Text 5750, 46 rader
Skriven 2006-04-18 18:49:28 av mark lewis (1:3634/12.0)
   Kommentar till text 5740 av Stephen Haffly (1:396/45.27)
Ärende: Getting PCMCIA modem to work
====================================
 ml> being PCMCIA, my first thought is if it can ride on the normal com
 ml> ports... 3f8, 2f8, 3e8, and 2e8 or if it is riding on another
 ml> address... one way to find out is to boot an unadulterated DOS floppy
 ml> and use debug to check the hardware table in the BIOS...

 SH> WinXP reported it as com4, which would be 2e8 IIRC.

yes, that is the normal address for COM4... however, please note a situation
that i do not think is widely known...

the original specs put out by IBM do not really differentiate between ports...
their default COM1 and COM2 are/were 3f8 and 2f8... i do not recall how COM3
and COM4 became 3e8 and 2e8... however (here's the situation)... there is
nothing (that i recall or can find right now) that specifically states that the
first com port in a machine has to be at 3f8... i've several machines that the
first com port is 2f8 and when the BIOS goes thru it's setup, 2f8 gets loaded
into the COM1 position... so, in that case, COM1 is 2f8 according to the
BIOS... the next one found is put into the COM2 slot and the system knows it as
COM2... 

 SH> <SNIP>

 ml> FWIW: this is also the section of the system that programs used to use
 ml> to determine what ports are installed on a machine and what addresses
 ml> they are using... you remember those old programs? like telix and
 ml> procomm and such? you'd just tell it to use COM2 and this is where
 ml> it'd go to get the address to access the hardware with ;)

 SH> Yes.  I do remember such programs, although I never used either of
 SH> the two you mentioned at home.  I used to use ripterm a lot.

wow... i remember ripterm... it is but a child compared to those others i
mentioned... they were two of the most popular that i know of and that is why i
mentioned them... there is/was also {COMMO} and several others that i'm drawing
a blank on right now...

 SH> It appears the problem is changes in the way 2.6 kernels handle
 SH> pcmcia. I am looking at kernel.org's howto to try to make heads or
 SH> tails of it now.

yes, the creation of PCMCIA modems did throw a spanner into the gearbox in many
cases... hopefully you will find the fix and be able to take advantage of it ;)

)\/(ark

 * Origin:  (1:3634/12)