More information on this oddity…
(makes a little more sense to me now, also)
Some history…
Prior to the reported event, I was running with one PCI Tulip card, and a
Ne2000 (Realtek 8019as based) ISA card.
The ISA card was not detected by the enumerators. Only the tulip driver was
running.
More information about the reported case.
All three additional PCI cards added to the system were NE2000 types (1
Winbond 89C940F, 2 Realtek 8029AS )
New information:
This morning I removed one of the NE2000 based PCI cards, and added another
tulip-based PCI card.
When I re-booted the system I got:
- 5 complaints from “netmanager” about invalid arguments
- All PCI cards with appropriate nicinfo
- Only 2 duplicate “drivers” running for the ISA based NE2000 compatible
card.
My theory (broad strokes):
- enumeration does not detect the ISA Ne2000-like card.
- enumeration does detect all of the PCI NE2000-like cards
- each time the NE2000 driver is started for (or detects) a PCI card, it
also detects the ISA card, and starts a “driver” for it also
- For some reason, the NE2000 driver code is unable to detect that it has a
driver already running for the ISA card, and starts another “driver” for it.
just some more obversations to use for debugging this weird situation…
Steve Munnings, Corman Technologies <steve@cormantech.com> wrote in message
news:988cls$pin$1@inn.qnx.com…
Hi,
As I am developing some networking software here, I stuffed a test machine
with 5 (yes, 5) networking cards.
4 were various PCI cards, and the 5th was an ISA (not PnP) card.
When I did “ls /dev/io-net” I saw seven network drivers started (en0
through
en6)!
When I used nicinfo to figure out who was who, I got the following
results:
- en0 - one of the PCI cards - cool
- en1 - another PCI card - cool
- en2 - another PCI card - cool
- en3 - the ISA card - first driver
- en4 - the ISA card - a “duplicate driver” - same MAC and all (except
for
“lan number”)
- en5 - the ISA card - a “duplicate driver” - same MAC and all (except
for
“lan number”)
- en6 - the last PCI card - cool - good stats, etc.,
The stats were all zeroes for the ISA card (it is plugged into a real
network, as is the en6 card. The other cards are unconnected)
SInce it is the ISA card that has the “weird drivers”, I did not include
the
output from pci - the pci drivers seem fine!
This is not a show stopper for me, but I thought you would like to know,
and
maybe figure out what went wrong…
\