Re: 3Com 3C05 adapter on 43P running 4.3
From: Jon Parmet (jon_at_parmetpc.volpe.dot.gov)
Date: 03/08/04
- Next message: mark taylor: "Re: kexit panic"
- Previous message: Gitte Togersen: "Fileset form Frensh language"
- In reply to: Dan Foster: "Re: 3Com 3C05 adapter on 43P running 4.3"
- Next in thread: Dan Foster: "Re: 3Com 3C05 adapter on 43P running 4.3"
- Reply: Dan Foster: "Re: 3Com 3C05 adapter on 43P running 4.3"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: 8 Mar 2004 07:00:27 -0800
Dan Foster <usenet@evilphb.org> wrote in message news:<slrnc4mv3j.rab.usenet@gaia.roc2.gblx.net>...
> In article <f0701dba.0403071044.7c163434@posting.google.com>, Jon Parmet <jon@parmetpc.volpe.dot.gov> wrote:
> > FlossyFloss <flossyflossmail-news@yahoo.com> wrote in message news:<c2d9us$5d5$1@charly.heeg.de>...
> >> Jon Parmet wrote:
> >>
> > My other reply hasn't appeared on Google. I went and pulled the 3Com
> > card out, and the ent0 PCI adapter still shows up. I think I just
> > missed it before. My bad. So the on-board adapter goes through the PCI
> > bus, fair enough.
> >
> > I guess the next question then is since the 3C905-T4 isn't plug N play
> > with this box, does the 43P support it at all?
>
> Well, it *is* plug-and-play... the 43P will support it fine *if* you have a
> device driver for the OS to make it usable.
>
> This is what you are missing -- the driver. To which, there was none made
> for the type of card you have. All hardware with drivers are essentially
> the ones that has an IBM feature code number and sometimes has IBM firmware
> or the vendor otherwise certified it for operation on AIX boxes along with
> providing an AIX device driver.
>
> In a nutshell, you will have to either write an AIX device driver for the
> ethernet card, or get a different ethernet adapter...
So in effect, it's not plug and play ;)
> and one that is
> specifically certified for AIX since it will have a device driver available
> (either in the base AIX OS or as a separate fileset provided by the
> manufacturer).
We have another 43P here (a newer 7043-150) that came with 2 cards.
Here's the output from it:
ent1 Available 10-b0 10/100 Mbps Ethernet PCI Adapter II (1410ff01)
ent0 Available 10-60 IBM 10/100 Mbps Ethernet PCI Adapter (23100020)
Are there any issues trying this card in the older 7248-132 box?
> For example, you can get the feature code (FC) 2968 ethernet adapter which
> is an IBM 10/100 Mbps fast ethernet PCI card on ebay with a starting bid of
> $75:
>
> http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=3082911798&category=51195
>
> (They've even got 250 available!! You could buy just one or many...)
>
> The device driver is already built into the OS, so it's just a matter of
> shutting down the machine (if this is not a newer pSeries machine that
> supports turning off PCI slot power like the 6H1), inserting the new enet
> card, and then powering it on. Boot will run cfgmgr automatically (does so
> on every boot) which will detect the device.
>
> You asked about en vs et... the *physical* card is, for example, ent0.
>
> What is en0? It is the logical ethernet interface that adheres to the DIX
> (aka 'Ethernet II') encapsulation standard. This is the standard that the
> rest of world ordinarily uses today.
>
> What is et0? It is the logical ethernet interface that adheres to the IEEE
> 802.3 LLC/SNAP encapsulation standard.
>
> It's slightly different than DIX, so if you have network hardware that ONLY
> understands this format and try to transmit DIX-encap'd Ethernet frames, it
> will be like speaking gibberish.
>
> Both are ethernet and refers to the same underlying physical device; the
> difference is that the card will transmit packets constructed differently,
> depending on if you configure en0 or et0.
>
> 99.9999999% of the time, you'll only use en0 and just disregard et0 unless
> you've got some *really* oddball/broken network that requires the
> alternative encap standard. This may have been slightly more common 10-15
> years ago, but just about nonexistent these days.
>
> IBM provides supports for both encap standards as a nicety, giving the
> customer options.
>
> -Dan
Ok, if I understand things, the layering goes something like this:
rj45-physical
ent0-link
en0-network
192.x.x.x-transport
Perhaps some of the boundaries between layers are fuzzier (e.g. Layers
0 and 1)?
Thanks to everyone for the replies (Flossy and Harry included).
Regards,
Jon
Thanks for
- Next message: mark taylor: "Re: kexit panic"
- Previous message: Gitte Togersen: "Fileset form Frensh language"
- In reply to: Dan Foster: "Re: 3Com 3C05 adapter on 43P running 4.3"
- Next in thread: Dan Foster: "Re: 3Com 3C05 adapter on 43P running 4.3"
- Reply: Dan Foster: "Re: 3Com 3C05 adapter on 43P running 4.3"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
