From 30935198b7d0be12b1c45c328b66a7fdefb16256 Mon Sep 17 00:00:00 2001 From: Haiqing Bai Date: Mon, 27 Aug 2018 09:32:26 +0800 Subject: tipc: fix the big/little endian issue in tipc_dest In function tipc_dest_push, the 32bit variables 'node' and 'port' are stored separately in uppper and lower part of 64bit 'value'. Then this value is assigned to dst->value which is a union like: union { struct { u32 port; u32 node; }; u64 value; } This works on little-endian machines like x86 but fails on big-endian machines. The fix remove the 'value' stack parameter and even the 'value' member of the union in tipc_dest, assign the 'node' and 'port' member directly with the input parameter to avoid the endian issue. Fixes: a80ae5306a73 ("tipc: improve destination linked list") Signed-off-by: Zhenbo Gao Acked-by: Jon Maloy Signed-off-by: Haiqing Bai Signed-off-by: David S. Miller --- net/tipc/name_table.c | 10 ++++------ net/tipc/name_table.h | 9 ++------- 2 files changed, 6 insertions(+), 13 deletions(-) (limited to 'net/tipc') diff --git a/net/tipc/name_table.c b/net/tipc/name_table.c index 88f027b502f6..66d5b2c5987a 100644 --- a/net/tipc/name_table.c +++ b/net/tipc/name_table.c @@ -980,20 +980,17 @@ int tipc_nl_name_table_dump(struct sk_buff *skb, struct netlink_callback *cb) struct tipc_dest *tipc_dest_find(struct list_head *l, u32 node, u32 port) { - u64 value = (u64)node << 32 | port; struct tipc_dest *dst; list_for_each_entry(dst, l, list) { - if (dst->value != value) - continue; - return dst; + if (dst->node == node && dst->port == port) + return dst; } return NULL; } bool tipc_dest_push(struct list_head *l, u32 node, u32 port) { - u64 value = (u64)node << 32 | port; struct tipc_dest *dst; if (tipc_dest_find(l, node, port)) @@ -1002,7 +999,8 @@ bool tipc_dest_push(struct list_head *l, u32 node, u32 port) dst = kmalloc(sizeof(*dst), GFP_ATOMIC); if (unlikely(!dst)) return false; - dst->value = value; + dst->node = node; + dst->port = port; list_add(&dst->list, l); return true; } diff --git a/net/tipc/name_table.h b/net/tipc/name_table.h index 0febba41da86..892bd750b85f 100644 --- a/net/tipc/name_table.h +++ b/net/tipc/name_table.h @@ -133,13 +133,8 @@ void tipc_nametbl_stop(struct net *net); struct tipc_dest { struct list_head list; - union { - struct { - u32 port; - u32 node; - }; - u64 value; - }; + u32 port; + u32 node; }; struct tipc_dest *tipc_dest_find(struct list_head *l, u32 node, u32 port); -- cgit v1.2.3 From bd583fe30427500a2d0abe25724025b1cb5e2636 Mon Sep 17 00:00:00 2001 From: Cong Wang Date: Thu, 23 Aug 2018 16:19:44 -0700 Subject: tipc: fix a missing rhashtable_walk_exit() rhashtable_walk_exit() must be paired with rhashtable_walk_enter(). Fixes: 40f9f4397060 ("tipc: Fix tipc_sk_reinit race conditions") Cc: Herbert Xu Cc: Ying Xue Signed-off-by: Cong Wang Signed-off-by: David S. Miller --- net/tipc/socket.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'net/tipc') diff --git a/net/tipc/socket.c b/net/tipc/socket.c index c1e93c9515bc..c9a50b62c738 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -2672,6 +2672,8 @@ void tipc_sk_reinit(struct net *net) rhashtable_walk_stop(&iter); } while (tsk == ERR_PTR(-EAGAIN)); + + rhashtable_walk_exit(&iter); } static struct tipc_sock *tipc_sk_lookup(struct net *net, u32 portid) -- cgit v1.2.3 From 9a07efa9aea2f4a59f35da0785a4e6a6b5a96192 Mon Sep 17 00:00:00 2001 From: Cong Wang Date: Fri, 24 Aug 2018 12:28:06 -0700 Subject: tipc: switch to rhashtable iterator syzbot reported a use-after-free in tipc_group_fill_sock_diag(), where tipc_group_fill_sock_diag() still reads tsk->group meanwhile tipc_group_delete() just deletes it in tipc_release(). tipc_nl_sk_walk() aims to lock this sock when walking each sock in the hash