; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -regalloc=fast -optimize-regalloc=0 -mtriple=i686-- -mattr=+mmx | FileCheck %s
; PR2082
; Local register allocator was refusing to use ESI, EDI, and EBP so it ran out of
; registers.
define void @transpose4x4(ptr %dst, ptr %src, i32 %dst_stride, i32 %src_stride) {
; CHECK-LABEL: transpose4x4:
; CHECK: # %bb.0: # %entry
; CHECK-NEXT: pushl %ebx
; CHECK-NEXT: .cfi_def_cfa_offset 8
; CHECK-NEXT: pushl %edi
; CHECK-NEXT: .cfi_def_cfa_offset 12
; CHECK-NEXT: pushl %esi
; CHECK-NEXT: .cfi_def_cfa_offset 16
; CHECK-NEXT: .cfi_offset %esi, -16
; CHECK-NEXT: .cfi_offset %edi, -12
; CHECK-NEXT: .cfi_offset %ebx, -8
; CHECK-NEXT: movl {{[0-9]+}}(%esp), %eax
; CHECK-NEXT: movl {{[0-9]+}}(%esp), %ecx
; CHECK-NEXT: leal (%ecx,%ecx,2), %edx
; CHECK-NEXT: movl {{[0-9]+}}(%esp), %esi
; CHECK-NEXT: movl {{[0-9]+}}(%esp), %edi
; CHECK-NEXT: leal (%edi,%edi,2), %ebx
; CHECK-NEXT: #APP
; CHECK-NEXT: movd (%esi), %mm0
; CHECK-NEXT: movd (%esi,%edi), %mm1
; CHECK-NEXT: movd (%esi,%edi,2), %mm2
; CHECK-NEXT: movd (%esi,%ebx), %mm3
; CHECK-NEXT: punpcklbw %mm1, %mm0 # mm0 = mm0[0],mm1[0],mm0[1],mm1[1],mm0[2],mm1[2],mm0[3],mm1[3]
; CHECK-NEXT: punpcklbw %mm3, %mm2 # mm2 = mm2[0],mm3[0],mm2[1],mm3[1],mm2[2],mm3[2],mm2[3],mm3[3]
; CHECK-NEXT: movq %mm0, %mm1
; CHECK-NEXT: punpcklwd %mm2, %mm0 # mm0 = mm0[0],mm2[0],mm0[1],mm2[1]
; CHECK-NEXT: punpckhwd %mm2, %mm1 # mm1 = mm1[2],mm2[2],mm1[3],mm2[3]
; CHECK-NEXT: movd %mm0, (%eax)
; CHECK-NEXT: punpckhdq %mm0, %mm0 # mm0 = mm0[1,1]
; CHECK-NEXT: movd %mm0, (%eax,%ecx)
; CHECK-NEXT: movd %mm1, (%eax,%ecx,2)
; CHECK-NEXT: punpckhdq %mm1, %mm1 # mm1 = mm1[1,1]
; CHECK-NEXT: movd %mm1, (%eax,%edx)
; CHECK-EMPTY:
; CHECK-NEXT: #NO_APP
; CHECK-NEXT: popl %esi
; CHECK-NEXT: .cfi_def_cfa_offset 12
; CHECK-NEXT: popl %edi
; CHECK-NEXT: .cfi_def_cfa_offset 8
; CHECK-NEXT: popl %ebx
; CHECK-NEXT: .cfi_def_cfa_offset 4
; CHECK-NEXT: retl
entry:
%dst_addr = alloca ptr ; <ptr> [#uses=5]
%src_addr = alloca ptr ; <ptr> [#uses=5]
%dst_stride_addr = alloca i32 ; <ptr> [#uses=4]
%src_stride_addr = alloca i32 ; <ptr> [#uses=4]
%"alloca point" = bitcast i32 0 to i32 ; <i32> [#uses=0]
store ptr %dst, ptr %dst_addr
store ptr %src, ptr %src_addr
store i32 %dst_stride, ptr %dst_stride_addr
store i32 %src_stride, ptr %src_stride_addr
