1 files changed, 303 insertions, 86 deletions

diff --git a/drivers/scsi/libiscsi.c b/drivers/scsi/libiscsi.c
index f1a4246f890c..b7689f3d05f5 100644
--- a/drivers/scsi/libiscsi.c
+++ b/drivers/scsi/libiscsi.c
@@ -266,6 +266,88 @@ static int iscsi_prep_bidi_ahs(struct iscsi_task *task)
 }
 
 /**
+ * iscsi_check_tmf_restrictions - check if a task is affected by TMF
+ * @task: iscsi task
+ * @opcode: opcode to check for
+ *
+ * During TMF a task has to be checked if it's affected.
+ * All unrelated I/O can be passed through, but I/O to the
+ * affected LUN should be restricted.
+ * If 'fast_abort' is set we won't be sending any I/O to the
+ * affected LUN.
+ * Otherwise the target is waiting for all TTTs to be completed,
+ * so we have to send all outstanding Data-Out PDUs to the target.
+ */
+static int iscsi_check_tmf_restrictions(struct iscsi_task *task, int opcode)
+{
+        struct iscsi_conn *conn = task->conn;
+        struct iscsi_tm *tmf = &conn->tmhdr;
+        unsigned int hdr_lun;
+
+        if (conn->tmf_state == TMF_INITIAL)
+                return 0;
+
+        if ((tmf->opcode & ISCSI_OPCODE_MASK) != ISCSI_OP_SCSI_TMFUNC)
+                return 0;
+
+        switch (ISCSI_TM_FUNC_VALUE(tmf)) {
+        case ISCSI_TM_FUNC_LOGICAL_UNIT_RESET:
+                /*
+                 * Allow PDUs for unrelated LUNs
+                 */
+                hdr_lun = scsilun_to_int((struct scsi_lun *)tmf->lun);
+                if (hdr_lun != task->sc->device->lun)
+                        return 0;
+                /* fall through */
+        case ISCSI_TM_FUNC_TARGET_WARM_RESET:
+                /*
+                 * Fail all SCSI cmd PDUs
+                 */
+                if (opcode != ISCSI_OP_SCSI_DATA_OUT) {
+                        iscsi_conn_printk(KERN_INFO, conn,
+                                          "task [op %x/%x itt "
+                                          "0x%x/0x%x] "
+                                          "rejected.\n",
+                                          task->hdr->opcode, opcode,
+                                          task->itt, task->hdr_itt);
+                        return -EACCES;
+                }
+                /*
+                 * And also all data-out PDUs in response to R2T
+                 * if fast_abort is set.
+                 */
+                if (conn->session->fast_abort) {
+                        iscsi_conn_printk(KERN_INFO, conn,
+                                          "task [op %x/%x itt "
+                                          "0x%x/0x%x] fast abort.\n",
+                                          task->hdr->opcode, opcode,
+                                          task->itt, task->hdr_itt);
+                        return -EACCES;
+                }
+                break;
+        case ISCSI_TM_FUNC_ABORT_TASK:
+                /*
+                 * the caller has already checked if the task
+                 * they want to abort was in the pending queue so if
+                 * we are here the cmd pdu has gone out already, and
+                 * we will only hit this for data-outs
+                 */
+                if (opcode == ISCSI_OP_SCSI_DATA_OUT &&
+                    task->hdr_itt == tmf->rtt) {
+                        ISCSI_DBG_SESSION(conn->session,
+                                          "Preventing task %x/%x from sending "
+                                          "data-out due to abort task in "
+                                          "progress\n", task->itt,
+                                          task->hdr_itt);
+                        return -EACCES;
+                }
+                break;
+        }
+
+        return 0;
+}
+
+/**
  * iscsi_prep_scsi_cmd_pdu - prep iscsi scsi cmd pdu
  * @task: iscsi task
  *
@@ -282,6 +364,10 @@ static int iscsi_prep_scsi_cmd_pdu(struct iscsi_task *task)
         itt_t itt;
         int rc;
 
+        rc = iscsi_check_tmf_restrictions(task, ISCSI_OP_SCSI_CMD);
+        if (rc)
+                return rc;
+
         if (conn->session->tt->alloc_pdu) {
                 rc = conn->session->tt->alloc_pdu(task, ISCSI_OP_SCSI_CMD);
                 if (rc)
@@ -577,12 +663,12 @@ static int iscsi_prep_mgmt_task(struct iscsi_conn *conn,
         struct iscsi_session *session = conn->session;
         struct iscsi_hdr *hdr = task->hdr;
         struct iscsi_nopout *nop = (struct iscsi_nopout *)hdr;
+        uint8_t opcode = hdr->opcode & ISCSI_OPCODE_MASK;
 
         if (conn->session->state == ISCSI_STATE_LOGGING_OUT)
                 return -ENOTCONN;
 
-        if (hdr->opcode != (ISCSI_OP_LOGIN | ISCSI_OP_IMMEDIATE) &&
-            hdr->opcode != (ISCSI_OP_TEXT | ISCSI_OP_IMMEDIATE))
+        if (opcode != ISCSI_OP_LOGIN && opcode != ISCSI_OP_TEXT)
                 nop->exp_statsn = cpu_to_be32(conn->exp_statsn);
         /*
          * pre-format CmdSN for outgoing PDU.
@@ -590,9 +676,12 @@ static int iscsi_prep_mgmt_task(struct iscsi_conn *conn,
         nop->cmdsn = cpu_to_be32(session->cmdsn);
         if (hdr->itt != RESERVED_ITT) {
                 /*
-                 * TODO: We always use immediate, so we never hit this.
+                 * TODO: We always use immediate for normal session pdus.
                  * If we start to send tmfs or nops as non-immediate then
                  * we should start checking the cmdsn numbers for mgmt tasks.
+                 *
+                 * During discovery sessions iscsid sends TEXT as non immediate,
+                 * but we always only send one PDU at a time.
                  */
                 if (conn->c_stage == ISCSI_CONN_STARTED &&
                     !(hdr->opcode & ISCSI_OP_IMMEDIATE)) {
@@ -620,22 +709,28 @@ __iscsi_conn_send_pdu(struct iscsi_conn *conn, struct iscsi_hdr *hdr,
 {
         struct iscsi_session *session = conn->session;
         struct iscsi_host *ihost = shost_priv(session->host);
+        uint8_t opcode = hdr->opcode & ISCSI_OPCODE_MASK;
         struct iscsi_task *task;
         itt_t itt;
 
         if (session->state == ISCSI_STATE_TERMINATE)
                 return NULL;
 
-        if (hdr->opcode == (ISCSI_OP_LOGIN | ISCSI_OP_IMMEDIATE) ||
-            hdr->opcode == (ISCSI_OP_TEXT | ISCSI_OP_IMMEDIATE))
+        if (opcode == ISCSI_OP_LOGIN || opcode == ISCSI_OP_TEXT) {
                 /*
                  * Login and Text are sent serially, in
                  * request-followed-by-response sequence.
                  * Same task can be used. Same ITT must be used.
                  * Note that login_task is preallocated at conn_create().
                  */
+                if (conn->login_task->state != ISCSI_TASK_FREE) {
+                        iscsi_conn_printk(KERN_ERR, conn, "Login/Text in "
+                                          "progress. Cannot start new task.\n");
+                        return NULL;
+                }
+
                 task = conn->login_task;
-        else {
+        } else {
                 if (session->state != ISCSI_STATE_LOGGED_IN)
                         return NULL;
 
@@ -1357,6 +1452,7 @@ EXPORT_SYMBOL_GPL(iscsi_requeue_task);
  **/
 static int iscsi_data_xmit(struct iscsi_conn *conn)
 {
+        struct iscsi_task *task;
         int rc = 0;
 
         spin_lock_bh(&conn->session->lock);
@@ -1394,11 +1490,8 @@ check_mgmt:
 
