Skip to content

[MachinePipeliner] Fix MachineMemOperand to be cleared when update fails #80841

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.

Sign up for GitHub

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

Closed
wants to merge 1 commit into from
Closed

[MachinePipeliner] Fix MachineMemOperand to be cleared when update fails #80841

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
4 changes: 2 additions & 2 deletions llvm/lib/CodeGen/ModuloSchedule.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -979,8 +979,8 @@ void ModuloScheduleExpander::updateMemOperands(MachineInstr &NewMI,
NewMMOs.push_back(
MF.getMachineMemOperand(MMO, AdjOffset, MMO->getSize()));
} else {
NewMMOs.push_back(
MF.getMachineMemOperand(MMO, 0, MemoryLocation::UnknownSize));
NewMMOs.clear();
break;
}
}
NewMI.setMemRefs(MF, NewMMOs);
Expand Down
126 changes: 126 additions & 0 deletions llvm/test/CodeGen/AArch64/sms-fix-memoperand.mir
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,126 @@
# RUN: llc --verify-machineinstrs -mtriple=aarch64 -o - %s -start-before=pipeliner -aarch64-enable-pipeliner
# REQUIRES: asserts

# Check that assertions do not fail

--- |
define dso_local void @f(ptr noalias nocapture noundef writeonly %a, ptr nocapture noundef readonly %b, i32 noundef %n) local_unnamed_addr #0 {
entry:
%cmp6 = icmp sgt i32 %n, 0
br i1 %cmp6, label %for.body.preheader, label %for.cond.cleanup

for.body.preheader: ; preds = %entry
%wide.trip.count = zext nneg i32 %n to i64
%0 = icmp eq i32 %n, 1
br i1 %0, label %for.cond.cleanup.loopexit.unr-lcssa, label %for.body.preheader.new

for.body.preheader.new: ; preds = %for.body.preheader
%unroll_iter = and i64 %wide.trip.count, 2147483646
br label %for.body

for.cond.cleanup.loopexit.unr-lcssa: ; preds = %for.body, %for.body.preheader
%indvars.iv.unr = phi i64 [ 0, %for.body.preheader ], [ %indvars.iv.next.1, %for.body ]
%1 = and i64 %wide.trip.count, 1
%lcmp.mod.not = icmp eq i64 %1, 0
br i1 %lcmp.mod.not, label %for.cond.cleanup, label %for.body.epil

for.body.epil: ; preds = %for.cond.cleanup.loopexit.unr-lcssa
%arrayidx.epil = getelementptr inbounds double, ptr %b, i64 %indvars.iv.unr
%2 = load double, ptr %arrayidx.epil, align 8
%mul.epil = fmul double %2, 2.000000e+00
%arrayidx2.epil = getelementptr inbounds double, ptr %a, i64 %indvars.iv.unr
store double %mul.epil, ptr %arrayidx2.epil, align 8
br label %for.cond.cleanup

for.cond.cleanup: ; preds = %for.body.epil, %for.cond.cleanup.loopexit.unr-lcssa, %entry
ret void

for.body: ; preds = %for.body, %for.body.preheader.new
%indvars.iv = phi i64 [ 0, %for.body.preheader.new ], [ %indvars.iv.next.1, %for.body ]
%3 = shl i64 %indvars.iv, 3
%scevgep13 = getelementptr i8, ptr %b, i64 %3
%4 = load double, ptr %scevgep13, align 8
%mul = fmul double %4, 2.000000e+00
%5 = shl i64 %indvars.iv, 3
%scevgep12 = getelementptr i8, ptr %a, i64 %5
store double %mul, ptr %scevgep12, align 8
%6 = shl i64 %indvars.iv, 3
%scevgep = getelementptr i8, ptr %b, i64 %6
%scevgep9 = getelementptr i8, ptr %scevgep, i64 8
%7 = load double, ptr %scevgep9, align 8
%mul.1 = fmul double %7, 2.000000e+00
%scevgep10 = getelementptr i8, ptr %a, i64 %6
%scevgep11 = getelementptr i8, ptr %scevgep10, i64 8
store double %mul.1, ptr %scevgep11, align 8
%indvars.iv.next.1 = add nuw i64 %indvars.iv, 2
%niter.ncmp.1 = icmp eq i64 %unroll_iter, %indvars.iv.next.1
br i1 %niter.ncmp.1, label %for.cond.cleanup.loopexit.unr-lcssa, label %for.body
}

...
---
name: f
tracksRegLiveness: true
liveins:
- { reg: '$x0', virtual-reg: '%5' }
- { reg: '$x1', virtual-reg: '%6' }
- { reg: '$w2', virtual-reg: '%7' }
body: |
bb.0.entry:
liveins: $x0, $x1, $w2

%7:gpr32common = COPY $w2
%6:gpr64common = COPY $x1
%5:gpr64common = COPY $x0
dead $wzr = SUBSWri %7, 1, 0, implicit-def $nzcv
Bcc 11, %bb.5, implicit $nzcv
B %bb.1

bb.1.for.body.preheader:
%10:gpr32 = ORRWrs $wzr, %7, 0
%0:gpr64 = SUBREG_TO_REG 0, killed %10, %subreg.sub_32
dead $wzr = SUBSWri %7, 1, 0, implicit-def $nzcv
Bcc 1, %bb.2, implicit $nzcv

bb.7:
%11:gpr64all = COPY $xzr
%9:gpr64all = COPY %11
B %bb.3

bb.2.for.body.preheader.new:
%14:gpr64common = ANDXri %0, 8157
%15:gpr64all = COPY $xzr
%13:gpr64all = COPY %15
B %bb.6

bb.3.for.cond.cleanup.loopexit.unr-lcssa:
%2:gpr64 = PHI %9, %bb.7, %4, %bb.6
%25:gpr32 = COPY %0.sub_32
TBZW killed %25, 0, %bb.5
B %bb.4

bb.4.for.body.epil:
%26:gpr64 = UBFMXri %2, 61, 60
%27:fpr64 = LDRDroX %6, %26, 0, 0 :: (load (s64) from %ir.arrayidx.epil)
%28:fpr64 = nofpexcept FADDDrr %27, %27, implicit $fpcr
STRDroX killed %28, %5, %26, 0, 0 :: (store (s64) into %ir.arrayidx2.epil)

bb.5.for.cond.cleanup:
RET_ReallyLR

bb.6.for.body:
%3:gpr64common = PHI %13, %bb.2, %4, %bb.6
%16:gpr64 = UBFMXri %3, 61, 60
%17:gpr64common = ADDXrr %6, %16
%18:fpr64 = LDRDui %17, 0 :: (load (s64) from %ir.scevgep13)
%19:fpr64 = nofpexcept FADDDrr %18, %18, implicit $fpcr
%20:gpr64common = ADDXrr %5, %16
%21:fpr64 = LDRDui %17, 1 :: (load (s64) from %ir.scevgep9)
%22:fpr64 = nofpexcept FADDDrr %21, %21, implicit $fpcr
%23:gpr64common = nuw ADDXri %3, 2, 0
%4:gpr64all = COPY %23
dead $xzr = SUBSXrr %14, %23, implicit-def $nzcv
Bcc 0, %bb.3, implicit $nzcv
B %bb.6

...