scsi: zfcp: Trace when request remove fails after qdio send fails

When we fail to send a FSF request in 'zfcp_fsf_req_send()' when calling
'zfcp_qdio_send()' we try to remove the request object from our internal
hash table again to prevent keeping a stale memory reference. This removal
might still - very much theoretically - fail.

To store some evidence of when this happens add a new trace record for
this case; tag: 'fsrsrmf'.

We reuse the 'ZFCP_DBF_HBA_RES' trace ID for this, but mark all fields
other then the request ID with ~0, to make fairly obvious that these are
invalid values. This faking has to be done because we don't have a valid
request object at this point, and can not safely access any of the memory
of the old object - we just failed to find it in our hash table, so it
might be gone already.

Here is an example of a decoded trace record:

    Timestamp      : 2023-02-17-13:09:12:748140
    Area           : HBA
    Subarea        : 1
    Level          : -
    Exception      : 000003ff7ff500c2
    CPU ID         : 0011
    Caller         : 0x0
    Record ID      : 1
    Tag            : fsrsrmf
    Request ID     : 0x0000000080126ab6
    Request status : 0xffffffff
    FSF cmnd       : 0xffffffff
    FSF sequence no: 0xffffffff
    FSF issued     : 2042-09-18-01:53:47:370495
    FSF stat       : 0xffffffff
    FSF stat qual  : ffffffff ffffffff ffffffff ffffffff
    Prot stat      : 0xffffffff
    Prot stat qual : ffffffff ffffffff ffffffff ffffffff
    Port handle    : 0xffffffff
    LUN handle     : 0xffffffff

This provides at least some basic evidence that this event happened, and
what object was affected.

Link: https://lore.kernel.org/r/99b8246b2d71b63fa4f9c56333e2037502f7f5af.1677000450.git.bblock@linux.ibm.com


Signed-off-by: Benjamin Block <bblock@linux.ibm.com>
Reviewed-by: Steffen Maier <maier@linux.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>