%tmp = load ptr, ptr %dst_addr, align 4 ; <ptr> [#uses=1]
%tmp1 = getelementptr i8, ptr %tmp, i32 0 ; <ptr> [#uses=1]
%tmp3 = load ptr, ptr %dst_addr, align 4 ; <ptr> [#uses=1]
%tmp4 = load i32, ptr %dst_stride_addr, align 4 ; <i32> [#uses=1]
%tmp5 = getelementptr i8, ptr %tmp3, i32 %tmp4 ; <ptr> [#uses=1]
%tmp7 = load i32, ptr %dst_stride_addr, align 4 ; <i32> [#uses=1]
%tmp8 = mul i32 %tmp7, 2 ; <i32> [#uses=1]
%tmp9 = load ptr, ptr %dst_addr, align 4 ; <ptr> [#uses=1]
%tmp10 = getelementptr i8, ptr %tmp9, i32 %tmp8 ; <ptr> [#uses=1]
%tmp13 = load i32, ptr %dst_stride_addr, align 4 ; <i32> [#uses=1]
%tmp14 = mul i32 %tmp13, 3 ; <i32> [#uses=1]
%tmp15 = load ptr, ptr %dst_addr, align 4 ; <ptr> [#uses=1]
%tmp16 = getelementptr i8, ptr %tmp15, i32 %tmp14 ; <ptr> [#uses=1]
%tmp18 = load ptr, ptr %src_addr, align 4 ; <ptr> [#uses=1]
%tmp19 = getelementptr i8, ptr %tmp18, i32 0 ; <ptr> [#uses=1]
%tmp21 = load ptr, ptr %src_addr, align 4 ; <ptr> [#uses=1]
%tmp22 = load i32, ptr %src_stride_addr, align 4 ; <i32> [#uses=1]
%tmp23 = getelementptr i8, ptr %tmp21, i32 %tmp22 ; <ptr> [#uses=1]
%tmp25 = load i32, ptr %src_stride_addr, align 4 ; <i32> [#uses=1]
%tmp26 = mul i32 %tmp25, 2 ; <i32> [#uses=1]
%tmp27 = load ptr, ptr %src_addr, align 4 ; <ptr> [#uses=1]
%tmp28 = getelementptr i8, ptr %tmp27, i32 %tmp26 ; <ptr> [#uses=1]
%tmp30 = load i32, ptr %src_stride_addr, align 4 ; <i32> [#uses=1]
%tmp31 = mul i32 %tmp30, 3 ; <i32> [#uses=1]
%tmp32 = load ptr, ptr %src_addr, align 4 ; <ptr> [#uses=1]
%tmp33 = getelementptr i8, ptr %tmp32, i32 %tmp31 ; <ptr> [#uses=1]
call void asm sideeffect "movd $4, %mm0 \0A\09movd $5, %mm1 \0A\09movd $6, %mm2 \0A\09movd $7, %mm3 \0A\09punpcklbw %mm1, %mm0 \0A\09punpcklbw %mm3, %mm2 \0A\09movq %mm0, %mm1 \0A\09punpcklwd %mm2, %mm0 \0A\09punpckhwd %mm2, %mm1 \0A\09movd %mm0, $0 \0A\09punpckhdq %mm0, %mm0 \0A\09movd %mm0, $1 \0A\09movd %mm1, $2 \0A\09punpckhdq %mm1, %mm1 \0A\09movd %mm1, $3 \0A\09", "=*m,=*m,=*m,=*m,*m,*m,*m,*m,~{dirflag},~{fpsr},~{flags}"( ptr elementtype( i32) %tmp1, ptr elementtype(i32) %tmp5, ptr elementtype(i32) %tmp10, ptr elementtype(i32) %tmp16, ptr elementtype(i32) %tmp19, ptr elementtype(i32) %tmp23, ptr elementtype(i32) %tmp28, ptr elementtype(i32) %tmp33 ) nounwind
br label %return
return: ; preds = %entry
ret void
}