Fix overly pessimistic shortcut in post-RA MachineLICM

Post-RA LICM keeps three sets of registers: PhysRegDefs, PhysRegClobbers
and TermRegs.  When it sees a definition of R it adds all aliases of R
to the corresponding set, so that when it needs to test for membership
it only needs to test a single register, rather than worrying about
aliases there too.  E.g. the final candidate loop just has:

    unsigned Def = Candidates[i].Def;
    if (!PhysRegClobbers.test(Def) && ...) {

to test whether register Def is multiply defined.

However, there was also a shortcut in ProcessMI to make sure we didn't
add candidates if we already knew that they would fail the final test.
This shortcut was more pessimistic than the final one because it
checked whether _any alias_ of the defined register was multiply defined.
This is too conservative for targets that define register pairs.
E.g. on z, R0 and R1 are sometimes used as a pair, so there is a
128-bit register that aliases both R0 and R1.  If a loop used
R0 and R1 independently, and the definition of R0 came first,
we would be able to hoist the R0 assignment (because that used
the final test quoted above) but not the R1 assignment (because
that meant we had two definitions of the paired R0/R1 register
and would fail the shortcut in ProcessMI).

This patch just uses the same check for the ProcessMI shortcut as
we use in the final candidate loop.

llvm-svn: 188774
This commit is contained in:
Richard Sandiford 2013-08-20 09:11:13 +00:00
parent f79c3a5aef
commit 96aa93d5f1
2 changed files with 26 additions and 4 deletions

View File

@ -468,12 +468,12 @@ void MachineLICM::ProcessMI(MachineInstr *MI,
for (MCRegAliasIterator AS(Reg, TRI, true); AS.isValid(); ++AS) {
if (PhysRegDefs.test(*AS))
PhysRegClobbers.set(*AS);
if (PhysRegClobbers.test(*AS))
PhysRegDefs.set(*AS);
}
if (PhysRegClobbers.test(Reg))
// MI defined register is seen defined by another instruction in
// the loop, it cannot be a LICM candidate.
RuledOut = true;
PhysRegDefs.set(*AS);
}
}
// Only consider reloads for now and remats which do not have register

View File

@ -80,3 +80,25 @@ define float @f7(float %in) {
call void asm sideeffect "blah", "~{f0},~{cc}"()
ret float %in
}
; Test that both registers in a GR128 pair get hoisted.
define void @f8(i32 %count) {
; CHECK-LABE: f8
; CHECK-DAG: lhi %r0, 0
; CHECK-DAG: lhi %r1, 1
; CHECK: %loop
; CHECK-NOT: %r
; CHECK: blah %r0, %r1
entry:
br label %loop
loop:
%this = phi i32 [ %count, %entry ], [ %next, %loop ]
call void asm sideeffect "blah $0, $1", "{r0},{r1}" (i32 0, i32 1)
%next = sub i32 %this, 1
%cmp = icmp ne i32 %next, 0
br i1 %cmp, label %loop, label %exit
exit:
ret void
}