[LU-8193] request mbits isn't set properly for EINPROGRESS resend

• Type: Bug
• Resolution: Fixed
• Priority: Major
• Fix Version/s: Lustre 2.9.0
• Severity: 3
• Assignee: Niu Yawei (Inactive)
• Reporter: Niu Yawei (Inactive)

Description

in ptl_send_rpc():

        /* If it's the first time to resend the request for EINPROGRESS,
         * we need to allocate a new XID (see after_reply()), it's different
         * from the resend for reply timeout. */
        if (request->rq_nr_resend != 0 &&
            list_empty(&request->rq_unreplied_list)) {
                __u64 min_xid = 0;
                /* resend for EINPROGRESS, allocate new xid to avoid reply
                 * reconstruction */
                spin_lock(&imp->imp_lock);
                ptlrpc_assign_next_xid_nolock(request);
                request->rq_mbits = request->rq_xid;
                min_xid = ptlrpc_known_replied_xid(imp);
                spin_unlock(&imp->imp_lock);

                lustre_msg_set_last_xid(request->rq_reqmsg, min_xid);
                DEBUG_REQ(D_RPCTRACE, request, "Allocating new xid for "
                          "resend on EINPROGRESS");
        } else if (request->rq_bulk != NULL) {
                ptlrpc_set_bulk_mbits(request);
                lustre_msg_set_mbits(request->rq_reqmsg, request->rq_mbits);
        }

lustre_msg_set_mbits() isn't called for first time EINPROGRESS resend, so the mbits won't be set properly in ptlrpc_body. This part of code needs be revised a bit.

Issue Links

• is related to: LU-7990 Large bulk IO support