[Mlir-commits] [mlir] [mlir][linalg] Add scalable vectorisation for depthwise convolutions (PR #81625)

[Cullen Rhodes]

================
@@ -3107,6 +3186,7 @@ struct Conv1DGenerator
     // Its possible we failed to create the Fma.
     if (!llvm::all_of(resVals, [](Value v) { return v; })) {
       // Manually revert (in reverse order) to avoid leaving a bad IR state.
+      // TODO: Replace with maybeMasked
----------------
c-rhodes wrote:

is there a reason this is not done in this patch?

https://github.com/llvm/llvm-project/pull/81625