kernel-ark/net/sched/cls_api.c
Daniel Borkmann c78e1746d3 net: sched: fix call_rcu() race on classifier module unloads
Vijay reported that a loop as simple as ...

  while true; do
    tc qdisc add dev foo root handle 1: prio
    tc filter add dev foo parent 1: u32 match u32 0 0  flowid 1
    tc qdisc del dev foo root
    rmmod cls_u32
  done

... will panic the kernel. Moreover, he bisected the change
apparently introducing it to 78fd1d0ab0 ("netlink: Re-add
locking to netlink_lookup() and seq walker").

The removal of synchronize_net() from the netlink socket
triggering the qdisc to be removed, seems to have uncovered
an RCU resp. module reference count race from the tc API.
Given that RCU conversion was done after e341694e3e ("netlink:
Convert netlink_lookup() to use RCU protected hash table")
which added the synchronize_net() originally, occasion of
hitting the bug was less likely (not impossible though):

When qdiscs that i) support attaching classifiers and,
ii) have at least one of them attached, get deleted, they
invoke tcf_destroy_chain(), and thus call into ->destroy()
handler from a classifier module.

After RCU conversion, all classifier that have an internal
prio list, unlink them and initiate freeing via call_rcu()
deferral.

Meanhile, tcf_destroy() releases already reference to the
tp->ops->owner module before the queued RCU callback handler
has been invoked.

Subsequent rmmod on the classifier module is then not prevented
since all module references are already dropped.

By the time, the kernel invokes the RCU callback handler from
the module, that function address is then invalid.

One way to fix it would be to add an rcu_barrier() to
unregister_tcf_proto_ops() to wait for all pending call_rcu()s
to complete.

synchronize_rcu() is not appropriate as under heavy RCU
callback load, registered call_rcu()s could be deferred
longer than a grace period. In case we don't have any pending
call_rcu()s, the barrier is allowed to return immediately.

Since we came here via unregister_tcf_proto_ops(), there
are no users of a given classifier anymore. Further nested
call_rcu()s pointing into the module space are not being
done anywhere.

Only cls_bpf_delete_prog() may schedule a work item, to
unlock pages eventually, but that is not in the range/context
of cls_bpf anymore.

Fixes: 25d8c0d55f ("net: rcu-ify tcf_proto")
Fixes: 9888faefe1 ("net: sched: cls_basic use RCU")
Reported-by: Vijay Subramanian <subramanian.vijay@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: John Fastabend <john.r.fastabend@intel.com>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Thomas Graf <tgraf@suug.ch>
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Cc: Alexei Starovoitov <ast@plumgrid.com>
Tested-by: Vijay Subramanian <subramanian.vijay@gmail.com>
Acked-by: Alexei Starovoitov <ast@plumgrid.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2015-05-21 18:48:18 -04:00