table to close race conditions with sock changes like this one, by acquiring tsk->sk.sk_lock.slock spinlock, unfortunately this doesn't work at all. All non-BH call path should take lock_sock() instead to make it work. tipc_nl_sk_walk() brutally iterates with raw rht_for_each_entry_rcu() where RCU read lock is required, this is the reason why lock_sock() can't be taken on this path. This could be resolved by switching to rhashtable iterator API's, where taking a sleepable lock is possible. Also, the iterator API's are friendly for restartable calls like diag dump, the last position is remembered behind the scence, all we need to do here is saving the iterator into cb->args[]. I tested this with parallel tipc diag dump and thousands of tipc socket creation and release, no crash or memory leak. Reported-by: syzbot+b9c8f3ab2994b7cd1625@syzkaller.appspotmail.com Cc: Jon Maloy Cc: Ying Xue Signed-off-by: Cong Wang Signed-off-by: David S. Miller --- net/tipc/diag.c | 2 ++ net/tipc/netlink.c | 2 ++ net/tipc/socket.c | 76 +++++++++++++++++++++++++++++++++++------------------- net/tipc/socket.h | 2 ++ 4 files changed, 56 insertions(+), 26 deletions(-) (limited to 'net/tipc') diff --git a/net/tipc/diag.c b/net/tipc/diag.c index aaabb0b776dd..73137f4aeb68 100644 --- a/net/tipc/diag.c +++ b/net/tipc/diag.c @@ -84,7 +84,9 @@ static int tipc_sock_diag_handler_dump(struct sk_buff *skb, if (h->nlmsg_flags & NLM_F_DUMP) { struct netlink_dump_control c = { + .start = tipc_dump_start, .dump = tipc_diag_dump, + .done = tipc_dump_done, }; netlink_dump_start(net->diag_nlsk, skb, h, &c); return 0; diff --git a/net/tipc/netlink.c b/net/tipc/netlink.c index 6ff2254088f6..99ee419210ba 100644 --- a/net/tipc/netlink.c +++ b/net/tipc/netlink.c @@ -167,7 +167,9 @@ static const struct genl_ops tipc_genl_v2_ops[] = { }, { .cmd = TIPC_NL_SOCK_GET, + .start = tipc_dump_start, .dumpit = tipc_nl_sk_dump, + .done = tipc_dump_done, .policy = tipc_nl_policy, }, { diff --git a/net/tipc/socket.c b/net/tipc/socket.c index c9a50b62c738..ab7a2a7178f7 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -3229,45 +3229,69 @@ int tipc_nl_sk_walk(struct sk_buff *skb, struct netlink_callback *cb, struct netlink_callback *cb, struct tipc_sock *tsk)) { - struct net *net = sock_net(skb->sk); - struct tipc_net *tn = tipc_net(net); - const struct bucket_table *tbl; - u32 prev_portid = cb->args[1]; - u32 tbl_id = cb->args[0]; - struct rhash_head *pos; + struct rhashtable_iter *iter = (void *)cb->args[0]; struct tipc_sock *tsk; int err; - rcu_read_lock(); - tbl = rht_dereference_rcu((&tn->sk_rht)->tbl, &tn->sk_rht); - for (; tbl_id < tbl->size; tbl_id++) { - rht_for_each_entry_rcu(tsk, pos, tbl, tbl_id, node) { - spin_lock_bh(&tsk->sk.sk_lock.slock); - if (prev_portid && prev_portid != tsk->portid) { - spin_unlock_bh(&tsk->sk.sk_lock.slock); + rhashtable_walk_start(iter); + while ((tsk = rhashtable_walk_next(iter)) != NULL) { + if (IS_ERR(tsk)) { + err = PTR_ERR(tsk); + if (err == -EAGAIN) { + err = 0; continue; } + break; + } - err = skb_handler(skb, cb, tsk); - if (err) { - prev_portid = tsk->portid; - spin_unlock_bh(&tsk->sk.sk_lock.slock); - goto out; - } - - prev_portid = 0; - spin_unlock_bh(&tsk->sk.sk_lock.slock); + sock_hold(&tsk->sk); + rhashtable_walk_stop(iter); + lock_sock(&tsk->sk); + err = skb_handler(skb, cb, tsk); + if (err) { + release_sock(&tsk->sk); + sock_put(&tsk->sk); + goto out; } + release_sock(&tsk->sk); + rhashtable_walk_start(iter); + sock_put(&tsk->sk); } + rhashtable_walk_stop(iter); out: - rcu_read_unlock(); - cb->args[0] = tbl_id; - cb->args[1] = prev_portid; - return skb->len; } EXPORT_SYMBOL(tipc_nl_sk_walk); +int tipc_dump_start(struct netlink_callback *cb) +{ + struct rhashtable_iter *iter = (void *)cb->args[0]; + struct net *net = sock_net(cb->skb->sk); + struct tipc_net *tn = tipc_net(net); + + if (!iter) { + iter = kmalloc(sizeof(*iter), GFP_KERNEL); + if (!iter) + return -ENOMEM; + + cb->args[0] = (long)iter; + } + + rhashtable_walk_enter(&tn->sk_rht, iter); + return 0; +} +EXPORT_SYMBOL(tipc_dump_start); + +int tipc_dump_done(struct netlink_callback *cb) +{ + struct rhashtable_iter *hti = (void *)cb->args[0]; + + rhashtable_walk_exit(hti); + kfree(hti); + return 0; +} +EXPORT_SYMBOL(tipc_dump_done); + int tipc_sk_fill_sock_diag(struct sk_buff *skb, struct netlink_callback *cb, struct tipc_sock *tsk, u32 sk_filter_state, u64 (*tipc_diag_gen_cookie)(struct sock *sk)) diff --git a/net/tipc/socket.h b/net/tipc/socket.h index aff9b2ae5a1f..d43032e26532 100644 --- a/net/tipc/socket.h +++ b/net/tipc/socket.h @@ -68,4 +68,6 @@ int tipc_nl_sk_walk(struct sk_buff *skb, struct netlink_callback *cb, int (*skb_handler)(struct sk_buff *skb, struct netlink_callback *cb, struct tipc_sock *tsk)); +int tipc_dump_start(struct netlink_callback *cb); +int tipc_dump_done(struct netlink_callback *cb); #endif -- cgit v1.2.3 From 9cc1bf3928b31e515ed15477b3c7eb653d0b3b42 Mon Sep 17 00:00:00 2001 From: Zhenbo Gao Date: Mon, 3 Sep 2018 16:36:45 +0800 Subject: tipc: correct spelling errors for struct tipc_bc_base's comment Trivial fix for two spelling mistakes. Signed-off-by: Zhenbo Gao Reviewed-by: Ying Xue Signed-off-by: David S. Miller --- net/tipc/bcast.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'net/tipc') diff --git a/net/tipc/bcast.c b/net/tipc/bcast.c index 9ee6cfea56dd..d8026543bf4c 100644 --- a/net/tipc/bcast.c +++ b/net/tipc/bcast.c @@ -51,12 +51,12 @@ const char tipc_bclink_name[] = "broadcast-link"; * struct tipc_bc_base - base structure for keeping broadcast send state * @link: broadcast send link structure * @inputq: data input queue; will only carry SOCK_WAKEUP messages - * @dest: array keeping number of reachable destinations per bearer + * @dests: array keeping number of reachable destinations per bearer * @primary_bearer: a bearer having links to all broadcast destinations, if any * @bcast_support: indicates if primary bearer, if any, supports broadcast * @rcast_support: indicates if all peer nodes support replicast * @rc_ratio: dest count as percentage of cluster size where send method changes - * @bc_threshold: calculated drom rc_ratio; if dests > threshold use broadcast + * @bc_threshold: calculated from rc_ratio; if dests > threshold use broadcast */ struct tipc_bc_base { struct tipc_link *link; -- cgit v1.2.3 From a484ef3442d2f05fa59edf4f6d14a8169d1b94a6 Mon Sep 17 00:00:00 2001 From: Zhenbo Gao Date: Mon, 3 Sep 2018 16:36:46 +0800 Subject: tipc: correct spelling errors for tipc_topsrv_queue_evt() comments tipc_conn_queue_evt -> tipc_topsrv_queue_evt tipc_send_work -> tipc_conn_send_work tipc_send_to_sock -> tipc_conn_send_to_sock Signed-off-by: Zhenbo Gao Reviewed-by: Ying Xue Signed-off-by: David S. Miller --- net/tipc/topsrv.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'net/tipc') diff --git a/net/tipc/topsrv.c b/net/tipc/topsrv.c index c8e34ef22c30..2627b5d812e9 100644 --- a/net/tipc/topsrv.c +++ b/net/tipc/topsrv.c @@ -313,8 +313,8 @@ static void tipc_conn_send_work(struct work_struct *work) conn_put(con); } -/* tipc_conn_queue_evt() - interrupt level call from a subscription instance - * The queued work is launched into tipc_send_work()->tipc_send_to_sock() +/* tipc_topsrv_queue_evt() - interrupt level call from a subscription instance + * The queued work is launched into tipc_conn_send_work()->tipc_conn_send_to_sock() */ void tipc_topsrv_queue_evt(struct net *net, int conid, u32 event, struct tipc_event *evt) -- cgit v1.2.3