Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[AArch64] Fix postinc operands for Cortex-A510 scheduling #68518

Closed
wants to merge 1 commit into from
Closed
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
68 changes: 34 additions & 34 deletions llvm/lib/Target/AArch64/AArch64SchedA510.td
Original file line number Diff line number Diff line change
Expand Up @@ -295,49 +295,49 @@ def : InstRW<[CortexA510WriteVLD2], (instregex "LD1Threev(16b|8h|4s|2d)$")>;
def : InstRW<[CortexA510WriteVLD2], (instregex "LD1Fourv(8b|4h|2s|1d)$")>;
def : InstRW<[CortexA510WriteVLD2], (instregex "LD1Fourv(16b|8h|4s|2d)$")>;

def : InstRW<[CortexA510WriteVLD1, WriteAdr], (instregex "LD1i(8|16|32|64)_POST$")>;
def : InstRW<[CortexA510WriteVLD1, WriteAdr], (instregex "LD1Rv(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
def : InstRW<[CortexA510WriteVLD1, WriteAdr], (instregex "LD1Onev(8b|4h|2s|1d)_POST$")>;
def : InstRW<[CortexA510WriteVLD2, WriteAdr], (instregex "LD1Onev(16b|8h|4s|2d)_POST$")>;
def : InstRW<[CortexA510WriteVLD2, WriteAdr], (instregex "LD1Twov(8b|4h|2s|1d)_POST$")>;
def : InstRW<[CortexA510WriteVLD2, WriteAdr], (instregex "LD1Twov(16b|8h|4s|2d)_POST$")>;
def : InstRW<[CortexA510WriteVLD2, WriteAdr], (instregex "LD1Threev(8b|4h|2s|1d)_POST$")>;
def : InstRW<[CortexA510WriteVLD2, WriteAdr], (instregex "LD1Threev(16b|8h|4s|2d)_POST$")>;
def : InstRW<[CortexA510WriteVLD2, WriteAdr], (instregex "LD1Fourv(8b|4h|2s|1d)_POST$")>;
def : InstRW<[CortexA510WriteVLD2, WriteAdr], (instregex "LD1Fourv(16b|8h|4s|2d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD1], (instregex "LD1i(8|16|32|64)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD1], (instregex "LD1Rv(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD1], (instregex "LD1Onev(8b|4h|2s|1d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD2], (instregex "LD1Onev(16b|8h|4s|2d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD2], (instregex "LD1Twov(8b|4h|2s|1d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD2], (instregex "LD1Twov(16b|8h|4s|2d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD2], (instregex "LD1Threev(8b|4h|2s|1d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD2], (instregex "LD1Threev(16b|8h|4s|2d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD2], (instregex "LD1Fourv(8b|4h|2s|1d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD2], (instregex "LD1Fourv(16b|8h|4s|2d)_POST$")>;

// 2-element structures
def : InstRW<[CortexA510WriteVLD2], (instregex "LD2i(8|16|32|64)$")>;
def : InstRW<[CortexA510WriteVLD2], (instregex "LD2Rv(8b|4h|2s|1d|16b|8h|4s|2d)$")>;
def : InstRW<[CortexA510WriteVLD2], (instregex "LD2Twov(8b|4h|2s)$")>;
def : InstRW<[CortexA510WriteVLD4], (instregex "LD2Twov(16b|8h|4s|2d)$")>;

def : InstRW<[CortexA510WriteVLD2, WriteAdr], (instregex "LD2i(8|16|32|64)(_POST)?$")>;
def : InstRW<[CortexA510WriteVLD2, WriteAdr], (instregex "LD2Rv(8b|4h|2s|1d|16b|8h|4s|2d)(_POST)?$")>;
def : InstRW<[CortexA510WriteVLD2, WriteAdr], (instregex "LD2Twov(8b|4h|2s)(_POST)?$")>;
def : InstRW<[CortexA510WriteVLD4, WriteAdr], (instregex "LD2Twov(16b|8h|4s|2d)(_POST)?$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD2], (instregex "LD2i(8|16|32|64)(_POST)?$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD2], (instregex "LD2Rv(8b|4h|2s|1d|16b|8h|4s|2d)(_POST)?$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD2], (instregex "LD2Twov(8b|4h|2s)(_POST)?$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD4], (instregex "LD2Twov(16b|8h|4s|2d)(_POST)?$")>;

