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>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Pine.LNX.4.64.0711031711080.10266@blonde.wat.veritas.com>
Date:	Sat, 3 Nov 2007 17:11:55 +0000 (GMT)
From:	Hugh Dickins <hugh@...itas.com>
To:	Christoph Lameter <clameter@....com>
cc:	Linus Torvalds <torvalds@...ux-foundation.org>,
	Andrew Morton <akpm@...ux-foundation.org>,
	linux-kernel@...r.kernel.org
Subject: [PATCH 2/2] slub: fix Objects count

The count of active Objects shown by Slub's slabinfo is too approximate,
because each cpu slab is counted as all in use, even when lots are free.
That makes tracing leaks harder than it need be.

Add a free count into kmem_cache_cpu (which doesn't enlarge it on 64-bit),
to keep that count in the hot and dirty per-cpu cacheline.

Signed-off-by: Hugh Dickins <hugh@...itas.com>
---
I recommend this for 2.6.24-rc2, and 2.6.23-stable.  But perhaps you've
intentionally avoided this little extra overhead (which isn't necessary
for any functional correctness).  I do think it helps to see accurate
numbers there - which Slab gives, doesn't it?

I believe I've seen a futures patch go by in which you fit something
else into kmem_cache_cpu here; then another in which you remove more.
Not sure where you stand now!

 include/linux/slub_def.h |    1 +
 mm/slub.c                |    9 +++++++--
 2 files changed, 8 insertions(+), 2 deletions(-)

--- 2.6.24-rc1+/include/linux/slub_def.h	2007-10-24 07:16:02.000000000 +0100
+++ linux/include/linux/slub_def.h	2007-11-03 13:22:48.000000000 +0000
@@ -14,6 +14,7 @@
 struct kmem_cache_cpu {
 	void **freelist;
 	struct page *page;
+	unsigned int free;
 	int node;
 	unsigned int offset;
 	unsigned int objsize;
--- 2.6.24-rc1+/mm/slub.c	2007-11-03 13:22:31.000000000 +0000
+++ linux/mm/slub.c	2007-11-03 13:22:48.000000000 +0000
@@ -1392,6 +1392,7 @@ static void deactivate_slab(struct kmem_
 		page->freelist = object;
 		page->inuse--;
 	}
+	c->free = 0;
 	c->page = NULL;
 	unfreeze_slab(s, page);
 }
@@ -1483,8 +1484,8 @@ load_freelist:
 	if (unlikely(SlabDebug(c->page)))
 		goto debug;
 
-	object = c->page->freelist;
 	c->freelist = object[c->offset];
+	c->free = s->objects - c->page->inuse - 1;
 	c->page->inuse = s->objects;
 	c->page->freelist = NULL;
 	c->node = page_to_nid(c->page);
@@ -1528,6 +1529,7 @@ new_slab:
 				if (c->freelist) {
 					object = c->freelist;
 					c->freelist = object[c->offset];
+					c->free--;
 					return object;
 				}
 				slab_lock(c->page);
@@ -1580,6 +1582,7 @@ static void __always_inline *slab_alloc(
 	else {
 		object = c->freelist;
 		c->freelist = object[c->offset];
+		c->free--;
 	}
 	local_irq_restore(flags);
 
@@ -1685,6 +1688,7 @@ static void __always_inline slab_free(st
 	if (likely(page == c->page && c->node >= 0)) {
 		object[c->offset] = c->freelist;
 		c->freelist = object;
+		c->free++;
 	} else
 		__slab_free(s, page, x, addr, c->offset);
 
@@ -1866,6 +1870,7 @@ static void init_kmem_cache_cpu(struct k
 {
 	c->page = NULL;
 	c->freelist = NULL;
+	c->free = 0;
 	c->node = 0;
 	c->offset = s->offset / sizeof(void *);
 	c->objsize = s->objsize;
@@ -3534,7 +3539,7 @@ static unsigned long slab_objects(struct
 				int x = 0;
 
 				if (flags & SO_OBJECTS)
-					x = page->inuse;
+					x = page->inuse - c->free;
 				else
 					x = 1;
 				total += x;
-
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/

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