netdev
[Top] [All Lists]

Re: [RFC] enhanced version of net_random()

To: Stephen Hemminger <shemminger@xxxxxxxx>
Subject: Re: [RFC] enhanced version of net_random()
From: Ben Greear <greearb@xxxxxxxxxxxxxxx>
Date: Thu, 12 Aug 2004 13:02:15 -0700
Cc: "David S. Miller" <davem@xxxxxxxxxx>, Alan Cox <alan@xxxxxxxxxxxxxxxxxxx>, "Theodore Ts'o" <tytso@xxxxxxx>, netdev@xxxxxxxxxxx, linux-kernel@xxxxxxxxxxxxxxx
In-reply-to: <20040812104835.3b179f5a@xxxxxxxxxxxxxxxxxxxxx>
Organization: Candela Technologies
References: <20040812104835.3b179f5a@xxxxxxxxxxxxxxxxxxxxx>
Sender: netdev-bounce@xxxxxxxxxxx
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.2) Gecko/20040803
Stephen Hemminger wrote:
While doing the network emulator, I discovered that the default net_random()
is too stupid, and get_random_bytes() is more than needed. Rather than put
another function in just for sch_netem, how about making net_random() smarter?
The tin-hat crowd already replace net_random() with get_random_bytes anyway.

Here is a proposed alternative to use a longer period PRNG for net_random().
The choice of TT800 was because it was freely available, had a long period,
was fast and relatively small footprint. The existing net_random() was not
really thread safe, but was immune to thread corruption.

Is it really worth the extra spin lock & math?  Maybe we could have a
net_more_random() method instead that encompasses this improved random logic?

Ben

--
Ben Greear <greearb@xxxxxxxxxxxxxxx>
Candela Technologies Inc  http://www.candelatech.com


<Prev in Thread] Current Thread [Next in Thread>