The semantics of tail predication loops means that the value of LR as an instruction is executed determines the predicate. In other words: mov r3, #3 DLSTP lr, r3 // Start tail predication, lr==3 VADD.s32 q0, q1, q2 // Lanes 0,1 and 2 are updated in q0. mov lr, #1 VADD.s32 q0, q1, q2 // Only first lane is updated. This means that the value of lr cannot be spilled and re-used in tail predication regions without potentially altering the behaviour of the program. More lanes than required could be stored, for example, and in the case of a gather those lanes might not have been setup, leading to alignment exceptions. This patch adds a new lr predicate operand to MVE instructions in order to keep a reference to the lr that they use as a tail predicate. It will usually hold the zeroreg meaning not predicated, being set to the LR phi value in the MVETPAndVPTOptimisationsPass. This will prevent it from being spilled anywhere that it needs to be used. A lot of tests needed updating. Differential Revision: https://reviews.llvm.org/D107638
101 lines
4.7 KiB
YAML
101 lines
4.7 KiB
YAML
# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
|
|
# RUN: llc -mtriple=thumbv8.1m.main-none-eabi -mattr=+mve -simplify-mir -run-pass=phi-node-elimination %s -o - | FileCheck %s
|
|
--- |
|
|
; ModuleID = '<stdin>'
|
|
target datalayout = "e-m:e-p:32:32-Fi8-i64:64-v128:64:128-a:0:32-n32-S64"
|
|
target triple = "arm-arm-none-eabi"
|
|
|
|
; Function Attrs: nofree norecurse nounwind
|
|
define void @test(i8* noalias nocapture %X, i8* noalias nocapture readonly %Y, i32 %n) {
|
|
entry:
|
|
%cmp6 = icmp sgt i32 %n, 0
|
|
br i1 %cmp6, label %for.body.preheader, label %for.cond.cleanup
|
|
|
|
for.body.preheader: ; preds = %entry
|
|
call void @llvm.memcpy.p0i8.p0i8.i32(i8* align 4 %X, i8* align 4 %Y, i32 %n, i1 false)
|
|
br label %for.cond.cleanup
|
|
|
|
for.cond.cleanup: ; preds = %for.body.preheader, %entry
|
|
ret void
|
|
}
|
|
|
|
; Function Attrs: argmemonly nofree nosync nounwind willreturn
|
|
declare void @llvm.memcpy.p0i8.p0i8.i32(i8* noalias nocapture writeonly, i8* noalias nocapture readonly, i32, i1 immarg)
|
|
|
|
...
|
|
---
|
|
name: test
|
|
tracksRegLiveness: true
|
|
body: |
|
|
; CHECK-LABEL: name: test
|
|
; CHECK: bb.0.entry:
|
|
; CHECK: successors: %bb.1(0x50000000), %bb.3(0x30000000)
|
|
; CHECK: liveins: $r0, $r1, $r2
|
|
; CHECK: [[COPY:%[0-9]+]]:rgpr = COPY killed $r2
|
|
; CHECK: [[COPY1:%[0-9]+]]:rgpr = COPY killed $r1
|
|
; CHECK: [[COPY2:%[0-9]+]]:rgpr = COPY killed $r0
|
|
; CHECK: t2CMPri [[COPY]], 1, 14 /* CC::al */, $noreg, implicit-def $cpsr
|
|
; CHECK: t2Bcc %bb.3, 11 /* CC::lt */, killed $cpsr
|
|
; CHECK: t2B %bb.1, 14 /* CC::al */, $noreg
|
|
; CHECK: bb.1.for.body.preheader:
|
|
; CHECK: [[t2ADDri:%[0-9]+]]:rgpr = t2ADDri [[COPY]], 15, 14 /* CC::al */, $noreg, $noreg
|
|
; CHECK: [[t2BICri:%[0-9]+]]:rgpr = t2BICri killed [[t2ADDri]], 16, 14 /* CC::al */, $noreg, $noreg
