[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sat, 3 Mar 2012 02:44:38 -0500
From: Neal Cardwell <ncardwell@...gle.com>
To: Yuchung Cheng <ycheng@...gle.com>
Cc: Ilpo Järvinen <ilpo.jarvinen@...sinki.fi>,
David Miller <davem@...emloft.net>,
Netdev <netdev@...r.kernel.org>,
Nandita Dukkipati <nanditad@...gle.com>,
Tom Herbert <therbert@...gle.com>
Subject: Re: [PATCH] tcp: fix tcp_retransmit_skb() to maintain MSS invariant
On Fri, Mar 2, 2012 at 10:05 PM, Yuchung Cheng <ycheng@...gle.com> wrote:
> On Fri, Mar 2, 2012 at 12:25 PM, Ilpo Järvinen
> <ilpo.jarvinen@...sinki.fi> wrote:
>> Why would we be splitting SACKed segment there?!? ...Ah, it seems that
>> it indeed is possible now after the non-FACK mode fragment, however, it
>> is totally useless work and seems to be breaking stuff. I think that the
>> proper solution is to prevent splitting of already SACKed stuff. ...Nice
>> find btw. :-)
>
> Bingo! I totally agree this is the bug. looks like a one-line patch now.
Agreed. I sent out a proposal:
http://patchwork.ozlabs.org/patch/144408/
Thanks, Ilpo, for filling in the details. The comment in
tcp_match_skb_to_sack() that says "Round if necessary so that SACKs
cover only full MSSes and/or the remaining small portion (if
present)", plus the analogous MSS alignment logic in
tcp_mss_split_point(), had me convinced that this MSS alignment
invariant that was needed for tcp_match_skb_to_sack() to work
correctly was actually intended.
neal
--
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
Powered by blists - more mailing lists