netdev - Re: [RFC net v7] net: skb_expand_head() adjust skb->truesize incorrectly

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: <20210920173949.7e060848@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
Date:   Mon, 20 Sep 2021 17:39:49 -0700
From:   Jakub Kicinski <kuba@...nel.org>
To:     Vasily Averin <vvs@...tuozzo.com>
Cc:     eric.dumazet@...il.com, netdev@...r.kernel.org,
        Christoph Paasch <christoph.paasch@...il.com>,
        Hao Sun <sunhao.th@...il.com>
Subject: Re: [RFC net v7] net: skb_expand_head() adjust skb->truesize
 incorrectly

On Tue, 21 Sep 2021 00:41:15 +0300 Vasily Averin wrote:
> > Thanks for taking a look. I would prefer not to bake any ideas about
> > the skb's function into generic functions. Enumerating every destructor
> > callback in generic code is impossible (technically so, since the code
> > may reside in modules).
> > 
> > Let me think about it. Perhaps we can extend sock callbacks with
> > skb_sock_inherit, and skb_adjust_trusize? That'd transfer the onus of
> > handling the adjustments done on splitting to the protocols. I'll see
> > if that's feasible unless someone can immediately call this path
> > ghastly.  
> 
> This is similar to Alexey Kuznetsov's suggestion for me, 
> see https://lkml.org/lkml/2021/8/27/460

Interesting, I wasn't thinking of keeping the ops pointer in every skb.

> However I think we can do it later,
> right now we need to fix somehow broken skb_expand_head(),
> please take look at v8.

I think v8 still has the issue that Eric was explaining over and over.