linux-kernel - Re: [PATCH 5.10 038/146] mptcp: do not warn on bad input from the network

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:   Wed, 23 Jun 2021 09:45:49 -0700 (PDT)
From:   Mat Martineau <mathew.j.martineau@...ux.intel.com>
To:     Paolo Abeni <pabeni@...hat.com>
cc:     Pavel Machek <pavel@....cz>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        linux-kernel@...r.kernel.org, stable@...r.kernel.org,
        "David S. Miller" <davem@...emloft.net>,
        Sasha Levin <sashal@...nel.org>
Subject: Re: [PATCH 5.10 038/146] mptcp: do not warn on bad input from the
 network

On Wed, 23 Jun 2021, Paolo Abeni wrote:

> Hello,
>
> On Wed, 2021-06-23 at 16:22 +0200, Pavel Machek wrote:
>> Hi!
>>
>>> From: Paolo Abeni <pabeni@...hat.com>
>>>
>>> [ Upstream commit 61e710227e97172355d5f150d5c78c64175d9fb2 ]
>>>
>>> warn_bad_map() produces a kernel WARN on bad input coming
>>> from the network. Use pr_debug() to avoid spamming the system
>>> log.
>>
>> So... we switched from WARN _ONCE_ to pr_debug, as many times as we
>> detect the problem.
>>
>> Should this be pr_debug_once?
>
> Thank you for double checking this!
>
> In the MPTCP code, we use pr_debug() statements as a debug tool, e.g.
> when enabled, it could print per-packet info with no restriction.
>
> There are (a few) similar use in the plain TCP code.
>
> pr_debug() is not supposed to be enabled on any production system,
> while the WARN_ONCE could trigger automated tools for irrelevant
> network noise.
>
> I thing pr_debug() is fine here.
>

Hi Pavel -

I agree with Paolo. This is not a frequently encountered condition, even 
when turning on this specific pr_debug(). I'd say the previous use of the 
_ONCE() variant was mostly because the effects were not optional (not 
because of frequency).

For protocol development, pr_debug() is both off by default and more
useful when it _is_ enabled. I'd prefer to keep this patch as-is.

Thanks,


--
Mat Martineau
Intel