This patch introduces a list of pending module requests. This new module
list is composed of nft_module_request objects that contain the module
name and one status field that tells if the module has been already
loaded (the 'done' field).
In the first pass, from the preparation phase, the netlink command finds
that a module is missing on this list. Then, a module request is
allocated and added to this list and nft_request_module() returns
-EAGAIN. This triggers the abort path with the autoload parameter set on
from nfnetlink, request_module() is called and the module request enters
the 'done' state. Since the mutex is released when loading modules from
the abort phase, the module list is zapped so this is iteration occurs
over a local list. Therefore, the request_module() calls happen when
object lists are in consistent state (after fulling aborting the
transaction) and the commit list is empty.
On the second pass, the netlink command will find that it already tried
to load the module, so it does not request it again and
nft_request_module() returns 0. Then, there is a look up to find the
object that the command was missing. If the module was successfully
loaded, the command proceeds normally since it finds the missing object
in place, otherwise -ENOENT is reported to userspace.
This patch also updates nfnetlink to include the reason to enter the
abort phase, which is required for this new autoload module rationale.
Fixes: ec7470b834fe ("netfilter: nf_tables: store transaction list locally while requesting module")
Reported-by: syzbot+29125d208b3dae9a7019@syzkaller.appspotmail.com
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
const struct nfnl_callback *cb; /* callback for individual types */
struct module *owner;
int (*commit)(struct net *net, struct sk_buff *skb);
const struct nfnl_callback *cb; /* callback for individual types */
struct module *owner;
int (*commit)(struct net *net, struct sk_buff *skb);
- int (*abort)(struct net *net, struct sk_buff *skb);
+ int (*abort)(struct net *net, struct sk_buff *skb, bool autoload);
void (*cleanup)(struct net *net);
bool (*valid_genid)(struct net *net, u32 genid);
};
void (*cleanup)(struct net *net);
bool (*valid_genid)(struct net *net, u32 genid);
};
struct netns_nftables {
struct list_head tables;
struct list_head commit_list;
struct netns_nftables {
struct list_head tables;
struct list_head commit_list;
+ struct list_head module_list;
struct mutex commit_mutex;
unsigned int base_seq;
u8 gencursor;
struct mutex commit_mutex;
unsigned int base_seq;
u8 gencursor;
-/*
- * Loading a module requires dropping mutex that guards the transaction.
- * A different client might race to start a new transaction meanwhile. Zap the
- * list of pending transaction and then restore it once the mutex is grabbed
- * again. Users of this function return EAGAIN which implicitly triggers the
- * transaction abort path to clean up the list of pending transactions.
- */
+struct nft_module_request {
+ struct list_head list;
+ char module[MODULE_NAME_LEN];
+ bool done;
+};
+
-static void nft_request_module(struct net *net, const char *fmt, ...)
+static int nft_request_module(struct net *net, const char *fmt, ...)
{
char module_name[MODULE_NAME_LEN];
{
char module_name[MODULE_NAME_LEN];
- LIST_HEAD(commit_list);
+ struct nft_module_request *req;
- list_splice_init(&net->nft.commit_list, &commit_list);
-
va_start(args, fmt);
ret = vsnprintf(module_name, MODULE_NAME_LEN, fmt, args);
va_end(args);
if (ret >= MODULE_NAME_LEN)
va_start(args, fmt);
ret = vsnprintf(module_name, MODULE_NAME_LEN, fmt, args);
va_end(args);
if (ret >= MODULE_NAME_LEN)
- mutex_unlock(&net->nft.commit_mutex);
- request_module("%s", module_name);
- mutex_lock(&net->nft.commit_mutex);
+ list_for_each_entry(req, &net->nft.module_list, list) {
+ if (!strcmp(req->module, module_name)) {
+ if (req->done)
+ return 0;
- WARN_ON_ONCE(!list_empty(&net->nft.commit_list));
