linux-kernel - [PATCH] Vmscan.c: Account for memory already freed in seeking to free more.

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Date:	Tue, 12 Dec 2006 21:54:41 +1100
From:	Nigel Cunningham <nigel@...pend2.net>
To:	LKML <linux-kernel@...r.kernel.org>
Subject: [PATCH] Vmscan.c: Account for memory already freed in seeking to
	free more.

The current versions of shrink_all_zones and shrink_all_memory don't
take account of memory already freed when making multiple calls to seek
to free memory. As a result, we can end up freeing far more memory than
was asked for. This can in turn result in more (unnecessary) paging
if/when the data is later needed.

These modifications seek to alleviate this situation by modifying
swap_cluster_max by the number of pages freed by shrink_inactive_list in
shrink_all_zones before proceeding to the next zone, and in
shrink_all_memory before shrinking slab and going to the next priority.

Signed-off-by: Nigel Cunningham <nigel@...pend2.net> 

diff -ruNp 930-vmscan.patch-old/mm/vmscan.c 930-vmscan.patch-new/mm/vmscan.c
@@ -1469,9 +1469,12 @@ static unsigned long shrink_all_zones(un
 
 		zone->nr_scan_inactive += (zone->nr_inactive >> prio) + 1;
 		if (zone->nr_scan_inactive >= nr_pages || pass > 3) {
+			int freed;
 			zone->nr_scan_inactive = 0;
 			nr_to_scan = min(nr_pages, zone->nr_inactive);
-			ret += shrink_inactive_list(nr_to_scan, zone, sc);
+			freed = shrink_inactive_list(nr_to_scan, zone, sc);
+			ret += freed;
+			sc->swap_cluster_max -= freed;
 			if (ret >= nr_pages)
 				return ret;
 		}
@@ -1550,9 +1553,14 @@ unsigned long shrink_all_memory(unsigned
 
 		for (prio = DEF_PRIORITY; prio >= 0; prio--) {
 			unsigned long nr_to_scan = nr_pages - ret;
+			int freed;
 
 			sc.nr_scanned = 0;
-			ret += shrink_all_zones(nr_to_scan, prio, pass, &sc);
+			sc.swap_cluster_max = nr_pages - ret;
+			freed = shrink_all_zones(nr_to_scan, prio, pass, &sc);
+			ret += freed;
+			lru_pages =- freed;
+			nr_to_scan = nr_pages - ret;
 			if (ret >= nr_pages)
 				goto out;
 


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/