// 3-element structures
def : InstRW<[CortexA510WriteVLD2], (instregex "LD3i(8|16|32|64)$")>;
def : InstRW<[CortexA510WriteVLD2], (instregex "LD3Rv(8b|4h|2s|1d|16b|8h|4s|2d)$")>;
def : InstRW<[CortexA510WriteVLD3], (instregex "LD3Threev(8b|4h|2s|1d)$")>;
def : InstRW<[CortexA510WriteVLD6], (instregex "LD3Threev(16b|8h|4s|2d)$")>;

def : InstRW<[CortexA510WriteVLD2, WriteAdr], (instregex "LD3i(8|16|32|64)_POST$")>;
def : InstRW<[CortexA510WriteVLD2, WriteAdr], (instregex "LD3Rv(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
def : InstRW<[CortexA510WriteVLD3, WriteAdr], (instregex "LD3Threev(8b|4h|2s|1d)_POST$")>;
def : InstRW<[CortexA510WriteVLD6, WriteAdr], (instregex "LD3Threev(16b|8h|4s|2d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD2], (instregex "LD3i(8|16|32|64)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD2], (instregex "LD3Rv(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD3], (instregex "LD3Threev(8b|4h|2s|1d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD6], (instregex "LD3Threev(16b|8h|4s|2d)_POST$")>;

// 4-element structures
def : InstRW<[CortexA510WriteVLD2], (instregex "LD4i(8|16|32|64)$")>; // load single 4-el structure to one lane of 4 regs.
def : InstRW<[CortexA510WriteVLD2], (instregex "LD4Rv(8b|4h|2s|1d|16b|8h|4s|2d)$")>; // load single 4-el structure, replicate to all lanes of 4 regs.
def : InstRW<[CortexA510WriteVLD4], (instregex "LD4Fourv(8b|4h|2s|1d)$")>; // load multiple 4-el structures to 4 regs.
def : InstRW<[CortexA510WriteVLD8], (instregex "LD4Fourv(16b|8h|4s|2d)$")>;

def : InstRW<[CortexA510WriteVLD2, WriteAdr], (instregex "LD4i(8|16|32|64)_POST$")>;
def : InstRW<[CortexA510WriteVLD2, WriteAdr], (instregex "LD4Rv(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
def : InstRW<[CortexA510WriteVLD4, WriteAdr], (instregex "LD4Fourv(8b|4h|2s|1d)_POST$")>;
def : InstRW<[CortexA510WriteVLD8, WriteAdr], (instregex "LD4Fourv(16b|8h|4s|2d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD2], (instregex "LD4i(8|16|32|64)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD2], (instregex "LD4Rv(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD4], (instregex "LD4Fourv(8b|4h|2s|1d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVLD8], (instregex "LD4Fourv(16b|8h|4s|2d)_POST$")>;

//---
// Vector Stores
Expand All @@ -347,28 +347,28 @@ def : InstRW<[CortexA510WriteVST1], (instregex "ST1Onev(8b|4h|2s|1d|16b|8h|4s|2d
def : InstRW<[CortexA510WriteVST1], (instregex "ST1Twov(8b|4h|2s|1d|16b|8h|4s|2d)$")>;
def : InstRW<[CortexA510WriteVST2], (instregex "ST1Threev(8b|4h|2s|1d|16b|8h|4s|2d)$")>;
def : InstRW<[CortexA510WriteVST4], (instregex "ST1Fourv(8b|4h|2s|1d|16b|8h|4s|2d)$")>;
def : InstRW<[CortexA510WriteVST1, WriteAdr], (instregex "ST1i(8|16|32|64)_POST$")>;
def : InstRW<[CortexA510WriteVST1, WriteAdr], (instregex "ST1Onev(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
def : InstRW<[CortexA510WriteVST1, WriteAdr], (instregex "ST1Twov(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
def : InstRW<[CortexA510WriteVST2, WriteAdr], (instregex "ST1Threev(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
def : InstRW<[CortexA510WriteVST4, WriteAdr], (instregex "ST1Fourv(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVST1], (instregex "ST1i(8|16|32|64)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVST1], (instregex "ST1Onev(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVST1], (instregex "ST1Twov(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVST2], (instregex "ST1Threev(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVST4], (instregex "ST1Fourv(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;

