lib/nvme: Don't re-queue I/O while disconnecting.

This can cause infinite loops if the callback tries to
queue an additional I/O.

Signed-off-by: Seth Howell <seth.howell@intel.com>
Change-Id: I4b80b97d334082465d9228b799ef901645fa968e
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1854
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This commit is contained in:
Seth Howell 2020-04-14 15:12:31 -07:00 committed by Jim Harris
parent b874f65743
commit af2d56ed94

View File

@ -725,6 +725,11 @@ nvme_qpair_submit_request(struct spdk_nvme_qpair *qpair, struct nvme_request *re
{
int rc;
/* This prevents us from entering an infinite loop when freeing queued I/O in disconnect. */
if (spdk_unlikely(nvme_qpair_get_state(qpair) == NVME_QPAIR_DISCONNECTING)) {
return -ENXIO;
}
if (spdk_unlikely(!STAILQ_EMPTY(&qpair->queued_req) && req->num_children == 0)) {
/*
* requests that have no children should be sent to the transport after all