linux-kernel - Re: Problem with USB driver using two devices

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-next>] [day] [month] [year] [list]

Message-ID: <20161123163545.GB10750@kroah.com>
Date:   Wed, 23 Nov 2016 17:35:45 +0100
From:   Greg KH <greg@...ah.com>
To:     Wolfgang Wilhelm <wilhelm@...tcomtec.com>
Cc:     linux-usb@...r.kernel.org,
        Austin S Hemmelgarn <ahferroin7@...il.com>,
        linux-kernel@...r.kernel.org
Subject: Re: Problem with USB driver using two devices

On Wed, Nov 23, 2016 at 05:17:35PM +0100, Wolfgang Wilhelm wrote:
> Thankyou very much for the really fast answer.
> 
> I don't get any error messages and I can communicate with
> the driver for the second device via ioctrl and write functions,
> i.e. write registers and read registers via the RBUF ioctrl function,
> only the read function for the second device does not work,
> i.e. no data is obtained from the mcs6_read function for the
> second device.

Hm, let me go look at the driver again, maybe something's odd with it.

> Thankyou for your hint using libusb, I will have a look on it.
> 
> Our USB device has three endpoints, two with 64 kb packet size
> for reading and writing registers (I know this is not standard
> for high speed) and one with 512 kb packet size for reading data.
> Do you think the problem could arise from this deviation
> from the USB standard?

No, it should be fine, the USB standard only care about endpoint sizes,
not logical "packet" sizes, right?

> Of course we would be happy if our driver could be merged into
> the Linux kernel.

If you can use libusb, I'd strongly recommend using that and not a
kernel driver at all, as we don't like adding kernel drivers where they
are not needed.

> Our device uses the same vendor/product id's as a
> "D-Link DSB-R100 USB FM radio" dsbr100, that has a built-in
> driver in some Linux distributions, so that dsbr100 must be
> blacklisted in /etc/modprobe.d/fbdev-blacklist.conf for
> using our driver.

Does your device work like the USB FM radio device?  Why is the same
vendor/device id being used here?  They are supposed to be unique for
different types of devices.

> Which security problems do you see in the code?

No checking that the values given to you by userspace are actually valid
and within "sane" ranges :)

thanks,

greg k-h