Re: [PATCH 1/3] driver for mcs7830 (aka DeLOCK) USB ethernet adapter

[Arnd Bergmann <arnd@arndb.de>]

On Thursday 28 September 2006 03:28, David Brownell wrote:
> ... yes, I'd assume it's a hardware issue too.  Try different
> cables; if you have a fast 'scope, you might see what kind of
> eye diagram you get.

Unfortunately, I don't have anything I could use to test this.

> Do you know how the remote wakeup mechanism works for this chip?
> It'd be interesting to see "usbnet" be taught how to autosuspend
> chips which will wake up the USB host when they get the right
> kind of packet ... for example, passing the multicast/broadcast
> filter, or addressed directly to that adapter.
> 
> Such an autosuspend mechanism would let the host controller stop
> polling a mostly-idle network link, getting rid of one source of
> periodic DMA transfers and thus allowing deeper sleep states for
> many x86 family CPUs.  And also, I'd expect, giving fewer
> opportunities for those broken RX packets.  :)

No, don't know. I did not find anything about this in the public
spec.

	Arnd <><