NVMeQueuePair::reqs has length NVME_NUM_REQS, which less than
NVME_QUEUE_SIZE by 1.
Fixes: 1086e95da17050 ("block/nvme: switch to a NVMeRequest freelist")
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsement...@yandex-team.ru>
---
Cc: stefa...@redhat.com
Cc: alex.c...@huawei.com
Cc: euler.ro...@huawei.com
Note, that there was similar patch in the past:
https://patchew.org/QEMU/20201208144452.91172-1-alex.c...@huawei.com/
I still think, that using NVME_NUM_REQS is better here.
block/nvme.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/block/nvme.c b/block/nvme.c
index b6e95f0b7e..0faedf3072 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -416,9 +416,10 @@ static bool nvme_process_completion(NVMeQueuePair *q)
q->cq_phase = !q->cq_phase;
}
cid = le16_to_cpu(c->cid);
- if (cid == 0 || cid > NVME_QUEUE_SIZE) {
- warn_report("NVMe: Unexpected CID in completion queue: %"PRIu32", "
- "queue size: %u", cid, NVME_QUEUE_SIZE);
+ if (cid == 0 || cid > NVME_NUM_REQS) {
+ warn_report("NVMe: Unexpected CID in completion queue: %" PRIu32
+ ", should be within: 1..%u inclusively", cid,
+ NVME_NUM_REQS);
continue;
}
trace_nvme_complete_command(s, q->index, cid);
--
2.34.1