def : InstRW<[CortexA510WriteVST2], (instregex "ST2i(8|16|32|64)$")>;
def : InstRW<[CortexA510WriteVST2], (instregex "ST2Twov(8b|4h|2s)$")>;
def : InstRW<[CortexA510WriteVST4], (instregex "ST2Twov(16b|8h|4s|2d)$")>;
def : InstRW<[CortexA510WriteVST2, WriteAdr], (instregex "ST2i(8|16|32|64)_POST$")>;
def : InstRW<[CortexA510WriteVST2, WriteAdr], (instregex "ST2Twov(8b|4h|2s)_POST$")>;
def : InstRW<[CortexA510WriteVST4, WriteAdr], (instregex "ST2Twov(16b|8h|4s|2d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVST2], (instregex "ST2i(8|16|32|64)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVST2], (instregex "ST2Twov(8b|4h|2s)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVST4], (instregex "ST2Twov(16b|8h|4s|2d)_POST$")>;

def : InstRW<[CortexA510WriteVST2], (instregex "ST3i(8|16|32|64)$")>;
def : InstRW<[CortexA510WriteVST4], (instregex "ST3Threev(8b|4h|2s|1d|16b|8h|4s|2d)$")>;
def : InstRW<[CortexA510WriteVST2, WriteAdr], (instregex "ST3i(8|16|32|64)_POST$")>;
def : InstRW<[CortexA510WriteVST4, WriteAdr], (instregex "ST3Threev(8b|4h|2s|1d|2d|16b|8h|4s|4d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVST2], (instregex "ST3i(8|16|32|64)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVST4], (instregex "ST3Threev(8b|4h|2s|1d|2d|16b|8h|4s|4d)_POST$")>;

def : InstRW<[CortexA510WriteVST2], (instregex "ST4i(8|16|32|64)$")>;
def : InstRW<[CortexA510WriteVST4], (instregex "ST4Fourv(8b|4h|2s|1d|16b|8h|4s|2d)$")>;
def : InstRW<[CortexA510WriteVST2, WriteAdr], (instregex "ST4i(8|16|32|64)_POST$")>;
def : InstRW<[CortexA510WriteVST4, WriteAdr], (instregex "ST4Fourv(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVST2], (instregex "ST4i(8|16|32|64)_POST$")>;
def : InstRW<[WriteAdr, CortexA510WriteVST4], (instregex "ST4Fourv(8b|4h|2s|1d|16b|8h|4s|2d)_POST$")>;

