OSchip
/
llvm-project
mirror of https://github.com/THU-DSP-LAB/llvm-project.git
13
0
Fork 6
Code Issues Packages Projects Releases Wiki Activity

[AArch64][GlobalISel] Don't emit a branch for a fallthrough G_BR at -O0. 

With optimizations we leave the decision to eliminate fallthrough branches to
bock placement, but at -O0 we should do it in the selector to save code size.

This regressed -O0 with a recent change to a combiner.
This commit is contained in:
Amara Emerson 2020-09-10 14:57:16 -07:00
parent 4e3edef4b8
commit 0448d11a06
6 changed files with 202 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
llvm/lib/Target/AArch64/GISel/AArch64InstructionSelector.cpp
View File

@ -35,6 +35,7 @@
#include "llvm/IR/Constants.h"
#include "llvm/IR/Type.h"
#include "llvm/IR/IntrinsicsAArch64.h"
#include "llvm/Pass.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/raw_ostream.h"
@ -1755,6 +1756,18 @@ bool AArch64InstructionSelector::earlySelect(MachineInstr &I) const {
MachineRegisterInfo &MRI = MF.getRegInfo();
switch (I.getOpcode()) {
case TargetOpcode::G_BR: {
// If the branch jumps to the fallthrough block, don't bother emitting it.
// Only do this for -O0 for a good code size improvement, because when
// optimizations are enabled we want to leave this choice to
// MachineBlockPlacement.
Function &F = MF.getFunction();
bool EnableOpt = MF.getTarget().getOptLevel() != CodeGenOpt::None;
if (EnableOpt || !MBB.isLayoutSuccessor(I.getOperand(0).getMBB()))
return false;
I.eraseFromParent();
return true;
}
case TargetOpcode::G_SHL:
return earlySelectSHL(I, MRI);
case TargetOpcode::G_CONSTANT: {

1
llvm/test/CodeGen/AArch64/GlobalISel/select-binop.mir
View File

@ -330,7 +330,6 @@ body: |
; CHECK: bb.0:
; CHECK: successors: %bb.1(0x80000000)
; CHECK: [[COPY:%[0-9]+]]:gpr32sp = COPY $w0
; CHECK: B %bb.1
; CHECK: bb.1:
; CHECK: [[ADDWri:%[0-9]+]]:gpr32sp = ADDWri [[COPY]], 1, 0
; CHECK: $w0 = COPY [[ADDWri]]

1
llvm/test/CodeGen/AArch64/GlobalISel/select-jump-table-brjt-constrain.mir
View File

@ -35,7 +35,6 @@ body: |
; CHECK: BR %6
; CHECK: bb.2:
; CHECK: successors: %bb.3(0x80000000)
; CHECK: B %bb.3
; CHECK: bb.3:
; CHECK: RET_ReallyLR
bb.1:

3
llvm/test/CodeGen/AArch64/GlobalISel/select-returnaddress-liveins.mir
View File

@ -19,7 +19,6 @@ body: |
; CHECK: successors: %bb.1(0x80000000)
; CHECK: liveins: $w0, $x0, $lr
; CHECK: [[COPY:%[0-9]+]]:gpr64sp = COPY $lr
; CHECK: B %bb.1
; CHECK: bb.1:
; CHECK: [[COPY1:%[0-9]+]]:gpr64 = COPY [[COPY]]
; CHECK: $x0 = COPY [[COPY1]]
@ -47,7 +46,6 @@ body: |
; CHECK: successors: %bb.1(0x80000000)
; CHECK: liveins: $w0, $x0, $lr
; CHECK: [[COPY:%[0-9]+]]:gpr64sp = COPY $lr
; CHECK: B %bb.1
; CHECK: bb.1:
; CHECK: [[COPY1:%[0-9]+]]:gpr64 = COPY [[COPY]]
; CHECK: $x0 = COPY [[COPY1]]
@ -78,7 +76,6 @@ body: |
; CHECK: liveins: $w0, $x0, $lr
; CHECK: [[COPY:%[0-9]+]]:gpr64sp = COPY $lr
; CHECK: [[COPY1:%[0-9]+]]:gpr64 = COPY [[COPY]]
; CHECK: B %bb.1
; CHECK: bb.1:
; CHECK: $x0 = COPY [[COPY1]]
; CHECK: [[COPY2:%[0-9]+]]:gpr64 = COPY [[COPY]]

