1 2005-06-02 Cluster File Systems, Inc. <info@clusterfs.com>
6 Frequency : occasional (large-scale events, cluster reboot, network failure)
8 Description: too many error messages on console obscure actual problem and
9 can slow down/panic server, or cause recovery to fail repeatedly
10 Details : enable rate-limiting of console error messages, and some messages
11 that were console errors now only go to the kernel log
13 Severity : enhancement
15 Description: add /proc/sys/portals/catastrophe entry which will report if
16 that node has previously LBUGged
18 2005-04-06 Cluster File Systems, Inc. <info@clusterfs.com>
20 - update gmnal to use PTL_MTU, fix module refcounting (b=5786)
22 2005-04-04 Cluster File Systems, Inc. <info@clusterfs.com>
24 - handle error return code in kranal_check_fma_rx() (5915,6054)
26 2005-02-04 Cluster File Systems, Inc. <info@clusterfs.com>
28 - update vibnal (Voltaire IB NAL)
29 - update gmnal (Myrinet NAL), gmnalid
31 2005-02-04 Eric Barton <eeb@bartonsoftware.com>
33 * Landed portals:b_port_step as follows...
35 - removed CFS_DECL_SPIN*
36 just use 'spinlock_t' and initialise with spin_lock_init()
38 - removed CFS_DECL_MUTEX*
39 just use 'struct semaphore' and initialise with init_mutex()
41 - removed CFS_DECL_RWSEM*
42 just use 'struct rw_semaphore' and initialise with init_rwsem()
44 - renamed cfs_sleep_chan -> cfs_waitq
45 cfs_sleep_link -> cfs_waitlink
47 - fixed race in linux version of arch-independent socknal
48 (the ENOMEM/EAGAIN decision).
50 - Didn't fix problems in Darwin version of arch-independent socknal
51 (resetting socket callbacks, eager ack hack, ENOMEM/EAGAIN decision)
53 - removed libcfs types from non-socknal header files (only some types
54 in the header files had been changed; the .c files hadn't been