         /* process pending command queue */
         while (!list_empty(&conn->cmdqueue)) {
-                if (conn->tmf_state == TMF_QUEUED)
-                        break;
-
-                conn->task = list_entry(conn->cmdqueue.next,
-                                         struct iscsi_task, running);
+                conn->task = list_entry(conn->cmdqueue.next, struct iscsi_task,
+                                        running);
                 list_del_init(&conn->task->running);
                 if (conn->session->state == ISCSI_STATE_LOGGING_OUT) {
                         fail_scsi_task(conn->task, DID_IMM_RETRY);
@@ -1406,7 +1499,7 @@ check_mgmt:
                 }
                 rc = iscsi_prep_scsi_cmd_pdu(conn->task);
                 if (rc) {
-                        if (rc == -ENOMEM) {
+                        if (rc == -ENOMEM || rc == -EACCES) {
                                 list_add_tail(&conn->task->running,
                                               &conn->cmdqueue);
                                 conn->task = NULL;
@@ -1428,17 +1521,18 @@ check_mgmt:
         }
 
         while (!list_empty(&conn->requeue)) {
-                if (conn->session->fast_abort && conn->tmf_state != TMF_INITIAL)
-                        break;
-
                 /*
                  * we always do fastlogout - conn stop code will clean up.
                  */
                 if (conn->session->state == ISCSI_STATE_LOGGING_OUT)
                         break;
 
-                conn->task = list_entry(conn->requeue.next,
-                                         struct iscsi_task, running);
+                task = list_entry(conn->requeue.next, struct iscsi_task,
+                                  running);
+                if (iscsi_check_tmf_restrictions(task, ISCSI_OP_SCSI_DATA_OUT))
+                        break;
+
+                conn->task = task;
                 list_del_init(&conn->task->running);
                 conn->task->state = ISCSI_TASK_RUNNING;
                 rc = iscsi_xmit_task(conn);
@@ -1591,7 +1685,7 @@ int iscsi_queuecommand(struct scsi_cmnd *sc, void (*done)(struct scsi_cmnd *))
         if (!ihost->workq) {
                 reason = iscsi_prep_scsi_cmd_pdu(task);
                 if (reason) {
-                        if (reason == -ENOMEM) {
+                        if (reason == -ENOMEM ||  reason == -EACCES) {
                                 reason = FAILURE_OOM;
                                 goto prepd_reject;
                         } else {
@@ -1643,9 +1737,21 @@ fault:
 }
 EXPORT_SYMBOL_GPL(iscsi_queuecommand);
 
-int iscsi_change_queue_depth(struct scsi_device *sdev, int depth)
+int iscsi_change_queue_depth(struct scsi_device *sdev, int depth, int reason)
 {
-        scsi_adjust_queue_depth(sdev, scsi_get_tag_type(sdev), depth);
+        switch (reason) {
+        case SCSI_QDEPTH_DEFAULT:
+                scsi_adjust_queue_depth(sdev, scsi_get_tag_type(sdev), depth);
+                break;
+        case SCSI_QDEPTH_QFULL:
+                scsi_track_queue_full(sdev, depth);
+                break;
+        case SCSI_QDEPTH_RAMP_UP:
+                scsi_adjust_queue_depth(sdev, scsi_get_tag_type(sdev), depth);
+                break;
+        default:
+                return -EOPNOTSUPP;
+        }
         return sdev->queue_depth;
 }
 EXPORT_SYMBOL_GPL(iscsi_change_queue_depth);
@@ -1660,72 +1766,6 @@ int iscsi_target_alloc(struct scsi_target *starget)
 }
 EXPORT_SYMBOL_GPL(iscsi_target_alloc);
 