635 lines
14 KiB
C

/*
* net/sched/cls_api.c Packet classifier API.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
*
* Changes:
*
* Eduardo J. Blanco <ejbs@netlabs.com.uy> :990222: kmod support
*
*/
#include <linux/module.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/string.h>
#include <linux/errno.h>
#include <linux/skbuff.h>
#include <linux/init.h>
#include <linux/kmod.h>
#include <linux/err.h>
#include <linux/slab.h>
#include <net/net_namespace.h>
#include <net/sock.h>
#include <net/netlink.h>
#include <net/pkt_sched.h>
#include <net/pkt_cls.h>
/* The list of all installed classifier types */
static LIST_HEAD(tcf_proto_base);
/* Protects list of registered TC modules. It is pure SMP lock. */
static DEFINE_RWLOCK(cls_mod_lock);
/* Find classifier type by string name */
static const struct tcf_proto_ops *tcf_proto_lookup_ops(struct nlattr *kind)
{
const struct tcf_proto_ops *t, *res = NULL;
if (kind) {
read_lock(&cls_mod_lock);
list_for_each_entry(t, &tcf_proto_base, head) {
if (nla_strcmp(kind, t->kind) == 0) {
if (try_module_get(t->owner))
res = t;
break;
}
}
read_unlock(&cls_mod_lock);
}
return res;
}
/* Register(unregister) new classifier type */
int register_tcf_proto_ops(struct tcf_proto_ops *ops)
{
struct tcf_proto_ops *t;
int rc = -EEXIST;
write_lock(&cls_mod_lock);
list_for_each_entry(t, &tcf_proto_base, head)
if (!strcmp(ops->kind, t->kind))
goto out;
list_add_tail(&ops->head, &tcf_proto_base);
rc = 0;
out:
write_unlock(&cls_mod_lock);
return rc;
}
EXPORT_SYMBOL(register_tcf_proto_ops);
int unregister_tcf_proto_ops(struct tcf_proto_ops *ops)
{
struct tcf_proto_ops *t;
int rc = -ENOENT;
/* Wait for outstanding call_rcu()s, if any, from a
* tcf_proto_ops's destroy() handler.
*/
rcu_barrier();
write_lock(&cls_mod_lock);
list_for_each_entry(t, &tcf_proto_base, head) {
if (t == ops) {
list_del(&t->head);
rc = 0;
break;
}
}
write_unlock(&cls_mod_lock);
return rc;
}
EXPORT_SYMBOL(unregister_tcf_proto_ops);
static int tfilter_notify(struct net *net, struct sk_buff *oskb,
struct nlmsghdr *n, struct tcf_proto *tp,
unsigned long fh, int event);
/* Select new prio value from the range, managed by kernel. */
static inline u32 tcf_auto_prio(struct tcf_proto *tp)
{
u32 first = TC_H_MAKE(0xC0000000U, 0U);
if (tp)
first = tp->prio - 1;
return first;
}
/* Add/change/delete/get a filter node */
static int tc_ctl_tfilter(struct sk_buff *skb, struct nlmsghdr *n)
{
struct net *net = sock_net(skb->sk);
struct nlattr *tca[TCA_MAX + 1];
struct tcmsg *t;
u32 protocol;
u32 prio;
u32 nprio;
u32 parent;
struct net_device *dev;
struct Qdisc *q;
struct tcf_proto __rcu **back;
struct tcf_proto __rcu **chain;
struct tcf_proto *tp;
const struct tcf_proto_ops *tp_ops;
const struct Qdisc_class_ops *cops;
unsigned long cl;
unsigned long fh;
int err;
int tp_created = 0;
if ((n->nlmsg_type != RTM_GETTFILTER) &&
!netlink_ns_capable(skb, net->user_ns, CAP_NET_ADMIN))
return -EPERM;
replay:
err = nlmsg_parse(n, sizeof(*t), tca, TCA_MAX, NULL);
if (err < 0)
return err;
t = nlmsg_data(n);
protocol = TC_H_MIN(t->tcm_info);
prio = TC_H_MAJ(t->tcm_info);
nprio = prio;
parent = t->tcm_parent;
cl = 0;
if (prio == 0) {
/* If no priority is given, user wants we allocated it. */
if (n->nlmsg_type != RTM_NEWTFILTER ||
!(n->nlmsg_flags & NLM_F_CREATE))
return -ENOENT;
prio = TC_H_MAKE(0x80000000U, 0U);
}
/* Find head of filter chain. */