|
|
; CHECK: [[t2LSRri:%[0-9]+]]:gprlr = t2LSRri killed [[t2BICri]], 4, 14 /* CC::al */, $noreg, $noreg
|
|
; CHECK: [[COPY3:%[0-9]+]]:rgpr = COPY [[COPY1]]
|
|
; CHECK: [[COPY4:%[0-9]+]]:rgpr = COPY [[COPY2]]
|
|
; CHECK: [[COPY5:%[0-9]+]]:rgpr = COPY [[COPY]]
|
|
; CHECK: [[t2WhileLoopStartLR:%[0-9]+]]:gprlr = t2WhileLoopStartLR killed [[t2LSRri]], %bb.3, implicit-def dead $cpsr
|
|
; CHECK: bb.2:
|
|
; CHECK: [[COPY6:%[0-9]+]]:rgpr = COPY [[COPY5]]
|
|
; CHECK: [[COPY7:%[0-9]+]]:gprlr = COPY [[t2WhileLoopStartLR]]
|
|
; CHECK: [[COPY8:%[0-9]+]]:rgpr = COPY [[COPY4]]
|
|
; CHECK: [[COPY9:%[0-9]+]]:rgpr = COPY [[COPY3]]
|
|
; CHECK: [[MVE_VCTP8_:%[0-9]+]]:vccr = MVE_VCTP8 [[COPY6]], 0, $noreg, $noreg
|
|
; CHECK: [[t2SUBri:%[0-9]+]]:rgpr = t2SUBri killed [[COPY6]], 16, 14 /* CC::al */, $noreg, $noreg
|
|
; CHECK: [[MVE_VLDRBU8_post:%[0-9]+]]:rgpr, [[MVE_VLDRBU8_post1:%[0-9]+]]:mqpr = MVE_VLDRBU8_post killed [[COPY9]], 16, 1, [[MVE_VCTP8_]], $noreg
|
|
; CHECK: [[MVE_VSTRBU8_post:%[0-9]+]]:rgpr = MVE_VSTRBU8_post killed [[MVE_VLDRBU8_post1]], killed [[COPY8]], 16, 1, killed [[MVE_VCTP8_]], $noreg
|
|
; CHECK: [[COPY10:%[0-9]+]]:rgpr = COPY [[MVE_VLDRBU8_post]]
|
|
; CHECK: [[COPY10:%[0-9]+]]:rgpr = COPY [[MVE_VSTRBU8_post]]
|
|
; CHECK: [[COPY10:%[0-9]+]]:rgpr = COPY [[t2SUBri]]
|
|
; CHECK: [[t2LoopEndDec:%[0-9]+]]:gprlr = t2LoopEndDec killed [[COPY7]], %bb.2, implicit-def dead $cpsr
|
|
; CHECK: t2B %bb.3, 14 /* CC::al */, $noreg
|
|
; CHECK: bb.3.for.cond.cleanup:
|
|
; CHECK: tBX_RET 14 /* CC::al */, $noreg
|
|
bb.0.entry:
|
|
successors: %bb.1(0x50000000), %bb.3(0x30000000)
|
|
liveins: $r0, $r1, $r2
|
|
|
|
%2:rgpr = COPY killed $r2
|
|
%1:rgpr = COPY killed $r1
|
|
%0:rgpr = COPY killed $r0
|
|
t2CMPri %2, 1, 14 /* CC::al */, $noreg, implicit-def $cpsr
|
|
t2Bcc %bb.3, 11 /* CC::lt */, killed $cpsr
|
|
t2B %bb.1, 14 /* CC::al */, $noreg
|
|
|
|
bb.1.for.body.preheader:
|
|
successors: %bb.3(0x40000000), %bb.2(0x40000000)
|
|
|
|
%3:rgpr = t2ADDri %2, 15, 14 /* CC::al */, $noreg, $noreg
|
|
%4:rgpr = t2BICri killed %3, 16, 14 /* CC::al */, $noreg, $noreg
|
|
%5:gprlr = t2LSRri killed %4, 4, 14 /* CC::al */, $noreg, $noreg
|
|
%6:gprlr = t2WhileLoopStartLR killed %5, %bb.3, implicit-def dead $cpsr
|
|
|
|
bb.2:
|
|
successors: %bb.2(0x40000000), %bb.3(0x40000000)
|
|
|
|
%7:rgpr = PHI %1, %bb.1, %8, %bb.2
|
|
%9:rgpr = PHI %0, %bb.1, %10, %bb.2
|
|
%11:gprlr = PHI %6, %bb.1, %12, %bb.2
|
|
%13:rgpr = PHI %2, %bb.1, %14, %bb.2
|
|
%15:vccr = MVE_VCTP8 %13, 0, $noreg, $noreg
|
|
%14:rgpr = t2SUBri killed %13, 16, 14 /* CC::al */, $noreg, $noreg
|
|
%8:rgpr, %16:mqpr = MVE_VLDRBU8_post killed %7, 16, 1, %15, $noreg
|
|
%10:rgpr = MVE_VSTRBU8_post killed %16, killed %9, 16, 1, killed %15, $noreg
|
|
%12:gprlr = t2LoopEndDec killed %11, %bb.2, implicit-def dead $cpsr
|
|
t2B %bb.3, 14 /* CC::al */, $noreg
|
|
|
|
bb.3.for.cond.cleanup:
|
|
tBX_RET 14 /* CC::al */, $noreg
|
|
|
|
...
|