[SCSI] qla2xxx: Don't perform a BIG_HAMMER if Get-ID (0x20) mailbox command fails...

The semantics we employ now in the driver, performing a
BIG_HAMMER in the event of Get-ID (0x20) mailbox command
failing, should only be done for FC.  On FC configurations, it
makes sense since advertising is only really performed once,
so a BIG_HAMMER to reinitiate the process is needed to
restart.  Under FCoE, this is not needed, as there's a
continous stream of advertisements/ACks at the protocol layer
to initiate a relogin/reinitialization process.

Signed-off-by: Andrew Vasquez <andrew.vasquez@qlogic.com>
Signed-off-by: Madhuranath Iyengar <Madhu.Iyengar@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

diff --git a/drivers/scsi/qla2xxx/qla_init.c b/drivers/scsi/qla2xxx/qla_init.c
index d0a0dc9d6d36fcb102adc46a2ce5eb3af71850a0..f948e1a73aeca69bcb2a418db87ab4aa84327eab 100644 (file)
--- a/drivers/scsi/qla2xxx/qla_init.c
+++ b/drivers/scsi/qla2xxx/qla_init.c
@@ -2024,6 +2024,7 @@ qla2x00_configure_hba(scsi_qla_host_t *vha)
            &loop_id, &al_pa, &area, &domain, &topo, &sw_cap);
        if (rval != QLA_SUCCESS) {
                if (LOOP_TRANSITION(vha) || atomic_read(&ha->loop_down_timer) ||
+                   IS_QLA8XXX_TYPE(ha) ||
                    (rval == QLA_COMMAND_ERROR && loop_id == 0x7)) {
                        DEBUG2(printk("%s(%ld) Loop is in a transition state\n",
                            __func__, vha->host_no));