PROBLEM: (CLD MGO103853, CLD MGO103295, CLD MGO103959) (Patch ID: OSF440-005)
********
This patch fixes a kmf for a bucket 2 (64 byte corruptor) problem when the
type of SCSI device dynamically changes.

A typical stack trace:
       panic
       trap
       _XentMM
       malloc(size = 64, bucket 2)
       falloc
       accept1
       oaccept
       syscall
       _Xsyscall

PROBLEM:  (QAR 74308 QAR 73808)    (Patch ID: OSF440-224)
********

The wrong EEI status is returned when a tape device reports
ASC/ ASCQ 0x2903 after Bus D has been detected.  The correct EEI status
for the this condition is EEI_DEVPATH_RESET; the status being returned
is EEI_DEVHARD_FAILURE.

PROBLEM:  (74839)    (Patch ID: OSF440-238)
********

This patch increases the outstanding i/o tag queue depth for an HSG80
unit from 5 to 25. This allows more i/o to be queued in the HSG80. With large
numbers of random i/o access, the HSG80 can better optimize the i/o requests,
resulting in better performance. Due to the likelyhood of the queue on
the g80 becoming congested, the scsi disk driver's queue full backoff
algorithm has been enhanced.

To determine the current queue depth for the HSG80, issue the following
command "ddr_config -s disk DEC HSG80" and look at the value of the
"TagQueueDepth" attribute. Prior to the patch, this would be 0x5. With
the patch, this will be 0x19.

PROBLEM:  (QAR 73967 QAR 73853)    (Patch ID: OSF440-255)
********

This patch fixes a problem in which the system can crash with
a kernel memory fault and the following stack trace:

  10 panic(s =  "kernel memory fault")
  11 trap()
  12 _XentMM)
  13 strncpy()
  14 ccmn_errlog3()
  15 cdisk_devt_to_trans()
  16 cdisk_io_barrier()
  17 drd_do_iob()
  18 drd_iob_thread()

PROBLEM:  (BCGM10J56, BCGM4180B, BCGM41HLM, MGO36744A, 7XRB90123)
          (PATCH ID: OSF440-319)
********

This patch fixes the following 3  problems:

1.)When  a SCSI I/O command results in a bus reset, which
causes the command to be re-issued, the system sometimes
can not recover and ends up in a tight loop of bus resets.

2.) After a recovery operation is performed, the read
disk capacity sizing of the disk sometime fails to match
the original value, which causes the recovery to fail.

3.) A BBR (bad block replacement) informational message
received from an HSZ is mistakenly classified as an error.

PROBLEM:  (QAR 76145) (PATCH ID:  OSF440-298)
********

This patch fixes a problem where programs that read, analyze and monitor
disk statistics (such as "collect") will occasionally display incorrect
results.  This is especially seen when calculating quantities related to the
percentage of time a disk is "busy" servicing I/O requests.

PROBLEM:  (71242, 73860) (PATCH ID: OSF440-308)
********


This patch fixes a problem in which the system can panic with a
kernel memory fault during an installation with an
HSZ70 or HSZ80 connected to the system.


PROBLEM:  (MGO27465A) (PATCH ID: OSF440-406)
********


This patch fixes a problem when the type of SCSI device dynamically changes,
which can result in a kernel memory fault or memory corruption panic.  The
sense inquiry buffer size is allocated too small.


A typical stack trace:
>  0 stop_secondary_cpu   src/kernel/arch/alpha/cpu.c : 608
   1 panic                src/kernel/bsd/subr_prf.c : 751
   2 event_timeout        src/kernel/arch/alpha/cpu.c : 1183
   3 xcpu_puts            src/kernel/bsd/subr_prf.c : 895
   4 printf               src/kernel/bsd/subr_prf.c : 423
   5 panic                src/kernel/bsd/subr_prf.c : 804
   6 closef               src/kernel/bsd/kern_descrip.c : 1973
   7 exit                 src/kernel/bsd/kern_exit.c : 1021
   8 rexit                src/kernel/bsd/kern_exit.c : 822
   9 syscall              src/kernel/arch/alpha/syscall_trap.c : 627
  10 _Xsyscall            src/kernel/arch/alpha/locore.s : 1505

Panic "closef: f_count not 1" due to corruption of 'file' structure