Author: Pengcheng Wang Date: 2025-01-13T11:34:54+08:00 New Revision: ad21c47ef81ed15630d556521d0900ff658b9108
URL: https://github.com/llvm/llvm-project/commit/ad21c47ef81ed15630d556521d0900ff658b9108 DIFF: https://github.com/llvm/llvm-project/commit/ad21c47ef81ed15630d556521d0900ff658b9108.diff LOG: Revert "[RISCV] Rework memcpy test (#120364)" This reverts commit 59bba39a692fd371d0dd0e6baba49a414bf7d855. Added: Modified: llvm/test/CodeGen/RISCV/memcpy.ll Removed: ################################################################################ diff --git a/llvm/test/CodeGen/RISCV/memcpy.ll b/llvm/test/CodeGen/RISCV/memcpy.ll index ce47476de9ce88..1ab3722080f700 100644 --- a/llvm/test/CodeGen/RISCV/memcpy.ll +++ b/llvm/test/CodeGen/RISCV/memcpy.ll @@ -7,935 +7,406 @@ ; RUN: | FileCheck %s --check-prefixes=RV32-BOTH,RV32-FAST ; RUN: llc < %s -mtriple=riscv64 -mattr=+unaligned-scalar-mem \ ; RUN: | FileCheck %s --check-prefixes=RV64-BOTH,RV64-FAST +%struct.x = type { i8, i8, i8, i8, i8, i8, i8, i8, i8, i8, i8 } -; ---------------------------------------------------------------------- -; Fully unaligned cases +@src = external dso_local global %struct.x +@dst = external dso_local global %struct.x -define void @unaligned_memcpy0(ptr nocapture %dest, ptr %src) nounwind { -; RV32-BOTH-LABEL: unaligned_memcpy0: -; RV32-BOTH: # %bb.0: # %entry -; RV32-BOTH-NEXT: ret -; -; RV64-BOTH-LABEL: unaligned_memcpy0: -; RV64-BOTH: # %bb.0: # %entry -; RV64-BOTH-NEXT: ret -entry: - tail call void @llvm.memcpy.p0.p0.i64(ptr %dest, ptr %src, i64 0, i1 false) - ret void -} - -define void @unaligned_memcpy1(ptr nocapture %dest, ptr %src) nounwind { -; RV32-BOTH-LABEL: unaligned_memcpy1: -; RV32-BOTH: # %bb.0: # %entry -; RV32-BOTH-NEXT: lbu a1, 0(a1) -; RV32-BOTH-NEXT: sb a1, 0(a0) -; RV32-BOTH-NEXT: ret -; -; RV64-BOTH-LABEL: unaligned_memcpy1: -; RV64-BOTH: # %bb.0: # %entry -; RV64-BOTH-NEXT: lbu a1, 0(a1) -; RV64-BOTH-NEXT: sb a1, 0(a0) -; RV64-BOTH-NEXT: ret -entry: - tail call void @llvm.memcpy.p0.p0.i64(ptr %dest, ptr %src, i64 1, i1 false) - ret void -} - -define void @unaligned_memcpy2(ptr nocapture %dest, ptr %src) nounwind { -; RV32-LABEL: unaligned_memcpy2: -; RV32: # %bb.0: # %entry -; RV32-NEXT: lbu a2, 1(a1) -; RV32-NEXT: sb a2, 1(a0) -; RV32-NEXT: lbu a1, 0(a1) -; RV32-NEXT: sb a1, 0(a0) -; RV32-NEXT: ret -; -; RV64-LABEL: unaligned_memcpy2: -; RV64: # %bb.0: # %entry -; RV64-NEXT: lbu a2, 1(a1) -; RV64-NEXT: sb a2, 1(a0) -; RV64-NEXT: lbu a1, 0(a1) -; RV64-NEXT: sb a1, 0(a0) -; RV64-NEXT: ret -; -; RV32-FAST-LABEL: unaligned_memcpy2: -; RV32-FAST: # %bb.0: # %entry -; RV32-FAST-NEXT: lh a1, 0(a1) -; RV32-FAST-NEXT: sh a1, 0(a0) -; RV32-FAST-NEXT: ret -; -; RV64-FAST-LABEL: unaligned_memcpy2: -; RV64-FAST: # %bb.0: # %entry -; RV64-FAST-NEXT: lh a1, 0(a1) -; RV64-FAST-NEXT: sh a1, 0(a0) -; RV64-FAST-NEXT: ret -entry: - tail call void @llvm.memcpy.p0.p0.i64(ptr %dest, ptr %src, i64 2, i1 false) - ret void -} - -define void @unaligned_memcpy3(ptr nocapture %dest, ptr %src) nounwind { -; RV32-LABEL: unaligned_memcpy3: -; RV32: # %bb.0: # %entry -; RV32-NEXT: lbu a2, 2(a1) -; RV32-NEXT: sb a2, 2(a0) -; RV32-NEXT: lbu a2, 1(a1) -; RV32-NEXT: sb a2, 1(a0) -; RV32-NEXT: lbu a1, 0(a1) -; RV32-NEXT: sb a1, 0(a0) -; RV32-NEXT: ret -; -; RV64-LABEL: unaligned_memcpy3: -; RV64: # %bb.0: # %entry -; RV64-NEXT: lbu a2, 2(a1) -; RV64-NEXT: sb a2, 2(a0) -; RV64-NEXT: lbu a2, 1(a1) -; RV64-NEXT: sb a2, 1(a0) -; RV64-NEXT: lbu a1, 0(a1) -; RV64-NEXT: sb a1, 0(a0) -; RV64-NEXT: ret -; -; RV32-FAST-LABEL: unaligned_memcpy3: -; RV32-FAST: # %bb.0: # %entry -; RV32-FAST-NEXT: lbu a2, 2(a1) -; RV32-FAST-NEXT: sb a2, 2(a0) -; RV32-FAST-NEXT: lh a1, 0(a1) -; RV32-FAST-NEXT: sh a1, 0(a0) -; RV32-FAST-NEXT: ret -; -; RV64-FAST-LABEL: unaligned_memcpy3: -; RV64-FAST: # %bb.0: # %entry -; RV64-FAST-NEXT: lbu a2, 2(a1) -; RV64-FAST-NEXT: sb a2, 2(a0) -; RV64-FAST-NEXT: lh a1, 0(a1) -; RV64-FAST-NEXT: sh a1, 0(a0) -; RV64-FAST-NEXT: ret -entry: - tail call void @llvm.memcpy.p0.p0.i64(ptr %dest, ptr %src, i64 3, i1 false) - ret void -} - -define void @unaligned_memcpy4(ptr nocapture %dest, ptr %src) nounwind { -; RV32-LABEL: unaligned_memcpy4: -; RV32: # %bb.0: # %entry -; RV32-NEXT: lbu a2, 3(a1) -; RV32-NEXT: sb a2, 3(a0) -; RV32-NEXT: lbu a2, 2(a1) -; RV32-NEXT: sb a2, 2(a0) -; RV32-NEXT: lbu a2, 1(a1) -; RV32-NEXT: sb a2, 1(a0) -; RV32-NEXT: lbu a1, 0(a1) -; RV32-NEXT: sb a1, 0(a0) -; RV32-NEXT: ret -; -; RV64-LABEL: unaligned_memcpy4: -; RV64: # %bb.0: # %entry -; RV64-NEXT: lbu a2, 3(a1) -; RV64-NEXT: sb a2, 3(a0) -; RV64-NEXT: lbu a2, 2(a1) -; RV64-NEXT: sb a2, 2(a0) -; RV64-NEXT: lbu a2, 1(a1) -; RV64-NEXT: sb a2, 1(a0) -; RV64-NEXT: lbu a1, 0(a1) -; RV64-NEXT: sb a1, 0(a0) -; RV64-NEXT: ret -; -; RV32-FAST-LABEL: unaligned_memcpy4: -; RV32-FAST: # %bb.0: # %entry -; RV32-FAST-NEXT: lw a1, 0(a1) -; RV32-FAST-NEXT: sw a1, 0(a0) -; RV32-FAST-NEXT: ret -; -; RV64-FAST-LABEL: unaligned_memcpy4: -; RV64-FAST: # %bb.0: # %entry -; RV64-FAST-NEXT: lw a1, 0(a1) -; RV64-FAST-NEXT: sw