- list_splice(&commit_list, &net->nft.commit_list);
+ /* A request to load this module already exists. */
+ return -EAGAIN;
+ }
+ }
+
+ req = kmalloc(sizeof(*req), GFP_KERNEL);
+ if (!req)
+ return -ENOMEM;
+
+ req->done = false;
+ strlcpy(req->module, module_name, MODULE_NAME_LEN);
+ list_add_tail(&req->list, &net->nft.module_list);
+
+ return -EAGAIN;
lockdep_nfnl_nft_mutex_not_held();
#ifdef CONFIG_MODULES
if (autoload) {
lockdep_nfnl_nft_mutex_not_held();
#ifdef CONFIG_MODULES
if (autoload) {
- nft_request_module(net, "nft-chain-%u-%.*s", family,
- nla_len(nla), (const char *)nla_data(nla));
- type = __nf_tables_chain_type_lookup(nla, family);
- if (type != NULL)
+ if (nft_request_module(net, "nft-chain-%u-%.*s", family,
+ nla_len(nla),
+ (const char *)nla_data(nla)) == -EAGAIN)
return ERR_PTR(-EAGAIN);
}
#endif
return ERR_PTR(-EAGAIN);
}
#endif
static int nft_expr_type_request_module(struct net *net, u8 family,
struct nlattr *nla)
{
static int nft_expr_type_request_module(struct net *net, u8 family,
struct nlattr *nla)
{
- nft_request_module(net, "nft-expr-%u-%.*s", family,
- nla_len(nla), (char *)nla_data(nla));
- if (__nft_expr_type_get(family, nla))
+ if (nft_request_module(net, "nft-expr-%u-%.*s", family,
+ nla_len(nla), (char *)nla_data(nla)) == -EAGAIN)
return -EAGAIN;
return 0;
return -EAGAIN;
return 0;
if (nft_expr_type_request_module(net, family, nla) == -EAGAIN)
return ERR_PTR(-EAGAIN);
if (nft_expr_type_request_module(net, family, nla) == -EAGAIN)
return ERR_PTR(-EAGAIN);
- nft_request_module(net, "nft-expr-%.*s",
- nla_len(nla), (char *)nla_data(nla));
- if (__nft_expr_type_get(family, nla))
+ if (nft_request_module(net, "nft-expr-%.*s",
+ nla_len(nla),
+ (char *)nla_data(nla)) == -EAGAIN)
return ERR_PTR(-EAGAIN);
}
#endif
return ERR_PTR(-EAGAIN);
}
#endif
err = PTR_ERR(ops);
#ifdef CONFIG_MODULES
if (err == -EAGAIN)
err = PTR_ERR(ops);
#ifdef CONFIG_MODULES
if (err == -EAGAIN)
- nft_expr_type_request_module(ctx->net,
- ctx->family,
- tb[NFTA_EXPR_NAME]);
+ if (nft_expr_type_request_module(ctx->net,
+ ctx->family,
+ tb[NFTA_EXPR_NAME]) != -EAGAIN)
+ err = -ENOENT;
lockdep_nfnl_nft_mutex_not_held();
#ifdef CONFIG_MODULES
if (list_empty(&nf_tables_set_types)) {
lockdep_nfnl_nft_mutex_not_held();
#ifdef CONFIG_MODULES
if (list_empty(&nf_tables_set_types)) {
- nft_request_module(ctx->net, "nft-set");
- if (!list_empty(&nf_tables_set_types))
+ if (nft_request_module(ctx->net, "nft-set") == -EAGAIN)
return ERR_PTR(-EAGAIN);
}
#endif
return ERR_PTR(-EAGAIN);
}
#endif
lockdep_nfnl_nft_mutex_not_held();
#ifdef CONFIG_MODULES
if (type == NULL) {
lockdep_nfnl_nft_mutex_not_held();
#ifdef CONFIG_MODULES
if (type == NULL) {
- nft_request_module(net, "nft-obj-%u", objtype);
- if (__nft_obj_type_get(objtype))
+ if (nft_request_module(net, "nft-obj-%u", objtype) == -EAGAIN)
return ERR_PTR(-EAGAIN);
}
#endif
return ERR_PTR(-EAGAIN);
}
#endif
lockdep_nfnl_nft_mutex_not_held();
#ifdef CONFIG_MODULES
if (type == NULL) {
lockdep_nfnl_nft_mutex_not_held();
#ifdef CONFIG_MODULES
if (type == NULL) {
- nft_request_module(net, "nf-flowtable-%u", family);
- if (__nft_flowtable_type_get(family))
+ if (nft_request_module(net, "nf-flowtable-%u", family) == -EAGAIN)
return ERR_PTR(-EAGAIN);
}
#endif
return ERR_PTR(-EAGAIN);
}
#endif
list_del_rcu(&chain->list);
}
list_del_rcu(&chain->list);
}
+static void nf_tables_module_autoload_cleanup(struct net *net)
+{
+ struct nft_module_request *req, *next;
+
+ WARN_ON_ONCE(!list_empty(&net->nft.commit_list));
+ list_for_each_entry_safe(req, next, &net->nft.module_list, list) {
+ WARN_ON_ONCE(!req->done);
+ list_del(&req->list);
+ kfree(req);
+ }
+}
+
static void nf_tables_commit_release(struct net *net)
{
struct nft_trans *trans;
static void nf_tables_commit_release(struct net *net)
{
struct nft_trans *trans;
* to prevent expensive synchronize_rcu() in commit phase.
*/
if (list_empty(&net->nft.commit_list)) {
* to prevent expensive synchronize_rcu() in commit phase.