/* Find link */
dev = __dev_get_by_index(net, t->tcm_ifindex);
if (dev == NULL)
return -ENODEV;
/* Find qdisc */
if (!parent) {
q = dev->qdisc;
parent = q->handle;
} else {
q = qdisc_lookup(dev, TC_H_MAJ(t->tcm_parent));
if (q == NULL)
return -EINVAL;
}
/* Is it classful? */
cops = q->ops->cl_ops;
if (!cops)
return -EINVAL;
if (cops->tcf_chain == NULL)
return -EOPNOTSUPP;
/* Do we search for filter, attached to class? */
if (TC_H_MIN(parent)) {
cl = cops->get(q, parent);
if (cl == 0)
return -ENOENT;
}
/* And the last stroke */
chain = cops->tcf_chain(q, cl);
err = -EINVAL;
if (chain == NULL)
goto errout;
/* Check the chain for existence of proto-tcf with this priority */
for (back = chain;
(tp = rtnl_dereference(*back)) != NULL;
back = &tp->next) {
if (tp->prio >= prio) {
if (tp->prio == prio) {
if (!nprio ||
(tp->protocol != protocol && protocol))
goto errout;
} else
tp = NULL;
break;
}
}
if (tp == NULL) {
/* Proto-tcf does not exist, create new one */
if (tca[TCA_KIND] == NULL || !protocol)
goto errout;
err = -ENOENT;
if (n->nlmsg_type != RTM_NEWTFILTER ||
!(n->nlmsg_flags & NLM_F_CREATE))
goto errout;
/* Create new proto tcf */
err = -ENOBUFS;
tp = kzalloc(sizeof(*tp), GFP_KERNEL);
if (tp == NULL)
goto errout;
err = -ENOENT;
tp_ops = tcf_proto_lookup_ops(tca[TCA_KIND]);
if (tp_ops == NULL) {
#ifdef CONFIG_MODULES
struct nlattr *kind = tca[TCA_KIND];
char name[IFNAMSIZ];
if (kind != NULL &&
nla_strlcpy(name, kind, IFNAMSIZ) < IFNAMSIZ) {
rtnl_unlock();
request_module("cls_%s", name);
rtnl_lock();
tp_ops = tcf_proto_lookup_ops(kind);
/* We dropped the RTNL semaphore in order to
* perform the module load. So, even if we
* succeeded in loading the module we have to
* replay the request. We indicate this using
* -EAGAIN.
*/
if (tp_ops != NULL) {
module_put(tp_ops->owner);
err = -EAGAIN;
}
}
#endif
kfree(tp);
goto errout;
}
tp->ops = tp_ops;
tp->protocol = protocol;
tp->prio = nprio ? :
TC_H_MAJ(tcf_auto_prio(rtnl_dereference(*back)));
tp->q = q;
tp->classify = tp_ops->classify;
tp->classid = parent;
err = tp_ops->init(tp);
if (err != 0) {
module_put(tp_ops->owner);
kfree(tp);
goto errout;
}
tp_created = 1;
} else if (tca[TCA_KIND] && nla_strcmp(tca[TCA_KIND], tp->ops->kind))
goto errout;
fh = tp->ops->get(tp, t->tcm_handle);
if (fh == 0) {
if (n->nlmsg_type == RTM_DELTFILTER && t->tcm_handle == 0) {
struct tcf_proto *next = rtnl_dereference(tp->next);
RCU_INIT_POINTER(*back, next);
tfilter_notify(net, skb, n, tp, fh, RTM_DELTFILTER);
tcf_destroy(tp, true);
err = 0;
goto errout;
}
err = -ENOENT;
if (n->nlmsg_type != RTM_NEWTFILTER ||
!(n->nlmsg_flags & NLM_F_CREATE))
goto errout;
} else {
switch (n->nlmsg_type) {
case RTM_NEWTFILTER:
err = -EEXIST;
if (n->nlmsg_flags & NLM_F_EXCL) {
if (tp_created)
tcf_destroy(tp, true);
goto errout;
}
break;
case RTM_DELTFILTER:
err = tp->ops->delete(tp, fh);
if (err == 0) {
struct tcf_proto *next = rtnl_dereference(tp->next);
tfilter_notify(net, skb, n, tp, fh, RTM_DELTFILTER);
if (tcf_destroy(tp, false))
RCU_INIT_POINTER(*back, next);
}
goto errout;
case RTM_GETTFILTER:
err = tfilter_notify(net, skb, n, tp, fh, RTM_NEWTFILTER);
goto errout;
default:
err = -EINVAL;
goto errout;
}
}
err = tp->ops->change(net, skb, tp, cl, t->tcm_handle, tca, &fh,
n->nlmsg_flags & NLM_F_CREATE ? TCA_ACT_NOREPLACE : TCA_ACT_REPLACE);
if (err == 0) {
if (tp_created) {
