[Mlir-commits] [mlir] [MLIR][SCF] Add support for pipelining dynamic loops (PR #74350)

[Quinn Dawkins]

================
@@ -508,8 +569,24 @@ LoopPipelinerInternal::emitEpilogue(RewriterBase &rewriter) {
   // Emit different versions of the induction variable. They will be
   // removed by dead code if not used.
   for (int64_t i = 0; i < maxStage; i++) {
-    Value newlastIter = rewriter.create<arith::ConstantIndexOp>(
-        forOp.getLoc(), lb + step * ((((ub - 1) - lb) / step) - i));
+    Location loc = forOp.getLoc();
+    Type t = lb.getType();
+    Value minusOne =
+        rewriter.create<arith::ConstantOp>(loc, rewriter.getIntegerAttr(t, -1));
+    // number of iterations = ((ub - 1) - lb) / step
+    Value totlaNumIteration = rewriter.create<arith::DivUIOp>(
----------------
qedawkins wrote:

typo: totlaNumIteration -> totalNumIteration

https://github.com/llvm/llvm-project/pull/74350