[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <95ae7576-fa52-4ef0-91f7-14aa0bb91208@redhat.com>
Date: Thu, 29 Aug 2024 11:33:33 +0200
From: Paolo Abeni <pabeni@...hat.com>
To: Krzysztof Kozlowski <krzk@...nel.org>,
Aleksandr Mishin <amishin@...rgos.ru>, Samuel Ortiz <sameo@...ux.intel.com>
Cc: netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
lvc-project@...uxtesting.org
Subject: Re: [PATCH v2] nfc: pn533: Add poll mod list filling check
On 8/29/24 11:06, Krzysztof Kozlowski wrote:
> On 29/08/2024 10:26, Paolo Abeni wrote:
>>
>>
>> On 8/27/24 10:48, Aleksandr Mishin wrote:
>>> In case of im_protocols value is 1 and tm_protocols value is 0 this
>>> combination successfully passes the check
>>> 'if (!im_protocols && !tm_protocols)' in the nfc_start_poll().
>>> But then after pn533_poll_create_mod_list() call in pn533_start_poll()
>>> poll mod list will remain empty and dev->poll_mod_count will remain 0
>>> which lead to division by zero.
>>>
>>> Normally no im protocol has value 1 in the mask, so this combination is
>>> not expected by driver. But these protocol values actually come from
>>> userspace via Netlink interface (NFC_CMD_START_POLL operation). So a
>>> broken or malicious program may pass a message containing a "bad"
>>> combination of protocol parameter values so that dev->poll_mod_count
>>> is not incremented inside pn533_poll_create_mod_list(), thus leading
>>> to division by zero.
>>> Call trace looks like:
>>> nfc_genl_start_poll()
>>> nfc_start_poll()
>>> ->start_poll()
>>> pn533_start_poll()
>>>
>>> Add poll mod list filling check.
>>>
>>> Found by Linux Verification Center (linuxtesting.org) with SVACE.
>>>
>>> Fixes: dfccd0f58044 ("NFC: pn533: Add some polling entropy")
>>> Signed-off-by: Aleksandr Mishin <amishin@...rgos.ru>
>>
>> The issue looks real to me and the proposed fix the correct one, but
>> waiting a little more for Krzysztof feedback, as he expressed concerns
>> on v1.
>
> There was one month delay between my reply and clarifications from
> Fedor, so original patch is neither in my mailbox nor in my brain.
>
>
> Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@...aro.org>
>
> However different problem is: shouldn't as well or instead
> nfc_genl_start_poll() validate the attributes received by netlink?
>
> We just pass them directly to the drivers and several other drivers
> might not expect random stuff there.
FTR, I had a similar thought and skimmed over other nfc drivers. I did
not see similar issues there.
Additionally I fear that existing user-space could feed to the kernel
such random stuff and work happily because the kernel is currently
ignoring it - on other drivers. Such cases will suddenly stop working.
I think we could/should merge the patch as-is, please LMK your thought.
Thanks,
Paolo
Powered by blists - more mailing lists