a1, 0(a0) -; RV64-FAST-NEXT: ret -entry: - tail call void @llvm.memcpy.p0.p0.i64(ptr %dest, ptr %src, i64 4, i1 false) - ret void -} +@.str1 = private unnamed_addr constant [31 x i8] c"DHRYSTONE PROGRAM, SOME STRING\00", align 1 +@.str2 = private unnamed_addr constant [36 x i8] c"DHRYSTONE PROGRAM, SOME STRING BLAH\00", align 1 +@.str3 = private unnamed_addr constant [24 x i8] c"DHRYSTONE PROGRAM, SOME\00", align 1 +@.str4 = private unnamed_addr constant [18 x i8] c"DHRYSTONE PROGR \00", align 1 +@.str5 = private unnamed_addr constant [7 x i8] c"DHRYST\00", align 1 +@.str6 = private unnamed_addr constant [14 x i8] c"/tmp/rmXXXXXX\00", align 1 +@spool.splbuf = internal global [512 x i8] zeroinitializer, align 16 -define void @unaligned_memcpy7(ptr nocapture %dest, ptr %src) nounwind { -; RV32-LABEL: unaligned_memcpy7: +define i32 @t0() { +; RV32-LABEL: t0: ; RV32: # %bb.0: # %entry -; RV32-NEXT: lbu a2, 6(a1) -; RV32-NEXT: sb a2, 6(a0) -; RV32-NEXT: lbu a2, 5(a1) -; RV32-NEXT: sb a2, 5(a0) -; RV32-NEXT: lbu a2, 4(a1) -; RV32-NEXT: sb a2, 4(a0) -; RV32-NEXT: lbu a2, 3(a1) -; RV32-NEXT: sb a2, 3(a0) -; RV32-NEXT: lbu a2, 2(a1) -; RV32-NEXT: sb a2, 2(a0) -; RV32-NEXT: lbu a2, 1(a1) -; RV32-NEXT: sb a2, 1(a0) -; RV32-NEXT: lbu a1, 0(a1) -; RV32-NEXT: sb a1, 0(a0) -; RV32-NEXT: ret -; -; RV64-LABEL: unaligned_memcpy7: -; RV64: # %bb.0: # %entry -; RV64-NEXT: lbu a2, 6(a1) -; RV64-NEXT: sb a2, 6(a0) -; RV64-NEXT: lbu a2, 5(a1) -; RV64-NEXT: sb a2, 5(a0) -; RV64-NEXT: lbu a2, 4(a1) -; RV64-NEXT: sb a2, 4(a0) -; RV64-NEXT: lbu a2, 3(a1) -; RV64-NEXT: sb a2, 3(a0) -; RV64-NEXT: lbu a2, 2(a1) -; RV64-NEXT: sb a2, 2(a0) -; RV64-NEXT: lbu a2, 1(a1) -; RV64-NEXT: sb a2, 1(a0) -; RV64-NEXT: lbu a1, 0(a1) -; RV64-NEXT: sb a1, 0(a0) -; RV64-NEXT: ret -; -; RV32-FAST-LABEL: unaligned_memcpy7: -; RV32-FAST: # %bb.0: # %entry -; RV32-FAST-NEXT: lw a2, 3(a1) -; RV32-FAST-NEXT: sw a2, 3(a0) -; RV32-FAST-NEXT: lw a1, 0(a1) -; RV32-FAST-NEXT: sw a1, 0(a0) -; RV32-FAST-NEXT: ret -; -; RV64-FAST-LABEL: unaligned_memcpy7: -; RV64-FAST: # %bb.0: # %entry -; RV64-FAST-NEXT: lw a2, 3(a1) -; RV64-FAST-NEXT: sw a2, 3(a0) -; RV64-FAST-NEXT: lw a1, 0(a1) -; RV64-FAST-NEXT: sw a1, 0(a0) -; RV64-FAST-NEXT: ret -entry: - tail call void @llvm.memcpy.p0.p0.i64(ptr %dest, ptr %src, i64 7, i1 false) - ret void -} - -define void @unaligned_memcpy8(ptr nocapture %dest, ptr %src) nounwind { -; RV32-LABEL: unaligned_memcpy8: -; RV32: # %bb.0: # %entry -; RV32-NEXT: lbu a2, 7(a1) -; RV32-NEXT: sb a2, 7(a0) -; RV32-NEXT: lbu a2, 6(a1) -; RV32-NEXT: sb a2, 6(a0) -; RV32-NEXT: lbu a2, 5(a1) -; RV32-NEXT: sb a2, 5(a0) -; RV32-NEXT: lbu a2, 4(a1) -; RV32-NEXT: sb a2, 4(a0) -; RV32-NEXT: lbu a2, 3(a1) -; RV32-NEXT: sb a2, 3(a0) -; RV32-NEXT: lbu a2, 2(a1) -; RV32-NEXT: sb a2, 2(a0) -; RV32-NEXT: lbu a2, 1(a1) -; RV32-NEXT: sb a2, 1(a0) -; RV32-NEXT: lbu a1, 0(a1) -; RV32-NEXT: sb a1, 0(a0) +; RV32-NEXT: lui a0, %hi(src) +; RV32-NEXT: lw a1, %lo(src)(a0) +; RV32-NEXT: lui a2, %hi(dst) +; RV32-NEXT: addi a0, a0, %lo(src) +; RV32-NEXT: sw a1, %lo(dst)(a2) +; RV32-NEXT: lw a1, 4(a0) +; RV32-NEXT: lh a3, 8(a0) +; RV32-NEXT: lbu a0, 10(a0) +; RV32-NEXT: addi a2, a2, %lo(dst) +; RV32-NEXT: sw a1, 4(a2) +; RV32-NEXT: sh a3, 8(a2) +; RV32-NEXT: sb a0, 10(a2) +; RV32-NEXT: li a0, 0 ; RV32-NEXT: ret ; -; RV64-LABEL: unaligned_memcpy8: +; RV64-LABEL: t0: ; RV64: # %bb.0: # %entry -; RV64-NEXT: lbu a2, 7(a1) -; RV64-NEXT: sb a2, 7(a0) -; RV64-NEXT: lbu a2, 6(a1) -; RV64-NEXT: sb a2, 6(a0) -; RV64-NEXT: lbu a2, 5(a1) -; RV64-NEXT: sb a2, 5(a0) -; RV64-NEXT: lbu a2, 4(a1) -; RV64-NEXT: sb a2, 4(a0) -; RV64-NEXT: lbu a2, 3(a1) -; RV64-NEXT: sb a2, 3(a0) -; RV64-NEXT: lbu a2, 2(a1) -; RV64-NEXT: sb a2, 2(a0) -; RV64-NEXT: lbu a2, 1(a1) -; RV64-NEXT: sb a2, 1(a0) -; RV64-NEXT: lbu a1, 0(a1) -; RV64-NEXT: sb a1, 0(a0) +; RV64-NEXT: lui a0, %hi(src) +; RV64-NEXT: lui a1, %hi(dst) +; RV64-NEXT: ld a2, %lo(src)(a0) +; RV64-NEXT: addi a0, a0, %lo(src) +; RV64-NEXT: lh a3, 8(a0) +; RV64-NEXT: lbu a0, 10(a0) +; RV64-NEXT: sd a2, %lo(dst)(a1) +; RV64-NEXT: addi a1, a1, %lo(dst) +; RV64-NEXT: sh a3, 8(a1) +; RV64-NEXT: sb a0, 10(a1) +; RV64-NEXT: li a0, 0 ; RV64-NEXT: ret ; -; RV32-FAST-LABEL: unaligned_memcpy8: +; RV32-FAST-LABEL: t0: ; RV32-FAST: # %bb.0: # %entry -; RV32-FAST-NEXT: lw a2, 4(a1) -; RV32-FAST-NEXT: sw a2, 4(a0) -; RV32-FAST-NEXT: lw a1, 0(a1) -; RV32-FAST-NEXT: sw a1, 0(a0) +; RV32-FAST-NEXT: lui a0, %hi(src) +; RV32-FAST-NEXT: lw a1, %lo(src)(a0) +; RV32-FAST-NEXT: addi a0, a0, %lo(src) +; RV32-FAST-NEXT: lw a2, 4(a0) +; RV32-FAST-NEXT: lw a0, 7(a0) +; RV32-FAST-NEXT: lui a3, %hi(dst) +; RV32-FAST-NEXT: sw a1, %lo(dst)(a3) +; RV32-FAST-NEXT: addi a1, a3, %lo(dst) +; RV32-FAST-NEXT: sw a0, 7(a1) +; RV32-FAST-NEXT: sw a2, 4(a1) +; RV32-FAST-NEXT: li a0, 0 ; RV32-FAST-NEXT: ret ; -; RV64-FAST-LABEL: unaligned_memcpy8: +; RV64-FAST-LABEL: t0: ; RV64-FAST: # %bb.0: # %entry -; RV64-FAST-NEXT: ld a1, 0(a1) -; RV64-FAST-NEXT: sd a1, 0(a0) +; RV64-FAST-NEXT: lui a0, %hi(src) +; RV64-FAST-NEXT: ld