[SCSI] bnx2i: Added chip cleanup for the remove module path
authorEddie Wai <eddie.wai@broadcom.com>
Thu, 12 Aug 2010 23:44:30 +0000 (16:44 -0700)
committerJames Bottomley <James.Bottomley@suse.de>
Sun, 5 Sep 2010 17:51:16 +0000 (14:51 -0300)
In the case when an ep_connect request is interrupted due to route
request stall, if the iSCSI daemon is terminated by the user, the chip
will be left in a state which will not get cleaned up upon module
removal.  Upon module reload, when the same context id is used for a
new connection, chip panic would occur.

This patch adds chip cleanup in the module removal path.

Signed-off-by: Eddie Wai <eddie.wai@broadcom.com>
Acked-by: Benjamin Li <benli@broadcom.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
drivers/scsi/bnx2i/bnx2i_init.c
drivers/scsi/bnx2i/bnx2i_iscsi.c

index a796f56..1294936 100644 (file)
@@ -167,6 +167,38 @@ void bnx2i_start(void *handle)
 
 
 /**
+ * bnx2i_chip_cleanup - local routine to handle chip cleanup
+ * @hba:       Adapter instance to register
+ *
+ * Driver checks if adapter still has any active connections before
+ *     executing the cleanup process
+ */
+static void bnx2i_chip_cleanup(struct bnx2i_hba *hba)
+{
+       struct bnx2i_endpoint *bnx2i_ep;
+       struct list_head *pos, *tmp;
+
+       if (hba->ofld_conns_active) {
+               /* Stage to force the disconnection
+                * This is the case where the daemon is either slow or
+                * not present
+                */
+               printk(KERN_ALERT "bnx2i: (%s) chip cleanup for %d active "
+                       "connections\n", hba->netdev->name,
+                       hba->ofld_conns_active);
+               mutex_lock(&hba->net_dev_lock);
+               list_for_each_safe(pos, tmp, &hba->ep_active_list) {
+                       bnx2i_ep = list_entry(pos, struct bnx2i_endpoint, link);
+                       /* Clean up the chip only */
+                       bnx2i_hw_ep_disconnect(bnx2i_ep);
+                       bnx2i_ep->cm_sk = NULL;
+               }
+               mutex_unlock(&hba->net_dev_lock);
+       }
+}
+
+
+/**
  * bnx2i_stop - cnic callback to shutdown adapter instance
  * @handle:    transparent handle pointing to adapter structure
  *
@@ -176,8 +208,6 @@ void bnx2i_start(void *handle)
 void bnx2i_stop(void *handle)
 {
        struct bnx2i_hba *hba = handle;
-       struct list_head *pos, *tmp;
-       struct bnx2i_endpoint *bnx2i_ep;
        int conns_active;
 
        /* check if cleanup happened in GOING_DOWN context */
@@ -198,24 +228,7 @@ void bnx2i_stop(void *handle)
                if (hba->ofld_conns_active == conns_active)
                        break;
        }
-       if (hba->ofld_conns_active) {
-               /* Stage to force the disconnection
-                * This is the case where the daemon is either slow or
-                * not present
-                */
-               printk(KERN_ALERT "bnx2i: Wait timeout, force all eps "
-                       "to disconnect (%d)\n", hba->ofld_conns_active);
-               mutex_lock(&hba->net_dev_lock);
-               list_for_each_safe(pos, tmp, &hba->ep_active_list) {
-                       bnx2i_ep = list_entry(pos, struct bnx2i_endpoint, link);
-                       /* Clean up the chip only */
-                       bnx2i_hw_ep_disconnect(bnx2i_ep);
-               }
-               mutex_unlock(&hba->net_dev_lock);
-               if (hba->ofld_conns_active)
-                       printk(KERN_ERR "bnx2i: EP disconnect timeout (%d)!\n",
-                               hba->ofld_conns_active);
-       }
+       bnx2i_chip_cleanup(hba);
 
        /* This flag should be cleared last so that ep_disconnect() gracefully
         * cleans up connection context
@@ -457,6 +470,7 @@ static void __exit bnx2i_mod_exit(void)
                adapter_count--;
 
                if (test_bit(BNX2I_CNIC_REGISTERED, &hba->reg_with_cnic)) {
+                       bnx2i_chip_cleanup(hba);
                        hba->cnic->unregister_device(hba->cnic, CNIC_ULP_ISCSI);
                        clear_bit(BNX2I_CNIC_REGISTERED, &hba->reg_with_cnic);
                }
index 0425540..fb50efb 100644 (file)
@@ -1953,6 +1953,9 @@ int bnx2i_hw_ep_disconnect(struct bnx2i_endpoint *bnx2i_ep)
        if (!cnic)
                return 0;
 
+       if (bnx2i_ep->state == EP_STATE_IDLE)
+               return 0;
+
        if (!bnx2i_ep_tcp_conn_active(bnx2i_ep))
                goto destroy_conn;