-void iscsi_session_recovery_timedout(struct iscsi_cls_session *cls_session)
-{
-        struct iscsi_session *session = cls_session->dd_data;
-
-        spin_lock_bh(&session->lock);
-        if (session->state != ISCSI_STATE_LOGGED_IN) {
-                session->state = ISCSI_STATE_RECOVERY_FAILED;
-                if (session->leadconn)
-                        wake_up(&session->leadconn->ehwait);
-        }
-        spin_unlock_bh(&session->lock);
-}
-EXPORT_SYMBOL_GPL(iscsi_session_recovery_timedout);
-
-int iscsi_eh_target_reset(struct scsi_cmnd *sc)
-{
-        struct iscsi_cls_session *cls_session;
-        struct iscsi_session *session;
-        struct iscsi_conn *conn;
-
-        cls_session = starget_to_session(scsi_target(sc->device));
-        session = cls_session->dd_data;
-        conn = session->leadconn;
-
-        mutex_lock(&session->eh_mutex);
-        spin_lock_bh(&session->lock);
-        if (session->state == ISCSI_STATE_TERMINATE) {
-failed:
-                ISCSI_DBG_EH(session,
-                             "failing target reset: Could not log back into "
-                             "target [age %d]\n",
-                             session->age);
-                spin_unlock_bh(&session->lock);
-                mutex_unlock(&session->eh_mutex);
-                return FAILED;
-        }
-
-        spin_unlock_bh(&session->lock);
-        mutex_unlock(&session->eh_mutex);
-        /*
-         * we drop the lock here but the leadconn cannot be destoyed while
-         * we are in the scsi eh
-         */
-        iscsi_conn_failure(conn, ISCSI_ERR_CONN_FAILED);
-
-        ISCSI_DBG_EH(session, "wait for relogin\n");
-        wait_event_interruptible(conn->ehwait,
-                                 session->state == ISCSI_STATE_TERMINATE ||
-                                 session->state == ISCSI_STATE_LOGGED_IN ||
-                                 session->state == ISCSI_STATE_RECOVERY_FAILED);
-        if (signal_pending(current))
-                flush_signals(current);
-
-        mutex_lock(&session->eh_mutex);
-        spin_lock_bh(&session->lock);
-        if (session->state == ISCSI_STATE_LOGGED_IN) {
-                ISCSI_DBG_EH(session,
-                             "target reset succeeded\n");
-        } else
-                goto failed;
-        spin_unlock_bh(&session->lock);
-        mutex_unlock(&session->eh_mutex);
-        return SUCCESS;
-}
-EXPORT_SYMBOL_GPL(iscsi_eh_target_reset);
-
 static void iscsi_tmf_timedout(unsigned long data)
 {
         struct iscsi_conn *conn = (struct iscsi_conn *)data;
@@ -2108,6 +2148,7 @@ int iscsi_eh_abort(struct scsi_cmnd *sc)
                 spin_lock_bh(&session->lock);
                 fail_scsi_task(task, DID_ABORT);
                 conn->tmf_state = TMF_INITIAL;
+                memset(hdr, 0, sizeof(*hdr));
                 spin_unlock_bh(&session->lock);
                 iscsi_start_tx(conn);
                 goto success_unlocked;
@@ -2118,6 +2159,7 @@ int iscsi_eh_abort(struct scsi_cmnd *sc)
         case TMF_NOT_FOUND:
                 if (!sc->SCp.ptr) {
                         conn->tmf_state = TMF_INITIAL;
+                        memset(hdr, 0, sizeof(*hdr));
                         /* task completed before tmf abort response */
                         ISCSI_DBG_EH(session, "sc completed while abort in "
                                               "progress\n");
@@ -2212,6 +2254,7 @@ int iscsi_eh_device_reset(struct scsi_cmnd *sc)
         iscsi_suspend_tx(conn);
 
         spin_lock_bh(&session->lock);
+        memset(hdr, 0, sizeof(*hdr));
         fail_scsi_tasks(conn, sc->device->lun, DID_ERROR);
         conn->tmf_state = TMF_INITIAL;
         spin_unlock_bh(&session->lock);
@@ -2229,6 +2272,172 @@ done:
 }
 EXPORT_SYMBOL_GPL(iscsi_eh_device_reset);
 