RCU_INIT_POINTER(tp->next, rtnl_dereference(*back));
rcu_assign_pointer(*back, tp);
}
tfilter_notify(net, skb, n, tp, fh, RTM_NEWTFILTER);
} else {
if (tp_created)
tcf_destroy(tp, true);
}
errout:
if (cl)
cops->put(q, cl);
if (err == -EAGAIN)
/* Replay the request. */
goto replay;
return err;
}
static int tcf_fill_node(struct net *net, struct sk_buff *skb, struct tcf_proto *tp,
unsigned long fh, u32 portid, u32 seq, u16 flags, int event)
{
struct tcmsg *tcm;
struct nlmsghdr *nlh;
unsigned char *b = skb_tail_pointer(skb);
nlh = nlmsg_put(skb, portid, seq, event, sizeof(*tcm), flags);
if (!nlh)
goto out_nlmsg_trim;
tcm = nlmsg_data(nlh);
tcm->tcm_family = AF_UNSPEC;
tcm->tcm__pad1 = 0;
tcm->tcm__pad2 = 0;
tcm->tcm_ifindex = qdisc_dev(tp->q)->ifindex;
tcm->tcm_parent = tp->classid;
tcm->tcm_info = TC_H_MAKE(tp->prio, tp->protocol);
if (nla_put_string(skb, TCA_KIND, tp->ops->kind))
goto nla_put_failure;
tcm->tcm_handle = fh;
if (RTM_DELTFILTER != event) {
tcm->tcm_handle = 0;
if (tp->ops->dump && tp->ops->dump(net, tp, fh, skb, tcm) < 0)
goto nla_put_failure;
}
nlh->nlmsg_len = skb_tail_pointer(skb) - b;
return skb->len;
out_nlmsg_trim:
nla_put_failure:
nlmsg_trim(skb, b);
return -1;
}
static int tfilter_notify(struct net *net, struct sk_buff *oskb,
struct nlmsghdr *n, struct tcf_proto *tp,
unsigned long fh, int event)
{
struct sk_buff *skb;
u32 portid = oskb ? NETLINK_CB(oskb).portid : 0;
skb = alloc_skb(NLMSG_GOODSIZE, GFP_KERNEL);
if (!skb)
return -ENOBUFS;
if (tcf_fill_node(net, skb, tp, fh, portid, n->nlmsg_seq, 0, event) <= 0) {
kfree_skb(skb);
return -EINVAL;
}
return rtnetlink_send(skb, net, portid, RTNLGRP_TC,
n->nlmsg_flags & NLM_F_ECHO);
}
struct tcf_dump_args {
struct tcf_walker w;
struct sk_buff *skb;
struct netlink_callback *cb;
};
static int tcf_node_dump(struct tcf_proto *tp, unsigned long n,
struct tcf_walker *arg)
{
struct tcf_dump_args *a = (void *)arg;
struct net *net = sock_net(a->skb->sk);
return tcf_fill_node(net, a->skb, tp, n, NETLINK_CB(a->cb->skb).portid,
a->cb->nlh->nlmsg_seq, NLM_F_MULTI, RTM_NEWTFILTER);
}
/* called with RTNL */
static int tc_dump_tfilter(struct sk_buff *skb, struct netlink_callback *cb)
{
struct net *net = sock_net(skb->sk);
int t;
int s_t;
struct net_device *dev;
struct Qdisc *q;
struct tcf_proto *tp, __rcu **chain;
struct tcmsg *tcm = nlmsg_data(cb->nlh);
unsigned long cl = 0;
const struct Qdisc_class_ops *cops;
struct tcf_dump_args arg;
if (nlmsg_len(cb->nlh) < sizeof(*tcm))
return skb->len;
dev = __dev_get_by_index(net, tcm->tcm_ifindex);
if (!dev)
return skb->len;
if (!tcm->tcm_parent)
q = dev->qdisc;
else
q = qdisc_lookup(dev, TC_H_MAJ(tcm->tcm_parent));
if (!q)
goto out;
cops = q->ops->cl_ops;
if (!cops)
goto errout;
if (cops->tcf_chain == NULL)
goto errout;
if (TC_H_MIN(tcm->tcm_parent)) {
cl = cops->get(q, tcm->tcm_parent);
if (cl == 0)
goto errout;
}
chain = cops->tcf_chain(q, cl);
if (chain == NULL)
goto errout;
s_t = cb->args[0];
for (tp = rtnl_dereference(*chain), t = 0;
tp; tp = rtnl_dereference(tp->next), t++) {
if (t < s_t)
continue;
if (TC_H_MAJ(tcm->tcm_info) &&
TC_H_MAJ(tcm->tcm_info) != tp->prio)
continue;
if (TC_H_MIN(tcm->tcm_info) &&
TC_H_MIN(tcm->tcm_info) != tp->protocol)
continue;
if (t > s_t)
memset(&cb->args[1], 0, sizeof(cb->args)-sizeof(cb->args[0]));
if (cb->args[1] == 0) {
if (tcf_fill_node(net, skb, tp, 0, NETLINK_CB(cb->skb).portid,
