netdev - Re: [PATCH] net: qmi_wwan: Add USB IDs for MDM6600 modem on Motorola Droid 4

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:   Sun, 19 Mar 2017 18:20:59 +0100
From:   Bjørn Mork <bjorn@...k.no>
To:     Tony Lindgren <tony@...mide.com>
Cc:     David Miller <davem@...emloft.net>, netdev@...r.kernel.org,
        linux-omap@...r.kernel.org, Marcel Partap <mpartap@....net>,
        Michael Scott <michael.scott@...aro.org>
Subject: Re: [PATCH] net: qmi_wwan: Add USB IDs for MDM6600 modem on Motorola Droid 4

Tony Lindgren <tony@...mide.com> writes:
> * Bjørn Mork <bjorn@...k.no> [170319 09:33]:
>> Tony Lindgren <tony@...mide.com> writes:
>> 
>> > This gets qmicli working with the MDM6600 modem.
>> >
>> > Cc: Bjørn Mork <bjorn@...k.no>
>> > Reviewed-by: Sebastian Reichel <sre@...nel.org>
>> > Tested-by: Sebastian Reichel <sre@...nel.org>
>> > Signed-off-by: Tony Lindgren <tony@...mide.com>
>> > ---
>> >  drivers/net/usb/qmi_wwan.c | 4 ++++
>> >  1 file changed, 4 insertions(+)
>> >
>> > diff --git a/drivers/net/usb/qmi_wwan.c b/drivers/net/usb/qmi_wwan.c
>> > --- a/drivers/net/usb/qmi_wwan.c
>> > +++ b/drivers/net/usb/qmi_wwan.c
>> > @@ -580,6 +580,10 @@ static const struct usb_device_id products[] = {
>> >  		USB_VENDOR_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, USB_CLASS_VENDOR_SPEC, 0x01, 0x69),
>> >  		.driver_info        = (unsigned long)&qmi_wwan_info,
>> >  	},
>> > +	{	/* Motorola Mapphone devices with MDM6600 */
>> > +		USB_VENDOR_AND_INTERFACE_INFO(0x22b8, USB_CLASS_VENDOR_SPEC, 0xfb, 0xff),
>> 
>> 
>> This is a bit unusual, so I'd like to verify that it is correct.  Do you
>> happen to have a "lsusb -v" or /sys/kernel/debug/usb/devices dump for
>> this device?  Is this usage of vendor + class consistent with the
>> Windows driver *.inf data?  Are you sure that the ff/fb/ff class is only
>> used for QMI functions by this vendor ID? 
>
> Well this is based on what the earlier Motorola mapphone v3.8 kernel has
> for in drivers/net/usb/qcusbnet/qcusbnet.c driver:
>
> +       { /* Motorola Xoom */
> +       USB_DEVICE_AND_INTERFACE_INFO(0x22B8, 0x2A70, 0xff, 0xfb, 0xff),
> +       .driver_info = (unsigned long)&xoom_qc_netinfo
> +       },
> +       { /* MDM9600 */
> +       USB_DEVICE_AND_INTERFACE_INFO(0x22B8, 0x2E0A, 0xff, 0xfb, 0xff),
> +       .driver_info = (unsigned long)&mdm9600_qc_netinfo
> +       },

That helps a lot. Was not aware of this driver.

> Where the "Motorola Xoom" id is also used for all mapphone devices
> with MDM6600 variants.
>
> Then xoom_qc_netinfo in the mapphone kernel has:
>
> +static const struct driver_info xoom_qc_netinfo = {
> +       .description   = "Xoom QCUSBNet Ethernet Device",
> +       .flags         = FLAG_ETHER|FLAG_SEND_ZLP,
> +       .bind          = xoom_qcnet_bind,
> +       .unbind        = qcnet_unbind,
> +       .data          = 0,
> +       .manage_power  = qcnet_manage_power,
> +};
>
> And the v3.8 kernel also has drivers/usb/serial/mdm6600.c:
>
> +static const struct usb_device_id mdm6600_id_table[] = {
> +       { USB_DEVICE_AND_INTERFACE_INFO(0x22b8, 0x2a70, 0xff, 0xff, 0xff) },
> +        /* MDM9600 */
> +       { USB_DEVICE_AND_INTERFACE_INFO(0x22b8, 0x2e0a, 0xff, 0xff, 0xff) },
> +       { USB_DEVICE_AND_INTERFACE_INFO(0x22b8, 0x900e, 0xff, 0xff, 0xff) },
> +       { },
> +};


Yes, looks like they consitently use ff/ff/ff for serial functions and
ff/fb/ff for QMI functions.  So adding a vendor rule seems appropriate.


> And then in drivers/usb/serial/moto_flashqsc.c:
>
> +static struct usb_device_id id_table[] = {
> +       {USB_DEVICE_AND_INTERFACE_INFO(0x22b8, 0x2a63, 0x0a, 0, 0)},
> +       {USB_DEVICE_AND_INTERFACE_INFO(0x22b8, 0x4281, 0x0a, 0, 0xfc)},
> +       {USB_DEVICE_AND_INTERFACE_INFO(0x22b8, 0x2db4, 0x0a, 0, 0xfc)},
> +       {USB_DEVICE(0x22b8, 0x4260)},
> +       {USB_DEVICE(0x22b8, 0x426D)},
> +       {},
> +};

This on the other hand, is something I hope I don't have to review :)
The 0x0a class (CDC Data) is always part of a multi-interface function,
and you would normally match on the control interface.

> Where the 0x4260 and 0x426d seem to be for the flash mode of the
> Wrigley3GLTE modem.
>
> See also lsusb -v output below. No idea if there's a Windows driver
> .inf file for this. Most likely whatever Windows driver is just using the
> generic Android USB driver(s). I know the USB on droid 4 can be multiplexed
> to have MDM6600 directly accessed, but I think that's only used for
> debugging the modem as that mode needs to be selected in the bootloader
> temporarily using volume keys.
>
> With the configuration in my patch, modprobe of qmi_wwan produces four
> wwan interfaces if that matters.

And I assume they all work?

>> The reason I ask is that I'd hate to have reports of other Motorola
>> devices where ff/fb/ff was used for some other USB function.  Yes, that
>> would be stupid.  But still... Experience shows that we cannot rule out
>> stupid when we consider USB descriptors.
>
> Yes thanks for checking. If you prefer to set it up in some other
> way, or need more info, please let me know.

No, it all looks very sane to me, given your explanation. Just wanted to
be absolutely sure. Thanks a lot.


Acked-by: Bjørn Mork <bjorn@...k.no>



Bjørn