[LU-7760] Incorrect pid printing in ptlrpc_send_new_req

• Type: Bug
• Resolution: Fixed
• Priority: Trivial
• Fix Version/s: Lustre 2.11.0
• Labels: easy hackathon2017
• Severity: 3
• Assignee: Sebastien Buisson (Inactive)
• Reporter: Oleg Drokin

Description

It appears that ptlrpc_send_new_req message prints pid as always 0:

00000100:00080000:5.0:1454959735.275832:0:27539:0:(client.c:1385:ptlrpc_send_new
_req()) @@@ req from PID 0 waiting for recovery: (FULL != DISCONN)  req@ffff8800
3f386b80 x1525630506782200/t0(0) o13->lustre-OST0000-osc-ffff8800cdc0f000@192.16
8.10.227@tcp:7/4 lens 224/368 e 0 to 0 dl 0 ref 1 fl Rpc:W/0/ffffffff rc 0/-1

This is because the message itself is printed before we set the value in the request:

                DEBUG_REQ(D_HA, req, "req from PID %d waiting for recovery: "
                          "(%s != %s)", lustre_msg_get_status(req->rq_reqmsg),
                          ptlrpc_import_state_name(req->rq_send_state),
                          ptlrpc_import_state_name(imp->imp_state));
                LASSERT(list_empty(&req->rq_list));
                list_add_tail(&req->rq_list, &imp->imp_delayed_list);
                atomic_inc(&req->rq_import->imp_inflight);
                spin_unlock(&imp->imp_lock);
                RETURN(0);
        }
...
        lustre_msg_set_status(req->rq_reqmsg, current_pid());

As such we should either move the set_status before that printing or just do current_pid() in that message which is likely even better.