Fix DropQDQ not dropping Q/DQ around opset-22 MaxPool

[tairenpiao]

Description

ONNX added a MaxPool-22 schema, but the DropQDQ selectors still matched only MaxPool-12, so the DequantizeLinear / QuantizeLinear around an opset-22 MaxPool were no longer dropped. This updates both DropQDQ selector op-version maps to {12, 22}, matching the QDQ propagation pass. (MaxPool stays pinned to integer-capable versions ≥ 12.)

Motivation and Context

Fixes #28770. Models that optimized to full integer at opset ≤ 21 regressed at opset 22+, blocking users from upgrading their opset.

[tairenpiao]

#27059 already proposes the same qdq_selector_action_transformer.cc {12} → {12, 22} change and is the original PR for this — it should take priority.

The thing this PR adds on top is updating GetMiscOpVersionsMap in selectors_actions/shared/utils.cc to {12, 22} as well, so the runtime / EP / minimal-build QDQ node-unit selectors also recognize the opset-22 MaxPool (otherwise that path still misses it).

Happy to close this in favor of #27059 if that one-line shared/utils.cc addition is folded in there.