Author: alc
Date: Fri Apr 17 15:35:16 2009
New Revision: 191204
URL: http://svn.freebsd.org/changeset/base/191204
Log:
MFC r175055
Defer setting either PG_CACHED or PG_FREE until after the free page
queues lock is acquired. Otherwise, the state of a reservation's
pages' flags and its population count can be inconsistent. That could
result in a page being freed twice.
(This change should have been included in the MFC of the superpages
support.)
Thanks to: pho for an illuminating crash
Approved by: re (kib)
Modified:
stable/7/sys/ (props changed)
stable/7/sys/contrib/pf/ (props changed)
stable/7/sys/dev/ath/ath_hal/ (props changed)
stable/7/sys/dev/cxgb/ (props changed)
stable/7/sys/vm/vm_page.c
Modified: stable/7/sys/vm/vm_page.c
==============================================================================
--- stable/7/sys/vm/vm_page.c Fri Apr 17 14:58:02 2009 (r191203)
+++ stable/7/sys/vm/vm_page.c Fri Apr 17 15:35:16 2009 (r191204)
@@ -1402,8 +1402,8 @@ vm_page_free_toq(vm_page_t m)
m->flags &= ~PG_ZERO;
vm_page_enqueue(PQ_HOLD, m);
} else {
- m->flags |= PG_FREE;
mtx_lock(&vm_page_queue_free_mtx);
+ m->flags |= PG_FREE;
cnt.v_free_count++;
#if VM_NRESERVLEVEL > 0
if (!vm_reserv_free_page(m))
@@ -1654,9 +1654,9 @@ vm_page_cache(vm_page_t m)
* Insert the page into the object's collection of cached pages
* and the physical memory allocator's cache/free page queues.
*/
- vm_page_flag_set(m, PG_CACHED);
vm_page_flag_clear(m, PG_ZERO);
mtx_lock(&vm_page_queue_free_mtx);
+ m->flags |= PG_CACHED;
cnt.v_cache_count++;
root = object->cache;
if (root == NULL) {
_______________________________________________
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