a1, %lo(src)(a0) +; RV64-FAST-NEXT: addi a0, a0, %lo(src) +; RV64-FAST-NEXT: lw a0, 7(a0) +; RV64-FAST-NEXT: lui a2, %hi(dst) +; RV64-FAST-NEXT: sd a1, %lo(dst)(a2) +; RV64-FAST-NEXT: addi a1, a2, %lo(dst) +; RV64-FAST-NEXT: sw a0, 7(a1) +; RV64-FAST-NEXT: li a0, 0 ; RV64-FAST-NEXT: ret entry: - tail call void @llvm.memcpy.p0.p0.i64(ptr %dest, ptr %src, i64 8, i1 false) - ret void + call void @llvm.memcpy.p0.p0.i32(ptr align 8 @dst, ptr align 8 @src, i32 11, i1 false) + ret i32 0 } -define void @unaligned_memcpy15(ptr nocapture %dest, ptr %src) nounwind { -; RV32-LABEL: unaligned_memcpy15: +define void @t1(ptr nocapture %C) nounwind { +; RV32-LABEL: t1: ; RV32: # %bb.0: # %entry -; RV32-NEXT: lbu a2, 14(a1) -; RV32-NEXT: sb a2, 14(a0) -; RV32-NEXT: lbu a2, 13(a1) -; RV32-NEXT: sb a2, 13(a0) -; RV32-NEXT: lbu a2, 12(a1) -; RV32-NEXT: sb a2, 12(a0) -; RV32-NEXT: lbu a2, 11(a1) -; RV32-NEXT: sb a2, 11(a0) -; RV32-NEXT: lbu a2, 10(a1) -; RV32-NEXT: sb a2, 10(a0) -; RV32-NEXT: lbu a2, 9(a1) -; RV32-NEXT: sb a2, 9(a0) -; RV32-NEXT: lbu a2, 8(a1) -; RV32-NEXT: sb a2, 8(a0) -; RV32-NEXT: lbu a2, 7(a1) -; RV32-NEXT: sb a2, 7(a0) -; RV32-NEXT: lbu a2, 6(a1) -; RV32-NEXT: sb a2, 6(a0) -; RV32-NEXT: lbu a2, 5(a1) -; RV32-NEXT: sb a2, 5(a0) -; RV32-NEXT: lbu a2, 4(a1) -; RV32-NEXT: sb a2, 4(a0) -; RV32-NEXT: lbu a2, 3(a1) -; RV32-NEXT: sb a2, 3(a0) -; RV32-NEXT: lbu a2, 2(a1) -; RV32-NEXT: sb a2, 2(a0) -; RV32-NEXT: lbu a2, 1(a1) -; RV32-NEXT: sb a2, 1(a0) -; RV32-NEXT: lbu a1, 0(a1) -; RV32-NEXT: sb a1, 0(a0) -; RV32-NEXT: ret +; RV32-NEXT: lui a1, %hi(.L.str1) +; RV32-NEXT: addi a1, a1, %lo(.L.str1) +; RV32-NEXT: li a2, 31 +; RV32-NEXT: tail memcpy ; -; RV64-LABEL: unaligned_memcpy15: +; RV64-LABEL: t1: ; RV64: # %bb.0: # %entry -; RV64-NEXT: lbu a2, 14(a1) -; RV64-NEXT: sb a2, 14(a0) -; RV64-NEXT: lbu a2, 13(a1) -; RV64-NEXT: sb a2, 13(a0) -; RV64-NEXT: lbu a2, 12(a1) -; RV64-NEXT: sb a2, 12(a0) -; RV64-NEXT: lbu a2, 11(a1) -; RV64-NEXT: sb a2, 11(a0) -; RV64-NEXT: lbu a2, 10(a1) -; RV64-NEXT: sb a2, 10(a0) -; RV64-NEXT: lbu a2, 9(a1) -; RV64-NEXT: sb a2, 9(a0) -; RV64-NEXT: lbu a2, 8(a1) -; RV64-NEXT: sb a2, 8(a0) -; RV64-NEXT: lbu a2, 7(a1) -; RV64-NEXT: sb a2, 7(a0) -; RV64-NEXT: lbu a2, 6(a1) -; RV64-NEXT: sb a2, 6(a0) -; RV64-NEXT: lbu a2, 5(a1) -; RV64-NEXT: sb a2, 5(a0) -; RV64-NEXT: lbu a2, 4(a1) -; RV64-NEXT: sb a2, 4(a0) -; RV64-NEXT: lbu a2, 3(a1) -; RV64-NEXT: sb a2, 3(a0) -; RV64-NEXT: lbu a2, 2(a1) -; RV64-NEXT: sb a2, 2(a0) -; RV64-NEXT: lbu a2, 1(a1) -; RV64-NEXT: sb a2, 1(a0) -; RV64-NEXT: lbu a1, 0(a1) -; RV64-NEXT: sb a1, 0(a0) -; RV64-NEXT: ret +; RV64-NEXT: lui a1, %hi(.L.str1) +; RV64-NEXT: addi a1, a1, %lo(.L.str1) +; RV64-NEXT: li a2, 31 +; RV64-NEXT: tail memcpy ; -; RV32-FAST-LABEL: unaligned_memcpy15: +; RV32-FAST-LABEL: t1: ; RV32-FAST: # %bb.0: # %entry -; RV32-FAST-NEXT: lw a2, 11(a1) -; RV32-FAST-NEXT: sw a2, 11(a0) -; RV32-FAST-NEXT: lw a2, 8(a1) -; RV32-FAST-NEXT: sw a2, 8(a0) -; RV32-FAST-NEXT: lw a2, 4(a1) -; RV32-FAST-NEXT: sw a2, 4(a0) -; RV32-FAST-NEXT: lw a1, 0(a1) -; RV32-FAST-NEXT: sw a1, 0(a0) +; RV32-FAST-NEXT: lui a1, 1141 +; RV32-FAST-NEXT: lui a2, 300325 +; RV32-FAST-NEXT: lui a3, 132181 +; RV32-FAST-NEXT: lui a4, 340483 +; RV32-FAST-NEXT: lui a5, 267556 +; RV32-FAST-NEXT: lui a6, 337154 +; RV32-FAST-NEXT: addi a1, a1, -439 +; RV32-FAST-NEXT: sw a1, 27(a0) +; RV32-FAST-NEXT: lui a1, 320757 +; RV32-FAST-NEXT: addi a2, a2, 1107 +; RV32-FAST-NEXT: addi a3, a3, -689 +; RV32-FAST-NEXT: addi a4, a4, -947 +; RV32-FAST-NEXT: sw a4, 16(a0) +; RV32-FAST-NEXT: sw a3, 20(a0) +; RV32-FAST-NEXT: sw a2, 24(a0) +; RV32-FAST-NEXT: lui a2, 365861 +; RV32-FAST-NEXT: addi a3, a5, 1871 +; RV32-FAST-NEXT: addi a4, a6, 69 +; RV32-FAST-NEXT: addi a1, a1, 1107 +; RV32-FAST-NEXT: addi a2, a2, -1980 +; RV32-FAST-NEXT: sw a2, 0(a0) +; RV32-FAST-NEXT: sw a1, 4(a0) +; RV32-FAST-NEXT: sw a4, 8(a0) +; RV32-FAST-NEXT: sw a3, 12(a0) ; RV32-FAST-NEXT: ret ; -; RV64-FAST-LABEL: unaligned_memcpy15: +; RV64-FAST-LABEL: t1: ; RV64-FAST: # %bb.0: # %entry -; RV64-FAST-NEXT: ld a2, 7(a1) -; RV64-FAST-NEXT: sd a2, 7(a0) -; RV64-FAST-NEXT: ld a1, 0(a1) +; RV64-FAST-NEXT: lui a1, %hi(.L.str1) +; RV64-FAST-NEXT: addi a2, a1, %lo(.L.str1) +; RV64-FAST-NEXT: ld a3, 23(a2) +; RV64-FAST-NEXT: ld a1, %lo(.L.str1)(a1) +; RV64-FAST-NEXT: ld a4, 8(a2) +; RV64-FAST-NEXT: ld a2, 16(a2) +; RV64-FAST-NEXT: sd a3, 23(a0) ; RV64-FAST-NEXT: sd a1, 0(a0) +; RV64-FAST-NEXT: sd a4, 8(a0) +; RV64-FAST-NEXT: sd a2, 16(a0) ; RV64-FAST-NEXT: ret entry: - tail call void @llvm.memcpy.p0.p0.i64(ptr %dest, ptr %src, i64 15, i1 false) + tail call void @llvm.memcpy.p0.p0.i64(ptr %C, ptr @.str1, i64 31, i1 false) ret void } -define void @unaligned_memcpy16(ptr nocapture %dest, ptr %src) nounwind { -; RV32-LABEL: unaligned_memcpy16: -; RV32: # %bb.0: # %entry -; RV32-NEXT: lbu a2, 15(a1) -; RV32-NEXT: sb a2, 15(a0) -; RV32-NEXT: lbu a2, 14(a1) -; RV32-NEXT: sb a2, 14(a0) -; RV32-NEXT: lbu a2, 13(a1) -; RV32-NEXT: sb a2, 13(a0) -; RV32-NEXT: lbu a2, 12(a1) -; RV32-NEXT: sb a2, 12(a0) -; RV32-NEXT: lbu a2, 11(a1) -; RV32-NEXT: sb a2, 11(a0) -; RV32-NEXT: lbu a2, 10(a1) -; RV32-NEXT: sb a2, 10(a0) -; RV32-NEXT: lbu a2, 9(a1) -; RV32-NEXT: sb a2, 9(a0) -; RV32-NEXT: lbu a2, 8(a1) -; RV32-NEXT: sb a2, 8(a0) -; RV32-NEXT: lbu a2, 7(a1) -; RV32-NEXT: sb a2, 7(a0) -; RV32-NEXT: lbu a2, 6(a1) -; RV32-NEXT: sb a2, 6(a0) -; RV32-NEXT: lbu a2, 5(a1) -; RV32-NEXT: sb a2, 5(a0) -; RV32-NEXT: lbu a2, 4(a1) -; RV32-NEXT: sb a2, 4(a0) -; RV32-NEXT: lbu a2, 3(a1) -; RV32-NEXT: sb a2, 3(a0) -; RV32-NEXT: lbu a2, 2(a1) -; RV32-NEXT: sb a2, 2(a0) -; RV32-NEXT: lbu a2, 1(a1) -; RV32-NEXT: sb a2, 1(a0) -; RV32-NEXT: lbu a1, 0(a1) -; RV32-NEXT: sb a1, 0(a0) -; RV32-NEXT: ret +define void @t2(ptr nocapture %C) nounwind { +; RV32-BOTH-LABEL: t2: +; RV32-BOTH: # %bb.0: # %entry +; RV32-BOTH-NEXT: lui a1, %hi(.L.str2) +; RV32-BOTH-NEXT: addi a1, a1, %lo(.L.str2) +; RV32-BOTH-NEXT: li a2, 36 +; RV32-BOTH-NEXT: tail memcpy ; -; RV64-LABEL: unaligned_memcpy16: +; RV64-LABEL: t2: ; RV64: # %bb.0: # %entry -; RV64-NEXT: lbu a2, 15(a1) -; RV64-NEXT: sb a2, 15(a0) -; RV64-NEXT: lbu a2, 14(a1) -; RV64-NEXT: sb a2, 14(a0) -; RV64-NEXT: lbu a2, 13(a1) -; RV64-NEXT: sb a2, 13(a0) -; RV64-NEXT: lbu a2, 12(a1) -; RV64-NEXT: sb a2, 12(a0) -; RV64-NEXT: lbu a2, 11(a1) -; RV64-NEXT: sb a2, 11(a0) -; RV64-NEXT: lbu a2, 10(a1) -; RV64-NEXT: sb a2, 10(a0) -; RV64-NEXT: lbu a2, 9(a1) -; RV64-NEXT: sb a2, 9(a0) -; RV64-NEXT: lbu a2, 8(a1) -; RV64-NEXT: sb a2, 8(a0) -; RV64-NEXT: lbu a2, 7(a1) -; RV64-NEXT: sb a2, 7(a0) -; RV64-NEXT: lbu a2, 6(a1) -; RV64-NEXT: sb a2, 6(a0) -; RV64-NEXT: lbu a2, 5(a1) -; RV64-NEXT: sb a2, 5(a0) -; RV64-NEXT: lbu a2, 4(a1) -; RV64-NEXT: sb a2, 4(a0) -; RV64-NEXT: lbu a2, 3(a1) -; RV64-NEXT: sb a2, 3(a0) -; RV64-NEXT: lbu a2, 2(a1) -; RV64-NEXT: sb a2, 2(a0) -; RV64-NEXT: lbu a2, 1(a1) -; RV64-NEXT: sb a2, 1(a0) -; RV64-NEXT: lbu a1, 0(a1) -; RV64-NEXT: sb a1, 0(a0) -; RV64-NEXT: ret -; -; RV32-FAST-LABEL: unaligned_memcpy16: -; RV32-FAST: # %bb.0: # %entry -; RV32-FAST-NEXT: lw a2, 12(a1) -; RV32-FAST-NEXT: sw a2, 12(a0) -; RV32-FAST-NEXT: lw a2, 8(a1) -; RV32-FAST-NEXT: sw a2, 8(a0) -; RV32-FAST-NEXT: lw a2, 4(a1) -; RV32-FAST-NEXT: sw a2, 4(a0) -; RV32-FAST-NEXT: lw a1, 0(a1) -; RV32-FAST-NEXT: sw a1, 0(a0) -; RV32-FAST-NEXT: ret +; RV64-NEXT: lui a1, %hi(.L.str2) +; RV64-NEXT: addi a1, a1, %lo(.L.str2) +; RV64-NEXT: li a2, 36 +; RV64-NEXT: tail memcpy ; -; RV64-FAST-LABEL: unaligned_memcpy16: +; RV64-FAST-LABEL: t2: ; RV64-FAST: # %bb.0: # %entry +; RV64-FAST-NEXT: lui a1, %hi(.L.str2) +; RV64-FAST-NEXT: lui a2, 1156 +; RV64-FAST-NEXT: ld a3, %lo(.L.str2)(a1) +; RV64-FAST-NEXT: addi a2, a2, 332 +; RV64-FAST-NEXT: addi a1, a1, %lo(.L.str2) +; RV64-FAST-NEXT: sw a2, 32(a0) ; RV64-FAST-NEXT: ld a2, 8(a1) +; RV64-FAST-NEXT: ld a4, 16(a1) +; RV64-FAST-NEXT: ld a1, 24(a1) +; RV64-FAST-NEXT: sd a3, 0(a0) ; RV64-FAST-NEXT: sd a2, 8(a0) -; RV64-FAST-NEXT: ld a1, 0(a1) -; RV64-FAST-NEXT: sd a1, 0(a0) +; RV64-FAST-NEXT: sd a4, 16(a0) +; RV64-FAST-NEXT: sd a1, 24(a0) ; RV64-FAST-NEXT: ret entry: - tail call void @llvm.memcpy.p0.p0.i64(ptr %dest, ptr %src, i64 16, i1 false) + tail call void @llvm.memcpy.p0.p0.i64(ptr %C, ptr @.str2, i64 36, i1 false) ret void } -define void @unaligned_memcpy31(ptr nocapture %dest, ptr %src) nounwind { -; RV32-LABEL: unaligned_memcpy31: +define void @t3(ptr nocapture %C) nounwind { +; RV32-LABEL: t3: ; RV32: # %bb.0: # %entry -; RV32-NEXT: lbu a2, 30(a1) -; RV32-NEXT: sb a2, 30(a0) -; RV32-NEXT: lbu a2, 29(a1) -; RV32-NEXT: sb a2, 29(a0) -; RV32-NEXT: lbu a2, 28(a1) -; RV32-NEXT: sb a2, 28(a0) -; RV32-NEXT: lbu a2, 27(a1) -; RV32-NEXT: sb a2, 27(a0) -; RV32-NEXT: lbu a2, 26(a1) -; RV32-NEXT: sb a2, 26(a0) -; RV32-NEXT: lbu a2, 25(a1) -; RV32-NEXT: sb a2, 25(a0) -; RV32-NEXT: lbu a2, 24(a1) -; RV32-NEXT: sb a2, 24(a0) -; RV32-NEXT: lbu a2, 23(a1) -; RV32-NEXT: sb a2, 23(a0) -; RV32-NEXT: lbu a2, 22(a1) -; RV32-NEXT: sb a2, 22(a0) -; RV32-NEXT: lbu a2, 21(a1) -; RV32-NEXT: sb a2, 21(a0) -; RV32-NEXT: lbu a2, 20(a1) -; RV32-NEXT: sb a2, 20(a0) -; RV32-NEXT: lbu a2, 19(a1) -; RV32-NEXT: sb a2, 19(a0) -; RV32-NEXT: lbu a2, 18(a1) -; RV32-NEXT: sb a2, 18(a0) -; RV32-NEXT: lbu a2, 17(a1) -; RV32-NEXT: sb a2, 17(a0) -; RV32-NEXT: lbu a2, 16(a1) -; RV32-NEXT: sb a2, 16(a0) -; RV32-NEXT: lbu a2, 15(a1) -; RV32-NEXT: sb a2, 15(a0) -; RV32-NEXT: lbu a2, 14(a1) -; RV32-NEXT: sb a2, 14(a0) -; RV32-NEXT: lbu a2, 13(a1) -; RV32-NEXT: sb a2, 13(a0) -; RV32-NEXT: lbu a2, 12(a1) -; RV32-NEXT: sb a2, 12(a0) -; RV32-NEXT: lbu a2, 11(a1) -; RV32-NEXT: sb a2, 11(a0) -; RV32-NEXT: lbu a2, 10(a1) -; RV32-NEXT: sb a2, 10(a0) -; RV32-NEXT: lbu a2, 9(a1) -; RV32-NEXT: sb a2, 9(a0) -; RV32-NEXT: lbu a2, 8(a1) -; RV32-NEXT: sb a2, 8(a0) -; RV32-NEXT: lbu a2, 7(a1) -; RV32-NEXT: sb a2, 7(a0) -; RV32-NEXT: lbu a2, 6(a1) -; RV32-NEXT: sb a2, 6(a0) -; RV32-NEXT: lbu a2, 5(a1) -; RV32-NEXT: sb a2, 5(a0) -; RV32-NEXT: lbu a2, 4(a1) -; RV32-NEXT: sb a2, 4(a0) -; RV32-NEXT: lbu a2, 3(a1) -; RV32-NEXT: sb a2, 3(a0) -; RV32-NEXT: lbu a2, 2(a1) -; RV32-NEXT: sb a2, 