From: "Rich" 

This is a multi-part message in MIME format.

------=_NextPart_000_004B_01C68F94.813DF190
Content-Type: text/plain;
        charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

   PCI IDs are more then five bytes.  There is a 2 byte vendor ID, 2 =
byte device ID, 1 byte revision, 2 byte subsystem vendor ID, 2 byte =
subsystem ID, as well as the 3 byte class code.  =
http://www.pcidatabase.com/ has some of these IDs documented.

Rich

  "Don Hills"  wrote in message =
news:Jv5jEtgaXOSY092yn{at}attglobal.net...
  In article , "Rich"  wrote:
  >   64K more than enough?  Maybe for mca which failed.

  EISA had a 4 byte ID. How many EISA IDs were assigned?

  PCI has a 5 byte ID. How many unique IDs have been generated to date?

  In all 3 cases, the address space is much more sparsely populated =
than,
  for example, the IPv4 address space.

  >   As for identifying compatible devices, mca lacks this mechanism.
  >It's entirely a guessing game.  The PCI approach is that the device
  >identifies with what it is compatible.

  That wasn't allowed. The device driver had to recognise the adapter =
ID.
  This ensured that the device driver had actually been tested with the
  adapter and proven to work reliably. Mainframe mindset.

  --=20
  Don Hills
------=_NextPart_000_004B_01C68F94.813DF190
Content-Type: text/html;
        charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable








   PCI IDs
are more then five =

bytes.  There is a 2 byte vendor ID, 2 byte device ID, 1 byte =
revision, 2=20
byte subsystem vendor ID, 2 byte subsystem ID, as well as the 3 byte = class=20
code.  http://www.pcidatabase.com/ " target="new">http://www.pcidatabase.com/ =">http://www.pcidatabase.com/">http://www.pcidatabase.com/ =
;has some=20
of these IDs documented.
 
Rich
 

  "Don Hills" <black.hole.4.spam{at}gmail.commailto:black.hole.4.spam{at}gmail.com">black.hole.4.spam{at}gmail.com
A>>=20
  wrote in message news:Jv5jEtgaXOSY092yn{at}attg=
lobal.net...In=20
  article <448f5ccf{at}w3.nls.net>,=20">mailto:448f5ccf{at}w3.nls.net">448f5ccf{at}w3.nls.net>,=20
  "Rich" <{at}>
wrote:>   64K more than =
enough?  Maybe=20
  for mca which failed.EISA had a 4 byte ID. How many EISA IDs =
were=20
  assigned?PCI has a 5 byte ID. How many unique IDs have been =
generated=20
  to date?In all 3 cases, the address space is much more =
sparsely=20
  populated than,for example, the IPv4 address=20
  space.>   As for identifying
compatible devices, =
mca=20
  lacks this mechanism.>It's entirely a guessing
game.  The =
PCI=20
  approach is that the device>identifies with what it is=20
  compatible.That wasn't allowed. The device driver had to =
recognise the=20
  adapter ID.This ensured that the device driver had actually been =
tested=20
  with theadapter and proven to work reliably. Mainframe =
mindset.--=20
  Don Hills 

------=_NextPart_000_004B_01C68F94.813DF190--

--- BBBS/NT v4.01 Flag-5