Enable split-KV for paged FlashAttention decode (#32102)
### Summary
Enable`FlashAttention`'s existing split-KV path for paged decode.
Split-KV divides long KV sequences across multiple CUDA thread blocks
and combines their partial results. This improves GPU utilization during
low-batch, long-context decoding.
The optimization is limited to safe cases:
- decode with one new token per request;
- global attention without attention sinks;
- KV sequences longer than 512 tokens.
Short contexts, local-window attention, prefill, and mixed-length query
batches retain their existing paths.