*/
if (list_empty(&net->nft.commit_list)) {
+ nf_tables_module_autoload_cleanup(net);
mutex_unlock(&net->nft.commit_mutex);
return;
}
mutex_unlock(&net->nft.commit_mutex);
return;
}
list_splice_tail_init(&net->nft.commit_list, &nf_tables_destroy_list);
spin_unlock(&nf_tables_destroy_list_lock);
list_splice_tail_init(&net->nft.commit_list, &nf_tables_destroy_list);
spin_unlock(&nf_tables_destroy_list_lock);
+ nf_tables_module_autoload_cleanup(net);
mutex_unlock(&net->nft.commit_mutex);
schedule_work(&trans_destroy_work);
mutex_unlock(&net->nft.commit_mutex);
schedule_work(&trans_destroy_work);
+static void nf_tables_module_autoload(struct net *net)
+{
+ struct nft_module_request *req, *next;
+ LIST_HEAD(module_list);
+
+ list_splice_init(&net->nft.module_list, &module_list);
+ mutex_unlock(&net->nft.commit_mutex);
+ list_for_each_entry_safe(req, next, &module_list, list) {
+ if (req->done) {
+ list_del(&req->list);
+ kfree(req);
+ } else {
+ request_module("%s", req->module);
+ req->done = true;
+ }
+ }
+ mutex_lock(&net->nft.commit_mutex);
+ list_splice(&module_list, &net->nft.module_list);
+}
+
static void nf_tables_abort_release(struct nft_trans *trans)
{
switch (trans->msg_type) {
static void nf_tables_abort_release(struct nft_trans *trans)
{
switch (trans->msg_type) {
-static int __nf_tables_abort(struct net *net)
+static int __nf_tables_abort(struct net *net, bool autoload)
{
struct nft_trans *trans, *next;
struct nft_trans_elem *te;
{
struct nft_trans *trans, *next;
struct nft_trans_elem *te;
nf_tables_abort_release(trans);
}
nf_tables_abort_release(trans);
}
+ if (autoload)
+ nf_tables_module_autoload(net);
+ else
+ nf_tables_module_autoload_cleanup(net);
+
nft_validate_state_update(net, NFT_VALIDATE_SKIP);
}
nft_validate_state_update(net, NFT_VALIDATE_SKIP);
}
-static int nf_tables_abort(struct net *net, struct sk_buff *skb)
+static int nf_tables_abort(struct net *net, struct sk_buff *skb, bool autoload)
- int ret = __nf_tables_abort(net);
+ int ret = __nf_tables_abort(net, autoload);
mutex_unlock(&net->nft.commit_mutex);
mutex_unlock(&net->nft.commit_mutex);
{
INIT_LIST_HEAD(&net->nft.tables);
INIT_LIST_HEAD(&net->nft.commit_list);
{
INIT_LIST_HEAD(&net->nft.tables);
INIT_LIST_HEAD(&net->nft.commit_list);
+ INIT_LIST_HEAD(&net->nft.module_list);
mutex_init(&net->nft.commit_mutex);
net->nft.base_seq = 1;
net->nft.validate_state = NFT_VALIDATE_SKIP;
mutex_init(&net->nft.commit_mutex);
net->nft.base_seq = 1;
net->nft.validate_state = NFT_VALIDATE_SKIP;
{
mutex_lock(&net->nft.commit_mutex);
if (!list_empty(&net->nft.commit_list))
{
mutex_lock(&net->nft.commit_mutex);
if (!list_empty(&net->nft.commit_list))
- __nf_tables_abort(net);
+ __nf_tables_abort(net, false);
__nft_release_tables(net);
mutex_unlock(&net->nft.commit_mutex);
WARN_ON_ONCE(!list_empty(&net->nft.tables));
__nft_release_tables(net);
mutex_unlock(&net->nft.commit_mutex);
WARN_ON_ONCE(!list_empty(&net->nft.tables));
}
done:
if (status & NFNL_BATCH_REPLAY) {
}
done:
if (status & NFNL_BATCH_REPLAY) {
+ ss->abort(net, oskb, true);
nfnl_err_reset(&err_list);
kfree_skb(skb);
module_put(ss->owner);
nfnl_err_reset(&err_list);
kfree_skb(skb);
module_put(ss->owner);
status |= NFNL_BATCH_REPLAY;
goto done;
} else if (err) {
status |= NFNL_BATCH_REPLAY;
goto done;
} else if (err) {
+ ss->abort(net, oskb, false);
netlink_ack(oskb, nlmsg_hdr(oskb), err, NULL);
}
} else {
netlink_ack(oskb, nlmsg_hdr(oskb), err, NULL);
}
} else {
+ ss->abort(net, oskb, false);
}
if (ss->cleanup)
ss->cleanup(net);
}
if (ss->cleanup)
ss->cleanup(net);