+void iscsi_session_recovery_timedout(struct iscsi_cls_session *cls_session)
+{
+        struct iscsi_session *session = cls_session->dd_data;
+
+        spin_lock_bh(&session->lock);
+        if (session->state != ISCSI_STATE_LOGGED_IN) {
+                session->state = ISCSI_STATE_RECOVERY_FAILED;
+                if (session->leadconn)
+                        wake_up(&session->leadconn->ehwait);
+        }
+        spin_unlock_bh(&session->lock);
+}
+EXPORT_SYMBOL_GPL(iscsi_session_recovery_timedout);
+
+/**
+ * iscsi_eh_session_reset - drop session and attempt relogin
+ * @sc: scsi command
+ *
+ * This function will wait for a relogin, session termination from
+ * userspace, or a recovery/replacement timeout.
+ */
+static int iscsi_eh_session_reset(struct scsi_cmnd *sc)
+{
+        struct iscsi_cls_session *cls_session;
+        struct iscsi_session *session;
+        struct iscsi_conn *conn;
+
+        cls_session = starget_to_session(scsi_target(sc->device));
+        session = cls_session->dd_data;
+        conn = session->leadconn;
+
+        mutex_lock(&session->eh_mutex);
+        spin_lock_bh(&session->lock);
+        if (session->state == ISCSI_STATE_TERMINATE) {
+failed:
+                ISCSI_DBG_EH(session,
+                             "failing session reset: Could not log back into "
+                             "%s, %s [age %d]\n", session->targetname,
+                             conn->persistent_address, session->age);
+                spin_unlock_bh(&session->lock);
+                mutex_unlock(&session->eh_mutex);
+                return FAILED;
+        }
+
+        spin_unlock_bh(&session->lock);
+        mutex_unlock(&session->eh_mutex);
+        /*
+         * we drop the lock here but the leadconn cannot be destoyed while
+         * we are in the scsi eh
+         */
+        iscsi_conn_failure(conn, ISCSI_ERR_CONN_FAILED);
+
+        ISCSI_DBG_EH(session, "wait for relogin\n");
+        wait_event_interruptible(conn->ehwait,
+                                 session->state == ISCSI_STATE_TERMINATE ||
+                                 session->state == ISCSI_STATE_LOGGED_IN ||
+                                 session->state == ISCSI_STATE_RECOVERY_FAILED);
+        if (signal_pending(current))
+                flush_signals(current);
+
+        mutex_lock(&session->eh_mutex);
+        spin_lock_bh(&session->lock);
+        if (session->state == ISCSI_STATE_LOGGED_IN) {
+                ISCSI_DBG_EH(session,
+                             "session reset succeeded for %s,%s\n",
+                             session->targetname, conn->persistent_address);
+        } else
+                goto failed;
+        spin_unlock_bh(&session->lock);
+        mutex_unlock(&session->eh_mutex);
+        return SUCCESS;
+}
+
+static void iscsi_prep_tgt_reset_pdu(struct scsi_cmnd *sc, struct iscsi_tm *hdr)
+{
+        memset(hdr, 0, sizeof(*hdr));
+        hdr->opcode = ISCSI_OP_SCSI_TMFUNC | ISCSI_OP_IMMEDIATE;
+        hdr->flags = ISCSI_TM_FUNC_TARGET_WARM_RESET & ISCSI_FLAG_TM_FUNC_MASK;
+        hdr->flags |= ISCSI_FLAG_CMD_FINAL;
+        hdr->rtt = RESERVED_ITT;
+}
+
+/**
+ * iscsi_eh_target_reset - reset target
+ * @sc: scsi command
+ *
+ * This will attempt to send a warm target reset. If that fails
+ * then we will drop the session and attempt ERL0 recovery.
+ */
+int iscsi_eh_target_reset(struct scsi_cmnd *sc)
+{
+        struct iscsi_cls_session *cls_session;
+        struct iscsi_session *session;
+        struct iscsi_conn *conn;
+        struct iscsi_tm *hdr;
+        int rc = FAILED;
+
+        cls_session = starget_to_session(scsi_target(sc->device));
+        session = cls_session->dd_data;
+
+        ISCSI_DBG_EH(session, "tgt Reset [sc %p tgt %s]\n", sc,
+                     session->targetname);
