mirror of
https://github.com/c64scene-ar/llvm-6502.git
synced 2024-12-13 20:32:21 +00:00
8e38e86266
The code in PPCTargetLowering::PerformDAGCombine() that handles unaligned Altivec vector loads generates a lvsl followed by a vperm. As we've seen in numerous other places, the vperm instruction has a big-endian bias, and this is fixed for little endian by complementing the permute control vector and swapping the input operands. In this case the lvsl is providing the permute control vector. Rather than generating an lvsl and a complement operation, it is sufficient to generate an lvsr instruction instead. Thus for LE code generation we will generate an lvsr rather than an lvsl, and swap the other input arguments on the vperm. The existing test/CodeGen/PowerPC/vec_misalign.ll is updated to test the code generation for PPC64 and PPC64LE, in addition to the existing PPC32/G5 testing. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@210493 91177308-0d34-0410-b5e6-96231b3b80d8
46 lines
1.9 KiB
LLVM
46 lines
1.9 KiB
LLVM
; RUN: llc < %s -march=ppc32 -mcpu=g5 | FileCheck %s
|
|
; RUN: llc < %s -mtriple=powerpc64-unknown-linux-gnu -mattr=+altivec | FileCheck %s
|
|
; RUN: llc < %s -mtriple=powerpc64le-unknown-linux-gnu -mattr=+altivec | FileCheck %s -check-prefix=CHECK-LE
|
|
|
|
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-darwin8"
|
|
%struct.S2203 = type { %struct.u16qi }
|
|
%struct.u16qi = type { <16 x i8> }
|
|
@s = weak global %struct.S2203 zeroinitializer ; <%struct.S2203*> [#uses=1]
|
|
|
|
define void @foo(i32 %x, ...) {
|
|
entry:
|
|
; CHECK: foo:
|
|
; CHECK-LE: foo:
|
|
%x_addr = alloca i32 ; <i32*> [#uses=1]
|
|
%ap = alloca i8* ; <i8**> [#uses=3]
|
|
%ap.0 = alloca i8* ; <i8**> [#uses=3]
|
|
%"alloca point" = bitcast i32 0 to i32 ; <i32> [#uses=0]
|
|
store i32 %x, i32* %x_addr
|
|
%ap1 = bitcast i8** %ap to i8* ; <i8*> [#uses=1]
|
|
call void @llvm.va_start( i8* %ap1 )
|
|
%tmp = load i8** %ap, align 4 ; <i8*> [#uses=1]
|
|
store i8* %tmp, i8** %ap.0, align 4
|
|
%tmp2 = load i8** %ap.0, align 4 ; <i8*> [#uses=1]
|
|
%tmp3 = getelementptr i8* %tmp2, i64 16 ; <i8*> [#uses=1]
|
|
store i8* %tmp3, i8** %ap, align 4
|
|
%tmp4 = load i8** %ap.0, align 4 ; <i8*> [#uses=1]
|
|
%tmp45 = bitcast i8* %tmp4 to %struct.S2203* ; <%struct.S2203*> [#uses=1]
|
|
%tmp6 = getelementptr %struct.S2203* @s, i32 0, i32 0 ; <%struct.u16qi*> [#uses=1]
|
|
%tmp7 = getelementptr %struct.S2203* %tmp45, i32 0, i32 0 ; <%struct.u16qi*> [#uses=1]
|
|
%tmp8 = getelementptr %struct.u16qi* %tmp6, i32 0, i32 0 ; <<16 x i8>*> [#uses=1]
|
|
%tmp9 = getelementptr %struct.u16qi* %tmp7, i32 0, i32 0 ; <<16 x i8>*> [#uses=1]
|
|
%tmp10 = load <16 x i8>* %tmp9, align 4 ; <<16 x i8>> [#uses=1]
|
|
; CHECK: lvsl
|
|
; CHECK: vperm
|
|
; CHECK-LE: lvsr
|
|
; CHECK-LE: vperm
|
|
store <16 x i8> %tmp10, <16 x i8>* %tmp8, align 4
|
|
br label %return
|
|
|
|
return: ; preds = %entry
|
|
ret void
|
|
}
|
|
|
|
declare void @llvm.va_start(i8*) nounwind
|