[Revert](fix) revert [ControlFlowOpt](feat) decompose pointers across control flow (#1344)

[245516766]

… (#1344)"

This reverts commit 6dab8f4.

New contributor declaration

• I am not making a trivial change, such as fixing a typo in a comment.

• I have written a PR description following these
  rules.

• I have run pre-commit run --from-ref origin/main --to-ref HEAD.

• Select one of the following.

  • I have added tests.
    • /test for lit tests
    • /unittest for C++ tests
    • /python/test for end-to-end tests
  • This PR does not need a test because FILL THIS IN.

• Select one of the following.

  • I have not added any lit tests.
  • The lit tests I have added follow these best practices,
    including the "tests should be minimal" section. (Usually running Python code
    and using the instructions it generates is not minimal.)

[github-actions]

🔍 OpenCodeReview found 6 issue(s) in this PR.

• ✅ Successfully posted inline: 5 comment(s)
• 📝 In summary (no line info): 1 comment(s)

---

[bug · critical]

📄 third_party/ascend/include/Utils/Utils.h

⚠️ GitHub could not post this as an inline comment: No line information provided

Same copy-paste bug in mulOpFoldResult: RHS side incorrectly uses lhsValue instead of rhsValue

Same pattern — RHS operand conversion uses lhsValue instead of rhsValue, causing arith::MulIOp to receive type-mismatched operands.

💡 Suggested Change

Before:

  if (rhsInt) {
    rhsValue = createConstIndexValueOp(loc, b, rhsInt.value());
  } else {
    lhsValue = convertToIndexIfNeeded(lhsValue, loc, b);
    assert(isa<IndexType>(lhsValue.getType()));
  }

After:

  if (rhsInt) {
    rhsValue = createConstIndexValueOp(loc, b, rhsInt.value());
  } else {
    rhsValue = convertToIndexIfNeeded(rhsValue, loc, b);
    assert(isa<IndexType>(rhsValue.getType()));
  }

[github-actions]

[bug · critical]
Same copy-paste bug in subOpFoldResult: RHS side incorrectly uses lhsValue instead of rhsValue

Same pattern as addOpFoldResult — when rhsInt is not set, lhsValue is converted again instead of rhsValue, causing rhsValue to retain its original non-index type, which will lead to MLIR verification failures for arith::SubIOp.

Suggestion:

Suggested change

	OpFoldResult subOpFoldResult(const OpFoldResult &lhs, const OpFoldResult &rhs,
	const Location &loc, OpBuilder &b);
	if (rhsInt) {
	rhsValue = createConstIndexValueOp(loc, b, rhsInt.value());
	} else {
	rhsValue = convertToIndexIfNeeded(rhsValue, loc, b);
	assert(isa<IndexType>(rhsValue.getType()));
	}

[github-actions]

[bug · critical]
Same copy-paste bug in divOpFoldResult: RHS side incorrectly uses lhsValue instead of rhsValue

Same pattern — RHS operand conversion uses lhsValue instead of rhsValue, causing arith::DivSIOp to receive type-mismatched operands.

Suggestion:

Suggested change

	OpFoldResult divOpFoldResult(const OpFoldResult &lhs, const OpFoldResult &rhs,
	const Location &loc, OpBuilder &b);
	if (rhsInt) {
	rhsValue = createConstIndexValueOp(loc, b, rhsInt.value());
	} else {
	rhsValue = convertToIndexIfNeeded(rhsValue, loc, b);
	assert(isa<IndexType>(rhsValue.getType()));
	}

[github-actions]

[bug · critical]
Copy-paste bug: RHS operand conversion incorrectly uses lhsValue instead of rhsValue

In the else branch handling the RHS (when rhsInt is falsy), the code converts lhsValue again instead of converting rhsValue. This means:

• rhsValue remains unconverted (e.g., i32/i64 type instead of index type)
• lhsValue gets re-converted (redundantly)
• The subsequent arith::AddIOp receives operands with mismatched types, which will fail MLIR verification

This bug exists in all similar functions (subOpFoldResult, mulOpFoldResult, divOpFoldResult, remOpFoldResult, minOpFoldResult) — only maxOpFoldResult is correctly implemented.

Suggestion:

Suggested change

	Value convertToIndexIfNeeded(Value intValue, const Location &loc, OpBuilder &b);
	auto rhsValue = dyn_cast<Value>(rhs);
	if (rhsInt) {
	rhsValue = createConstIndexValueOp(loc, b, rhsInt.value());
	} else {
	rhsValue = convertToIndexIfNeeded(rhsValue, loc, b);
	assert(isa<IndexType>(rhsValue.getType()));
	}

[github-actions]

[bug · critical]
Same copy-paste bug in remOpFoldResult: RHS side incorrectly uses lhsValue instead of rhsValue

Same pattern — RHS operand conversion uses lhsValue instead of rhsValue, causing arith::RemSIOp to receive type-mismatched operands.

Suggestion:

Suggested change

	Value convertToIndexIfNeeded(Value intValue, const Location &loc, OpBuilder &b);
	if (rhsInt) {
	rhsValue = createConstIndexValueOp(loc, b, rhsInt.value());
	} else {
	rhsValue = convertToIndexIfNeeded(rhsValue, loc, b);
	assert(isa<IndexType>(rhsValue.getType()));
	}

[github-actions]

[bug · critical]
Same copy-paste bug in minOpFoldResult: RHS side incorrectly uses lhsValue instead of rhsValue

Same pattern — RHS operand conversion uses lhsValue instead of rhsValue, causing arith::MinSIOp to receive type-mismatched operands.

Suggestion:

Suggested change

	Value convertToIndexIfNeeded(Value intValue, const Location &loc, OpBuilder &b);
	if (rhsInt) {
	rhsValue = createConstIndexValueOp(loc, b, rhsInt.value());
	} else {
	rhsValue = convertToIndexIfNeeded(rhsValue, loc, b);
	assert(isa<IndexType>(rhsValue.getType()));
	}