+
+        mutex_lock(&session->eh_mutex);
+        spin_lock_bh(&session->lock);
+        /*
+         * Just check if we are not logged in. We cannot check for
+         * the phase because the reset could come from a ioctl.
+         */
+        if (!session->leadconn || session->state != ISCSI_STATE_LOGGED_IN)
+                goto unlock;
+        conn = session->leadconn;
+
+        /* only have one tmf outstanding at a time */
+        if (conn->tmf_state != TMF_INITIAL)
+                goto unlock;
+        conn->tmf_state = TMF_QUEUED;
+
+        hdr = &conn->tmhdr;
+        iscsi_prep_tgt_reset_pdu(sc, hdr);
+
+        if (iscsi_exec_task_mgmt_fn(conn, hdr, session->age,
+                                    session->tgt_reset_timeout)) {
+                rc = FAILED;
+                goto unlock;
+        }
+
+        switch (conn->tmf_state) {
+        case TMF_SUCCESS:
+                break;
+        case TMF_TIMEDOUT:
+                spin_unlock_bh(&session->lock);
+                iscsi_conn_failure(conn, ISCSI_ERR_CONN_FAILED);
+                goto done;
+        default:
+                conn->tmf_state = TMF_INITIAL;
+                goto unlock;
+        }
+
+        rc = SUCCESS;
+        spin_unlock_bh(&session->lock);
+
+        iscsi_suspend_tx(conn);
+
+        spin_lock_bh(&session->lock);
+        memset(hdr, 0, sizeof(*hdr));
+        fail_scsi_tasks(conn, -1, DID_ERROR);
+        conn->tmf_state = TMF_INITIAL;
+        spin_unlock_bh(&session->lock);
+
+        iscsi_start_tx(conn);
+        goto done;
+
+unlock:
+        spin_unlock_bh(&session->lock);
+done:
+        ISCSI_DBG_EH(session, "tgt %s reset result = %s\n", session->targetname,
+                     rc == SUCCESS ? "SUCCESS" : "FAILED");
+        mutex_unlock(&session->eh_mutex);
+
+        if (rc == FAILED)
+                rc = iscsi_eh_session_reset(sc);
+        return rc;
+}
+EXPORT_SYMBOL_GPL(iscsi_eh_target_reset);
+
 /*
  * Pre-allocate a pool of @max items of @item_size. By default, the pool
  * should be accessed via kfifo_{get,put} on q->queue.
@@ -2495,6 +2704,7 @@ iscsi_session_setup(struct iscsi_transport *iscsit, struct Scsi_Host *shost,
         session->host = shost;
         session->state = ISCSI_STATE_FREE;
         session->fast_abort = 1;
+        session->tgt_reset_timeout = 30;
         session->lu_reset_timeout = 15;
         session->abort_timeout = 10;
         session->scsi_cmds_max = scsi_cmds;
@@ -2856,6 +3066,7 @@ static void iscsi_start_session_recovery(struct iscsi_session *session,
         spin_lock_bh(&session->lock);
         fail_scsi_tasks(conn, -1, DID_TRANSPORT_DISRUPTED);
         fail_mgmt_tasks(session, conn);
+        memset(&conn->tmhdr, 0, sizeof(conn->tmhdr));
         spin_unlock_bh(&session->lock);
         mutex_unlock(&session->eh_mutex);
 }
@@ -2932,6 +3143,9 @@ int iscsi_set_param(struct iscsi_cls_conn *cls_conn,
         case ISCSI_PARAM_LU_RESET_TMO:
                 sscanf(buf, "%d", &session->lu_reset_timeout);
                 break;
+        case ISCSI_PARAM_TGT_RESET_TMO:
+                sscanf(buf, "%d", &session->tgt_reset_timeout);
+                break;
         case ISCSI_PARAM_PING_TMO:
                 sscanf(buf, "%d", &conn->ping_timeout);
                 break;
@@ -3031,6 +3245,9 @@ int iscsi_session_get_param(struct iscsi_cls_session *cls_session,
         case ISCSI_PARAM_LU_RESET_TMO:
                 len = sprintf(buf, "%d\n", session->lu_reset_timeout);
                 break;
+        case ISCSI_PARAM_TGT_RESET_TMO:
+                len = sprintf(buf, "%d\n", session->tgt_reset_timeout);
+                break;
         case ISCSI_PARAM_INITIAL_R2T_EN:
                 len = sprintf(buf, "%d\n", session->initial_r2t_en);
                 break;