[PATCH] D80078: [LoopSimplify] don't separate nested loops with convergent calls

[Sameer Sahasrabuddhe via Phabricator via llvm-commits]

This revision was automatically updated to reflect the committed changes.
Closed by commit rG6c8488436644: [LoopSimplify] don't separate nested loops with convergent calls (authored by sameerds).

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D80078/new/

https://reviews.llvm.org/D80078

Files:
  llvm/lib/Transforms/Utils/LoopSimplify.cpp
  llvm/test/Transforms/LoopSimplify/convergent.ll


Index: llvm/test/Transforms/LoopSimplify/convergent.ll
===================================================================
--- /dev/null
+++ llvm/test/Transforms/LoopSimplify/convergent.ll
@@ -0,0 +1,25 @@
+; RUN: opt < %s -S -loop-simplify | FileCheck %s
+
+; Don't separate out nested loops if a convergent call is present
+
+; CHECK-NOT: BB1.outer
+; CHECK: BB1.backedge
+
+define i32 @test(i1 %loop_cond, i1 %exit_cond, i32 %init) {
+entry:
+  br label %BB1
+
+BB1:
+  %indvar = phi i32 [%indvar, %BB1], [%inc, %BB2], [%init, %entry]
+  call void @f() convergent
+  br i1 %loop_cond, label %BB1, label %BB2
+
+BB2:
+  %inc = add i32 %indvar, 1
+  br i1 %exit_cond, label %exit, label %BB1
+
+exit:
+  ret i32 %inc
+}
+
+declare void @f() convergent
Index: llvm/lib/Transforms/Utils/LoopSimplify.cpp
===================================================================
--- llvm/lib/Transforms/Utils/LoopSimplify.cpp
+++ llvm/lib/Transforms/Utils/LoopSimplify.cpp
@@ -230,6 +230,27 @@
   if (!Preheader)
     return nullptr;
 
+  // Treat the presence of convergent functions conservatively. The
+  // transformation is invalid if calls to certain convergent
+  // functions (like an AMDGPU barrier) get included in the resulting
+  // inner loop. But blocks meant for the inner loop will be
+  // identified later at a point where it's too late to abort the
+  // transformation. Also, the convergent attribute is not really
+  // sufficient to express the semantics of functions that are
+  // affected by this transformation. So we choose to back off if such
+  // a function call is present until a better alternative becomes
+  // available. This is similar to the conservative treatment of
+  // convergent function calls in GVNHoist and JumpThreading.
+  for (auto BB : L->blocks()) {
+    for (auto &II : *BB) {
+      if (auto CI = dyn_cast<CallBase>(&II)) {
+        if (CI->isConvergent()) {
+          return nullptr;
+        }
+      }
+    }
+  }
+
   // The header is not a landing pad; preheader insertion should ensure this.
   BasicBlock *Header = L->getHeader();
   assert(!Header->isEHPad() && "Can't insert backedge to EH pad");


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D80078.264792.patch
Type: text/x-patch
Size: 2152 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200519/ce054ccf/attachment.bin>