1
llvm/test/CodeGen/AArch64/GlobalISel/select-xor.mir
View File

@ -132,7 +132,6 @@ body: |
; CHECK-LABEL: name: xor_constant_n1_s32_gpr_2bb
; CHECK: bb.0:
; CHECK: successors: %bb.1(0x80000000)
; CHECK: B %bb.1
; CHECK: bb.1:
; CHECK: [[COPY:%[0-9]+]]:gpr32 = COPY $w0
; CHECK: [[ORNWrr:%[0-9]+]]:gpr32 = ORNWrr $wzr, [[COPY]]

190
llvm/test/CodeGen/AArch64/unwind-preserved.ll
View File

@ -1,6 +1,6 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --extra_scrub
; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sve -O0 -global-isel=0 -global-isel-abort=0 < %s | FileCheck %s
; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sve -O0 -global-isel=1 -global-isel-abort=0 < %s | FileCheck %s
; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sve -O0 -global-isel=1 -global-isel-abort=0 < %s | FileCheck %s --check-prefix=GISEL
; Test that z0 is saved/restored, as the unwinder may only retain the low 64bits (d0).
define <vscale x 4 x i32> @invoke_callee_may_throw_sve(<vscale x 4 x i32> %v) personality i8 0 {
@ -125,6 +125,128 @@ define <vscale x 4 x i32> @invoke_callee_may_throw_sve(<vscale x 4 x i32> %v) pe
; CHECK-NEXT: addvl sp, sp, #18
; CHECK-NEXT: ldp x29, x30, [sp], #16 // 16-byte Folded Reload
; CHECK-NEXT: ret
;
; GISEL-LABEL: invoke_callee_may_throw_sve:
; GISEL: .Lfunc_begin0:
; GISEL-NEXT: .cfi_startproc
; GISEL-NEXT: // %bb.0:
; GISEL-NEXT: stp x29, x30, [sp, #-16]! // 16-byte Folded Spill
; GISEL-NEXT: addvl sp, sp, #-18
; GISEL-NEXT: str p15, [sp, #4, mul vl] // 2-byte Folded Spill
; GISEL-NEXT: str p14, [sp, #5, mul vl] // 2-byte Folded Spill
; GISEL-NEXT: str p13, [sp, #6, mul vl] // 2-byte Folded Spill
; GISEL-NEXT: str p12, [sp, #7, mul vl] // 2-byte Folded Spill
; GISEL-NEXT: str p11, [sp, #8, mul vl] // 2-byte Folded Spill
; GISEL-NEXT: str p10, [sp, #9, mul vl] // 2-byte Folded Spill
; GISEL-NEXT: str p9, [sp, #10, mul vl] // 2-byte Folded Spill
; GISEL-NEXT: str p8, [sp, #11, mul vl] // 2-byte Folded Spill
; GISEL-NEXT: str p7, [sp, #12, mul vl] // 2-byte Folded Spill
; GISEL-NEXT: str p6, [sp, #13, mul vl] // 2-byte Folded Spill
; GISEL-NEXT: str p5, [sp, #14, mul vl] // 2-byte Folded Spill
; GISEL-NEXT: str p4, [sp, #15, mul vl] // 2-byte Folded Spill
; GISEL-NEXT: str z23, [sp, #2, mul vl] // 16-byte Folded Spill
; GISEL-NEXT: str z22, [sp, #3, mul vl] // 16-byte Folded Spill
; GISEL-NEXT: str z21, [sp, #4, mul vl] // 16-byte Folded Spill
; GISEL-NEXT: str z20, [sp, #5, mul vl] // 16-byte Folded Spill
; GISEL-NEXT: str z19, [sp, #6, mul vl] // 16-byte Folded Spill
; GISEL-NEXT: str z18, [sp, #7, mul vl] // 16-byte Folded Spill
; GISEL-NEXT: str z17, [sp, #8, mul vl] // 16-byte Folded Spill
; GISEL-NEXT: str z16, [sp, #9, mul vl] // 16-byte Folded Spill
; GISEL-NEXT: str z15, [sp, #10, mul vl] // 16-byte Folded Spill
; GISEL-NEXT: str z14, [sp, #11, mul vl] // 16-byte Folded Spill
; GISEL-NEXT: str z13, [sp, #12, mul vl] // 16-byte Folded Spill
; GISEL-NEXT: str z12, [sp, #13, mul vl] // 16-byte Folded Spill
; GISEL-NEXT: str z11, [sp, #14, mul vl] // 16-byte Folded Spill
; GISEL-NEXT: str z10, [sp, #15, mul vl] // 16-byte Folded Spill
; GISEL-NEXT: str z9, [sp, #16, mul vl] // 16-byte Folded Spill
