be2iscsi: Replace _bh with _irqsave/irqrestore

be2iscsi: Replace _bh with _irqsave/irqrestore

Kernel panic - not syncing: Watchdog detected hard LOCKUP on cpu 21
Pid: 13242, comm: flush-8:80 Tainted: G        W  -- ------------    2.6.32-573.el6.x86_64 #1
Call Trace:
<NMI>  [<ffffffff81537a84>] ? panic+0xa7/0x16f
[<ffffffff810149c9>] ? sched_clock+0x9/0x10
[<ffffffff810ed4bd>] ? watchdog_overflow_callback+0xcd/0xd0
[<ffffffff81124037>] ? __perf_event_overflow+0xa7/0x240
[<ffffffff8101dc54>] ? x86_perf_event_set_period+0xf4/0x180
[<ffffffff81124684>] ? perf_event_overflow+0x14/0x20
[<ffffffff81024a02>] ? intel_pmu_handle_irq+0x202/0x3f0
[<ffffffff8153cd89>] ? perf_event_nmi_handler+0x39/0xb0
[<ffffffff8153e845>] ? notifier_call_chain+0x55/0x80
[<ffffffff81389ba0>] ? scsi_done+0x0/0x60
[<ffffffff8153e8aa>] ? atomic_notifier_call_chain+0x1a/0x20
[<ffffffff810a788e>] ? notify_die+0x2e/0x30
[<ffffffff8153c503>] ? do_nmi+0x1c3/0x350
[<ffffffff8153bdc0>] ? nmi+0x20/0x30
[<ffffffff81389ba0>] ? scsi_done+0x0/0x60
[<ffffffff81389ba0>] ? scsi_done+0x0/0x60
[<ffffffff8153b62e>] ? _spin_lock+0x1e/0x30
<<EOE>>  [<ffffffffa00e2eaf>] ? iscsi_queuecommand+0x7f/0x4e0 [libiscsi]
[<ffffffff81389df5>] ? scsi_dispatch_cmd+0xe5/0x310
[<ffffffff813927be>] ? scsi_request_fn+0x5be/0x750
[<ffffffff81089bad>] ? del_timer+0x7d/0xe0
[<ffffffff81273542>] ? __generic_unplug_device+0x32/0x40
[<ffffffff8126e823>] ? elv_insert+0xd3/0x190
[<ffffffff8126e920>] ? __elv_add_request+0x40/0x90

In beiscsi_alloc_pdu, _bh versions of spin_lock are being used for
protecting SGLs and WRBs. _bh versions are needed as the function gets
invoked in process context and BLOCK_IOPOLL softirq.

In spin_unlock_bh, after releasing the lock and enabling BH, do_softirq
is called which executes till last SOFTIRQ.

beiscsi_alloc_pdu is called under session lock. Through block layer,
iSCSI stack in some cases send IOs with interrupts disabled. In such paths,
CPU will get stuck for a while for session lock with interrupts disabled
because in other CPU do_softirq is executing under session lock thus
causing hard lock up.

Use spin_lock_irqsave/spin_lock_irqrestore as the driver can't be sure
in all paths interrupts are enabled.

Signed-off-by: Jitendra Bhivare <jitendra.bhivare@broadcom.com>
Orabug: 23712824
Signed-off-by: Manjunath Govindashetty <manjunath.govindashetty@oracle.com>