[MLIR][OpenMP] Mark openmp.parallel and omp.wsloop as having recursive side effects

[wsmoses]

Summary

An OpenMP wsloop is simply a regular for loop with the bounds determined by the thread number, and the same justification to allow this for scf.for works for omp.wsloop.

An OpenMP parallel is a parallel for, per thread. Similarly the same justification for scf.parallel having recursive side effects applies here.

In both cases the general justification is that the ops themselves don't have side effects (besides inaccessible runtime-specific memory) and thus the side effects are simply that of the contained ops.

[bondhugula]

Can you please add a commit summary with a one-liner rationale here?

[wsmoses]

In D120853#3358585, @bondhugula wrote:

Can you please add a commit summary with a one-liner rationale here?

Done.

[bondhugula]

LGTM.

mlir/include/mlir/Dialect/OpenMP/OpenMPOps.td

def OpenMP_PointerLikeType : Type<
   "OpenMP-compatible variable type", "::mlir::omp::PointerLikeType">;
 
 //===----------------------------------------------------------------------===//
 // 2.6 parallel Construct
 //===----------------------------------------------------------------------===//
 
 def ParallelOp : OpenMP_Op<"parallel", [
                  AutomaticAllocationScope, AttrSizedOperandSegments,
-                 DeclareOpInterfaceMethods<OutlineableOpenMPOpInterface>]> {
+                 DeclareOpInterfaceMethods<OutlineableOpenMPOpInterface>,
+                 RecursiveSideEffects]> {
   let summary = "parallel construct";
   let description = [{
     The parallel construct includes a region of code which is to be executed
     by a team of threads.
 
     The optional $if_expr_var parameter specifies a boolean result of a
     conditional check. If this value is 1 or is not provided then the parallel
     region runs as normal, if it is 0 then the parallel region is executed with
def SectionsOp : OpenMP_Op<"sections", [AttrSizedOperandSegments]> {
   let hasVerifier = 1;
 }
 
 //===----------------------------------------------------------------------===//
 // 2.9.2 Workshare Loop Construct
 //===----------------------------------------------------------------------===//
 
 def WsLoopOp : OpenMP_Op<"wsloop", [AttrSizedOperandSegments,
-                         AllTypesMatch<["lowerBound", "upperBound", "step"]>]> {
+                         AllTypesMatch<["lowerBound", "upperBound", "step"]>,
+                         RecursiveSideEffects]> {
   let summary = "workshare loop construct";
   let description = [{
     The workshare loop construct specifies that the iterations of the loop(s)
     will be executed in parallel by threads in the current context. These
     iterations are spread across threads that already exist in the enclosing
     parallel region. The lower and upper bounds specify a half-open range: the
     range includes the lower bound but does not include the upper bound. If the
     `inclusive` attribute is specified then the upper bound is also included.