[PATCH] bluetooth: handle value within the ida range should not be handled in BIG

Edward Adam Davis posted 1 patch 1 year, 6 months ago
net/bluetooth/hci_event.c | 3 +++
1 file changed, 3 insertions(+)
[PATCH] bluetooth: handle value within the ida range should not be handled in BIG
Posted by Edward Adam Davis 1 year, 6 months ago
hci_le_big_sync_established_evt is necessary to filter out cases where the handle
value is belone to ida id range, otherwise ida will be erroneously released in 
hci_conn_cleanup. 

Fixes: 181a42edddf5 ("Bluetooth: Make handle of hci_conn be unique")
Reported-by: syzbot+b2545b087a01a7319474@syzkaller.appspotmail.com
Signed-off-by: Edward Adam Davis <eadavis@qq.com>
---
 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 a487f9df8145..4130d64d9a80 100644
--- a/net/bluetooth/hci_event.c
+++ b/net/bluetooth/hci_event.c
@@ -6893,6 +6893,9 @@ static void hci_le_big_sync_established_evt(struct hci_dev *hdev, void *data,
 
 		bis = hci_conn_hash_lookup_handle(hdev, handle);
 		if (!bis) {
+                        if (handle > HCI_CONN_HANDLE_MAX)
+                               continue;
+
 			bis = hci_conn_add(hdev, ISO_LINK, BDADDR_ANY,
 					   HCI_ROLE_SLAVE, handle);
 			if (IS_ERR(bis))
-- 
2.43.0
Re: [PATCH] bluetooth: handle value within the ida range should not be handled in BIG
Posted by Pavel Skripkin 1 year, 6 months ago

On 6/16/24 1:20 PM, 'Edward Adam Davis' via syzkaller-bugs wrote:
> hci_le_big_sync_established_evt is necessary to filter out cases where the handle
> value is belone to ida id range, otherwise ida will be erroneously released in
> hci_conn_cleanup.
> 
> Fixes: 181a42edddf5 ("Bluetooth: Make handle of hci_conn be unique")
> Reported-by: syzbot+b2545b087a01a7319474@syzkaller.appspotmail.com
> Signed-off-by: Edward Adam Davis <eadavis@qq.com>
> ---

There is one more user of  `hci_conn_add` which may pass too big handle 
which is `hci_le_cis_req_evt`.

I think, it should be resolved on API level as I tried to test here [0], 
but syzbot is feeling bad for some reason


[0] 
https://lore.kernel.org/all/31ac448d-2a21-4e93-8a00-5c7090970452@gmail.com/

>   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 a487f9df8145..4130d64d9a80 100644
> --- a/net/bluetooth/hci_event.c
> +++ b/net/bluetooth/hci_event.c
> @@ -6893,6 +6893,9 @@ static void hci_le_big_sync_established_evt(struct hci_dev *hdev, void *data,
>   
>   		bis = hci_conn_hash_lookup_handle(hdev, handle);
>   		if (!bis) {
> +                        if (handle > HCI_CONN_HANDLE_MAX)
> +                               continue;
> +
>   			bis = hci_conn_add(hdev, ISO_LINK, BDADDR_ANY,
>   					   HCI_ROLE_SLAVE, handle);
>   			if (IS_ERR(bis))


-- 
With regards,
Pavel Skripkin
Re: [syzbot] [bluetooth?] WARNING in hci_conn_del
Posted by Edward Adam Davis 1 year, 6 months ago
On Sun, 16 Jun 2024 13:37:17 +0300, Pavel Skripkin wrote:
> > hci_le_big_sync_established_evt is necessary to filter out cases where the handle
> > value is belone to ida id range, otherwise ida will be erroneously released in
> > hci_conn_cleanup.
> >
> > Fixes: 181a42edddf5 ("Bluetooth: Make handle of hci_conn be unique")
> > Reported-by: syzbot+b2545b087a01a7319474@syzkaller.appspotmail.com
> > Signed-off-by: Edward Adam Davis <eadavis@qq.com>
> > ---
> 
> There is one more user of  `hci_conn_add` which may pass too big handle
> which is `hci_le_cis_req_evt`.
This case only affect hci_le_big_sync_established_evt.

--
Edward