; GISEL-NEXT: str z8, [sp, #17, mul vl] // 16-byte Folded Spill
; GISEL-NEXT: addvl sp, sp, #-2
; GISEL-NEXT: .cfi_escape 0x0f, 0x0d, 0x8f, 0x00, 0x11, 0x10, 0x22, 0x11, 0xa0, 0x01, 0x92, 0x2e, 0x00, 0x1e, 0x22 // sp + 16 + 160 * VG
; GISEL-NEXT: .cfi_escape 0x10, 0x48, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x78, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d8 @ cfa - 16 - 8 * VG
; GISEL-NEXT: .cfi_escape 0x10, 0x49, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x70, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d9 @ cfa - 16 - 16 * VG
; GISEL-NEXT: .cfi_escape 0x10, 0x4a, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x68, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d10 @ cfa - 16 - 24 * VG
; GISEL-NEXT: .cfi_escape 0x10, 0x4b, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x60, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d11 @ cfa - 16 - 32 * VG
; GISEL-NEXT: .cfi_escape 0x10, 0x4c, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x58, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d12 @ cfa - 16 - 40 * VG
; GISEL-NEXT: .cfi_escape 0x10, 0x4d, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x50, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d13 @ cfa - 16 - 48 * VG
; GISEL-NEXT: .cfi_escape 0x10, 0x4e, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x48, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d14 @ cfa - 16 - 56 * VG
; GISEL-NEXT: .cfi_escape 0x10, 0x4f, 0x0a, 0x11, 0x70, 0x22, 0x11, 0x40, 0x92, 0x2e, 0x00, 0x1e, 0x22 // $d15 @ cfa - 16 - 64 * VG
; GISEL-NEXT: .cfi_offset w30, -8
; GISEL-NEXT: .cfi_offset w29, -16
; GISEL-NEXT: .Ltmp0:
; GISEL-NEXT: str z0, [sp, #1, mul vl] // 16-byte Folded Spill
; GISEL-NEXT: bl may_throw_sve
; GISEL-NEXT: .Ltmp1:
; GISEL-NEXT: str z0, [sp] // 16-byte Folded Spill
; GISEL-NEXT: b .LBB0_1
; GISEL-NEXT: .LBB0_1: // %.Lcontinue
; GISEL-NEXT: ldr z0, [sp] // 16-byte Folded Reload
; GISEL-NEXT: addvl sp, sp, #2
; GISEL-NEXT: ldr p15, [sp, #4, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p14, [sp, #5, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p13, [sp, #6, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p12, [sp, #7, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p11, [sp, #8, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p10, [sp, #9, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p9, [sp, #10, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p8, [sp, #11, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p7, [sp, #12, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p6, [sp, #13, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p5, [sp, #14, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p4, [sp, #15, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr