[V1] Support sliding window attention (vllm-project#9679)

Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
Signed-off-by: Linkun Chen <github+anyscale@lkchen.net>

Author: WoosukKwon

vllm/v1/attention/backends/flash_attn.py

@@ -82,8 +82,10 @@ def __init__(
         if alibi_slopes is not None:
             alibi_slopes = torch.tensor(alibi_slopes, dtype=torch.float32)
         self.alibi_slopes = alibi_slopes
-        self.sliding_window = ((sliding_window, sliding_window)
-                               if sliding_window is not None else (-1, -1))
+        if sliding_window is None:
+            self.sliding_window = (-1, -1)
+        else:
+            self.sliding_window = (sliding_window - 1, 0)
         self.kv_cache_dtype = kv_cache_dtype
         if logits_soft_cap is None:
             # In flash-attn, setting logits_soft_cap as 0 means no soft cap.
@@ -93,12 +95,6 @@ def __init__(
         assert self.num_heads % self.num_kv_heads == 0
         self.num_queries_per_kv = self.num_heads // self.num_kv_heads
 
-        if sliding_window is not None:
-            # NOTE(woosuk): flash-attn's sliding window does not work with
-            # paged KV cache.
-            raise ValueError(
-                "Sliding window is not supported in FlashAttention.")
-
         support_head_sizes = FlashAttentionBackend.get_supported_head_sizes()
         if head_size not in support_head_sizes:
             raise ValueError(