Skip to content

[LVA] Update ignore instructions RLE and DSE. #31134

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

Merged
merged 1 commit into from
May 8, 2020
Merged

[LVA] Update ignore instructions RLE and DSE. #31134

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
2 changes: 2 additions & 0 deletions lib/SILOptimizer/Transforms/DeadStoreElimination.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -161,6 +161,8 @@ static bool isDeadStoreInertInstruction(SILInstruction *Inst) {
case SILInstructionKind::DeallocRefInst:
case SILInstructionKind::CondFailInst:
case SILInstructionKind::FixLifetimeInst:
case SILInstructionKind::EndAccessInst:
case SILInstructionKind::SetDeallocatingInst:
return true;
default:
return false;
Expand Down
3 changes: 3 additions & 0 deletions lib/SILOptimizer/Transforms/RedundantLoadElimination.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -160,6 +160,9 @@ static bool isRLEInertInstruction(SILInstruction *Inst) {
case SILInstructionKind::IsEscapingClosureInst:
case SILInstructionKind::IsUniqueInst:
case SILInstructionKind::FixLifetimeInst:
case SILInstructionKind::EndAccessInst:
case SILInstructionKind::SetDeallocatingInst:
case SILInstructionKind::DeallocRefInst:
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I wonder if we can centralize this notion in a helper. Just a random thought.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

There is a lot that could be centralized between DSE and RLE. They both end up doing the work of the other pass anyway so, it might make sense to combine them.

return true;
default:
return false;
Expand Down
18 changes: 18 additions & 0 deletions test/SILOptimizer/dead_store_elim.sil
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1501,3 +1501,21 @@ bb0(%0 : $*foo):
%20 = tuple()
return %20 : $()
}

// Check that begin_access, end_access, strong_release, set_deallocating, and dealloc_ref don't prevent optimization.
// CHECK-LABEL: ignore_read_write
// CHECK: bb0
// CHECK-NOT: store
// CHECK-LABEL: end sil function 'ignore_read_write'
sil @ignore_read_write : $@convention(thin) () -> Int {
bb0:
%1 = alloc_ref [stack] $foo
%2 = integer_literal $Builtin.Int64, 0
%3 = struct $Int (%2 : $Builtin.Int64)
%4 = ref_element_addr %1 : $foo, #foo.a
store %3 to %4 : $*Int
set_deallocating %1 : $foo
dealloc_ref %1 : $foo
dealloc_ref [stack] %1 : $foo
return %3 : $Int
}
19 changes: 19 additions & 0 deletions test/SILOptimizer/redundant_load_elim.sil
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1225,3 +1225,22 @@ bb0(%0 : $*Int):
%3 = tuple()
return %3 : $()
}

// Check that begin_access, end_access, strong_release, set_deallocating, and dealloc_ref don't prevent optimization.
// CHECK-LABEL: ignore_read_write
// CHECK: bb0
// CHECK-NOT: load
// CHECK-LABEL: end sil function 'ignore_read_write'
sil @ignore_read_write : $@convention(thin) () -> Int32 {
bb0:
%1 = alloc_ref [stack] $AX
%2 = integer_literal $Builtin.Int32, 0
%3 = struct $Int32 (%2 : $Builtin.Int32)
%4 = ref_element_addr %1 : $AX, #AX.current
store %3 to %4 : $*Int32
%5 = load %4 : $*Int32
set_deallocating %1 : $AX
dealloc_ref %1 : $AX
dealloc_ref [stack] %1 : $AX
return %5 : $Int32
}