netdev - Re: [PATCH v2 net-next] tcp: avoid expensive pskb_expand

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: <alpine.DEB.2.00.1204191443040.735@wel-95.cs.helsinki.fi>
Date:	Thu, 19 Apr 2012 14:57:08 +0300 (EEST)
From:	"Ilpo Järvinen" <ilpo.jarvinen@...sinki.fi>
To:	Eric Dumazet <eric.dumazet@...il.com>
cc:	Neal Cardwell <ncardwell@...gle.com>,
	David Miller <davem@...emloft.net>,
	netdev <netdev@...r.kernel.org>,
	Tom Herbert <therbert@...gle.com>,
	"Maciej Żenczykowski" <maze@...gle.com>,
	Yuchung Cheng <ycheng@...gle.com>
Subject: Re: [PATCH v2 net-next] tcp: avoid expensive pskb_expand_head()
 calls

On Thu, 19 Apr 2012, Eric Dumazet wrote:

> On Thu, 2012-04-19 at 13:30 +0200, Eric Dumazet wrote:
> > On Thu, 2012-04-19 at 14:10 +0300, Ilpo Järvinen wrote:
> > 
> > > Now that you have non-zero offset_ack, are the tcp_fragment() callsites 
> > > safe and working? ...I'm mostly worried about tcp_mark_head_lost which 
> > > does some assumptions about tp->snd_una and TCP_SKB_CB(skb)->seq, however, 
> > > also other fragmenting does not preserve offset_ack properly (which might 
> > > not be end of world though)?
> > 
> > Good point, I'll take a look.
> 
> Hmm, the only point this could matter is if a packet is retransmitted.
>
> For other packets, offset_ack = 0 (default value on skb allocation)
> 
> And tcp_retransmit_skb() first call tcp_trim_head(sk, skb) if needed so
> tcp_fragment() is called with == 0
> 
>         if (before(TCP_SKB_CB(skb)->seq, tp->snd_una)) {
>                 if (before(TCP_SKB_CB(skb)->end_seq, tp->snd_una))
>                         BUG();
>                 if (tcp_trim_head(sk, skb))
>                         return -ENOMEM;
>         }
> 
> ...
> if (skb->len > cur_mss) {
> 	if (tcp_fragment(sk, skb, cur_mss, cur_mss))
> 
> 
> 
> I could add a BUG_ON(offset_ack == 0) to make sure this assertion is
> true.

If you end up putting something like that make sure you use WARN_ON 
instead as this surely isn't fatal enough to warrant full stop of the
box :-).

> What do you think ?

I'm not concerned of the output side, that seems to work because 
of the in tcp_retransmit_skb getting rid of the extra first.

The ACK input stuff is more interesting, e.g., this one in 
tcp_mark_head_lost:

	err = tcp_fragment(sk, skb, (packets - oldcnt) * mss, mss);

It splits from TCP_SKB_CB(skb)->seq + (packets - oldcnt) * mss whereas
I think the desired point would be: TCP_SKB_CB(skb)->seq + offset_ack + 
(packets - oldcnt) * mss?

...There is similar case in sacktag code too while it's aligning to mss 
boundaries in tcp_match_skb_to_sack.

-- 
 i.