Be conservative about unordered accesses for the moment

Background: As described in https://reviews.llvm.org/D57601, I'm working towards separating volatile and atomic in the MMO uses for atomic instructions.

In https://reviews.llvm.org/D57593, I fixed a bug where isUnordered was returning the wrong result, but didn't account for the fact I was getting slightly ahead of myself. While both uses of isUnordered are correct (as far as I can tell), we don't have tests to demonstrate this and being aggressive gets in the way of having the removal of volatile truly be non-functional. Once D57601 lands, I will return to these call sites, revert this patch, and add the appropriate tests to show the expected behaviour.

Differential Revision: https://reviews.llvm.org/D57802

llvm-svn: 353766
This commit is contained in:
Philip Reames 2019-02-11 23:34:33 +00:00
parent e3cd735ea6
commit 5906a6591c
2 changed files with 8 additions and 3 deletions

View File

@ -235,8 +235,11 @@ bool ImplicitNullChecks::canHandle(const MachineInstr *MI) {
assert(!llvm::any_of(MI->operands(), IsRegMask) &&
"Calls were filtered out above!");
auto IsUnordered = [](MachineMemOperand *MMO) { return MMO->isUnordered(); };
return llvm::all_of(MI->memoperands(), IsUnordered);
// TODO: This should be isUnordered (see D57601) once test cases are written
// demonstrating that.
auto IsSimple = [](MachineMemOperand *MMO) {
return !MMO->isVolatile() && !MMO->isAtomic(); };
return llvm::all_of(MI->memoperands(), IsSimple);
}
ImplicitNullChecks::DependenceResult

View File

@ -1291,8 +1291,10 @@ bool MachineInstr::hasOrderedMemoryRef() const {
return true;
// Check if any of our memory operands are ordered.
// TODO: This should probably be be isUnordered (see D57601), but the callers
// need audited and test cases written to be sure.
return llvm::any_of(memoperands(), [](const MachineMemOperand *MMO) {
return !MMO->isUnordered();
return MMO->isVolatile() || MMO->isAtomic();
});
}