2(a0) -; RV32-NEXT: lbu a2, 1(a1) -; RV32-NEXT: sb a2, 1(a0) -; RV32-NEXT: lbu a1, 0(a1) -; RV32-NEXT: sb a1, 0(a0) -; RV32-NEXT: ret +; RV32-NEXT: lui a1, %hi(.L.str3) +; RV32-NEXT: addi a1, a1, %lo(.L.str3) +; RV32-NEXT: li a2, 24 +; RV32-NEXT: tail memcpy ; -; RV64-LABEL: unaligned_memcpy31: +; RV64-LABEL: t3: ; RV64: # %bb.0: # %entry -; RV64-NEXT: lbu a2, 30(a1) -; RV64-NEXT: sb a2, 30(a0) -; RV64-NEXT: lbu a2, 29(a1) -; RV64-NEXT: sb a2, 29(a0) -; RV64-NEXT: lbu a2, 28(a1) -; RV64-NEXT: sb a2, 28(a0) -; RV64-NEXT: lbu a2, 27(a1) -; RV64-NEXT: sb a2, 27(a0) -; RV64-NEXT: lbu a2, 26(a1) -; RV64-NEXT: sb a2, 26(a0) -; RV64-NEXT: lbu a2, 25(a1) -; RV64-NEXT: sb a2, 25(a0) -; RV64-NEXT: lbu a2, 24(a1) -; RV64-NEXT: sb a2, 24(a0) -; RV64-NEXT: lbu a2, 23(a1) -; RV64-NEXT: sb a2, 23(a0) -; RV64-NEXT: lbu a2, 22(a1) -; RV64-NEXT: sb a2, 22(a0) -; RV64-NEXT: lbu a2, 21(a1) -; RV64-NEXT: sb a2, 21(a0) -; RV64-NEXT: lbu a2, 20(a1) -; RV64-NEXT: sb a2, 20(a0) -; RV64-NEXT: lbu a2, 19(a1) -; RV64-NEXT: sb a2, 19(a0) -; RV64-NEXT: lbu a2, 18(a1) -; RV64-NEXT: sb a2, 18(a0) -; RV64-NEXT: lbu a2, 17(a1) -; RV64-NEXT: sb a2, 17(a0) -; RV64-NEXT: lbu a2, 16(a1) -; RV64-NEXT: sb a2, 16(a0) -; RV64-NEXT: lbu a2, 15(a1) -; RV64-NEXT: sb a2, 15(a0) -; RV64-NEXT: lbu a2, 14(a1) -; RV64-NEXT: sb a2, 14(a0) -; RV64-NEXT: lbu a2, 13(a1) -; RV64-NEXT: sb a2, 13(a0) -; RV64-NEXT: lbu a2, 12(a1) -; RV64-NEXT: sb a2, 12(a0) -; RV64-NEXT: lbu a2, 11(a1) -; RV64-NEXT: sb a2, 11(a0) -; RV64-NEXT: lbu a2, 10(a1) -; RV64-NEXT: sb a2, 10(a0) -; RV64-NEXT: lbu a2, 9(a1) -; RV64-NEXT: sb a2, 9(a0) -; RV64-NEXT: lbu a2, 8(a1) -; RV64-NEXT: sb a2, 8(a0) -; RV64-NEXT: lbu a2, 7(a1) -; RV64-NEXT: sb a2, 7(a0) -; RV64-NEXT: lbu a2, 6(a1) -; RV64-NEXT: sb a2, 6(a0) -; RV64-NEXT: lbu a2, 5(a1) -; RV64-NEXT: sb a2, 5(a0) -; RV64-NEXT: lbu a2, 4(a1) -; RV64-NEXT: sb a2, 4(a0) -; RV64-NEXT: lbu a2, 3(a1) -; RV64-NEXT: sb a2, 3(a0) -; RV64-NEXT: lbu a2, 2(a1) -; RV64-NEXT: sb a2, 2(a0) -; RV64-NEXT: lbu a2, 1(a1) -; RV64-NEXT: sb a2, 1(a0) -; RV64-NEXT: lbu a1, 0(a1) -; RV64-NEXT: sb a1, 0(a0) -; RV64-NEXT: ret +; RV64-NEXT: lui a1, %hi(.L.str3) +; RV64-NEXT: addi a1, a1, %lo(.L.str3) +; RV64-NEXT: li a2, 24 +; RV64-NEXT: tail memcpy ; -; RV32-FAST-LABEL: unaligned_memcpy31: +; RV32-FAST-LABEL: t3: ; RV32-FAST: # %bb.0: # %entry -; RV32-FAST-NEXT: lw a2, 27(a1) -; RV32-FAST-NEXT: sw a2, 27(a0) -; RV32-FAST-NEXT: lw a2, 24(a1) -; RV32-FAST-NEXT: sw a2, 24(a0) -; RV32-FAST-NEXT: lw a2, 20(a1) -; RV32-FAST-NEXT: sw a2, 20(a0) -; RV32-FAST-NEXT: lw a2, 16(a1) +; RV32-FAST-NEXT: lui a1, 1109 +; RV32-FAST-NEXT: lui a2, 340483 +; RV32-FAST-NEXT: lui a3, 267556 +; RV32-FAST-NEXT: lui a4, 337154 +; RV32-FAST-NEXT: lui a5, 320757 +; RV32-FAST-NEXT: addi a1, a1, -689 +; RV32-FAST-NEXT: addi a2, a2, -947 ; RV32-FAST-NEXT: sw a2, 16(a0) -; RV32-FAST-NEXT: lw a2, 12(a1) -; RV32-FAST-NEXT: sw a2, 12(a0) -; RV32-FAST-NEXT: lw a2, 8(a1) -; RV32-FAST-NEXT: sw a2, 8(a0) -; RV32-FAST-NEXT: lw a2, 4(a1) -; RV32-FAST-NEXT: sw a2, 4(a0) -; RV32-FAST-NEXT: lw a1, 0(a1) +; RV32-FAST-NEXT: sw a1, 20(a0) +; RV32-FAST-NEXT: lui a1, 365861 +; RV32-FAST-NEXT: addi a2, a3, 1871 +; RV32-FAST-NEXT: addi a3, a4, 69 +; RV32-FAST-NEXT: addi a4, a5, 1107 +; RV32-FAST-NEXT: addi a1, a1, -1980 ; RV32-FAST-NEXT: sw a1, 0(a0) +; RV32-FAST-NEXT: sw a4, 4(a0) +; RV32-FAST-NEXT: sw a3, 8(a0) +; RV32-FAST-NEXT: sw a2, 12(a0) ; RV32-FAST-NEXT: ret ; -; RV64-FAST-LABEL: unaligned_memcpy31: +; RV64-FAST-LABEL: t3: ; RV64-FAST: # %bb.0: # %entry -; RV64-FAST-NEXT: ld a2, 23(a1) -; RV64-FAST-NEXT: sd a2, 23(a0) -; RV64-FAST-NEXT: ld a2, 16(a1) -; RV64-FAST-NEXT: sd a2, 16(a0) -; RV64-FAST-NEXT: ld a2, 8(a1) -; RV64-FAST-NEXT: sd a2, 8(a0) -; RV64-FAST-NEXT: ld a1, 0(a1) -; RV64-FAST-NEXT: sd a1, 0(a0) +; RV64-FAST-NEXT: lui a1, %hi(.L.str3) +; RV64-FAST-NEXT: ld a2, %lo(.L.str3)(a1) +; RV64-FAST-NEXT: addi a1, a1, %lo(.L.str3) +; RV64-FAST-NEXT: ld a3, 8(a1) +; RV64-FAST-NEXT: ld a1, 16(a1) +; RV64-FAST-NEXT: sd a2, 0(a0) +; RV64-FAST-NEXT: sd a3, 8(a0) +; RV64-FAST-NEXT: sd a1, 16(a0) ; RV64-FAST-NEXT: ret entry: - tail call void @llvm.memcpy.p0.p0.i64(ptr %dest, ptr %src, i64 31, i1 false) - ret void -} - -; ---------------------------------------------------------------------- -; Fully aligned cases - -define void @aligned_memcpy0(ptr nocapture %dest, ptr %src) nounwind { -; RV32-BOTH-LABEL: aligned_memcpy0: -; RV32-BOTH: # %bb.0: # %entry -; RV32-BOTH-NEXT: ret -; -; RV64-BOTH-LABEL: aligned_memcpy0: -; RV64-BOTH: # %bb.0: # %entry -; RV64-BOTH-NEXT: ret -entry: - tail call void @llvm.memcpy.p0.p0.i64(ptr align 8 %dest, ptr align 8 %src, i64 0, i1 false) - ret void -} - -define void @aligned_memcpy1(ptr nocapture %dest, ptr %src) nounwind { -; RV32-BOTH-LABEL: aligned_memcpy1: -; RV32-BOTH: # %bb.0: # %entry -; RV32-BOTH-NEXT: lbu a1, 0(a1) -; RV32-BOTH-NEXT: sb a1, 0(a0) -; RV32-BOTH-NEXT: ret -; -; RV64-BOTH-LABEL: aligned_memcpy1: -; RV64-BOTH: # %bb.0: # %entry -; RV64-BOTH-NEXT: lbu a1, 0(a1) -; RV64-BOTH-NEXT: sb a1, 0(a0) -; RV64-BOTH-NEXT: ret -entry: - tail call void @llvm.memcpy.p0.p0.i64(ptr align 8 %dest, ptr align 8 %src, i64 1, i1 false) - ret void -} - -define void @aligned_memcpy2(ptr nocapture %dest, ptr %src) nounwind { -; RV32-BOTH-LABEL: aligned_memcpy2: -; RV32-BOTH: # %bb.0: # %entry -; RV32-BOTH-NEXT: lh a1, 0(a1) -; RV32-BOTH-NEXT: sh a1, 0(a0) -; RV32-BOTH-NEXT: ret -; -; RV64-BOTH-LABEL: aligned_memcpy2: -; RV64-BOTH: # %bb.0: # %entry -; RV64-BOTH-NEXT: lh a1, 0(a1) -; RV64-BOTH-NEXT: sh a1, 0(a0) -; RV64-BOTH-NEXT: ret -entry: - tail call void @llvm.memcpy.p0.p0.i64(ptr align 8 %dest, ptr align 8 %src, i64 2, i1 false) + tail call void @llvm.memcpy.p0.p0.i64(ptr %C, ptr @.str3, i64 24, i1 false) ret void } -define void @aligned_memcpy3(ptr nocapture %dest, ptr %src) nounwind { -; RV32-BOTH-LABEL: aligned_memcpy3: -; RV32-BOTH: # %bb.0: # %entry -; RV32-BOTH-NEXT: lbu a2, 2(a1) -; RV32-BOTH-NEXT: sb a2, 2(a0) -; RV32-BOTH-NEXT: lh a1, 0(a1) -; RV32-BOTH-NEXT: sh a1, 0(a0) -; RV32-BOTH-NEXT: ret -; -; RV64-BOTH-LABEL: aligned_memcpy3: -; RV64-BOTH: # %bb.0: # %entry -; RV64-BOTH-NEXT: lbu a2, 2(a1) -; RV64-BOTH-NEXT: sb a2, 2(a0) -; RV64-BOTH-NEXT: lh a1, 0(a1) -; RV64-BOTH-NEXT: sh a1, 0(a0) -; RV64-BOTH-NEXT: ret -entry: - tail call void @llvm.memcpy.p0.p0.i64(ptr align 8 %dest, ptr align 8 %src, i64 3, i1 false) - ret void -} - -define void @aligned_memcpy4(ptr nocapture %dest, ptr %src) nounwind { -; RV32-BOTH-LABEL: aligned_memcpy4: -; RV32-BOTH: # %bb.0: # %entry -; RV32-BOTH-NEXT: lw a1, 0(a1) -; RV32-BOTH-NEXT: sw a1, 0(a0) -; RV32-BOTH-NEXT: ret -; -; RV64-BOTH-LABEL: aligned_memcpy4: -; RV64-BOTH: # %bb.0: # %entry -; RV64-BOTH-NEXT: lw a1, 0(a1) -; RV64-BOTH-NEXT: sw a1, 0(a0) -; RV64-BOTH-NEXT: ret -entry: - tail call void @llvm.memcpy.p0.p0.i64(ptr align 8 %dest, ptr align 8 %src, i64 4, i1 false) - ret void -} - -define void @aligned_memcpy7(ptr nocapture %dest, ptr %src) nounwind { -; RV32-LABEL: aligned_memcpy7: +define void @t4(ptr nocapture %C) nounwind { +; RV32-LABEL: t4: ; RV32: # %bb.0: # %entry -; RV32-NEXT: lbu a2, 6(a1) -; RV32-NEXT: sb a2, 6(a0) -; RV32-NEXT: lh a2, 4(a1) -; RV32-NEXT: sh a2, 4(a0) -; RV32-NEXT: lw a1, 0(a1) -; RV32-NEXT: sw a1, 0(a0) -; RV32-NEXT: ret +; RV32-NEXT: lui a1, %hi(.L.str4) +; RV32-NEXT: addi a1, a1, %lo(.L.str4) +; RV32-NEXT: li a2, 18 +; RV32-NEXT: tail memcpy ; -; RV64-LABEL: aligned_memcpy7: +; RV64-LABEL: t4: ; RV64: # %bb.0: # %entry -; RV64-NEXT: lbu a2, 6(a1) -; RV64-NEXT: sb a2, 6(a0) -; RV64-NEXT: lh a2, 4(a1) -; RV64-NEXT: sh a2, 4(a0) -; RV64-NEXT: lw a1, 0(a1) -; RV64-NEXT: sw a1, 0(a0) -; RV64-NEXT: ret +; RV64-NEXT: lui a1, %hi(.L.str4) +; RV64-NEXT: addi a1, a1, %lo(.L.str4) +; RV64-NEXT: li a2, 18 +; RV64-NEXT: tail memcpy ; -; RV32-FAST-LABEL: aligned_memcpy7: +; RV32-FAST-LABEL: t4: ; RV32-FAST: # %bb.0: # %entry -; RV32-FAST-NEXT: lw a2, 3(a1) -; RV32-FAST-NEXT: sw a2, 3(a0) -; RV32-FAST-NEXT: lw a1, 0(a1) +; RV32-FAST-NEXT: li a1, 32 +; RV32-FAST-NEXT: lui a2, 132388 +; RV32-FAST-NEXT: lui a3, 337154 +; RV32-FAST-NEXT: lui a4, 320757 +; RV32-FAST-NEXT: sh a1, 16(a0) +; RV32-FAST-NEXT: lui a1, 365861 +; RV32-FAST-NEXT: addi a2, a2, 1871 +; RV32-FAST-NEXT: addi a3, a3, 69 +; RV32-FAST-NEXT: addi a4, a4, 1107 +; RV32-FAST-NEXT: addi a1, a1, -1980 ; RV32-FAST-NEXT: sw a1, 0(a0) +; RV32-FAST-NEXT: sw a4, 4(a0) +; RV32-FAST-NEXT: sw a3, 8(a0) +; RV32-FAST-NEXT: sw a2, 12(a0) ; RV32-FAST-NEXT: ret ; -; RV64-FAST-LABEL: aligned_memcpy7: +; RV64-FAST-LABEL: t4: ; RV64-FAST: # %bb.0: # %entry -; RV64-FAST-NEXT: lw a2, 3(a1) -; RV64-FAST-NEXT: sw a2, 3(a0) -; RV64-FAST-NEXT: lw a1, 0(a1) -; RV64-FAST-NEXT: sw a1, 0(a0) +; RV64-FAST-NEXT: lui a1, %hi(.L.str4) +; RV64-FAST-NEXT: ld a2, %lo(.L.str4)(a1) +; RV64-FAST-NEXT: addi a1, a1, %lo(.L.str4) +; RV64-FAST-NEXT: ld a1, 8(a1) +; RV64-FAST-NEXT: li a3, 32 +; RV64-FAST-NEXT: sd a2, 0(a0) +; RV64-FAST-NEXT: sd a1, 8(a0) +; RV64-FAST-NEXT: sh a3, 16(a0) ; RV64-FAST-NEXT: ret entry: - tail call void @llvm.memcpy.p0.p0.i64(ptr align 8 %dest, ptr align 8 %src, i64 7, i1 false) - ret void -} - -define void @aligned_memcpy8(ptr nocapture %dest, ptr %src) nounwind { -; RV32-BOTH-LABEL: aligned_memcpy8: -; RV32-BOTH: # %bb.0: # %entry -; RV32-BOTH-NEXT: lw a2, 4(a1) -; RV32-BOTH-NEXT: sw a2, 4(a0) -; RV32-BOTH-NEXT: lw a1, 0(a1) -; RV32-BOTH-NEXT: sw a1, 0(a0) -; RV32-BOTH-NEXT: ret -; -; RV64-BOTH-LABEL: aligned_memcpy8: -; RV64-BOTH: # %bb.0: # %entry -; RV64-BOTH-NEXT: ld a1, 0(a1) -; RV64-BOTH-NEXT: sd a1, 0(a0) -; RV64-BOTH-NEXT: ret -entry: - tail call void @llvm.memcpy.p0.p0.i64(ptr align 8 %dest, ptr align 8 %src, i64 8, i1 false) + tail call void @llvm.memcpy.p0.p0.i64(ptr %C, ptr @.str4, i64 18, i1 false) ret void } -define void @aligned_memcpy15(ptr nocapture %dest, ptr %src) nounwind { -; RV32-LABEL: aligned_memcpy15: +define void @t5(ptr nocapture %C) nounwind { +; RV32-LABEL: t5: ; RV32: # %bb.0: # %entry -; RV32-NEXT: lbu a2, 14(a1) -; RV32-NEXT: sb a2, 14(a0) -; RV32-NEXT: lh a2, 12(a1) -; RV32-NEXT: sh a2, 12(a0) -; RV32-NEXT: lw