Hey guys. I'm pretty sure my problem here isn't exactly CE related, but I
really don't have anywhere else to go for help right now. I'm hoping
someone here has run into a similar problem and is able to shine some light
on my situation.
We're working on a Mainstone II eval board from Intel. Our current OS
configuration is a very stripped down CE 4.20 platform with minimal GWES,
NDIS, PCMCIA, and the default NE2000 compatible ethernet driver. We're to
get a Socket LP-E compact flash ethernet card (with PCMCIA adapter) running
so we can do some more app development. We've had this exact same card,
with the exact same driver, running on our own hardware platform with no
problems. Unfortunately, we see some goofy things on the Mainstone II.
After booting, we see the card detect, we see PCMCIA run the detect routine
for the driver, and we see the correct driver get loaded. When using DHCP,
we even see an initial DHCP discovery packet go out from the card. However,
we never see the card respond to the DHCP response. It really looks like
we're not getting any interrupts from the card. We can remove the card,
re-insert it, see the driver reload, etc, but we never get a transmit or
receive interrupt.
We've been digging down deep into the OAL to verify all of the interrupts
are being properly enabled and disabled in the various OEMInterrupt*
functions, and everything seems to be conceptually fine. But, we're seeing
something strange when the driver tries to enable the interrupts on the
board. The Mainstone II board has a big-honkin' FPGA that controls all of
the PCMCIA interrupt sources. When we watch the driver try to unmask the
interrupts through the FPGA's "Platform Interrupt Mask/Enable Register" it
looks like it doesn't work. We see the correct bits getting sent to this
register, but then when we read them back, it looks like they don't change.
First, has anyone else actually used the PCMCIA driver for an NE2000
ethernet card on the Mainstone II board? (We're working with a modified
PPC2003 BSP, build 016.) If so, have you seen any problems like this?
We're starting to question the programming of the FPGA, and we're trying to
get the code for it from Intel, but that always takes forever.
Any help would be greatly appreciated guys. This one has us stumped. (When
our hardware guy starts suggesting things like physically replacing the
FPGA, I get worried.) ;)
Thanks in advance guys.
Mark Murawski
Vocollect, Inc.