[KV offload] Offloading connector async scheduling support (#27648)

Signed-off-by: KevinCheung2259 <2651309292@qq.com>
Co-authored-by: Nick Hill <nhill@redhat.com>
This commit is contained in:
Yue Zhang
2025-11-01 21:08:56 +00:00
committed by GitHub
co-authored by Nick Hill
parent 1e88fb751b
commit 685c99ee77
@@ -274,8 +274,8 @@ class OffloadingConnectorScheduler:
if num_new_blocks <= 0: if num_new_blocks <= 0:
continue continue
num_gpu_blocks = num_blocks * self.block_size_factor # NOTE: In async scheduling, placeholders may temporarily make
assert len(req.block_hashes) >= num_gpu_blocks # len(req.block_hashes) < num_blocks * self.block_size_factor.
new_block_hashes = self._get_block_hashes( new_block_hashes = self._get_block_hashes(
req, start_idx=start_block_idx, end_idx=num_blocks req, start_idx=start_block_idx, end_idx=num_blocks