[Cerowrt-devel] cerowrt-3.7.3-2 released
dpreed at reed.com
dpreed at reed.com
Mon Jan 21 16:47:39 EST 2013
How random do you need things to be? Security needs true randomness, but there is nothing else I know of where you need "entropy" sources to achieve sufficient unpredictability to meet the need of some algorithm. A simple pseudo-random sequence is good enough.
From: "Dave Taht" <dave.taht at gmail.com>
Sent: Monday, January 21, 2013 2:47pm
To: cerowrt-devel at lists.bufferbloat.net
Subject: [Cerowrt-devel] cerowrt-3.7.3-2 released
And is at:
+ resync with openwrt
+ uftp4 alpha
+ quagga update
+ Hopefully the last ipv6 instruction traps killed (thx ketan and robert!)
- I see rngd eating a ton of cpu on a given transfer. I'd been meaning to look at the quality of the new entropy stuff for ages, and have long hoped to be rid of that daemon
After we do a bit more testing I'll tag and push the source to this version, and then get cracking on higher layers in the stack again.
A bit about uftp4 support. My intent with getting uftp into cerowrt 18 months ago was to have an easy multicast test, which would let someone (for example) set a rate of 1Mbit and observe what happens to a wireless network, observe packet loss at various distances from the AP, etc, as well as get a grip on the timing delays induced by the power save mode in wifi (basically storing all multicast content and broadcasting it as "content after beacon", (CAP, also known colloquially as "crap after beacon"))...
Anyway, while I was thinking that as my basic use case in this project...
Dennis Bush was actually off making a new version do more useful things. We got to talking last week... I've tried the ipv4 support and it's nifty. Ipv6 didn't work but perhaps my x86 build is bad.
Fixing bufferbloat with cerowrt: [http://www.teklibre.com/cerowrt/subscribe.html] http://www.teklibre.com/cerowrt/subscribe.html
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Cerowrt-devel