linux-kernel - Re: [PATCH 1/1] net: tun: add XDP metadata support

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] [day] [month] [year] [list]

Message-ID: <67a01e615bdb1_3c12af2942f@willemb.c.googlers.com.notmuch>
Date: Sun, 02 Feb 2025 20:39:45 -0500
From: Willem de Bruijn <willemdebruijn.kernel@...il.com>
To: Willem de Bruijn <willemdebruijn.kernel@...il.com>, 
 Marcus Wichelmann <marcus.wichelmann@...zner-cloud.de>, 
 Stanislav Fomichev <stfomichev@...il.com>
Cc: netdev@...r.kernel.org, 
 linux-kernel@...r.kernel.org, 
 bpf@...r.kernel.org, 
 willemdebruijn.kernel@...il.com, 
 jasowang@...hat.com, 
 andrew+netdev@...n.ch, 
 davem@...emloft.net, 
 edumazet@...gle.com, 
 kuba@...nel.org, 
 pabeni@...hat.com, 
 ast@...nel.org, 
 daniel@...earbox.net, 
 hawk@...nel.org, 
 john.fastabend@...il.com
Subject: Re: [PATCH 1/1] net: tun: add XDP metadata support

> > >> +		metasize = metasize > 0 ? metasize : 0;
> > > 
> > > Why is this part needed?
> > 
> > When an xdp_buff was initialized withouth metadata support (meta_valid
> > argument of xdp_prepare_buff is false), then data_meta == data + 1.
> > So this check makes sure that metadata was supported for the given xdp_buff
> > and metasize is not -1 (data - data_meta).
> > 
> > But you have a good point here: Because we have control over the
> > initialization of xdp_buff in the tun_build_skb function (first code path),
> > we know, that metadata is always supported for that buffer and metasize
> > is never < 0. So this check is redundant and I'll remove it.
> > 
> > But in the tun_xdp_one function (second code path), I'd prefer to keep that
> > check, as the xdp_buff is externally passed to tun_sendmsg and the tun driver
> > should probably not make assumptions about the metadata support of buffers
> > created by other drivers (e.g. vhost_net).
> > 
> > Thank you for taking a look, I hope things are more clear now.
> 
> Please use min()
> 

Err.. max.