netdev - Re: [Patch net v2 1/3] bluetooth: validate HCI_EVENT

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 23 Apr 2019 21:42:40 +0200
From:   Marcel Holtmann <marcel@...tmann.org>
To:     Cong Wang <xiyou.wangcong@...il.com>
Cc:     netdev@...r.kernel.org, linux-bluetooth@...r.kernel.org,
        syzbot+cec7a50c412a2c03f8f5@...kaller.appspotmail.com,
        syzbot+660883c56e2fa65d4497@...kaller.appspotmail.com,
        Johan Hedberg <johan.hedberg@...il.com>,
        Dan Carpenter <dan.carpenter@...cle.com>,
        Tomas Bortoli <tomasbortoli@...il.com>
Subject: Re: [Patch net v2 1/3] bluetooth: validate HCI_EVENT_PKT packet
 carefully

Hi Cong,

> hci_event_packet() blindly assumes all packets are sane, at least
> for packets allocated via vhci_get_user() path this is not true.
> We have to check if we access skb data out-of-bound with
> pskb_may_pull() before each skb->data dereference on RX path.
> 
> Reported-and-tested-by: syzbot+cec7a50c412a2c03f8f5@...kaller.appspotmail.com
> Reported-and-tested-by: syzbot+660883c56e2fa65d4497@...kaller.appspotmail.com
> Cc: Marcel Holtmann <marcel@...tmann.org>
> Cc: Johan Hedberg <johan.hedberg@...il.com>
> Cc: Dan Carpenter <dan.carpenter@...cle.com>
> Reviewed-by: Tomas Bortoli <tomasbortoli@...il.com>
> Signed-off-by: Cong Wang <xiyou.wangcong@...il.com>
> ---
> net/bluetooth/hci_event.c | 262 +++++++++++++++++++++++++++++++-------
> 1 file changed, 218 insertions(+), 44 deletions(-)
> 
> diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c
> index 609fd6871c5a..2fef70c0bffe 100644
> --- a/net/bluetooth/hci_event.c
> +++ b/net/bluetooth/hci_event.c
> @@ -2331,10 +2331,13 @@ static void hci_cs_switch_role(struct hci_dev *hdev, u8 status)
> 
> static void hci_inquiry_complete_evt(struct hci_dev *hdev, struct sk_buff *skb)
> {
> -	__u8 status = *((__u8 *) skb->data);
> 	struct discovery_state *discov = &hdev->discovery;
> 	struct inquiry_entry *e;
> +	__u8 status;
> 
> +	if (unlikely(!pskb_may_pull(skb, 1)))
> +		return;
> +	status = *((__u8 *)skb->data);
> 	BT_DBG("%s status 0x%2.2x", hdev->name, status);
> 
> 	hci_conn_check_pending(hdev);
> @@ -2391,14 +2394,21 @@ static void hci_inquiry_complete_evt(struct hci_dev *hdev, struct sk_buff *skb)
> static void hci_inquiry_result_evt(struct hci_dev *hdev, struct sk_buff *skb)
> {
> 	struct inquiry_data data;
> -	struct inquiry_info *info = (void *) (skb->data + 1);
> -	int num_rsp = *((__u8 *) skb->data);
> +	struct inquiry_info *info;
> +	int num_rsp;
> 
> 	BT_DBG("%s num_rsp %d", hdev->name, num_rsp);
> 
> +	if (unlikely(!pskb_may_pull(skb, 1)))
> +		return;
> +	num_rsp = *((__u8 *)skb->data);
> 	if (!num_rsp)
> 		return;
> 
> +	if (unlikely(!pskb_may_pull(skb, 1 + num_rsp * sizeof(*info))))
> +		return;
> +	info = (void *)(skb->data + 1);
> +
> 	if (hci_dev_test_flag(hdev, HCI_PERIODIC_INQ))
> 		return;

this really looks like we better create a macro for this. It is repetitive code that can be turned into just a macro usage.

Regards

Marcel