scst.h, scst, device handlers: Fix scst_replace_port_info

[bmeagherix]

scst_pass_through_cmd_done can run in interrupt context, and call scst_replace_port_info, which in turn was calling scst_lookup_tg_id. Since scst_lookup_tg_id does a mutex_lock, we should not call it from interrupt context.

The call path was added in e72eae3 and could lead to:

BUG: scheduling while atomic: swapper/14/0/0x00000102

Preemption disabled at:
Call Trace:
[<ffffffffa1d1b146>] ___slab_alloc.constprop.0+0x4d6/0x750
 <IRQ>
 dump_stack_lvl+0x44/0x5c
 __schedule_bug.cold+0x80/0x8d
 __schedule+0x76c/0x860
 schedule+0x5a/0xb0
 schedule_preempt_disabled+0x14/0x30
 __mutex_lock.constprop.0+0x6ce/0x700
 scst_lookup_tg_id+0x19/0xa0 [scst]
 scst_pass_through_cmd_done+0x14f/0x2e0 [scst]
 scsi_end_async+0x44/0x70 [scst]
 __blk_mq_end_request+0x42/0x120
 scsi_end_request+0xdd/0x1c0 [scsi_mod]
 scsi_io_completion+0x56/0x880 [scsi_mod]
 ? scsi_device_unbusy+0x1e/0x90 [scsi_mod]
 blk_complete_reqs+0x3d/0x50
 __do_softirq+0xed/0x2fe
 ? ktime_get+0x38/0xa0
 do_softirq.part.0+0x95/0xd0
 </IRQ>
 <TASK>
 flush_smp_call_function_queue+0x78/0x90
 do_idle+0x177/0x2b0
 cpu_startup_entry+0x26/0x30
 start_secondary+0x130/0x150
 secondary_startup_64_no_verify+0xe5/0xeb
 </TASK>

(BTW, the fix suggested in this PR was influenced by the current location of scst_pass_through_cmd_done & scst_lookup_tg_id.)