2208c97c1b
Similar to 806761a7629df268c8aed49657aeccffa6bca449 -mtriple= specifies the full target triple while -march= merely sets the architecture part of the default target triple, leaving a target triple which may not make sense. Therefore, -march= is error-prone and not recommended for tests without a target triple. The issue has been benign as we recognize $unknown-apple-darwin as ELF instead of rejecting it outrightly.
34 lines
1004 B
LLVM
34 lines
1004 B
LLVM
; RUN: llc -mtriple=hexagon -mcpu=hexagonv5 < %s | FileCheck %s
|
|
; Generate MemOps for V4 and above.
|
|
|
|
|
|
define void @f(ptr %p) nounwind {
|
|
entry:
|
|
; CHECK: memw(r{{[0-9]+}}+#40) -= #1
|
|
%p.addr = alloca ptr, align 4
|
|
store ptr %p, ptr %p.addr, align 4
|
|
%0 = load ptr, ptr %p.addr, align 4
|
|
%add.ptr = getelementptr inbounds i32, ptr %0, i32 10
|
|
%1 = load i32, ptr %add.ptr, align 4
|
|
%sub = sub nsw i32 %1, 1
|
|
store i32 %sub, ptr %add.ptr, align 4
|
|
ret void
|
|
}
|
|
|
|
define void @g(ptr %p, i32 %i) nounwind {
|
|
entry:
|
|
; CHECK: memw(r{{[0-9]+}}+#40) -= #1
|
|
%p.addr = alloca ptr, align 4
|
|
%i.addr = alloca i32, align 4
|
|
store ptr %p, ptr %p.addr, align 4
|
|
store i32 %i, ptr %i.addr, align 4
|
|
%0 = load ptr, ptr %p.addr, align 4
|
|
%1 = load i32, ptr %i.addr, align 4
|
|
%add.ptr = getelementptr inbounds i32, ptr %0, i32 %1
|
|
%add.ptr1 = getelementptr inbounds i32, ptr %add.ptr, i32 10
|
|
%2 = load i32, ptr %add.ptr1, align 4
|
|
%sub = sub nsw i32 %2, 1
|
|
store i32 %sub, ptr %add.ptr1, align 4
|
|
ret void
|
|
}
|