- Resolves: rhbz#731460
- Resolves: rhbz#739083
- Resolves: rhbz#739086

- Add upstream revision 2178 - Fix problem where openais will segfault if there are gaps in recovery queue
- Add upstream revision 2179 - whitetank ipc: handle EINTR in msg_send
- Add upstream revision 2180 - whitetank ipc: Don't deadlock in ipc_disconnect
- Add upstream revision 2181 - Ignore memb_join messages during flush operations
- Add upstream revision 2183 - Deliver all messages from my_high_seq_recieved to the last gap

- Resolves: rhbz#638609
- Add upstream revision 2159 - fix segfault that can happen without null
  terminated checkpoint names.

- Resolves: rhbz#60107
- Add upstream revision 2144 - don't reset last_aru when token is retransmitted
- Add upstream revision 2152 - fix merge error with 2144

- Reverts: rhbz#601077
- Remove upstream revision 2144 - don't reset last_aru when token is
  retransmitted - patch was incomplete

- Resolves: rhbz#588897
- Resolves: rhbz#588895
- Resolves: rhbz#588628
- Resolves: rhbz#588627
- Add upstream revision 2130 - Resolve problem where ring id is not saved
  and restored when recovery is interrupted.
- Add upstream revision 2129 - Resolve problem where highly lossy networks
  will sigabort (assert) if 512 or more entries are originated in recovery.
- Add upstream revision 2128 - Use proper typing for retransmit token storage
  area.
- Add upstream revision 2127 - Fix stack corruption that happens because of
  invalid length in memmove operation on highly lossy networks.