cb->nlh->nlmsg_seq, NLM_F_MULTI,
RTM_NEWTFILTER) <= 0)
break;
cb->args[1] = 1;
}
if (tp->ops->walk == NULL)
continue;
arg.w.fn = tcf_node_dump;
arg.skb = skb;
arg.cb = cb;
arg.w.stop = 0;
arg.w.skip = cb->args[1] - 1;
arg.w.count = 0;
tp->ops->walk(tp, &arg.w);
cb->args[1] = arg.w.count + 1;
if (arg.w.stop)
break;
}
cb->args[0] = t;
errout:
if (cl)
cops->put(q, cl);
out:
return skb->len;
}
void tcf_exts_destroy(struct tcf_exts *exts)
{
#ifdef CONFIG_NET_CLS_ACT
tcf_action_destroy(&exts->actions, TCA_ACT_UNBIND);
INIT_LIST_HEAD(&exts->actions);
#endif
}
EXPORT_SYMBOL(tcf_exts_destroy);
int tcf_exts_validate(struct net *net, struct tcf_proto *tp, struct nlattr **tb,
struct nlattr *rate_tlv, struct tcf_exts *exts, bool ovr)
{
#ifdef CONFIG_NET_CLS_ACT
{
struct tc_action *act;
INIT_LIST_HEAD(&exts->actions);
if (exts->police && tb[exts->police]) {
act = tcf_action_init_1(net, tb[exts->police], rate_tlv,
"police", ovr,
TCA_ACT_BIND);
if (IS_ERR(act))
return PTR_ERR(act);
act->type = exts->type = TCA_OLD_COMPAT;
list_add(&act->list, &exts->actions);
} else if (exts->action && tb[exts->action]) {
int err;
err = tcf_action_init(net, tb[exts->action], rate_tlv,
NULL, ovr,
TCA_ACT_BIND, &exts->actions);
if (err)
return err;
}
}
#else
if ((exts->action && tb[exts->action]) ||
(exts->police && tb[exts->police]))
return -EOPNOTSUPP;
#endif
return 0;
}
EXPORT_SYMBOL(tcf_exts_validate);
void tcf_exts_change(struct tcf_proto *tp, struct tcf_exts *dst,
struct tcf_exts *src)
{
#ifdef CONFIG_NET_CLS_ACT
LIST_HEAD(tmp);
tcf_tree_lock(tp);
list_splice_init(&dst->actions, &tmp);
list_splice(&src->actions, &dst->actions);
dst->type = src->type;
tcf_tree_unlock(tp);
tcf_action_destroy(&tmp, TCA_ACT_UNBIND);
#endif
}
EXPORT_SYMBOL(tcf_exts_change);
#define tcf_exts_first_act(ext) \
list_first_entry_or_null(&(exts)->actions, \
struct tc_action, list)
int tcf_exts_dump(struct sk_buff *skb, struct tcf_exts *exts)
{
#ifdef CONFIG_NET_CLS_ACT
struct nlattr *nest;
if (exts->action && !list_empty(&exts->actions)) {
/*
* again for backward compatible mode - we want
* to work with both old and new modes of entering
* tc data even if iproute2 was newer - jhs
*/
if (exts->type != TCA_OLD_COMPAT) {
nest = nla_nest_start(skb, exts->action);
if (nest == NULL)
goto nla_put_failure;
if (tcf_action_dump(skb, &exts->actions, 0, 0) < 0)
goto nla_put_failure;
nla_nest_end(skb, nest);
} else if (exts->police) {
struct tc_action *act = tcf_exts_first_act(exts);
nest = nla_nest_start(skb, exts->police);
if (nest == NULL || !act)
goto nla_put_failure;
if (tcf_action_dump_old(skb, act, 0, 0) < 0)
goto nla_put_failure;
nla_nest_end(skb, nest);
}
}
return 0;
nla_put_failure:
nla_nest_cancel(skb, nest);
return -1;
#else
return 0;
#endif
}
EXPORT_SYMBOL(tcf_exts_dump);
int tcf_exts_dump_stats(struct sk_buff *skb, struct tcf_exts *exts)
{
#ifdef CONFIG_NET_CLS_ACT
struct tc_action *a = tcf_exts_first_act(exts);
if (a != NULL && tcf_action_copy_stats(skb, a, 1) < 0)
return -1;
#endif
return 0;
}
EXPORT_SYMBOL(tcf_exts_dump_stats);
static int __init tc_filter_init(void)
{
rtnl_register(PF_UNSPEC, RTM_NEWTFILTER, tc_ctl_tfilter, NULL, NULL);
rtnl_register(PF_UNSPEC, RTM_DELTFILTER, tc_ctl_tfilter, NULL, NULL);
rtnl_register(PF_UNSPEC, RTM_GETTFILTER, tc_ctl_tfilter,
tc_dump_tfilter, NULL);
return 0;
}
subsys_initcall(tc_filter_init);