Re: [NTLK] Lantern, Get_Card_Info.pkg & IBM Ethernet Credit CardII

From: David Holmes (dh_ntlk_at_delphys.net)
Date: Sat Feb 01 2003 - 09:28:42 PST


Paul Guyot wrote:
>
> Did I mention that Lantern Data Link Layer Tools pdf is a good manual

Well I was already looking at this as well as the includes
but having another perspective on this really helped to clarify
things. What specifically helped was your explanation of the
card server and card handlers space. And having your example
of GetCardInfo is a great help. These two things help to fill in
some details that a newbie -reads- but may not -understand-.
 
> Very probably because parentheses aren't a problem. Hiroshi says
> there is an issue with the CIS string is longer than 64 or 32
> characters (or something like this), which seems weird to me as one
> could very probably convince Lantern to work with such cards.
> Still. If you have parentheses, just put them in the strings as in:
> Constant kDeviceCISProductName := "look, I have (parentheses)";

This I tried. As well as a number of other combinations (just
to see if there would be no install error - of course it wouldn't
match.) If the parentheses were in the string I would always get
an error on Install of the the package. I figured that
length was not a problem as the string is 32 chars:

Combo PCMCIA EthernetCard (EC2T)

but I will try replacing these instead of removing them.
I tried various escape and hex code things as well.

> It was able to get this information. But some other CIS records might
> be unreadable (by the Newton) and therefore the error in the parsing
<...>
> In particular, Hiroshi says that if the Newton accesses a 3V card at
> 5V, it can damage both the card and the Newton. You've be warned....

Yes - I believed that these two cards are both old enough to be
5V but this is worth confirming (I suspect I may have done this
years ago when I first ordered the Lantern CD from Apple but it
is worth rechecking.)

I should say that I'm somewhat familiar with PCMCIA from the Linux
environment as well as working with CF cards in assembler. The 2 cards
I have (IBM and Linksys) share the same module in David Hinds' work,
namely pcnet_cs. I'd really like to find some way to replicate his
kind of setup here - i.e. the card handler has an externally
available list of the card ids matched to the name of the driver.
Though, if I'm having trouble with parenthesis I have my work cut
out for me.
 
I guess my somewhat naive assumption is that because these cards
work in other environments (and because I'm assuming they're both 5V of
course) that it was just a "recognition" problem of some kind. Of
course there could be some broken things in CL chips I guess.

Here is a dump for the two cards (the IBM has a Vpp1 of
12V but this should just be for reprogramming I think.)

the IBM card:

  dev_info
    fn_specific 100ns, 16kb
  attr_dev_info
    FLASH 150ns, 4kb
  manfid 0x00a4, 0x0002
  funcid network_adapter [post] [rom]
  vers_1 4.1, "IBM Corp.", "Ethernet", "0934214",
    "PC-NIC ENCC 990010759-001A"
  config base 0x20000 mask 0x000b last_index 0x01
  cftable_entry 0x01 [default]
    [mwait]
    Vcc Vnom 5V Istatic 13mA Iavg 200mA
    Vpp1 Vnom 12V
    timing wait 2500ns
    io 0x0000-0x001f [lines=5] [16bit]
    irq mask 0xffff [level] [pulse]
    memory 0x11000-0x14fff @ 0x0000
  common_jedec 0x00 0x00, 0x89 0xb8
  attr_jedec 0x89 0xb8
  checksum 0x0000-0x0085 = 0x00

the Linksys card:

  dev_info
    fn_specific 100ns, 16kb
    FLASH 150ns, 60kb
  attr_dev_info
    FLASH 150ns, 4kb
  manfid 0x0149, 0xc1ab
  vers_1 4.1, "Linksys", "Combo PCMCIA EthernetCard (EC2T)", "2.0"
  config base 0x0fd0 mask 0x000b last_index 0x06
  cftable_entry 0x00 [default]
    [mwait]
    io 0x0000-0x001f [lines=5] [16bit]
    irq mask 0xbefc [level]
  cftable_entry 0x01
    io 0x0300-0x031f [lines=10] [16bit] [range]
  cftable_entry 0x02
    io 0x0320-0x033f [lines=10] [16bit] [range]
  cftable_entry 0x03
    io 0x0340-0x035f [lines=10] [16bit] [range]
  funcid network_adapter

> In case you really want to try to modify Newton Devices, change word
> at 0x2AEE from 00EB to 0182. It should plug ParseUnrecognizedCard to

Ah - "magic happens here". I will try this one - thanks!

--david
David Holmes
http://delphys.net/d.holmes

-- 
This is the NewtonTalk list - http://www.newtontalk.net/ for all inquiries
List FAQ/Etiquette/Terms: http://www.newtontalk.net/faq.html
Official Newton FAQ: http://www.chuma.org/newton/faq/


This archive was generated by hypermail 2.1.5 : Sun Feb 02 2003 - 20:30:00 PST