netdev - Re: hard_start_xmit struct sk

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: <166884.22908.qm@web52012.mail.re2.yahoo.com>
Date:	Sat, 22 Dec 2007 10:18:08 +0100 (CET)
From:	Matti Linnanvuori <mattilinnanvuori@...oo.com>
To:	Jeff Garzik <jeff@...zik.org>
Cc:	netdev@...r.kernel.org
Subject: Re: hard_start_xmit struct sk_buff read-only?

--- Jeff Garzik <jeff@...zik.org> schrieb:

> This begs the question, what are you trying to do? 
> ;-)

I have seen network drivers whose hard_start_xmit
function calls skb_push or skb_put to extend the used
data area of the struct skb_buff passed to the
function. That avoids allocating a new buffer and
copying the whole data area to it. The drivers
apparently do that because the device expects a
specific header or suffix in the DMA buffer it
handles.

> In terms of object lifetimes, control passes to the
> net driver when 
> ->hard_start_xmit() is called, but that does not
> mean you can freely 
> scribble over things -- the skb may have been
> cloned, its destructor 
> callback still needs to be called (via
> dev_kfree_skb), etc.

So I assume modifying the data or header area of
struct sk_buff is incorrect because it can break
cloned buffers.

      Heute schon einen Blick in die Zukunft von E-Mails wagen? www.yahoo.de/mail
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html