Re: [PATCH] USB: input: powermate: fix oops with malicious USB descriptors

From: Dmitry Torokhov
Date: Mon Mar 14 2016 - 12:48:44 EST


On Mon, Mar 14, 2016 at 9:46 AM, Josh Boyer <jwboyer@xxxxxxxxxxxxxxxxx> wrote:
> On Mon, Mar 14, 2016 at 12:15 PM, Greg Kroah-Hartman
> <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
>> On Mon, Mar 14, 2016 at 10:12:53AM -0400, Josh Boyer wrote:
>>> The powermate driver expects at least one valid USB endpoint in its
>>> probe function. If given malicious descriptors that specify 0 for
>>> the number of endpoints, it will crash. Validate the number of
>>> endpoints on the interface before using them.
>>>
>>> The full report for this issue can be found here:
>>> http://seclists.org/bugtraq/2016/Mar/85
>>>
>>> Reported-by: Ralf Spenneberg <ralf@xxxxxxxxxxxxxx>
>>> Cc: stable <stable@xxxxxxxxxxxxxxx>
>>> Signed-off-by: Josh Boyer <jwboyer@xxxxxxxxxxxxxxxxx>
>>> ---
>>> drivers/input/misc/powermate.c | 3 +++
>>> 1 file changed, 3 insertions(+)
>>
>> I'll queue these up after 4.6-rc1 is out as the merge window is closed
>> right now, but we might want to think about a better way to handle this
>> type of thing in the USB core. A way to keep from having to add checks
>> like this for every single driver, when the driver shouldn't even really
>> have their probe function called unless their expected endpoints are
>> going to be there.
>
> I thought this discussion came up a while ago, when something very
> similar was fixed in the whiteheat driver (commit cbb4be652d374). I
> can't find the thread at the moment, but I thought someone said this
> had to be per-driver for some reason. I'm more than happy to have a
> core subsystem fix if it's possible.
>
>> I'll think about that over the next few weeks...
>
> I have something around 8 drivers with issues like this. I think
> Oliver (now CC'd) is working from the same set of bugs. Should we
> hold off on submitting individual fixes until later, or should we go
> ahead and submit them?

I'll take input bits, there is no need to keep kernel oopsing while we
are working on a more general solution.

Thanks.

--
Dmitry