[Mlir-commits] [mlir] [MLIR][Linalg] Fix crash on duplicate dimensions in linalg.broadcast (PR #211203)

[Chibuoyim Ogbonna]

================
@@ -1196,6 +1196,30 @@ func.func @broadcast_size_1_extension_not_supported(
 
 // -----
 
+func.func @broadcast_duplicate_dims(
+    %input: tensor<i32>, %init: tensor<32x2xi32>) -> tensor<32x2xi32> {
+  // expected-error @+1 {{'linalg.broadcast' op dimensions should not contain duplicates}}
+  %bcast = linalg.broadcast
+      ins(%input:tensor<i32>)
+      outs(%init:tensor<32x2xi32>)
+      dimensions = [0, 0]
+  func.return %bcast : tensor<32x2xi32>
+}
+
+// -----
+
+func.func @broadcast_scalar_input(
+    %input: f32, %init: tensor<8x16xf32>) -> tensor<8x16xf32> {
+  // expected-error @+1 {{'linalg.broadcast' op operand #0 must be memref of any non-token type values or ranked tensor of any non-token type values, but got 'f32'}}
----------------
bruteforceboy wrote:

I think it does make sense to be able to broadcast a scalar just like we do for the OD case.

https://github.com/llvm/llvm-project/pull/211203