//---
// Floating Point Conversions, MAC, DIV, SQRT
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -47,10 +47,10 @@ define void @f_undef_1(<8 x i64> %a, ptr %dst) {
; CHECK-LABEL: f_undef_1:
; CHECK: // %bb.0: // %BB
; CHECK-NEXT: mov v16.16b, v0.16b
; CHECK-NEXT: mov x8, x0
; CHECK-NEXT: mov v5.16b, v2.16b
; CHECK-NEXT: // kill: def $q1 killed $q1 def $q1_q2
; CHECK-NEXT: // kill: def $q3 killed $q3 def $q3_q4
; CHECK-NEXT: mov x8, x0
; CHECK-NEXT: mov v2.16b, v1.16b
; CHECK-NEXT: mov v4.16b, v3.16b
; CHECK-NEXT: mov v17.16b, v16.16b
Expand Down
18 changes: 9 additions & 9 deletions llvm/test/CodeGen/AArch64/arm64-indexed-vector-ldst.ll
Original file line number Diff line number Diff line change
Expand Up @@ -14320,8 +14320,8 @@ define <8 x i8> @test_v8i8_post_imm_ld1lane(ptr %bar, ptr %ptr, <8 x i8> %A) {
; CHECK: ; %bb.0:
; CHECK-NEXT: ; kill: def $d0 killed $d0 def $q0
; CHECK-NEXT: ld1.b { v0 }[1], [x0], #1
; CHECK-NEXT: ; kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: str x0, [x1]
; CHECK-NEXT: ; kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: ret
;
; CHECK-GISEL-LABEL: test_v8i8_post_imm_ld1lane:
Expand All @@ -14345,8 +14345,8 @@ define <8 x i8> @test_v8i8_post_reg_ld1lane(ptr %bar, ptr %ptr, i64 %inc, <8 x i
; CHECK: ; %bb.0:
; CHECK-NEXT: ; kill: def $d0 killed $d0 def $q0
; CHECK-NEXT: ld1.b { v0 }[1], [x0], x2
; CHECK-NEXT: ; kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: str x0, [x1]
; CHECK-NEXT: ; kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: ret
;
; CHECK-GISEL-LABEL: test_v8i8_post_reg_ld1lane:
Expand Down Expand Up @@ -14413,8 +14413,8 @@ define <4 x i16> @test_v4i16_post_imm_ld1lane(ptr %bar, ptr %ptr, <4 x i16> %A)
; CHECK: ; %bb.0:
; CHECK-NEXT: ; kill: def $d0 killed $d0 def $q0
; CHECK-NEXT: ld1.h { v0 }[1], [x0], #2
; CHECK-NEXT: ; kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: str x0, [x1]
; CHECK-NEXT: ; kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: ret
;
; CHECK-GISEL-LABEL: test_v4i16_post_imm_ld1lane:
Expand All @@ -14439,8 +14439,8 @@ define <4 x i16> @test_v4i16_post_reg_ld1lane(ptr %bar, ptr %ptr, i64 %inc, <4 x
; CHECK-NEXT: lsl x8, x2, #1
; CHECK-NEXT: ; kill: def $d0 killed $d0 def $q0
; CHECK-NEXT: ld1.h { v0 }[1], [x0], x8
; CHECK-NEXT: ; kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: str x0, [x1]
; CHECK-NEXT: ; kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: ret
;
; CHECK-GISEL-LABEL: test_v4i16_post_reg_ld1lane:
Expand Down Expand Up @@ -14507,8 +14507,8 @@ define <2 x i32> @test_v2i32_post_imm_ld1lane(ptr %bar, ptr %ptr, <2 x i32> %A)
; CHECK: ; %bb.0:
; CHECK-NEXT: ; kill: def $d0 killed $d0 def $q0
; CHECK-NEXT: ld1.s { v0 }[1], [x0], #4
; CHECK-NEXT: ; kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: str x0, [x1]
; CHECK-NEXT: ; kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: ret
;
; CHECK-GISEL-LABEL: test_v2i32_post_imm_ld1lane:
Expand All @@ -14533,8 +14533,8 @@ define <2 x i32> @test_v2i32_post_reg_ld1lane(ptr %bar, ptr %ptr, i64 %inc, <2 x
; CHECK-NEXT: lsl x8, x2, #2
; CHECK-NEXT: ; kill: def $d0 killed $d0 def $q0
; CHECK-NEXT: ld1.s { v0 }[1], [x0], x8
; CHECK-NEXT: ; kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: str x0, [x1]
; CHECK-NEXT: ; kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: ret
;
; CHECK-GISEL-LABEL: test_v2i32_post_reg_ld1lane:
Expand Down Expand Up @@ -14644,8 +14644,8 @@ define <2 x float> @test_v2f32_post_imm_ld1lane(ptr %bar, ptr %ptr, <2 x float>
; CHECK: ; %bb.0:
; CHECK-NEXT: ; kill: def $d0 killed $d0 def $q0
; CHECK-NEXT: ld1.s { v0 }[1], [x0], #4
; CHECK-NEXT: ; kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: str x0, [x1]
; CHECK-NEXT: ; kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: ret
;
; CHECK-GISEL-LABEL: test_v2f32_post_imm_ld1lane:
Expand All @@ -14670,8 +14670,8 @@ define <2 x float> @test_v2f32_post_reg_ld1lane(ptr %bar, ptr %ptr, i64 %inc, <2
; CHECK-NEXT: lsl x8, x2, #2
; CHECK-NEXT: ; kill: def $d0 killed $d0 def $q0
; CHECK-NEXT: ld1.s { v0 }[1], [x0], x8
; CHECK-NEXT: ; kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: str x0, [x1]
; CHECK-NEXT: ; kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: ret
;
; CHECK-GISEL-LABEL: test_v2f32_post_reg_ld1lane:
Expand Down Expand Up @@ -14776,9 +14776,9 @@ define <4 x i16> @test_v4i16_post_reg_ld1lane_forced_narrow(ptr %bar, ptr %ptr,
; CHECK-NEXT: lsl x8, x2, #1
; CHECK-NEXT: ; kill: def $d0 killed $d0 def $q0
; CHECK-NEXT: ld1.h { v0 }[1], [x0], x8
; CHECK-NEXT: ; kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: str x0, [x1]
; CHECK-NEXT: ldr d1, [x3]
; CHECK-NEXT: ; kill: def $d0 killed $d0 killed $q0
; CHECK-NEXT: cnt.8b v1, v1
; CHECK-NEXT: uaddlp.4h v1, v1
; CHECK-NEXT: uaddlp.2s v1, v1
Expand Down
Loading