diff options
| author | Israel Rukshin <israelr@nvidia.com> | 2021-10-06 11:09:45 +0300 |
|---|---|---|
| committer | Greg Kroah-Hartman <gregkh@linuxfoundation.org> | 2021-11-18 21:16:18 +0300 |
| commit | 0703920365d306b9bee2760ce59f32a385404967 (patch) | |
| tree | 982aea5b59fcd843888a73af55ffa53ed30b5908 /drivers/nvme | |
| parent | 3c82292ee9bc0d3144d03ed0f96344ebf9e27453 (diff) | |
| download | linux-0703920365d306b9bee2760ce59f32a385404967.tar.xz | |
nvmet-tcp: fix use-after-free when a port is removed
[ Upstream commit 2351ead99ce9164fb42555aee3f96af84c4839e9 ]
When removing a port, all its controllers are being removed, but there
are queues on the port that doesn't belong to any controller (during
connection time). This causes a use-after-free bug for any command
that dereferences req->port (like in nvmet_alloc_ctrl). Those queues
should be destroyed before freeing the port via configfs. Destroy
the remaining queues after the accept_work was cancelled guarantees
that no new queue will be created.
Signed-off-by: Israel Rukshin <israelr@nvidia.com>
Reviewed-by: Max Gurtovoy <mgurtovoy@nvidia.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
Diffstat (limited to 'drivers/nvme')
| -rw-r--r-- | drivers/nvme/target/tcp.c | 16 |
1 files changed, 16 insertions, 0 deletions
diff --git a/drivers/nvme/target/tcp.c b/drivers/nvme/target/tcp.c index 46c3b3be7e03..84c387e4bf43 100644 --- a/drivers/nvme/target/tcp.c +++ b/drivers/nvme/target/tcp.c @@ -1740,6 +1740,17 @@ err_port: return ret; } +static void nvmet_tcp_destroy_port_queues(struct nvmet_tcp_port *port) +{ + struct nvmet_tcp_queue *queue; + + mutex_lock(&nvmet_tcp_queue_mutex); + list_for_each_entry(queue, &nvmet_tcp_queue_list, queue_list) + if (queue->port == port) + kernel_sock_shutdown(queue->sock, SHUT_RDWR); + mutex_unlock(&nvmet_tcp_queue_mutex); +} + static void nvmet_tcp_remove_port(struct nvmet_port *nport) { struct nvmet_tcp_port *port = nport->priv; @@ -1749,6 +1760,11 @@ static void nvmet_tcp_remove_port(struct nvmet_port *nport) port->sock->sk->sk_user_data = NULL; write_unlock_bh(&port->sock->sk->sk_callback_lock); cancel_work_sync(&port->accept_work); + /* + * Destroy the remaining queues, which are not belong to any + * controller yet. + */ + nvmet_tcp_destroy_port_queues(port); sock_release(port->sock); kfree(port); |