z23, [sp, #2, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z22, [sp, #3, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z21, [sp, #4, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z20, [sp, #5, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z19, [sp, #6, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z18, [sp, #7, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z17, [sp, #8, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z16, [sp, #9, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z15, [sp, #10, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z14, [sp, #11, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z13, [sp, #12, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z12, [sp, #13, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z11, [sp, #14, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z10, [sp, #15, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z9, [sp, #16, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z8, [sp, #17, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: addvl sp, sp, #18
; GISEL-NEXT: ldp x29, x30, [sp], #16 // 16-byte Folded Reload
; GISEL-NEXT: ret
; GISEL-NEXT: .LBB0_2: // %.Lunwind
; GISEL-NEXT: .Ltmp2:
; GISEL-NEXT: ldr z0, [sp, #1, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: addvl sp, sp, #2
; GISEL-NEXT: ldr p15, [sp, #4, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p14, [sp, #5, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p13, [sp, #6, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p12, [sp, #7, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p11, [sp, #8, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p10, [sp, #9, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p9, [sp, #10, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p8, [sp, #11, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p7, [sp, #12, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p6, [sp, #13, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p5, [sp, #14, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr p4, [sp, #15, mul vl] // 2-byte Folded Reload
; GISEL-NEXT: ldr z23, [sp, #2, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z22, [sp, #3, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z21, [sp, #4, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z20, [sp, #5, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z19, [sp, #6, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z18, [sp, #7, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z17, [sp, #8, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z16, [sp, #9, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z15, [sp, #10, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z14, [sp, #11, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z13, [sp, #12, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z12, [sp, #13, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z11, [sp, #14, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z10, [sp, #15, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z9, [sp, #16, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: ldr z8, [sp, #17, mul vl] // 16-byte Folded Reload
; GISEL-NEXT: addvl sp, sp, #18
; GISEL-NEXT: ldp x29, x30, [sp], #16 // 16-byte Folded Reload
; GISEL-NEXT: ret
%result = invoke <vscale x 4 x i32> @may_throw_sve(<vscale x 4 x i32> %v) to label %.Lcontinue unwind label %.Lunwind
.Lcontinue:
ret <vscale x 4 x i32> %result
@ -204,6 +326,72 @@ define aarch64_vector_pcs <4 x i32> @invoke_callee_may_throw_neon(<4 x i32> %v)
; CHECK-NEXT: ldp q23, q22, [sp, #32] // 32-byte Folded Reload
; CHECK-NEXT: add sp, sp, #304 // =304
