[Mlir-commits] [mlir] [mlir][SCFToGPU] Fix crash when converting affine.for with iter_args to GPU (PR #185073)

[Mehdi Amini]

================
@@ -184,6 +184,12 @@ AffineLoopToGpuConverter::collectBounds(AffineForOp forOp, unsigned numLoops) {
   steps.reserve(numLoops);
   AffineForOp currentLoop = forOp;
   for (unsigned i = 0; i < numLoops; ++i) {
+    if (currentLoop.getNumIterOperands() > 0) {
----------------
joker-eph wrote:

I thought about it, but it seems we don't have tests for it, so it does not seem like a behavior that the pass expects at the moment, I'll treat it as a "feature request" for the pass (which should come with appropriate testing)

https://github.com/llvm/llvm-project/pull/185073