From: Andi Kleen <ak@muc.de>
Date: Wed, 27 Jul 2005 18:43:50 +0000 (-0700)
Subject: [PATCH] Undo mempolicy shared policy rbtree microoptimization
X-Git-Tag: v2.6.13-rc4~25^2~141
X-Git-Url: https://err.no/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=90c5029e471636f21221bf66b9a46ada2ab79a22;p=linux-2.6

[PATCH] Undo mempolicy shared policy rbtree microoptimization

All mempolicy changes must be inside the spinlock and readding the rb_erase
prevents a crash while doing:

> echo "1" > /tmp/numatest
> numactl --length=0x4000 --shm /tmp/numatest --localalloc
> numactl --length=0x2000 --offset=0 --shm /tmp/numatest --membind=0
> numactl --length=0x2000 --offset=0x2000 --shm /tmp/numatest --membind=1
> ipcs
> ipcrm -M "the_key_value_of_this_shm_area"

Based on a patch by John Blackwood

Cc: <john.blackwood@ccur.com>
Cc: <andrea@suse.de>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
---

diff --git a/mm/mempolicy.c b/mm/mempolicy.c
index cb41c31e7c..1694845526 100644
--- a/mm/mempolicy.c
+++ b/mm/mempolicy.c
@@ -1138,11 +1138,11 @@ void mpol_free_shared_policy(struct shared_policy *p)
 	while (next) {
 		n = rb_entry(next, struct sp_node, nd);
 		next = rb_next(&n->nd);
+		rb_erase(&n->nd, &p->root);
 		mpol_free(n->policy);
 		kmem_cache_free(sn_cache, n);
 	}
 	spin_unlock(&p->lock);
-	p->root = RB_ROOT;
 }
 
 /* assumes fs == KERNEL_DS */