PROBLEM:  (SHGBS5101 & QAR 59405)    (Patch ID: TCR150-016)
********
This patch fixes a problem that can cause a cluster member to panic in
rcv_deqlk_msg() with the panic string set to:

        dlm_panic

A typical stack trace would look like:

  0 stop_secondary_cpu()
  1 panic("event_timeout: panic request")
  2 event_timeout()
  3 xcpu_puts()
  4 printf()
  5 panic("dlm_panic")
  6 dlm_panic()
  7 rcv_deqlk_msg()
  8 dlm_process_message()
  9 dlm_recv()

PROBLEM:  (None.)    (Patch ID: TCR150-022)
********
This patch provides performance enhancements that are required by Oracle
v8.0.5.

PROBLEM:  (MGO103549, QAR 62607)    (Patch ID: TCR150-025)
********
This patch fixes a system panic with the following message:

	snd_grantlk_msg: no memory for message

PROBLEM:  (MGO103933)    (Patch ID: TCR150-047)
********
This patch fixes a problem when a system in a TruCluster panics with the
following data:

panic string: "dlm_panic"

A typical stack trace would look like the following:

         5 panic
         6 dlm_panic
         7 deq_granted
         8 deq_one_lock
         9 deq_lock
        10 dlm_unlk
        11 syscall
        12 _Xsyscall
PROBLEM:  (QAR 58803)   (Patch ID: TCR150-006)
********
This patch fixes a problem in TruCluster Production Server Software that can
occur during multiple node failures.  The distributed lock manager (DLM) may
issue the following panic under rare conditions.

        dlm getch: illegal csid

The conditions required are that two nodes must fail during the period of time
that the DLM is rebuilding its lock database and that the connection manager
is starting its update operation.

One possible stack trace would be:

   5 panic("dlm getch: illegal csid")
   6 cnx_getch()
   7 snd_grantlk_msg()
   8 grant_lock()
   9 grant_one_cvt()
  10 grant_cvts()
  11 convert_one()
  12 convert_lock()
  13 dlm_convert()
  14 syscall()

All stack traces would include the call to cnx_getch.

PROBLEM:  (MGO103894)    (Patch ID: TCR150-041)
********
This patch fixes a deadlock condition between the DLM rebuild thread and the
Connection Manager ping daemon (cnxpingd).  The deadlock can cause users of
DLM (e.g. Oracle) to hang.

PROBLEM:  (QAR 70143, BCGM60LQH)    (Patch ID: TCR150-059)
********
This patch corrects a problem in the TruCluster DLM subsystem in which
a node can panic with the panic string "convert_lock: bad lock state"
and the stack trace as follows:

   0 stop_secondary_cpu()
   1 panic()
   2 event_timeout(func = 0xfffffc000028ee80, arg = 0xfffffc00008ad630)
  3 xcpu_puts()
   4 printf()
   5 panic()
   6 convert_lock()
   7 dlm_convert()
   8 syscall()
   9 _Xsyscall()

PROBLEM:  (DEKB70763)    (Patch ID: TCR150-074)
********

This patch corrects a problem in the Distributed Lock Manager (DLM) in the
TruCluster Production Server product in which an error in the session
layer on one member can result in sending a corrupted DLM message to the
receiving member. The corruption in the packets will be random, however if
the dlm_checksum_msg sysconfigtab parameter is set to 1, the receiving
member will consistently panic with a panic string of
"dlm_process_message: corrupt message" when a corrupt DLM message is received.

Some other panic strings that were seen are:

1) "rcv_cvt: wrong lock"
2) "rcv_resp: manage lock: bad lock state"
3) "rcv_cvt: lk_txid mismatch"
4) "rcv_resp: bad response status"

These indicate different corruptions in a DLM message.