linux-kernel - Re: [PATCH v3 2/4] Bluetooth: hci_qca: Retry btsoc initialize when it fails

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]

Message-ID: <20200103162750.GC89495@google.com>
Date:   Fri, 3 Jan 2020 08:27:50 -0800
From:   Matthias Kaehlcke <mka@...omium.org>
To:     rjliao@...eaurora.org
Cc:     marcel@...tmann.org, johan.hedberg@...il.com,
        linux-kernel@...r.kernel.org, linux-bluetooth@...r.kernel.org,
        linux-arm-msm@...r.kernel.org,
        linux-bluetooth-owner@...r.kernel.org
Subject: Re: [PATCH v3 2/4] Bluetooth: hci_qca: Retry btsoc initialize when
 it fails

On Fri, Jan 03, 2020 at 02:31:46PM +0800, rjliao@...eaurora.org wrote:
> 在 2020-01-03 02:41，Matthias Kaehlcke 写道：
> 
> > Hi Rocky,
> > 
> > On Fri, Dec 27, 2019 at 03:21:28PM +0800, Rocky Liao wrote:
> > 
> > > This patch adds the retry of btsoc initialization when it fails.
> > > There are
> > > reports that the btsoc initialization may fail on some platforms but
> > > the
> > > repro ratio is very low. The failure may be caused by UART, platform
> > > HW or
> > > the btsoc itself but it's very difficlut to root cause, given the
> > > repro
> > > ratio is very low. Add a retry for the btsoc initialization will
> > > resolve
> > > most of the failures and make Bluetooth finally works.
> > 
> > Is this problem specific to a certain chipset?
> > 
> > What are the symptoms?
> 
> It's reported on Rome so far but I think the patch is potentially helpful
> for
> wcn399x as well.
> 
> The symptoms is the firmware downloading failed due to the UART write timed
> out.

Working around this with retries seems ok for now if the repro rate is
really low, but it shouldn't necessarily be interpreted as "the problem is
fixed". Please mention the symptoms in the commit message for documentation,
then the retries can potentially be removed in the futures when the root
cause is fixed.

> > > enum qca_btsoc_type soc_type = qca_soc_type(hu);
> > > const char *firmware_name = qca_get_firmware_name(hu);
> > > int ret;
> > > @@ -1275,6 +1280,7 @@ static int qca_setup(struct hci_uart *hu)
> > > */
> > > set_bit(HCI_QUIRK_SIMULTANEOUS_DISCOVERY, &hdev->quirks);
> > > 
> > > +retry:
> > > if (qca_is_wcn399x(soc_type)) {
> > > bt_dev_info(hdev, "setting up wcn3990");
> > > 
> > > @@ -1293,6 +1299,12 @@ static int qca_setup(struct hci_uart *hu)
> > > return ret;
> > > } else {
> > > bt_dev_info(hdev, "ROME setup");
> > > +        if (hu->serdev) {
> > > +            qcadev = serdev_device_get_drvdata(hu->serdev);
> > > +            gpiod_set_value_cansleep(qcadev->bt_en, 1);
> > > +            /* Controller needs time to bootup. */
> > > +            msleep(150);
> > 
> > Shouldn't this be in qca_power_on(), analogous to the power off code
> > from
> > "[1/4]Bluetooth: hci_qca: Add QCA Rome power off support to the
> > qca_power_off()"?
> > 
> > qca_power_on() should then also be called for ROME. If you opt for this
> > it
> > should be done in a separate patch, or possibly merged into the one
> > mentioned above.
> > 
> 
> There is no qca_power_on() func and wcn399x is calling qca_wcn3990_init() to
> do power on, I prefer to not do this change this time.

I would say it's precisely the right time to add this function. Patch 1 of this
series adds handling of the BT_EN GPIO to qca_power_off(), now this patch
duplicates the code of the BT_EN handling in qca_open().

> If it's needed

'needed' is a relative term. It certainly isn't needed from a purely functional
POV. However it is desirable for encapsulation and to avoid code duplication.

> it should be a new patch to add qca_power_on() which supports both Rome and wcn399x.

Agreed

m.