[SSI] openssi ChangeLog,1.251,1.252

Update of /cvsroot/ssic-linux/openssi
In directory sfp-cvsdas-3.v30.ch3.sourceforge.com:/tmp/cvs-serv31032

Modified Files:
      Tag: OPENSSI-FC
	ChangeLog 
Log Message:
update


Index: ChangeLog
===================================================================
RCS file: /cvsroot/ssic-linux/openssi/ChangeLog,v
retrieving revision 1.251
retrieving revision 1.252
diff -u -d -r1.251 -r1.252
--- ChangeLog	2 Apr 2010 20:58:21 -0000	1.251
+++ ChangeLog	3 Apr 2010 19:48:15 -0000	1.252
@@ -1,5 +1,82 @@
 2010/??/??	*** Release of OpenSSI 2.0.0rc1 ***
 
+2010/04/03	rogertsang <[email protected]>
+	* CVS Tags: OPENSSI-FC-1-9-6-PRE34
+
+	* IPC SHM:
+	- sys_shmctl()
+	  - Move SSI code into new functions shm_find_svr_node() and
+	    ssi_shmctl_info().
+	  - Fix processing bogus data for SHM_INFO command on ssi_shmctl()
+	    error.
+	- ssi_shmctl()
+	  - Fix bogus return value on RPC error.
+	  - Handle ICS -EAGAIN error.
+	- do_ssi_shm_noclients() handle RIPC_SHM_NOCLIENTS() -EAGAIN error.
+
+	 cluster/ssi/ipc/ipcshm_svr.c |   28 ++--
+	 ipc/shm.c                    |  199 +++++++++++++++++------------------
+	 2 files changed, 111 insertions(+), 116 deletions(-)
+
+2010/04/03	rogertsang <[email protected]>
+	* CVS Tags: OPENSSI-FC-1-9-6-PRE33
+
+	* IPC Semaphores:
+	- sys_semctl()
+	  - Move SSI code into ssi_semctl().
+	  - Handle cli_ripc_semctl() error -EAGAIN.
+	  - Fix semctl_nolock() deadlock with destination node when both nodes
+	    are doing RPC for SEM_INFO command.
+	- semctl_nolock()
+	  - Move SSI code to ssi_semctl_nolock_info().
+	  - Handle cli_ripc_semctl() error -EAGAIN.
+	  - Fix race with sys_semctl() over global variable "remote_cmd".
+	- Fix alloc_semundo() sem_semundo struct memory leak; assigning
+	  sem_semundo struct to sem_array that is going away when lost race
+	  with ipc_rmid().
+	- Remove redundant in ssi_semop() which is mostly taken from original
+	  sys_semtimedop().
+	- Fix ssi_semop() dereferencing invalid sem_semundo struct if lost race
+	  with ssi_semexit() though this is unlikely to occur because in
+	  sem_exit() the last thread removes the the sem_semundo struct.
+	- Enable debug message in freeundos().
+	- sem_exit() handle cli_ripc_semexit() error -EAGAIN.
+	- Remove ugly goto's in __ssi_semexit().
+
+	 ipc/sem.c |  604 +++++++++++++++++++++++-------------------------------
+	 1 files changed, 268 insertions(+), 336 deletions(-)
+
+2010/04/03	rogertsang <[email protected]>
+	* CVS Tags: OPENSSI-FC-1-9-6-PRE32
+
+	* IPC Messages:
+	- Fix msg_get_svr_node() race with msg_set_svr_node() as not all code
+	  paths hold msg_ids.sem mutex. Use msg_nodehint_lock read/write
+	  spinlock.
+	- Consolidate redundant code into new function msg_find_svr_node().
+	- Remove redundant in ssi_msgsnd(). Most of the code is from
+	  sys_msgsnd().
+	- Fix sys_msgsnd()
+	  - Unhandled ssi_msgpack() error -ENOMEM.
+	  - Dereferencing freed "packedsegs" object.
+	  - Handle cli_ripc_msgsnd() error -EAGAIN.
+	- Remove redundant in ssi_msgrcv(). Most of the code is from
+	  sys_msgrcv().
+	- Fix sys_msgctl(MSG_INFO)
+	  - Might populate user buffer with bogus info on RPC error -EREMOTE.
+	  - Did not account for remote nodes when calculating max_id.
+	  - Deadlock with destination node when both nodes are doing RPC.
+	- Move remote code paths to new functions ssi_remote_msgctl(),
+	  ssi_msgctl_info(), ssi_remote_msgrcv(), ssi_remote_msgsnd().
+	- sys_msgctl() handle cli_ripc_msgctl() error -EAGAIN.
+	- sys_msgrcv() handle cli_ripc_msgrcv() error -EAGAIN.
+
+	 cluster/ssi/ipc/ipcmsg_svr.c |    8 
+	 include/linux/msg.h          |    3 
+	 ipc/msg.c                    |  762 +++++++++++++++--------------------
+	 ipc/msgutil.c                |   13 
+	 4 files changed, 356 insertions(+), 430 deletions(-)
+
 2010/04/02	rogertsang <[email protected]>
 	* CVS Tags: OPENSSI-FC-1-9-6-PRE31, OPENSSI-FC-2-0-0-PRE3
 


------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev