Complete the transition to RCU with RCU safe comment extension handling.
Signed-off-by: Jozsef Kadlecsik <kadlec@xxxxxxxxxxxxxxxxx>
---
include/linux/netfilter/ipset/ip_set.h | 7 ++++-
include/linux/netfilter/ipset/ip_set_comment.h | 38 ++++++++++++++++++--------
net/netfilter/ipset/ip_set_bitmap_gen.h | 19 +++++++++----
net/netfilter/ipset/ip_set_list_set.c | 17 ++++++++----
4 files changed, 57 insertions(+), 24 deletions(-)
diff --git a/include/linux/netfilter/ipset/ip_set.h b/include/linux/netfilter/ipset/ip_set.h
index d7215d5..344616d 100644
--- a/include/linux/netfilter/ipset/ip_set.h
+++ b/include/linux/netfilter/ipset/ip_set.h
@@ -108,8 +108,13 @@ struct ip_set_counter {
atomic64_t packets;
};
+struct ip_set_comment_rcu {
+ struct rcu_head rcu;
+ char str[0];
+};
+
struct ip_set_comment {
- char *str;
+ struct ip_set_comment_rcu __rcu *c;
};
struct ip_set_skbinfo {
diff --git a/include/linux/netfilter/ipset/ip_set_comment.h b/include/linux/netfilter/ipset/ip_set_comment.h
index 21217ea..8d02485 100644
--- a/include/linux/netfilter/ipset/ip_set_comment.h
+++ b/include/linux/netfilter/ipset/ip_set_comment.h
@@ -16,41 +16,57 @@ ip_set_comment_uget(struct nlattr *tb)
return nla_data(tb);
}
+/* Called from uadd only, protected by the set spinlock.
+ * The kadt functions don't use the comment extensions in any way.
+ */
static inline void
ip_set_init_comment(struct ip_set_comment *comment,
const struct ip_set_ext *ext)
{
+ struct ip_set_comment_rcu *c = rcu_dereference_protected(comment->c, 1);
size_t len = ext->comment ? strlen(ext->comment) : 0;
- if (unlikely(comment->str)) {
- kfree(comment->str);
- comment->str = NULL;
+ if (unlikely(c)) {
+ kfree_rcu(c, rcu);
+ rcu_assign_pointer(comment->c, NULL);
}
if (!len)
return;
if (unlikely(len > IPSET_MAX_COMMENT_SIZE))
len = IPSET_MAX_COMMENT_SIZE;
- comment->str = kzalloc(len + 1, GFP_ATOMIC);
- if (unlikely(!comment->str))
+ c = kzalloc(sizeof(*c) + len + 1, GFP_ATOMIC);
+ if (unlikely(!c))
return;
- strlcpy(comment->str, ext->comment, len + 1);
+ strlcpy(c->str, ext->comment, len + 1);
+ rcu_assign_pointer(comment->c, c);
}
+/* Used only when dumping a set, protected by rcu_read_lock_bh() */
static inline int
ip_set_put_comment(struct sk_buff *skb, struct ip_set_comment *comment)
{
- if (!comment->str)
+ struct ip_set_comment_rcu *c = rcu_dereference_bh(comment->c);
+
+ if (!c)
return 0;
- return nla_put_string(skb, IPSET_ATTR_COMMENT, comment->str);
+ return nla_put_string(skb, IPSET_ATTR_COMMENT, c->str);
}
+/* Called from uadd/udel, flush or the garbage collectors protected
+ * by the set spinlock.
+ * Called when the set is destroyed and when there can't be any user
+ * of the set data anymore.
+ */
static inline void
ip_set_comment_free(struct ip_set_comment *comment)
{
- if (unlikely(!comment->str))
+ struct ip_set_comment_rcu *c;
+
+ c = rcu_dereference_protected(comment->c, 1);
+ if (unlikely(!c))
return;
- kfree(comment->str);
- comment->str = NULL;
+ kfree_rcu(c, rcu);
+ rcu_assign_pointer(comment->c, NULL);
}
#endif
diff --git a/net/netfilter/ipset/ip_set_bitmap_gen.h b/net/netfilter/ipset/ip_set_bitmap_gen.h
index 6660f2a..86429f3 100644
--- a/net/netfilter/ipset/ip_set_bitmap_gen.h
+++ b/net/netfilter/ipset/ip_set_bitmap_gen.h
@@ -209,10 +209,13 @@ mtype_list(const struct ip_set *set,
struct nlattr *adt, *nested;
void *x;
u32 id, first = cb->args[IPSET_CB_ARG0];
+ int ret = 0;
adt = ipset_nest_start(skb, IPSET_ATTR_ADT);
if (!adt)
return -EMSGSIZE;
+ /* Extensions may be replaced */
+ rcu_read_lock();
for (; cb->args[IPSET_CB_ARG0] < map->elements;
cb->args[IPSET_CB_ARG0]++) {
id = cb->args[IPSET_CB_ARG0];
@@ -228,9 +231,11 @@ mtype_list(const struct ip_set *set,
if (!nested) {
if (id == first) {
nla_nest_cancel(skb, adt);
- return -EMSGSIZE;
- } else
- goto nla_put_failure;
+ ret = -EMSGSIZE;
+ goto out;
+ }
+
+ goto nla_put_failure;
}
if (mtype_do_list(skb, map, id, set->dsize))
goto nla_put_failure;
@@ -244,16 +249,18 @@ mtype_list(const struct ip_set *set,
/* Set listing finished */
cb->args[IPSET_CB_ARG0] = 0;
- return 0;
+ goto out;
nla_put_failure:
nla_nest_cancel(skb, nested);
if (unlikely(id == first)) {
cb->args[IPSET_CB_ARG0] = 0;
- return -EMSGSIZE;
+ ret = -EMSGSIZE;
}
ipset_nest_end(skb, adt);
- return 0;
+out:
+ rcu_read_unlock();
+ return ret;
}
static void
diff --git a/net/netfilter/ipset/ip_set_list_set.c b/net/netfilter/ipset/ip_set_list_set.c
index 507d1ab..76a1c5f 100644
--- a/net/netfilter/ipset/ip_set_list_set.c
+++ b/net/netfilter/ipset/ip_set_list_set.c
@@ -477,6 +477,7 @@ list_set_list(const struct ip_set *set,
struct nlattr *atd, *nested;
u32 i = 0, first = cb->args[IPSET_CB_ARG0];
struct set_elem *e;
+ int ret = 0;
atd = ipset_nest_start(skb, IPSET_ATTR_ADT);
if (!atd)
@@ -487,6 +488,7 @@ list_set_list(const struct ip_set *set,
i++;
}
+ rcu_read_lock();
list_for_each_entry_from(e, &map->members, list) {
i++;
if (SET_WITH_TIMEOUT(set) &&
@@ -496,9 +498,10 @@ list_set_list(const struct ip_set *set,
if (!nested) {
if (i == first) {
nla_nest_cancel(skb, atd);
- return -EMSGSIZE;
- } else
- goto nla_put_failure;
+ ret = -EMSGSIZE;
+ goto out;
+ }
+ goto nla_put_failure;
}
if (nla_put_string(skb, IPSET_ATTR_NAME,
ip_set_name_byindex(map->net, e->id)))
@@ -511,17 +514,19 @@ list_set_list(const struct ip_set *set,
ipset_nest_end(skb, atd);
/* Set listing finished */
cb->args[IPSET_CB_ARG0] = 0;
- return 0;
+ goto out;
nla_put_failure:
nla_nest_cancel(skb, nested);
if (unlikely(i == first)) {
cb->args[IPSET_CB_ARG0] = 0;
- return -EMSGSIZE;
+ ret = -EMSGSIZE;
}
cb->args[IPSET_CB_ARG0] = i - 1;
ipset_nest_end(skb, atd);
- return 0;
+out:
+ rcu_read_unlock();
+ return ret;
}
static bool
--
1.8.5.1
--
To unsubscribe from this list: send the line "unsubscribe netfilter-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
