linux-kernel - Re: [PATCH v5 2/5] Bluetooth: hci_qca: Deassert RTS while baudrate change command

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:   Mon, 21 Jan 2019 09:56:13 +0100
From:   Johan Hovold <johan@...nel.org>
To:     Matthias Kaehlcke <mka@...omium.org>
Cc:     Johan Hovold <johan@...nel.org>,
        Balakrishna Godavarthi <bgodavar@...eaurora.org>,
        marcel@...tmann.org, johan.hedberg@...il.com,
        linux-kernel@...r.kernel.org, linux-bluetooth@...r.kernel.org,
        hemantg@...eaurora.org, linux-arm-msm@...r.kernel.org,
        Johan Hovold <jhovold@...il.com>
Subject: Re: [PATCH v5 2/5] Bluetooth: hci_qca: Deassert RTS while baudrate
 change command

On Fri, Jan 18, 2019 at 04:31:09PM -0800, Matthias Kaehlcke wrote:
> On Fri, Jan 18, 2019 at 10:44:16AM +0100, Johan Hovold wrote:
> > On Thu, Jan 17, 2019 at 09:21:09AM -0800, Matthias Kaehlcke wrote:
> > 
> > > I observed that the qcom_geni_serial driver doesn't raise RTS with
> > > flow control disabled. It seems we have to investigate why that's the
> > > case. I agree that the driver should be platform agnostic.
> > 
> > Sounds like a driver bug, unless the hardware is just "odd". The
> > driver implementation of this looks very non-standard judging from a
> > quick peek.
> > 
> > In fact, qcom_geni_serial_get_mctrl() is currently a no-op if hardware
> > flow control is not enabled:
> > 
> > 	if (uart_console(uport) || !uart_cts_enabled(uport))
> > 		return;
> > 
> > Perhaps dropping the !uart_cts_enabled() test is sufficient.
> 
> Thanks for taking a look Johan, that was indeed the problem (also
> in set_mctrl()). I posted a fix: https://lore.kernel.org/patchwork/patch/1033611/

Nice (I did mean set_mctrl() above, as I think you noticed).

> Balakrishna, the following (applied on top of your patch) works for me
> with the UART patch above:
> 
> diff --git a/drivers/bluetooth/hci_qca.c b/drivers/bluetooth/hci_qca.c
> index 9d5e41f159c78f..60bfdf01f72841 100644
> --- a/drivers/bluetooth/hci_qca.c
> +++ b/drivers/bluetooth/hci_qca.c
> @@ -1080,7 +1080,7 @@ static int qca_set_speed(struct hci_uart *hu, enum qca_speed_type speed_type)
>  {
>  	unsigned int speed, qca_baudrate;
>  	struct qca_serdev *qcadev;
> -	int ret;
> +	int ret = 0;
>  
>  	if (speed_type == QCA_INIT_SPEED) {
>  		speed = qca_get_speed(hu, QCA_INIT_SPEED);
> @@ -1097,22 +1097,27 @@ static int qca_set_speed(struct hci_uart *hu, enum qca_speed_type speed_type)
>  		 * the old speed.
>  		 */
>  		qcadev = serdev_device_get_drvdata(hu->serdev);
> -		if (qcadev->btsoc_type == QCA_WCN3990)
> +		if (qcadev->btsoc_type == QCA_WCN3990) {
> +			hci_uart_set_flow_control(hu, true);

Wow, yeah, this parameter inversion is indeed confusing...

>  			serdev_device_set_rts(hu->serdev, false);
> +		}

But looking at hci_uart_set_flow_control() now, it actually also
deasserts RTS. So all you need here is the hci_uart_set_flow_control()
call.  

And that makes the inversion make a bit more sense too, even if the
naming is a bit unfortunate with respect to
serdev_device_set_flow_control() at least.

>  		qca_baudrate = qca_get_baudrate_value(speed);
>  		bt_dev_dbg(hu->hdev, "Set UART speed to %d", speed);
>  		ret = qca_set_baudrate(hu->hdev, qca_baudrate);
>  		if (ret)
> -			return ret;
> +			goto out;
>  
>  		host_set_baudrate(hu, speed);
>  
> -		if (qcadev->btsoc_type == QCA_WCN3990)
> +out:
> +		if (qcadev->btsoc_type == QCA_WCN3990) {
> +			hci_uart_set_flow_control(hu, false);
>  			serdev_device_set_rts(hu->serdev, true);
> +		}

And same here.

Johan