Received-SPF: pass (google.com: domain of linux-bluetooth-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <20200814190909.361764-1-sonnysasaka@chromium.org>
In-Reply-To: <20200814190909.361764-1-sonnysasaka@chromium.org>
From:   Sonny Sasaka <sonnysasaka@chromium.org>
Date:   Wed, 16 Sep 2020 15:39:58 -0700
Message-ID: <CAO271m=+2vqnmWK=M5ySfSYYBsKqAJbjmCDnjj4C_BxD3vQ5gQ@mail.gmail.com>
Subject: Re: [PATCH v2] Bluetooth: Fix auto-creation of hci_conn at Conn
 Complete event
To:     BlueZ <linux-bluetooth@vger.kernel.org>
Cc:     Abhishek Pandit-Subedi <abhishekpandit@chromium.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

Dear maintainers,

Friendly ping to review this patch. Thanks!

On Fri, Aug 14, 2020 at 12:09 PM Sonny Sasaka <sonnysasaka@chromium.org> wrote:
>
> Currently the code auto-creates hci_conn only if the remote address has
> been discovered before. This may not be the case. For example, the
> remote device may trigger connection after reboot at already-paired
> state so there is no inquiry result found, but it is still correct to
> create the hci_conn when Connection Complete event is received.
>
> A better guard is to check against bredr allowlist. Devices in the
> allowlist have been given permission to auto-connect.
>
> Fixes: 4f40afc6c764 ("Bluetooth: Handle BR/EDR devices during suspend")
> Signed-off-by: Sonny Sasaka <sonnysasaka@chromium.org>
> Reviewed-by: Abhishek Pandit-Subedi <abhishekpandit@chromium.org>
>
> ---
>  net/bluetooth/hci_event.c | 17 +++++++++++------
>  1 file changed, 11 insertions(+), 6 deletions(-)
>
> diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c
> index 33d8458fdd4a..6b83da1790ee 100644
> --- a/net/bluetooth/hci_event.c
> +++ b/net/bluetooth/hci_event.c
> @@ -2566,7 +2566,6 @@ static void hci_inquiry_result_evt(struct hci_dev *hdev, struct sk_buff *skb)
>  static void hci_conn_complete_evt(struct hci_dev *hdev, struct sk_buff *skb)
>  {
>         struct hci_ev_conn_complete *ev = (void *) skb->data;
> -       struct inquiry_entry *ie;
>         struct hci_conn *conn;
>
>         BT_DBG("%s", hdev->name);
> @@ -2575,13 +2574,19 @@ static void hci_conn_complete_evt(struct hci_dev *hdev, struct sk_buff *skb)
>
>         conn = hci_conn_hash_lookup_ba(hdev, ev->link_type, &ev->bdaddr);
>         if (!conn) {
> -               /* Connection may not exist if auto-connected. Check the inquiry
> -                * cache to see if we've already discovered this bdaddr before.
> -                * If found and link is an ACL type, create a connection class
> +               /* Connection may not exist if auto-connected. Check the bredr
> +                * allowlist to see if this device is allowed to auto connect.
> +                * If link is an ACL type, create a connection class
>                  * automatically.
> +                *
> +                * Auto-connect will only occur if the event filter is
> +                * programmed with a given address. Right now, event filter is
> +                * only used during suspend.
>                  */
> -               ie = hci_inquiry_cache_lookup(hdev, &ev->bdaddr);
> -               if (ie && ev->link_type == ACL_LINK) {
> +               if (ev->link_type == ACL_LINK &&
> +                   hci_bdaddr_list_lookup_with_flags(&hdev->whitelist,
> +                                                     &ev->bdaddr,
> +                                                     BDADDR_BREDR)) {
>                         conn = hci_conn_add(hdev, ev->link_type, &ev->bdaddr,
>                                             HCI_ROLE_SLAVE);
>                         if (!conn) {
> --
> 2.26.2
>