| From | Greg Kroah-Hartman <> | Subject | [PATCH 4.1 157/159] inet: fix races with reqsk timers | Date | Sat, 26 Sep 2015 13:56:43 -0700 |
| |
4.1-stable review patch. If anyone has any objections, please let me know.
------------------
From: Eric Dumazet <edumazet@google.com>
[ Upstream commit 2235f2ac75fd2501c251b0b699a9632e80239a6d ]
reqsk_queue_destroy() and reqsk_queue_unlink() should use del_timer_sync() instead of del_timer() before calling reqsk_put(), otherwise we could free a req still used by another cpu.
But before doing so, reqsk_queue_destroy() must release syn_wait_lock spinlock or risk a dead lock, as reqsk_timer_handler() might need to take this same spinlock from reqsk_queue_unlink() (called from inet_csk_reqsk_queue_drop())
Fixes: fa76ce7328b2 ("inet: get rid of central tcp/dccp listener timer") Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> --- net/core/request_sock.c | 8 +++++++- net/ipv4/inet_connection_sock.c | 2 +- 2 files changed, 8 insertions(+), 2 deletions(-)
--- a/net/core/request_sock.c +++ b/net/core/request_sock.c @@ -103,10 +103,16 @@ void reqsk_queue_destroy(struct request_ spin_lock_bh(&queue->syn_wait_lock); while ((req = lopt->syn_table[i]) != NULL) { lopt->syn_table[i] = req->dl_next; + /* Because of following del_timer_sync(), + * we must release the spinlock here + * or risk a dead lock. + */ + spin_unlock_bh(&queue->syn_wait_lock); atomic_inc(&lopt->qlen_dec); - if (del_timer(&req->rsk_timer)) + if (del_timer_sync(&req->rsk_timer)) reqsk_put(req); reqsk_put(req); + spin_lock_bh(&queue->syn_wait_lock); } spin_unlock_bh(&queue->syn_wait_lock); } --- a/net/ipv4/inet_connection_sock.c +++ b/net/ipv4/inet_connection_sock.c @@ -584,7 +584,7 @@ static bool reqsk_queue_unlink(struct re } spin_unlock(&queue->syn_wait_lock); - if (del_timer(&req->rsk_timer)) + if (del_timer_sync(&req->rsk_timer)) reqsk_put(req); return found; }
|