a2, 8(a1) -; RV32-NEXT: sw a2, 8(a0) -; RV32-NEXT: lw a2, 4(a1) -; RV32-NEXT: sw a2, 4(a0) -; RV32-NEXT: lw a1, 0(a1) -; RV32-NEXT: sw a1, 0(a0) +; RV32-NEXT: li a1, 84 +; RV32-NEXT: li a2, 83 +; RV32-NEXT: li a3, 89 +; RV32-NEXT: li a4, 82 +; RV32-NEXT: li a5, 72 +; RV32-NEXT: li a6, 68 +; RV32-NEXT: sb a2, 4(a0) +; RV32-NEXT: sb a1, 5(a0) +; RV32-NEXT: sb zero, 6(a0) +; RV32-NEXT: sb a6, 0(a0) +; RV32-NEXT: sb a5, 1(a0) +; RV32-NEXT: sb a4, 2(a0) +; RV32-NEXT: sb a3, 3(a0) ; RV32-NEXT: ret ; -; RV64-LABEL: aligned_memcpy15: +; RV64-LABEL: t5: ; RV64: # %bb.0: # %entry -; RV64-NEXT: lbu a2, 14(a1) -; RV64-NEXT: sb a2, 14(a0) -; RV64-NEXT: lh a2, 12(a1) -; RV64-NEXT: sh a2, 12(a0) -; RV64-NEXT: lw a2, 8(a1) -; RV64-NEXT: sw a2, 8(a0) -; RV64-NEXT: ld a1, 0(a1) -; RV64-NEXT: sd a1, 0(a0) +; RV64-NEXT: li a1, 84 +; RV64-NEXT: li a2, 83 +; RV64-NEXT: li a3, 89 +; RV64-NEXT: li a4, 82 +; RV64-NEXT: li a5, 72 +; RV64-NEXT: li a6, 68 +; RV64-NEXT: sb a2, 4(a0) +; RV64-NEXT: sb a1, 5(a0) +; RV64-NEXT: sb zero, 6(a0) +; RV64-NEXT: sb a6, 0(a0) +; RV64-NEXT: sb a5, 1(a0) +; RV64-NEXT: sb a4, 2(a0) +; RV64-NEXT: sb a3, 3(a0) ; RV64-NEXT: ret ; -; RV32-FAST-LABEL: aligned_memcpy15: +; RV32-FAST-LABEL: t5: ; RV32-FAST: # %bb.0: # %entry -; RV32-FAST-NEXT: lw a2, 11(a1) -; RV32-FAST-NEXT: sw a2, 11(a0) -; RV32-FAST-NEXT: lw a2, 8(a1) -; RV32-FAST-NEXT: sw a2, 8(a0) -; RV32-FAST-NEXT: lw a2, 4(a1) -; RV32-FAST-NEXT: sw a2, 4(a0) -; RV32-FAST-NEXT: lw a1, 0(a1) +; RV32-FAST-NEXT: lui a1, 1349 +; RV32-FAST-NEXT: addi a1, a1, 857 +; RV32-FAST-NEXT: sw a1, 3(a0) +; RV32-FAST-NEXT: lui a1, 365861 +; RV32-FAST-NEXT: addi a1, a1, -1980 ; RV32-FAST-NEXT: sw a1, 0(a0) ; RV32-FAST-NEXT: ret ; -; RV64-FAST-LABEL: aligned_memcpy15: +; RV64-FAST-LABEL: t5: ; RV64-FAST: # %bb.0: # %entry -; RV64-FAST-NEXT: ld a2, 7(a1) -; RV64-FAST-NEXT: sd a2, 7(a0) -; RV64-FAST-NEXT: ld a1, 0(a1) -; RV64-FAST-NEXT: sd a1, 0(a0) +; RV64-FAST-NEXT: lui a1, 1349 +; RV64-FAST-NEXT: addi a1, a1, 857 +; RV64-FAST-NEXT: sw a1, 3(a0) +; RV64-FAST-NEXT: lui a1, 365861 +; RV64-FAST-NEXT: addi a1, a1, -1980 +; RV64-FAST-NEXT: sw a1, 0(a0) ; RV64-FAST-NEXT: ret entry: - tail call void @llvm.memcpy.p0.p0.i64(ptr align 8 %dest, ptr align 8 %src, i64 15, i1 false) + tail call void @llvm.memcpy.p0.p0.i64(ptr %C, ptr @.str5, i64 7, i1 false) ret void } -define void @aligned_memcpy16(ptr nocapture %dest, ptr %src) nounwind { -; RV32-BOTH-LABEL: aligned_memcpy16: -; RV32-BOTH: # %bb.0: # %entry -; RV32-BOTH-NEXT: lw a2, 12(a1) -; RV32-BOTH-NEXT: sw a2, 12(a0) -; RV32-BOTH-NEXT: lw a2, 8(a1) -; RV32-BOTH-NEXT: sw a2, 8(a0) -; RV32-BOTH-NEXT: lw a2, 4(a1) -; RV32-BOTH-NEXT: sw a2, 4(a0) -; RV32-BOTH-NEXT: lw a1, 0(a1) -; RV32-BOTH-NEXT: sw a1, 0(a0) -; RV32-BOTH-NEXT: ret -; -; RV64-BOTH-LABEL: aligned_memcpy16: -; RV64-BOTH: # %bb.0: # %entry -; RV64-BOTH-NEXT: ld a2, 8(a1) -; RV64-BOTH-NEXT: sd a2, 8(a0) -; RV64-BOTH-NEXT: ld a1, 0(a1) -; RV64-BOTH-NEXT: sd a1, 0(a0) -; RV64-BOTH-NEXT: ret -entry: - tail call void @llvm.memcpy.p0.p0.i64(ptr align 8 %dest, ptr align 8 %src, i64 16, i1 false) - ret void -} - -define void @aligned_memcpy31(ptr nocapture %dest, ptr %src) nounwind { -; RV32-LABEL: aligned_memcpy31: +define void @t6() nounwind { +; RV32-LABEL: t6: ; RV32: # %bb.0: # %entry -; RV32-NEXT: lbu a2, 30(a1) -; RV32-NEXT: sb a2, 30(a0) -; RV32-NEXT: lh a2, 28(a1) -; RV32-NEXT: sh a2, 28(a0) -; RV32-NEXT: lw a2, 24(a1) -; RV32-NEXT: sw a2, 24(a0) -; RV32-NEXT: lw a2, 20(a1) -; RV32-NEXT: sw a2, 20(a0) -; RV32-NEXT: lw a2, 16(a1) -; RV32-NEXT: sw a2, 16(a0) -; RV32-NEXT: lw a2, 12(a1) -; RV32-NEXT: sw a2, 12(a0) -; RV32-NEXT: lw a2, 8(a1) -; RV32-NEXT: sw a2, 8(a0) -; RV32-NEXT: lw a2, 4(a1) -; RV32-NEXT: sw a2, 4(a0) -; RV32-NEXT: lw a1, 0(a1) -; RV32-NEXT: sw a1, 0(a0) +; RV32-NEXT: addi sp, sp, -16 +; RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill +; RV32-NEXT: lui a0, %hi(spool.splbuf) +; RV32-NEXT: addi a0, a0, %lo(spool.splbuf) +; RV32-NEXT: lui a1, %hi(.L.str6) +; RV32-NEXT: addi a1, a1, %lo(.L.str6) +; RV32-NEXT: li a2, 14 +; RV32-NEXT: call memcpy +; RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload +; RV32-NEXT: addi sp, sp, 16 ; RV32-NEXT: ret ; -; RV64-LABEL: aligned_memcpy31: +; RV64-LABEL: t6: ; RV64: # %bb.0: # %entry -; RV64-NEXT: lbu a2, 30(a1) -; RV64-NEXT: sb a2, 30(a0) -; RV64-NEXT: lh a2, 28(a1) -; RV64-NEXT: sh a2, 28(a0) -; RV64-NEXT: lw a2, 24(a1) -; RV64-NEXT: sw a2, 24(a0) -; RV64-NEXT: ld a2, 16(a1) -; RV64-NEXT: sd a2, 16(a0) -; RV64-NEXT: ld a2, 8(a1) -; RV64-NEXT: sd a2, 8(a0) -; RV64-NEXT: ld a1, 0(a1) -; RV64-NEXT: sd a1, 0(a0) +; RV64-NEXT: addi sp, sp, -16 +; RV64-NEXT: sd ra, 8(sp) # 8-byte Folded Spill +; RV64-NEXT: lui a0, %hi(spool.splbuf) +; RV64-NEXT: addi a0, a0, %lo(spool.splbuf) +; RV64-NEXT: lui a1, %hi(.L.str6) +; RV64-NEXT: addi a1, a1, %lo(.L.str6) +; RV64-NEXT: li a2, 14 +; RV64-NEXT: call memcpy +; RV64-NEXT: ld ra, 8(sp) # 8-byte Folded Reload +; RV64-NEXT: addi sp, sp, 16 ; RV64-NEXT: ret ; -; RV32-FAST-LABEL: aligned_memcpy31: +; RV32-FAST-LABEL: t6: ; RV32-FAST: # %bb.0: # %entry -; RV32-FAST-NEXT: lw a2, 27(a1) -; RV32-FAST-NEXT: sw a2, 27(a0) -; RV32-FAST-NEXT: lw a2, 24(a1) -; RV32-FAST-NEXT: sw a2, 24(a0) -; RV32-FAST-NEXT: lw a2, 20(a1) -; RV32-FAST-NEXT: sw a2, 20(a0) -; RV32-FAST-NEXT: lw a2, 16(a1) -; RV32-FAST-NEXT: sw a2, 16(a0) -; RV32-FAST-NEXT: lw a2, 12(a1) -; RV32-FAST-NEXT: sw a2, 12(a0) -; RV32-FAST-NEXT: lw a2, 8(a1) -; RV32-FAST-NEXT: sw a2, 8(a0) -; RV32-FAST-NEXT: lw a2, 4(a1) -; RV32-FAST-NEXT: sw a2, 4(a0) -; RV32-FAST-NEXT: lw a1, 0(a1) -; RV32-FAST-NEXT: sw a1, 0(a0) +; RV32-FAST-NEXT: lui a0, %hi(spool.splbuf) +; RV32-FAST-NEXT: li a1, 88 +; RV32-FAST-NEXT: sh a1, %lo(spool.splbuf+12)(a0) +; RV32-FAST-NEXT: lui a1, 361862 +; RV32-FAST-NEXT: addi a1, a1, -1960 +; RV32-FAST-NEXT: sw a1, %lo(spool.splbuf+8)(a0) +; RV32-FAST-NEXT: lui a1, 362199 +; RV32-FAST-NEXT: addi a1, a1, 559 +; RV32-FAST-NEXT: sw a1, %lo(spool.splbuf+4)(a0) +; RV32-FAST-NEXT: lui a1, 460503 +; RV32-FAST-NEXT: addi a1, a1, 1071 +; RV32-FAST-NEXT: sw a1, %lo(spool.splbuf)(a0) ; RV32-FAST-NEXT: ret ; -; RV64-FAST-LABEL: aligned_memcpy31: +; RV64-FAST-LABEL: t6: ; RV64-FAST: # %bb.0: # %entry -; RV64-FAST-NEXT: ld a2, 23(a1) -; RV64-FAST-NEXT: sd a2, 23(a0) -; RV64-FAST-NEXT: ld a2, 16(a1) -; RV64-FAST-NEXT: sd a2, 16(a0) -; RV64-FAST-NEXT: ld a2, 8(a1) -; RV64-FAST-NEXT: sd a2, 8(a0) -; RV64-FAST-NEXT: ld a1, 0(a1) -; RV64-FAST-NEXT: sd a1, 0(a0) +; RV64-FAST-NEXT: lui a0, %hi(.L.str6) +; RV64-FAST-NEXT: ld a1, %lo(.L.str6)(a0) +; RV64-FAST-NEXT: addi a0, a0, %lo(.L.str6) +; RV64-FAST-NEXT: ld a0, 6(a0) +; RV64-FAST-NEXT: lui a2, %hi(spool.splbuf) +; RV64-FAST-NEXT: sd a1, %lo(spool.splbuf)(a2) +; RV64-FAST-NEXT: sd a0, %lo(spool.splbuf+6)(a2) ; RV64-FAST-NEXT: ret entry: - tail call void @llvm.memcpy.p0.p0.i64(ptr align 8 %dest, ptr align 8 %src, i64 31, i1 false) + call void @llvm.memcpy.p0.p0.i64(ptr @spool.splbuf, ptr @.str6, i64 14, i1 false) ret void } -; ------------------------------------------------------------------------ -; A few partially aligned cases +%struct.Foo = type { i32, i32, i32, i32 } - -define void @memcpy16_align4(ptr nocapture %dest, ptr nocapture %src) nounwind { -; RV32-BOTH-LABEL: memcpy16_align4: +define void @t7(ptr nocapture %a, ptr nocapture %b) nounwind { +; RV32-BOTH-LABEL: t7: ; RV32-BOTH: # %bb.0: # %entry ; RV32-BOTH-NEXT: lw a2, 12(a1) ; RV32-BOTH-NEXT: sw a2, 12(a0) @@ -947,7 +418,7 @@ define void @memcpy16_align4(ptr nocapture %dest, ptr nocapture %src) nounwind { ; RV32-BOTH-NEXT: sw a1, 0(a0) ; RV32-BOTH-NEXT: ret ; -; RV64-LABEL: memcpy16_align4: +; RV64-LABEL: t7: ; RV64: # %bb.0: # %entry ; RV64-NEXT: lw a2, 12(a1) ; RV64-NEXT: sw a2, 12(a0) @@ -959,7 +430,7 @@ define void @memcpy16_align4(ptr nocapture %dest, ptr nocapture %src) nounwind { ; RV64-NEXT: sw a1, 0(a0) ; RV64-NEXT: ret ; -; RV64-FAST-LABEL: memcpy16_align4: +; RV64-FAST-LABEL: t7: ; RV64-FAST: # %bb.0: # %entry ; RV64-FAST-NEXT: ld a2, 8(a1) ; RV64-FAST-NEXT: sd a2, 8(a0) @@ -967,58 +438,11 @@ define void @memcpy16_align4(ptr nocapture %dest, ptr nocapture %src) nounwind { ; RV64-FAST-NEXT: sd a1, 0(a0) ; RV64-FAST-NEXT: ret entry: - tail call void @llvm.memcpy.p0.p0.i32(ptr align 4 %dest, ptr align 4 %src, i32 16, i1 false) + tail call void @llvm.memcpy.p0.p0.i32(ptr align 4 %a, ptr align 4 %b, i32 16, i1 false) ret void } -define i32 @memcpy11_align8(ptr nocapture %dest, ptr %src) { -; RV32-LABEL: memcpy11_align8: -; RV32: # %bb.0: # %entry -; RV32-NEXT: lbu a2, 10(a1) -; RV32-NEXT: sb a2, 10(a0) -; RV32-NEXT: lh a2, 8(a1) -; RV32-NEXT: sh a2, 8(a0) -; RV32-NEXT: lw a2, 4(a1) -; RV32-NEXT: sw a2, 4(a0) -; RV32-NEXT: lw a1, 0(a1) -; RV32-NEXT: sw a1, 0(a0) -; RV32-NEXT: li a0, 0 -; RV32-NEXT: ret -; -; RV64-LABEL: memcpy11_align8: -; RV64: # %bb.0: # %entry -; RV64-NEXT: lbu a2, 10(a1) -; RV64-NEXT: sb a2, 10(a0) -; RV64-NEXT: lh a2, 8(a1) -; RV64-NEXT: sh a2, 8(a0) -; RV64-NEXT: ld a1, 0(a1) -; RV64-NEXT: sd a1, 0(a0) -; RV64-NEXT: li a0, 0 -; RV64-NEXT: ret -; -; RV32-FAST-LABEL: memcpy11_align8: -; RV32-FAST: # %bb.0: # %entry -; RV32-FAST-NEXT: lw a2, 7(a1) -; RV32-FAST-NEXT: sw a2, 7(a0) -; RV32-FAST-NEXT: lw a2, 4(a1) -; RV32-FAST-NEXT: sw a2, 4(a0) -; RV32-FAST-NEXT: lw a1, 0(a1) -; RV32-FAST-NEXT: sw a1, 0(a0) -; RV32-FAST-NEXT: li a0, 0 -; RV32-FAST-NEXT: ret -; -; RV64-FAST-LABEL: memcpy11_align8: -; RV64-FAST: # %bb.0: # %entry -; RV64-FAST-NEXT: lw a2, 7(a1) -; RV64-FAST-NEXT: sw a2, 7(a0) -; RV64-FAST-NEXT: ld a1, 0(a1) -; RV64-FAST-NEXT: sd a1, 0(a0) -; RV64-FAST-NEXT: li a0, 0 -; RV64-FAST-NEXT: ret -entry: - call void @llvm.memcpy.p0.p0.i32(ptr align 8 %dest, ptr align 8 %src, i32 11, i1 false) - ret i32 0 -} - declare void @llvm.memcpy.p0.p0.i32(ptr nocapture, ptr nocapture, i32, i1) nounwind declare void @llvm.memcpy.p0.p0.i64(ptr nocapture, ptr nocapture, i64, i1) nounwind +;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line: +; RV64-BOTH: {{.*}} _______________________________________________ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
