2018-04-19 11:03:09

by Chriz Chow

Subject: [PATCH] Bluetooth: Prevent buffer overflow for large advertisement data

In the Bluetooth Core Specifications 5.0 it has no limitation of
the length of Advertising or Scan Response Data. Therefore,
there are some devices sending out advertising data longer than
HCI_MAX_AD_LENGTH, causing the buffer last_adv_data overflows.

It prevents the issue by checking the data length before copying.

Signed-off-by: Chriz Chow <[email protected]>
---
net/bluetooth/hci_event.c | 3 +++
1 file changed, 3 insertions(+)

diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c
index 139707c..1300bb0 100644
--- a/net/bluetooth/hci_event.c
+++ b/net/bluetooth/hci_event.c
@@ -1118,6 +1118,9 @@ static void store_pending_adv_report(struct hci_dev *hdev, bdaddr_t *bdaddr,
{
struct discovery_state *d = &hdev->discovery;

+ if (len > HCI_MAX_AD_LENGTH)
+ return;
+
bacpy(&d->last_adv_addr, bdaddr);
d->last_adv_addr_type = bdaddr_type;
d->last_adv_rssi = rssi;
--
2.7.4

2018-04-23 17:58:43

by Marcel Holtmann

Subject: Re: [PATCH v2] Bluetooth: Prevent buffer overflow for large advertisement data

Hi Chriz,

> There are some controllers sending out advertising data with illegal
> length value which is longer than HCI_MAX_AD_LENGTH, causing the
> buffer last_adv_data overflows. To avoid these controllers from
> overflowing the buffer, we do not process the advertisement data
> if its length is incorrect.
>
> Signed-off-by: Chriz Chow <[email protected]>
> ---
> net/bluetooth/hci_event.c | 12 ++++++++----
> 1 file changed, 8 insertions(+), 4 deletions(-)

patch has been applied to bluetooth-next tree.

Regards

Marcel

2018-04-20 07:46:24

by Chriz Chow

Subject: [PATCH v2] Bluetooth: Prevent buffer overflow for large advertisement data

There are some controllers sending out advertising data with illegal
length value which is longer than HCI_MAX_AD_LENGTH, causing the
buffer last_adv_data overflows. To avoid these controllers from
overflowing the buffer, we do not process the advertisement data
if its length is incorrect.

Signed-off-by: Chriz Chow <[email protected]>
---
net/bluetooth/hci_event.c | 12 ++++++++----
1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c
index 139707c..dc4cfd5 100644
--- a/net/bluetooth/hci_event.c
+++ b/net/bluetooth/hci_event.c
@@ -4942,10 +4942,14 @@ static void hci_le_adv_report_evt(struct hci_dev *hdev, struct sk_buff *skb)
struct hci_ev_le_advertising_info *ev = ptr;
s8 rssi;

- rssi = ev->data[ev->length];
- process_adv_report(hdev, ev->evt_type, &ev->bdaddr,
- ev->bdaddr_type, NULL, 0, rssi,
- ev->data, ev->length);
+ if (ev->length <= HCI_MAX_AD_LENGTH) {
+ rssi = ev->data[ev->length];
+ process_adv_report(hdev, ev->evt_type, &ev->bdaddr,
+ ev->bdaddr_type, NULL, 0, rssi,
+ ev->data, ev->length);
+ } else {
+ BT_ERR("Dropped invalid adv data");
+ }

ptr += sizeof(*ev) + ev->length + 1;
}
--
2.7.4

2018-04-19 12:28:52

by Szymon Janc

Subject: Re: [PATCH] Bluetooth: Prevent buffer overflow for large advertisement data

Hi Chriz,

On Thursday, 19 April 2018 13:03:09 CEST Chriz Chow wrote:
> In the Bluetooth Core Specifications 5.0 it has no limitation of
> the length of Advertising or Scan Response Data. Therefore,
> there are some devices sending out advertising data longer than
> HCI_MAX_AD_LENGTH, causing the buffer last_adv_data overflows.
>
> It prevents the issue by checking the data length before copying.
>
> Signed-off-by: Chriz Chow <[email protected]>
> ---
> net/bluetooth/hci_event.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c
> index 139707c..1300bb0 100644
> --- a/net/bluetooth/hci_event.c
> +++ b/net/bluetooth/hci_event.c
> @@ -1118,6 +1118,9 @@ static void store_pending_adv_report(struct hci_dev
> *hdev, bdaddr_t *bdaddr, {
> struct discovery_state *d = &hdev->discovery;
>
> + if (len > HCI_MAX_AD_LENGTH)
> + return;
> +
> bacpy(&d->last_adv_addr, bdaddr);
> d->last_adv_addr_type = bdaddr_type;
> d->last_adv_rssi = rssi;

This would indicate that controller is sending LE Advertising Report Event
with illegal length value (valid range is 0x00-0x1f). If there are such broken
controllers around we could check those but I think this should be done as
soon as possible ie. in hci_le_adv_report_evt() function.

Also, >31 bytes of advertising data is only for Extended Advertising which
would require using LE Set Extended Scan commands and would result in LE
Extended Advertising Reports being generated instead (none of those are yet
supported in kernel).

--
pozdrawiam
Szymon Janc