[SSI] openssi ChangeLog,1.243,1.244
Roger Tsang <[email protected]>
| Newsgroups | gmane.linux.cluster.ssic.cvs |
|---|---|
| Message-ID | <[email protected]> |
Update of /cvsroot/ssic-linux/openssi
In directory sfp-cvsdas-3.v30.ch3.sourceforge.com:/tmp/cvs-serv29028
Modified Files:
Tag: OPENSSI-FC
ChangeLog
Log Message:
update
Index: ChangeLog
===================================================================
RCS file: /cvsroot/ssic-linux/openssi/ChangeLog,v
retrieving revision 1.243
retrieving revision 1.244
diff -u -d -r1.243 -r1.244
--- ChangeLog 18 Nov 2009 06:19:40 -0000 1.243
+++ ChangeLog 17 Dec 2009 06:05:08 -0000 1.244
@@ -1,4 +1,67 @@
-2009/??/?? *** Release of OpenSSI 1.9.6 ***
+2009/12/?? *** Release of OpenSSI 1.9.6 ***
+
+2009/12/17 rogertsang <[email protected]>
+ * CFS:
+ - Re-enable lockless sct_flags. (#ifdef SVRCFSTOK_SCT_FLAGS_NOLOCK)
+
+ * VPROC:
+ - Fix select_bad_process() NULL p_vproc pointer dereference.
+ - Fix de_thread() related issues:
+ - Fix de_thread() calling VPOP_RECLAIM_CHILD() while holding
+ spin_lock(task->proc_lock). Code path might sleep.
+ - Fix NULL p_vproc pointer dereference when thread group leader is
+ exiting, has PF_FORKNOEXEC flag, and has lost race with
+ de_thread(). Affects exit_notify(), forget_original_parent(), and
+ will_become_orphaned_pgrp(). When p_vproc is NULL the exiting
+ process reaps itself.
+ - Fix dpvproc_nocldwait_async_handler() NULL pointer dereference due
+ to failed lookup for vproc of new parent when lost race with
+ de_thread(). (#ifdef RELEASE_TASK_LEADER_GONE__VPROC_FIX)
+ - New PVPOP_SWITCH_PID() to switch VProc PID of process doing
+ de_thread().
+ - New PV_DETHREAD_REAP flag is set during de_thread() to indicate to
+ child reaper to reap thread group leader without delay.
+ - Fix oops in do_notify_resume() path due to NULL pointer
+ dereference. p_vproc is NULL after bad fork.
+
+ * VPROC (#ifdef VPROC_RELEASE__REFCNT_RACE_FIX):
+ - Fix extra VPROC_RELE() on PVPOP_RMV_PGRP_LIST() -ESRCH error.
+ - Fix PVPOP_REAP() extra dpvproc_remove_pgrp_leader_member_care() on
+ PVPOP_RMV_PGRP_LIST() -EBUSY error; was typo in if-test.
+ - PVPOP_REAP() no longer does async PVPOP_RMV_PGRP_LIST(); now sync.
+ Don't see how reaping parent can hold PGRP LOCK in PVPOP_REAP() path.
+ - Fix exit_vproc() did not decrement vproc reference for process group
+ leader on member node. This is in do_fork() error path.
+
+ * VPROC (#ifdef VPROC_CONTINUE):
+ - Fix dpvproc_nocldwait_async_handler() deadlock on PVPOP_REAP() -EBUSY
+ due to delay_group_leader(). PV_REAPER_RETRY flag is set and group
+ leader is re-inserted into the queue so that the remaining threads in
+ thread group can be reaped by this handler to remove the
+ delay_group_leader() condition.
+ - Regression:
+ - Fix pproc_reap() should succeed, but returns -EBUSY when process is
+ a thread group leader in non-empty thread group and is already in
+ EXIT_DEAD state.
+
+ * Files in ci-linux project were modified (see ci/ChangeLog).
+
+ cluster/ssi/cfs/svrcfs.c | 4
+ cluster/ssi/vproc/dvp_async.c | 52 +++++++-
+ cluster/ssi/vproc/dvp_pvpops.c | 197 +++++++++++++++++++++++++++++--
+ cluster/ssi/vproc/dvp_vpops.c | 56 ++++++--
+ cluster/ssi/vproc/nd_carelist.c | 10 +
+ cluster/ssi/vproc/nd_cleanup.c | 4
+ cluster/ssi/vproc/vp_subr.c | 2
+ fs/exec.c | 67 ++++++++--
+ include/cluster/ssi/cfs/cfstok.h | 8 -
+ include/linux/config.h | 11 -
+ include/linux/dpvproc.h | 15 ++
+ include/linux/vproc.h | 4
+ kernel/exit.c | 81 ++++++++++--
+ kernel/pid.c | 7 -
+ mm/oom_kill.c | 3
+ 15 files changed, 454 insertions(+), 67 deletions(-)
2009/11/18 rogertsang <[email protected]>
* CFS (#ifdef CFS_PAGES_CACHE):
------------------------------------------------------------------------------
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev