[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20090611111412.C908733C023@tippex.mynet.homeunix.org>
Date: Thu, 11 Jun 2009 13:14:12 +0200
From: Anders Eriksson <aeriksson@...tmail.fm>
To: Francois Romieu <romieu@...zoreil.com>
cc: Anders Eriksson <aeriksson@...tmail.fm>, netdev@...r.kernel.org
Subject: Re: WoL and r8169
romieu@...zoreil.com said:
> Anders Eriksson <aeriksson@...tmail.fm> : [...]
>> Yes! That diff made the trick for WoL on magic package on my 2.6.30-rc6
>> kernel. Is it sent to Linus for -final?
> No but it does not preclude it being properly submitted thereafter then for
> -stable.
Great.
> [...]
>> I tried to wake up the machine using regular tcp traffic, on the theory that
>> it's unicast and phy activity. No luck though.
> Did the regular tcp traffic include a so-called magic (tcp) packet ?
Hmm,no. What is that? I'm looking at having what's best described as "wake on
directed packet" (ref 8 here, http://en.wikipedia.org/wiki/Wake-on-LAN).
Afaict, the box is expected to wake up on any traffic directed towards it (its
MAC and/or IP).
My reading of ethtool(8) says that 'u' should do the trick, and if I add 'b' I
should be able to wake on arp (as arp is eth broadcast). In my test, however,
I've made sure to have the arp cache filled, and tcpdump confirms that an ssh
session to a pm-suspend'd box results in a series of unicsat in-session tcp
packets, and no arp traffic. Eventullay the arp cache times out (or is pruned
because of the faild transmissions?) and arp requests happens, but no wakeup.
What's a magic tcp packet?
/Anders
--
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