Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 24 Jan 2001 12:06:53 -0800
From:      jayanth <jayanth@yahoo-inc.com>
To:        Paul Herman <pherman@frenchfries.net>
Cc:        Jonathan Lemon <jlemon@flugsvamp.com>, net@FreeBSD.ORG
Subject:   Re: I have delayed ACK problems
Message-ID:  <20010124120653.B65472@yahoo-inc.com>
In-Reply-To: <Pine.BSF.4.32.0101241854420.21262-100000@husten.security.at12.de>; from pherman@frenchfries.net on Wed, Jan 24, 2001 at 08:50:06PM %2B0100
References:  <200101241721.f0OHL2627523@prism.flugsvamp.com> <Pine.BSF.4.32.0101241854420.21262-100000@husten.security.at12.de>

next in thread | previous in thread | raw e-mail | index | archive | help
it looks like the tcp code is not acking the data when the PUSH
flag is set. It would be useful to generate an immediate ACK
when the PUSH flag is set. 

it would be interesting to see if the other sends two full sized
tcp segments if the freebsd stack will then immediately acknowledge
the data.

jayanth


Paul Herman (pherman@frenchfries.net) wrote:
> Hi Jonathan,
> 
> On Wed, 24 Jan 2001, Jonathan Lemon wrote:
> 
> >   >10:49:54.279497 16.1035 > 10.40438: . ack 19433 win 17520 (DF)
> >   >10:49:54.371025 16.1035 > 10.40438: . ack 21481 win 17520 (DF)
> >
> > Are you sure the trace has delayed ack turned off?
> 
> Sorry, wasn't clear there.  Yes, this trace is with delayed acks
> turned ON, exhibiting the performance problems I was describing.
> 
> > Try bumping up net.inet.tcp.slowstart_flightsize, which specifies
> > how many packets can be outstanding (unacked) at the beginning
> > of a TCP connection.
> 
> Do you mean lower?  Then it wouldn't wait for more segments to arrive
> and run up the delayed ack timer, correct?
> 
> Thanks, I'll try it out (both ways, higher & lower values) but this
> unfortunately seems like a workaround rather than a solution.
> 
> Something just doesn't sit well me.  This shouldn't happen in a
> 100Mbit LAN.  This seems like a plain vanilla network transaction, and
> FreeBSD is failing on something, where other OSes in the same
> environment don't.
> 
> Time to dig through sys/netinet/tcp* ...
> 
> -Paul.
> 
> 
> 
> 
> 
> 
> 
> To Unsubscribe: send mail to majordomo@FreeBSD.org
> with "unsubscribe freebsd-net" in the body of the message
> 
> 


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-net" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20010124120653.B65472>