[Mlir-commits] [mlir] [mlir][linalg] Extend `FuseElementwiseOps` pattern to work with named ops (PR #144922)

[Javed Absar]

================
@@ -215,16 +216,33 @@ bool mlir::linalg::areElementwiseOpsFusable(OpOperand *fusedOperand) {
 /// Generate the region of the fused tensor operation. The region of the fused
 /// op must be empty.
 static void generateFusedElementwiseOpRegion(
-    RewriterBase &rewriter, GenericOp fusedOp,
+    RewriterBase &rewriter, LinalgOp fusedOp,
     AffineMap consumerToProducerLoopsMap, OpOperand *fusedOperand,
     unsigned nloops, llvm::SmallDenseSet<int> &preservedProducerResults) {
-  auto producer = cast<GenericOp>(fusedOperand->get().getDefiningOp());
-  auto consumer = cast<GenericOp>(fusedOperand->getOwner());
+  auto producer = cast<LinalgOp>(fusedOperand->get().getDefiningOp());
+  auto consumer = cast<LinalgOp>(fusedOperand->getOwner());
   // Build the region of the fused op.
+
+  // Since some ops, like `linalg.map`, do not have block arguments for init
----------------
javedabsar1 wrote:

> I agree that it needs to be updated, or even better if we deprecate it in favour of `linalg.elementwise` as `linalg.map` is same semantically with `linalg.elementwise`, in fact `linalg.elementwise` seems more general, IIRC.

 `linalg.map` is more than `linalg.elementwise` . linalg.map e.g. this is valid linalg.map and not mappable to elementwise
 
 ```mlir
     %mapped = linalg.map outs(%1 : tensor<?xindex>)
      () {
        %2 = linalg.index 0 : index
        %dim = memref.dim %0, %2 : memref<*xf32>
        linalg.yield %dim : index
      }
```      

https://github.com/llvm/llvm-project/pull/144922