[llvm] 8c8074f - [Flang][OpenMP] Fix OpenMP static scheduling when trip count is zero (#170863)

[via llvm-commits]

Author: Michael Klemm
Date: 2025-12-08T18:29:02+01:00
New Revision: 8c8074f0fa516210011c7b3beb63807ffbe5c5f7

URL: https://github.com/llvm/llvm-project/commit/8c8074f0fa516210011c7b3beb63807ffbe5c5f7
DIFF: https://github.com/llvm/llvm-project/commit/8c8074f0fa516210011c7b3beb63807ffbe5c5f7.diff

LOG: [Flang][OpenMP] Fix OpenMP static scheduling when trip count is zero (#170863)

Code-gen produced incorrect code for cases when the trip count an
associated DO loop was zero. The generated code evaluated the trip count
of the loop and substracted 1 from it. When this was passed to
__kmpc_for_static_init_4u, the value was interpreted as unsigned, which
made the upper bound of the worksharing loop 2^32-1 and caused a
division by zero in the calculation of the loop bounds for the threads.

Added: 
    

Modified: 
    clang/test/OpenMP/irbuilder_for_unsigned_static_chunked.c
    llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp

Removed: 
    


################################################################################
diff  --git a/clang/test/OpenMP/irbuilder_for_unsigned_static_chunked.c b/clang/test/OpenMP/irbuilder_for_unsigned_static_chunked.c
index b937568ca9f11..dfa5e8a6e0943 100644
--- a/clang/test/OpenMP/irbuilder_for_unsigned_static_chunked.c
+++ b/clang/test/OpenMP/irbuilder_for_unsigned_static_chunked.c
@@ -50,7 +50,9 @@ extern "C" void workshareloop_unsigned_static_chunked(float *a, float *b, float
 // CHECK:       omp_loop.preheader:
 // CHECK-NEXT:    store i32 0, ptr [[P_LOWERBOUND]], align 4
 // CHECK-NEXT:    [[TMP3:%.*]] = sub i32 [[DOTCOUNT]], 1
-// CHECK-NEXT:    store i32 [[TMP3]], ptr [[P_UPPERBOUND]], align 4
+// CHECK-NEXT:    [[COUNTCOND:%.*]] = icmp eq i32 [[DOTCOUNT]], 0
+// CHECK-NEXT:    [[UPPERSEL:%.*]] = select i1 [[COUNTCOND]], i32 0, i32 [[TMP3]]
+// CHECK-NEXT:    store i32 [[UPPERSEL]], ptr [[P_UPPERBOUND]], align 4
 // CHECK-NEXT:    store i32 1, ptr [[P_STRIDE]], align 4
 // CHECK-NEXT:    [[OMP_GLOBAL_THREAD_NUM:%.*]] = call i32 @__kmpc_global_thread_num(ptr @[[GLOB1:[0-9]+]])
 // CHECK-NEXT:    call void @__kmpc_for_static_init_4u(ptr @[[GLOB1]], i32 [[OMP_GLOBAL_THREAD_NUM]], i32 33, ptr [[P_LASTITER]], ptr [[P_LOWERBOUND]], ptr [[P_UPPERBOUND]], ptr [[P_STRIDE]], i32 1, i32 5)

diff  --git a/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp b/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
index 553333d53a106..9c521828660c2 100644
--- a/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
+++ b/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
@@ -5140,7 +5140,10 @@ OpenMPIRBuilder::applyStaticChunkedWorkshareLoop(
       ConstantInt::get(I32Type, static_cast<int>(DistScheduleSchedType));
   Builder.CreateStore(Zero, PLowerBound);
   Value *OrigUpperBound = Builder.CreateSub(CastedTripCount, One);
-  Builder.CreateStore(OrigUpperBound, PUpperBound);
+  Value *IsTripCountZero = Builder.CreateICmpEQ(CastedTripCount, Zero);
+  Value *UpperBound =
+      Builder.CreateSelect(IsTripCountZero, Zero, OrigUpperBound);
+  Builder.CreateStore(UpperBound, PUpperBound);
   Builder.CreateStore(One, PStride);
 
   // Call the "init" function and update the trip count of the loop with the