[Bugfix] Keep the GLM-5.3 kpool tail and Qwen4 QSA ring out of the null block (#59528)

Signed-off-by: Yifan Qiao <yifanqiao@inferact.ai>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
Yifan Qiao
2026-10-08 20:44:04 +00:00
committed by GitHub
co-authored by Claude Opus 5.5
parent 2ddde19e8e
commit 240785b82c
5 changed files with 45 additions and 24 deletions
+5 -2
View File
@@ -127,7 +127,9 @@ def circular_qsa_slot_mapping(
valid = (requests >= 0) & (requests < block_table.shape[0]) & (positions >= 0)
safe_requests = requests.clamp(0, block_table.shape[0] - 1)
physical_blocks = block_table[safe_requests, 0].long()
valid &= physical_blocks >= 0
# Dummy and padding requests own no ring and sit on the null block,
# which must never be written.
valid &= physical_blocks > 0
slots = physical_blocks * compressor_state_size + positions.remainder(
compressor_state_size
)
@@ -288,7 +290,8 @@ def _build_qsa_metadata_kernel(
mask=valid,
other=-1,
)
valid &= physical_block >= 0
# Dummy and padding requests own no ring and sit on the null block.
valid &= physical_block > 0
slot = physical_block * circular_buffer_size + (
logical_position % circular_buffer_size
)