PROBLEM:  (CLD TKTR42550)    (Patch ID: OSF440-118)
********
Upon booting the following error message will be seen in the console log
mchan0 - mchan_reset failed - mcerr = _0x40001, retry_count = 10
mchan0 - mchan_reset failed - mcerr = _0x40001, retry_count =  9
                                                .
                                                .
                                                .
mchan0 - mchan_reset failed - mcerr = _0x40001, retry_count =  1
mchan0: failure in resetting _controller - Failed to configure device in
PCI _bus 0, slot 5 with driver mchan

PROBLEM:  (HPAQB15SF, HPAQA07WV, HPAQC1MMC, MGO78908A, VNO28218A, MGO49488A,
           ZPO036959) (PATCH ID: OSF440-403)
********


This patch fixes a problem in the memory channel driver which could
result in panics with rm - inconsistent local spinlock structures
being logged.

PROBLEM: (BCSMC0SLW, BCGM70KGG, BCGM10MCB, TKTBC0076,
          77220, 75146) (PATCH ID: OSF440-572)
********
This patch fixes an incorrect heartbeat timer within the memory channel driver
which caused rail failures to be incorrectly reported on memory channel version
2 cards.  With the heartbeat timer set too short, the system can be erroneously
led to believe a hardware failure has occurred.  Messages of the form
"rmerror_int: ..." are output to the messages file containing an error_type
which has bit 29 set in error_type (heartbeat timeout).  The binary errorlog
will also have this data.  Typically, the error_type data will be 0xe00000000.
The messages are followed by the system hanging or panicing.  When panicing,
the following message is produced:

panic (cpu 0): rm_failover_if_necessary, both rails bad

A real hardware failure produces the same symptoms and stack trace.  For
example, having an error_type of 0xe00000002 indicates a write transmit
hardware fatal failure.