; CHECK-NEXT: ret
;
; GISEL-LABEL: invoke_callee_may_throw_neon:
; GISEL: .Lfunc_begin1:
; GISEL-NEXT: .cfi_startproc
; GISEL-NEXT: // %bb.0:
; GISEL-NEXT: sub sp, sp, #304 // =304
; GISEL-NEXT: stp q23, q22, [sp, #32] // 32-byte Folded Spill
; GISEL-NEXT: stp q21, q20, [sp, #64] // 32-byte Folded Spill
; GISEL-NEXT: stp q19, q18, [sp, #96] // 32-byte Folded Spill
; GISEL-NEXT: stp q17, q16, [sp, #128] // 32-byte Folded Spill
; GISEL-NEXT: stp q15, q14, [sp, #160] // 32-byte Folded Spill
; GISEL-NEXT: stp q13, q12, [sp, #192] // 32-byte Folded Spill
; GISEL-NEXT: stp q11, q10, [sp, #224] // 32-byte Folded Spill
; GISEL-NEXT: stp q9, q8, [sp, #256] // 32-byte Folded Spill
; GISEL-NEXT: stp x29, x30, [sp, #288] // 16-byte Folded Spill
; GISEL-NEXT: .cfi_def_cfa_offset 304
; GISEL-NEXT: .cfi_offset w30, -8
; GISEL-NEXT: .cfi_offset w29, -16
; GISEL-NEXT: .cfi_offset b8, -32
; GISEL-NEXT: .cfi_offset b9, -48
; GISEL-NEXT: .cfi_offset b10, -64
; GISEL-NEXT: .cfi_offset b11, -80
; GISEL-NEXT: .cfi_offset b12, -96
; GISEL-NEXT: .cfi_offset b13, -112
; GISEL-NEXT: .cfi_offset b14, -128
; GISEL-NEXT: .cfi_offset b15, -144
; GISEL-NEXT: .cfi_offset b16, -160
; GISEL-NEXT: .cfi_offset b17, -176
; GISEL-NEXT: .cfi_offset b18, -192
; GISEL-NEXT: .cfi_offset b19, -208
; GISEL-NEXT: .cfi_offset b20, -224
; GISEL-NEXT: .cfi_offset b21, -240
; GISEL-NEXT: .cfi_offset b22, -256
; GISEL-NEXT: .cfi_offset b23, -272
; GISEL-NEXT: .Ltmp3:
; GISEL-NEXT: str q0, [sp, #16] // 16-byte Folded Spill
; GISEL-NEXT: bl may_throw_neon
; GISEL-NEXT: .Ltmp4:
; GISEL-NEXT: str q0, [sp] // 16-byte Folded Spill
; GISEL-NEXT: // %bb.1: // %.Lcontinue
; GISEL-NEXT: ldr q0, [sp] // 16-byte Folded Reload
; GISEL-NEXT: ldp x29, x30, [sp, #288] // 16-byte Folded Reload
; GISEL-NEXT: ldp q9, q8, [sp, #256] // 32-byte Folded Reload
; GISEL-NEXT: ldp q11, q10, [sp, #224] // 32-byte Folded Reload
; GISEL-NEXT: ldp q13, q12, [sp, #192] // 32-byte Folded Reload
; GISEL-NEXT: ldp q15, q14, [sp, #160] // 32-byte Folded Reload
; GISEL-NEXT: ldp q17, q16, [sp, #128] // 32-byte Folded Reload
; GISEL-NEXT: ldp q19, q18, [sp, #96] // 32-byte Folded Reload
; GISEL-NEXT: ldp q21, q20, [sp, #64] // 32-byte Folded Reload
; GISEL-NEXT: ldp q23, q22, [sp, #32] // 32-byte Folded Reload
; GISEL-NEXT: add sp, sp, #304 // =304
; GISEL-NEXT: ret
; GISEL-NEXT: .LBB1_2: // %.Lunwind
; GISEL-NEXT: .Ltmp5:
; GISEL-NEXT: ldr q0, [sp, #16] // 16-byte Folded Reload
; GISEL-NEXT: ldp x29, x30, [sp, #288] // 16-byte Folded Reload
; GISEL-NEXT: ldp q9, q8, [sp, #256] // 32-byte Folded Reload
; GISEL-NEXT: ldp q11, q10, [sp, #224] // 32-byte Folded Reload
; GISEL-NEXT: ldp q13, q12, [sp, #192] // 32-byte Folded Reload
; GISEL-NEXT: ldp q15, q14, [sp, #160] // 32-byte Folded Reload
; GISEL-NEXT: ldp q17, q16, [sp, #128] // 32-byte Folded Reload
; GISEL-NEXT: ldp q19, q18, [sp, #96] // 32-byte Folded Reload
; GISEL-NEXT: ldp q21, q20, [sp, #64] // 32-byte Folded Reload
; GISEL-NEXT: ldp q23, q22, [sp, #32] // 32-byte Folded Reload
; GISEL-NEXT: add sp, sp, #304 // =304
; GISEL-NEXT: ret
%result = invoke aarch64_vector_pcs <4 x i32> @may_throw_neon(<4 x i32> %v) to label %.Lcontinue unwind label %.Lunwind
.Lcontinue:
ret <4 x i32> %result