Subject: Re: Call for testers: ACPI suspend/resume, part 2
To: Iain Hibbert <plunky@rya-online.net>
From: John Nemeth <jnemeth@victoria.tc.ca>
List: current-users
Date: 12/22/2007 15:37:45
On Apr 9,  9:16am, Iain Hibbert wrote:
} On Fri, 21 Dec 2007, Jared D. McNeill wrote:
} > On Fri, 21 Dec 2007, Iain Hibbert wrote:
} > > My apologies, I've been working 12 hour days and ENOTIME until probably
} > > next week - wrt bthub etc, in the past power down events just caused USB
} > > devices to detach which would detach from the bluetooth stack (its safe to
} > > do that) and thus detach bthub etc, requiring an /etc/rc.d run through to
} > > reconfigure after resume.
} >
} > Is it possible to use a bluetooth keyboard or mouse as a wake device, or is
} > the firmware on my ppc mini playing tricks on me?
} 
} It could be, but I don't know the specifics.
} 
} A bluetooth connection can be parked (temporarily suspended) by the link
} master and I guess they could do this on suspend but I don't know what
} happens if you remove the power to the bluetooth controller because the
} connection state is inside the controller and there is no way to 'remind'
} it on powerup that there were parked connections - you might as well
} disconnect/reconnect I guess because its designed for short range mobile
} connections anyway and a disconnected keyboard/mouse will generally
} attempt to connect if you press something.
} 
} Also I don't know anything about the state of the computer when the system
} is suspended as I've only ever used shutdown;  is the processor alive and
} will it run if data is received on the USB port?  I don't know of any
} method to make an bluetooth controller provide a 'system wake' signal, it
} would only be able to send data to the host which would need to run up
} through the USB/bluetooth/wskbd stack before being handled by a wakeup
} process.
} 
} Is there a standard wakeup method?  How do wake-on-network devices work?

     You leave the card powered up and it looks at received packets for
a signature.  I don't recall the exact details, but I believe it is the
MAC address repeated a certain number of times and that it can be
anywhere within the packet.  There are WOL connectors on motherboards.
When the card sees the "signature" packet, it momentarily shorts the
pins of the WOL connector.  At that point, the computer powers on.
There may be some changes in the details with current systems, but this
is the gist of it.  Given this, it may be that pmf needs to have some
changes to leave the NIC powered up in order to support WOL (I don't
have the gory details at hand).

     The first time I dealt with one of these systems it was rather
weird.  I plugged a network cable into a system that was turned off and
the link light came on.  That made me scratch my head for a couple of
seconds.

}-- End of excerpt from Iain Hibbert