netdev - Re: [RFC PATCH 1/2] net: extend drop reasons for multiple subsystems

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]

Message-ID: <cb8ca010692920d909d0155aac9d66761bbf250c.camel@sipsolutions.net>
Date:   Thu, 30 Mar 2023 10:11:12 +0200
From:   Johannes Berg <johannes@...solutions.net>
To:     Jakub Kicinski <kuba@...nel.org>
Cc:     linux-wireless@...r.kernel.org, netdev@...r.kernel.org
Subject: Re: [RFC PATCH 1/2] net: extend drop reasons for multiple subsystems

On Wed, 2023-03-29 at 21:05 -0700, Jakub Kicinski wrote:
> On Wed, 29 Mar 2023 23:46:19 +0200 Johannes Berg wrote:
> > -	DEBUG_NET_WARN_ON_ONCE(reason <= 0 || reason >= SKB_DROP_REASON_MAX);
> > +	DEBUG_NET_WARN_ON_ONCE(reason == SKB_NOT_DROPPED_YET);
> 
> We can still validate that the top bits are within known range 
> of subsystems?

Yeah, I was being a bit sneaky here ;)

We could, for sure. Given that the users should probably be defensively
coded anyway (as I did in drop_monitor), I wasn't sure if we _should_.

It seemed to me that for experimentation, especially if your driver is a
module, it might be easier to allow this?

That said, I don't have any strong feelings about it, and I have some
bugs here anyway so I can just add that.

We _could_ also keep a check for the core subsystem, but not sure that's
worth it?

johannes