[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAHmME9pk8HEFRq_mBeatNbwXTx7UEfiQ_HG_+Lyz7E+80GmbSA@mail.gmail.com>
Date: Tue, 11 Feb 2020 10:59:37 +0100
From: "Jason A. Donenfeld" <Jason@...c4.com>
To: Florian Westphal <fw@...len.de>
Cc: Netdev <netdev@...r.kernel.org>, David Miller <davem@...emloft.net>
Subject: Re: [PATCH v2 net 1/5] icmp: introduce helper for NAT'd source
address in network device context
On Mon, Feb 10, 2020 at 11:33 PM Florian Westphal <fw@...len.de> wrote:
> I think this is a bit too conservative, f.e. i don't see how
> ndo-called skbs could be shared (tx path needs to be able to change skb
> list pointers)?
Are you *sure* about that? Dave - do you know?
I'm asking with *asterisks* because I see a few drivers in tree that
do call skb_share_check from their ndo_start_xmit function. If this
makes no sense and isn't needed, then I'll send a cleanup series for
these. And, I'll remove it from icmp_ndo_send for v3, replacing it
with a `if (WARN_ON(skb_shared(skb_in))) return;`.
Thanks,
Jason
Powered by blists - more mailing lists