[SSI] openssi ChangeLog,1.254,1.255
Roger Tsang <[email protected]> Sun, 18 Apr 2010 20:48:19 +0000
| 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-serv12614
Modified Files:
Tag: OPENSSI-FC
ChangeLog
Log Message:
update
Index: ChangeLog
===================================================================
RCS file: /cvsroot/ssic-linux/openssi/ChangeLog,v
retrieving revision 1.254
retrieving revision 1.255
diff -u -d -r1.254 -r1.255
--- ChangeLog 5 Apr 2010 05:17:05 -0000 1.254
+++ ChangeLog 18 Apr 2010 20:48:17 -0000 1.255
@@ -1,5 +1,58 @@
2010/??/?? *** Release of OpenSSI 2.0.0rc1 ***
+2010/04/18 rogertsang <[email protected]>
+ * CVS Tags: OPENSSI-FC-1-9-6-PRE38
+
+ * MOSIX (#ifdef SSI_BALANCE_MEMORY):
+ - Use cond_resched_lock() instead of run_over_maxchunk,
+ sort_age_maxchunk. Prevent needless re-traversing of the page table.
+ - run_over_dirty_pages(), sort_and_age_pages() support 4-level page
+ table.
+
+ * VPROC:
+ - as_do_pg() use pud_alloc() like install_arg_page() in Linux base;
+ based on John Hughes' findings.
+ - Fix inappropriate inlining of as_pull_pgdir(), as_count_pgdir().
+ - No need to hold page_table_lock while traversing struct mm->mmap list
+ of VMA's in as_count_mm_info().
+ - Fix array bounds violation due to as_saved_auxv array in as_info
+ struct is smaller than saved_auxv array in mm struct.
+ - Use cond_resched_lock() instead of run_over_maxchunk. Prevent
+ needless re-traversing of the page table.
+ - Fix memory corruption due to do_swap_page() unmapping unmapped PTE in
+ as_pull_pgdir() path. Memory corruption occurs when the PTE passed to
+ do_swap_page() is not the first PTE in the list of PTE's because in
+ as_pull_pgdir() only the first PTE is mapped. Bug only affects
+ kernels with CONFIG_HIGHPTE defined.
+ - Process migration to support 4-level page table.
+
+ cluster/ssi/mosixll/freemem.c | 67 +++++-
+ cluster/ssi/mosixll/service.c | 126 +++++++++---
+ cluster/ssi/vproc/as_xscribe.c | 318 ++++++++++++++++++-------------
+ include/cluster/gen/as_xscribe.x | 2
+ 4 files changed, 340 insertions(+), 173 deletions(-)
+
+2010/04/18 rogertsang <[email protected]>
+ * CVS Tags: OPENSSI-FC-1-9-6-PRE37
+
+ * IPC Semaphores:
+ - Avoid duplicate pid alive checks when there are multiple
+ semundo_nodedown_thread()'s. Last thread to reset
+ sem_semundo->pid_checked. This will prevent the process_is_alive()
+ storm but depending on timing may leave behind some sem_semundo
+ structs until the next node down event or until the shared semaphore
+ is removed.
+ - Fix semundo_nodedown_thread() skipped remaining sem_semundo structs
+ in sem_array after encountering sem_semundo with pid_checked and pid
+ was not dead. Bug introduced in OPENSSI-FC-1-9-6-PRE30.
+ - Remove redundant code in semundo_nodedown_thread(). Call
+ __ssi_semexit() instead.
+ - Move beginning of __ssi_semexit() into ssi_semexit() so that
+ semundo_nodedown_thread() can use __ssi_semexit().
+
+ ipc/sem.c | 137 ++++++++++++++++++++++++++++--------------------------
+ 1 files changed, 72 insertions(+), 65 deletions(-)
+
2010/04/05 rogertsang <[email protected]>
* CVS Tags: OPENSSI-FC-1-9-6-PRE36
------------------------------------------------------------------------------
Download Intel® 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