netdev - Re: [PATCH] netem: fix rate extension and drop accounting

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-ID: <20120716232600.GI3415@nuttenaction>
Date:	Tue, 17 Jul 2012 01:26:01 +0200
From:	Hagen Paul Pfeifer <hagen@...u.net>
To:	Mark Gordon <msg@...gle.com>
Cc:	Eric Dumazet <eric.dumazet@...il.com>,
	David Miller <davem@...emloft.net>,
	netdev <netdev@...r.kernel.org>,
	Yuchung Cheng <ycheng@...gle.com>,
	Andreas Terzis <aterzis@...gle.com>
Subject: Re: [PATCH] netem: fix rate extension and drop accounting

Sorry Eric for the delay! I am a little bit under project pressure. But I will
test patches where I can.

* Mark Gordon | 2012-07-16 15:15:07 [-0700]:

>Sorry for the late reply on this.  I'm still pretty unsure about the math
>we're doing on
>
>delay -= netem_skb_cb(skb_peek(list))->time_to_send - now;
>
>First, shouldn't it be skb_peek_tail(list)?  Even if you do that the math

The delta is between now and when the _next_ packet is to send. New packets
are enqueued on the tail.

>doesn't really seem to work out.  In my mind there ought to be at least one
>more non-linearity.  The code before/after this patch does not work as I
>would expect and seems to have fairly random effects on the bandwidth and
>that the below patch does work for me.  Here is what I'm suggesting
>(relative to the new patches)

As Eric said: there are problems in combination with a static delay. During
rate extension development we tested the raw/vanilla "rate" functionality.
The rate part works faultless[TM] - at least independet of any other
"delay-latency generator".

Hagen
--
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