[eepro100] Intel 815EFV

Alf elvin@elvin.net
Sat, 21 Jul 2001 03:18:41 +0800 (SGT)


Hello,

	Some quick background. My company (in Singapore) is developing
thin client solutions for kiosk-based systems. We were previously using
the Intel 815EEAL, which is now EOL (End of Line) [This board works 100%
with no problems with the eepro100 drivers]
	We have now obtained the Intel 815EFV (Micro-ATX form factor, the
normal ATX form factor is coming in this monday or so) which is using the
same Intel 815E chipset.

	The network interface is acting up though. (I've tested the 2.2.19
kernel eepro100, 2.4.6 eepro100, and the ftp.scyld.com's v1.15 drivers,
all the *same* results)
	When I *first* boot up cold, the network interface will be
detected, initialized, but not work at all. If I do warm reboot, the
network interface will be detected, initizliaed and *WORKS*. (Using
modules, I rmmod, insmod for the cold boot, but it does not work, a reboot
is needed to make it work)

	One observation is that on cold boot and the network interface is
not working, the output of ifconfig shows 0 RX packets, but many TX
packets. (I usually use ping to test, when it does not work, i get zero
replies, and the arp tables are incomplete, RX never increases)
	I am working in a 10/100 switched network environment (Tried a
10mb hub and another 10/100 switch, no diff)

	I noticed this text
	"Receiver lock-up bug exists - enablign work-around" for the v1.15
drivers. (Though the work-around does not seem to work on first bootup)

	And for v1.15, the General self-test and Serial sub-system
self-test will *FAIL* when it's cold booting. (After a reboot, they all
pass and the network card works) Note that the 2.2.19's v1.09j always
*PASSES* all the tests.

	Note that this happens only on linux. The windows drivers works
fine on the first boot,etc... *grrr*

	My apologies if this mail is long, and thanks in advance to
whomever can provide me any clues or diagnostic directions.

Elvin Tan
-