Hot-insert PCMCIA cards / change MAC-ID on older Thinkpads?

PCMCIA always stood for Personal Computer Memory Card Industry Association.

But the organization changed the name of the cards to "PC Cards" in the mid 1990's. ALL of the cards. PCMCIA remains the name of the organization. But it's incorrect for any of the cards, although it's still widely used.

A Cardbus card is still a PC Card.

Yes, a 32 Bit PC Card.

Clearly he doesn't understand the difference between Boot from LAN and Wake on LAN. They are two completely different animals and totally unrelated.

Wake on LAN, AKA Magic Packet, enables a computer to be powered up when a specific packet addressed to it is received. The NIC remains on, and is able to yank on a PCI signal that wakes the machine if it sees the proper packet.

Boot from Net allows a diskless workstation or thin client to be booted up from the network server. It requires a boot ROM on the network card. Boot from net was popular in the days of Netware, and is still popular on thin clients, though most thin clients now can boot from internal Flash as well, depending on the OS.

With PXE, the code is usually in the system boot ROM, though you could have a NIC card with a PXE boot ROM. PXE makes everything a lot easier since it's a pretty common BIOS feature, at least on the higher-end machines.

That is not what these guys were talking about.

On Fri, 13 Oct 2006 16:54:25 -0700, SMS wrote in :

They are related in that a machine has to be woken to boot. ( For extra credit, tell us how you boot a sleeping remote machine without WOL. :)

Whereupon it boots. ( You've picked up on the PCI signal, so I guess it is possible to teach an old dog new tricks. :)

I'm guessing you're actually thinking of RIPL (Remote Initial Program Load). ;)

Neither RIPL nor PXE, nor for that matter WOL, have anything to do with the actual issue at hand: where the MAC address comes from; i.e., this is all just a smoke screen.

Nope. See: |

It's not the card. It's the driver.

On Sat, 14 Oct 2006 01:14:18 GMT, Jeff Liebermann wrote in :

I'll give you at most 1/2 point, since the driver and the card are essentially the same product, as distinct from the host computer, minus

I hope you do not support computers or networks.

You really have no idea what these guys are talking about. So why are you intruding on their discussion while showing your ignorance.

Well to be fair, not many novices understand boot from net, especially if they weren't around in the days of Novell Netware. I've done a lot of it on networks of thin clients, but it's not something that the typical user would encounter.

Typical Navas. He doesn't know when to stop after he's caught showing his lack of knowledge.

No, the PC doesn't have to be woken to boot, because it's in S5 where you can't WOL anyway. Boot from net clients are turned on the old fashioned way, with a power button. I guess it would be possible to do both, first WOL then boot from net, but this isn't commonly done.

Wake-on-LAN is unrelated. A PC, typically with a disk drive, is woken up from S1, S2, S3, or S4 state via the NIC (integrated or PCI card) that yanks on a dedicated PCI signal. The NIC is powered by the STBY rail out of the power supply.

On Fri, 13 Oct 2006 18:58:25 -0700, SMS wrote in :

Yet you didn't apparently know the correct term, which is RIPL. ;)

Whatever you say.

Actually it is. And again, it's either RIPL or PXE, not "boot from net". ( No extra credit for you, I'm afraid. :)

On the contrary, as I explained.

Nope. I want my full +1 point, with prime rate interest added.

The original Orinoco drivers supported promiscuous mode and some other goodies including hot swapping the cards without changing the MAC. That was more of a bug than a feature, but it was there. The clue was that Ethereal could sniff the entire WLAN on the old drivers, but could only sniff traffic to/from the card in the later drivers.

Lucent/Agere initially removed promiscuous mode and probably hot swap from the drivers. Along with the drivers came firmware changes. As I recall, the 4.x firmware did hot swap and promiscuous mode just fine. The later version 6.x firmware broke both of these.

Once Proxim got its hands on the product, not only was the driver rescribbled for NDIS5, but the chipset and hardware changed radically from the original Hermes chipset. They called it Orinoco but that only went as deep as the label.

And you had no idea of what they were talking about. Just like when you tried to claim that the MAC address comes from ARP. You may be great on throwing around terminology, but you have no idea what they mean.

No, it is what you are saying.

The only thing you explained is that you have no idea what these guys were discussing when you jumped in with your foot in your mouth.

On Fri, 13 Oct 2006 19:08:04 -0800, "Dana" wrote in :

FYI, I pay no attention to content-free ad hominem responses (a sure sign of a bankrupt position), so you might as well save the typing.

Only calling on you for saying Mac addresses come from ARP, and then you jumped into a discussion that you had no clue about, so I called you on that as well.

On Fri, 13 Oct 2006 22:45:46 -0800, "Dana" wrote in :

You were wrong, rude, and devoid of meaningful content.

And as always, he doesn't have the grace to admit that he's in over his head and apologize for his errors. Nothing ever changes with Navas, no matter what group he tries to show off in.

WOL is totally unrelated for Boot from Net. By now he probably understands this, but he'll never admit his mistake.

On Sat, 14 Oct 2006 07:01:22 -0700, SMS wrote in :

Such ad hominems only serve to discredit you and your claims.

No mistake, as I've already shown. ( And how would you really know in any event, since you claim to be filtering out all of my posts? ;)

One last time:

What got this silly topic drift started was the claim was that storing the MAC address in EEPROM on a NIC was essential, particularly for "boot-from-LAN". That simply isn't true -- both RIPL (the correct term you still don't seem to know) and PXE are indifferent to where the MAC address is stored, as in the case of hosts that store the MAC address in their own non-volatile memory. Case in Point: The original Everex All-in-One PC, which I conceived and engineered, where Ethernet (including the MAC address) is done by the system BIOS. ( Your experience again? ;)

I mentioned Wake on LAN (WOL) because: [a] WOL is needed for remote boot, no matter what the boot method; [b] RIPL and PXE don't care about the MAC address, as noted above; and [c] WOL is dependent on the MAC address, unlike either RIPL or PXE, since the Magic Packet contains the MAC address, and since the NIC has to be functional while the host is sleeping.

Apology accepted.

Actually I was not.

Possibly, but so were you for jumping in on that discussion that was obvious, that you had no idea of what they were talking about.

And again you are in denial. He states facts and you want to call them ad hominems. Do you even know what an ad hominem is?

What you have shown is you did not understand what they were discussing.

And again you are wrong. Have you ever heard of a diskless workstation/dumb terminal, that is what these guys are talking about. The terminal do not have to have a wake on lan function, because the user is going to push the power button. What the issue is here is you have no idea of what they are talking about, either you have never heard of diskless workstations/dumb terminals and bootp (which is more than likely the case) or two, you do not understand this type of technology and are confusing it with wol.