Commit ac3a42b1 authored by Jim Harris's avatar Jim Harris Committed by Tomasz Zawadzki
Browse files

nvmf: retry connect commands internally when subsys not ready 



It is better to not fail connect commands when a subsystem
is not ready.  The host will not be expecting that and will
typically treat it as a catastrophic failure (i.e. it won't
retry the connect).

So instead when this situation occurs, start a poller for
the connect request.  We will continue to retry processing
it until the subsystem is ready to handle it.

Fixes issue #1985.

Signed-off-by: default avatarJim Harris <james.r.harris@intel.com>
Change-Id: Id8835df8f0edf1e889fdd7e754e261c2a880cbb6
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/8571


Tested-by: default avatarSPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: default avatarZiye Yang <ziye.yang@intel.com>
Reviewed-by: default avatarChangpeng Liu <changpeng.liu@intel.com>
Reviewed-by: default avatarShuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
parent 65ef1f32

lib/nvmf/ctrlr.c

+40 −4
Original line number Diff line number Diff line
@@ -801,6 +801,25 @@ out: 
	return status;

}



static int nvmf_ctrlr_cmd_connect(struct spdk_nvmf_request *req);



static int

retry_connect(void *arg)

{

	struct spdk_nvmf_request *req = arg;

	struct spdk_nvmf_subsystem_poll_group *sgroup;

	int rc;



	sgroup = nvmf_subsystem_pg_from_connect_cmd(req);

	sgroup->mgmt_io_outstanding++;

	spdk_poller_unregister(&req->poller);

	rc = nvmf_ctrlr_cmd_connect(req);

	if (rc == SPDK_NVMF_REQUEST_EXEC_STATUS_COMPLETE) {

		_nvmf_request_complete(req);

	}

	return SPDK_POLLER_BUSY;

}



static int

nvmf_ctrlr_cmd_connect(struct spdk_nvmf_request *req)

{
@@ -825,12 +844,29 @@ nvmf_ctrlr_cmd_connect(struct spdk_nvmf_request *req) 
	    (subsystem->state == SPDK_NVMF_SUBSYSTEM_PAUSING) ||

	    (subsystem->state == SPDK_NVMF_SUBSYSTEM_PAUSED) ||

	    (subsystem->state == SPDK_NVMF_SUBSYSTEM_DEACTIVATING)) {

		SPDK_ERRLOG("Subsystem '%s' is not ready\n", subsystem->subnqn);

		struct spdk_nvmf_subsystem_poll_group *sgroup;



		if (req->timeout_tsc == 0) {

			/* We will only retry the request up to 1 second. */

			req->timeout_tsc = spdk_get_ticks() + spdk_get_ticks_hz();

		} else if (spdk_get_ticks() > req->timeout_tsc) {

			SPDK_ERRLOG("Subsystem '%s' was not ready for 1 second\n", subsystem->subnqn);

			rsp->status.sct = SPDK_NVME_SCT_COMMAND_SPECIFIC;

			rsp->status.sc = SPDK_NVMF_FABRIC_SC_CONTROLLER_BUSY;

			return SPDK_NVMF_REQUEST_EXEC_STATUS_COMPLETE;

		}



		/* Subsystem is not ready to handle a connect. Use a poller to retry it

		 * again later. Decrement the mgmt_io_outstanding to avoid the

		 * subsystem waiting for this command to complete before unpausing.

		 */

		sgroup = nvmf_subsystem_pg_from_connect_cmd(req);

		sgroup->mgmt_io_outstanding--;

		SPDK_DEBUGLOG(nvmf, "Subsystem '%s' is not ready for connect, retrying...\n", subsystem->subnqn);

		req->poller = SPDK_POLLER_REGISTER(retry_connect, req, 100);

		return SPDK_NVMF_REQUEST_EXEC_STATUS_ASYNCHRONOUS;

	}



	/* Ensure that hostnqn is null terminated */

	if (!memchr(data->hostnqn, '\0', SPDK_NVMF_NQN_MAX_LEN + 1)) {

		SPDK_ERRLOG("Connect HOSTNQN is not null terminated\n");