linux-kernel - Re: [PATCH v2] Bluetooth: btrtl: HCI reset on close for Realtek BT chip

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, 25 Jun 2019 13:10:20 +0800
From:   Daniel Drake <drake@...lessm.com>
To:     Jian-Hong Pan <jian-hong@...lessm.com>
Cc:     Johan Hedberg <johan.hedberg@...il.com>,
        Marcel Holtmann <marcel@...tmann.org>,
        Linux Bluetooth mailing list 
        <linux-bluetooth@...r.kernel.org>,
        Linux Kernel <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2] Bluetooth: btrtl: HCI reset on close for Realtek BT chip

On Mon, Jun 24, 2019 at 2:24 PM Jian-Hong Pan <jian-hong@...lessm.com> wrote:
> Realtek RTL8822BE BT chip on ASUS X420FA cannot be turned on correctly
> after on-off several times. Bluetooth daemon sets BT mode failed when
> this issue happens.

You could also mention that scanning must be active while turning off
for this bug to be hit.

> bluetoothd[1576]: Failed to set mode: Failed (0x03)
>
> If BT is tunred off, then turned on again, it works correctly again.

Typo: turned

> According to the vendor driver, the HCI_QUIRK_RESET_ON_CLOSE flag is set
> during probing. So, this patch makes Realtek's BT reset on close to fix
> this issue.

Checked the vendor driver - I see what you are referring to, so the
change seems correct.

#if HCI_VERSION_CODE >= KERNEL_VERSION(3, 7, 1)
    if (!reset)
        set_bit(HCI_QUIRK_RESET_ON_CLOSE, &hdev->quirks);
    RTKBT_DBG("set_bit(HCI_QUIRK_RESET_ON_CLOSE, &hdev->quirks);");
#endif

However I'm pretty sure this is not saying that kernel 3.7.0 did not
need the reset. I think it just means that the flag did not exist
before Linux-3.7.1, so they added the ifdef to add some level of
compatibility with older kernel versions. I think you can remove
"since kernel v3.7.1." from the comment.

After those changes you can add:

Link: https://bugzilla.kernel.org/show_bug.cgi?id=203429
Reviewed-by: Daniel Drake <drake@...lessm.com>


> Signed-off-by: Jian-Hong Pan <jian-hong@...lessm.com>
> ---
> v2:
>  - According to the vendor driver, it makes "all" Realtek's BT reset on
>    close. So, this version makes it the same.
>  - Change to the new subject for all Realtek BT chips.
>
>  drivers/bluetooth/btrtl.c | 4 ++++
>  1 file changed, 4 insertions(+)
>
> diff --git a/drivers/bluetooth/btrtl.c b/drivers/bluetooth/btrtl.c
> index 208feef63de4..be6d5f7e1e44 100644
> --- a/drivers/bluetooth/btrtl.c
> +++ b/drivers/bluetooth/btrtl.c
> @@ -630,6 +630,10 @@ int btrtl_setup_realtek(struct hci_dev *hdev)
>                 return PTR_ERR(btrtl_dev);
>
>         ret = btrtl_download_firmware(hdev, btrtl_dev);
> +       /* According to the vendor driver, BT must be reset on close to avoid
> +        * firmware crash since kernel v3.7.1.
> +        */
> +       set_bit(HCI_QUIRK_RESET_ON_CLOSE, &hdev->quirks);
>
>         btrtl_free(btrtl_dev);
>
> --
> 2.22.0
>