2 ; RUN: llc -mtriple=x86_64-unknown-linux-gnu < %s | FileCheck -check-prefix=X86 %s
3 ; RUN: llc -mtriple=x86_64-unknown-linux-gnu -debug-only=isel < %s -o /dev/null 2>&1 | FileCheck -check-prefix=DBGDAG %s
5 ; It's OK to merge the load / store of the first 2 components, but
6 ; they must not be placed on the same chain after merging.
8 ; X86-LABEL: {{^}}merge_store_partial_overlap_load:
9 ; X86-DAG: movzwl ([[BASEREG:%[a-z]+]]), %e[[LO2:[a-z]+]]
10 ; X86-DAG: movzbl 2([[BASEREG]]), %e[[HI1:[a-z]]]
12 ; X86-NEXT: movw %[[LO2]], 1([[BASEREG]])
13 ; X86-NEXT: movb %[[HI1]]l, 3([[BASEREG]])
16 ; DBGDAG-LABEL: Optimized legalized selection DAG: %bb.0 'merge_store_partial_overlap_load:'
17 ; DBGDAG: [[ENTRYTOKEN:t[0-9]+]]: ch,glue = EntryToken
18 ; DBGDAG-DAG: [[BASEPTR:t[0-9]+]]: i64,ch = CopyFromReg [[ENTRYTOKEN]],
19 ; DBGDAG-DAG: [[ADDPTR:t[0-9]+]]: i64 = add {{(nuw )?}}[[BASEPTR]], Constant:i64<2>
21 ; DBGDAG-DAG: [[LD2:t[0-9]+]]: i16,ch = load<(load (s16) from %ir.tmp81, align 1)> [[ENTRYTOKEN]], [[BASEPTR]], undef:i64
22 ; DBGDAG-DAG: [[LD1:t[0-9]+]]: i8,ch = load<(load (s8) from %ir.tmp12)> [[ENTRYTOKEN]], [[ADDPTR]], undef:i64
24 ; DBGDAG-DAG: [[ST1:t[0-9]+]]: ch = store<(store (s8) into %ir.tmp14)> [[ENTRYTOKEN]], [[LD1]], t{{[0-9]+}}, undef:i64
25 ; DBGDAG-DAG: [[LOADTOKEN:t[0-9]+]]: ch = TokenFactor [[LD2]]:1, [[LD1]]:1
26 ; DBGDAG-DAG: [[ST2:t[0-9]+]]: ch = store<(store (s16) into %ir.tmp10, align 1)> [[LOADTOKEN]], [[LD2]], t{{[0-9]+}}, undef:i64
28 ; DBGDAG: X86ISD::RET_GLUE t{{[0-9]+}},
30 ; DBGDAG-LABEL: Instruction selection begins
31 define void @merge_store_partial_overlap_load(ptr %tmp) {
32 %tmp8 = getelementptr inbounds [4 x i8], ptr %tmp, i32 0, i8 0
33 %tmp10 = getelementptr inbounds [4 x i8], ptr %tmp, i32 0, i8 1
34 %tmp12 = getelementptr inbounds [4 x i8], ptr %tmp, i32 0, i8 2
35 %tmp14 = getelementptr [4 x i8], ptr %tmp, i32 0, i8 3
37 %tmp9 = load i8, ptr %tmp8, align 1 ; base + 0
38 %tmp11 = load i8, ptr %tmp10, align 1 ; base + 1
39 %tmp13 = load i8, ptr %tmp12, align 1 ; base + 2
41 store i8 %tmp9, ptr %tmp10, align 1 ; base + 1
42 store i8 %tmp11, ptr %tmp12, align 1 ; base + 2
43 store i8 %tmp13, ptr %tmp14, align 1 ; base + 3
46 ; load base + 0, base + 1
47 ; store base + 1, base + 2