[SCSI] pm8001: Fix possible racing condition.

There is a possble racing scenario.

'process_oq' is called by two routines, as shown below.

pm8001_8001_dispatch = {
         .......

        .isr             = pm8001_chip_isr --> process_oq,// A
        .isr_process_oq  = process_oq,                   //  B
        .....
}

process_oq() --> process_one_iomb() --> mpi_sata_completion()

In 'mpi_sata_completion', "pm8001_ha->lock" is first released.
It means lock is taken before,  which is true for
the context A, as 'pm8001_ha->lock' is taken in 'pm8001_chip_isr()'

But for context B there is no lock taken before and pm8001_ha->lock
is unlocked in 'mpi_sata_completion()'. This may unlock the lock
taken in context A. Possible racing ??

If 'pm8001_ha->lock' is taken in 'process_oq()' instead of
'pm8001_chip_isr' then the above issue can be avoided.

Signed-off-by: Santosh Nayak <santoshprasadnayak@gmail.com>
Acked-by: Jack Wang <jack_wang@usish.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>

drivers/scsi/pm8001/pm8001_hwi.c

@@ -3869,7 +3869,9 @@ static int process_oq(struct pm8001_hba_info *pm8001_ha)
 	void *pMsg1 = NULL;
 	u8 bc = 0;
 	u32 ret = MPI_IO_STATUS_FAIL;
+	unsigned long flags;
 
+	spin_lock_irqsave(&pm8001_ha->lock, flags);
 	circularQ = &pm8001_ha->outbnd_q_tbl[0];
 	do {
 		ret = mpi_msg_consume(pm8001_ha, circularQ, &pMsg1, &bc);
@@ -3890,6 +3892,7 @@ static int process_oq(struct pm8001_hba_info *pm8001_ha)
 				break;
 		}
 	} while (1);
+	spin_unlock_irqrestore(&pm8001_ha->lock, flags);
 	return ret;
 }
 
@@ -4295,12 +4298,9 @@ static u32 pm8001_chip_is_our_interupt(struct pm8001_hba_info *pm8001_ha)
 static irqreturn_t
 pm8001_chip_isr(struct pm8001_hba_info *pm8001_ha)
 {
-	unsigned long flags;
-	spin_lock_irqsave(&pm8001_ha->lock, flags);
 	pm8001_chip_interrupt_disable(pm8001_ha);
 	process_oq(pm8001_ha);
 	pm8001_chip_interrupt_enable(pm8001_ha);
-	spin_unlock_irqrestore(&pm8001_ha->lock, flags);
 	return IRQ_HANDLED;
 }