bnx2i: Updated the connection shutdown/cleanup timeout
authorEddie Wai <eddie.wai@broadcom.com>
Mon, 16 May 2011 18:13:19 +0000 (11:13 -0700)
committerSteve Conklin <sconklin@canonical.com>
Fri, 15 Jul 2011 17:20:59 +0000 (12:20 -0500)
BugLink: http://bugs.launchpad.net/bugs/793702

commit d5307a078bb0288945c900c6f4a2fd77ba6d0817 upstream.

Modified the 10s wait time for inflight offload connections to
advance to the next state to 2s based on test result.
Modified the 20s shutdown timeout to 30s based on test result.

Signed-off-by: Eddie Wai <eddie.wai@broadcom.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <jbottomley@parallels.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

drivers/scsi/bnx2i/bnx2i_init.c
drivers/scsi/bnx2i/bnx2i_iscsi.c

index 72a7b2d..dd4622e 100644 (file)
@@ -227,7 +227,7 @@ void bnx2i_stop(void *handle)
        wait_event_interruptible_timeout(hba->eh_wait,
                                         (list_empty(&hba->ep_ofld_list) &&
                                         list_empty(&hba->ep_destroy_list)),
-                                        10 * HZ);
+                                        2 * HZ);
        /* Wait for all endpoints to be torn down, Chip will be reset once
         *  control returns to network driver. So it is required to cleanup and
         * release all connection resources before returning from this routine.
index 204d962..0a46832 100644 (file)
@@ -858,7 +858,7 @@ struct bnx2i_hba *bnx2i_alloc_hba(struct cnic_dev *cnic)
        mutex_init(&hba->net_dev_lock);
        init_waitqueue_head(&hba->eh_wait);
        if (test_bit(BNX2I_NX2_DEV_57710, &hba->cnic_dev_type)) {
-               hba->hba_shutdown_tmo = 20 * HZ;
+               hba->hba_shutdown_tmo = 30 * HZ;
                hba->conn_teardown_tmo = 20 * HZ;
                hba->conn_ctx_destroy_tmo = 6 * HZ;
        } else {        /* 5706/5708/5709 */