[0-9]+}}, 162(1)
+; CHECK: lwz {{[0-9]+}}, 147(1)
+; CHECK: lwz {{[0-9]+}}, 154(1)
+; CHECK: lwz {{[0-9]+}}, 161(1)
}
projects / oota-llvm.git / blobdiff