Hello,
the problem looks like this:
PING www.ee (212.107.32.146): 56 data bytes
64 bytes from 212.107.32.146: icmp_seq=0 ttl=249 time=3.7 ms
wrong data byte #34 should be 0x22 but was 0x28
14 15 16 17 18 19 1a 1b 1c 1d 1e 1f 20 21 28 4 75 0 22 23 24 21 22
23 24 25 22 23 24 25 22 23
24 25 ff ff 0 0 0 0 0 0 0 0 0 0 0 0
64 bytes from 212.107.32.146: icmp_seq=1 ttl=249 time=3.7 ms
wrong data byte #34 should be 0x22 but was 0x0
14 15 16 17 18 19 1a 1b 1c 1d 1e 1f 20 21 0 0 0 5 22 23 24 25 22
23 24 25 22 23 24 25 0 6
7 74 0 6 0 0 0 0 0 0 0 0 0 0 0 0
64 bytes from 212.107.32.146: icmp_seq=2 ttl=249 time=3.2 ms
wrong data byte #34 should be 0x22 but was 0xff
14 15 16 17 18 19 1a 1b 1c 1d 1e 1f 20 21 ff ff ff ff 22 23 24 25
22 23 24 25 22 23 24 25 22 23
24 25 ff 5 0 0 0 0 0 0 0 0 0 0 0 0
64 bytes from 212.107.32.146: icmp_seq=3 ttl=249 time=2.5 ms
wrong data byte #34 should be 0x22 but was 0x0
14 15 16 17 18 19 1a 1b 1c 1d 1e 1f 20 21 0 0 0 0 22 23 24 21 22
23 24 25 22 23 24 25 22 23
24 25 0 0 0 0 0 0 0 0 0 0 0 0 0 0
--- www.ee ping statistics ---
4 packets transmitted, 4 packets received, 0% packet loss
round-trip min/avg/max = 2.5/3.2/3.7 ms
the card is SMC1255TX which features an Accton EN5251B chip. It works fine
in x86 boxes with PCI 2.0 and PCI 2.1 (have not tried with 2.2 but should
work according to documentation). My tests were done with on x86 and alpha
with linux 2.4.18 kernel and with single and multiple cards of the same
and different types. The distribution is Debian unstable/woody.
I have a Ruffian Alpha box (which has PCI bus version 2.0, due to which I
could not use the currently sold 3com cards, which require pci 2.1). So,
having tested the SMC1255TX card on x86 boxes I tried to use them on the
Alpha box, but got the above error messages (the box appears dumb from
across the net).
Does the problem seem to be a hardware compatibility issue of SMC1255TX or
does it have anything to do with the Alpha port of the kernel?
I tried both the kernel version of the tulip driver and the unmodified
version of Donald Becker's pci-scan and tulip modules, but the result was
similar.
The worrying line in tulip-diag output (attached) is the following:
Interrupt sources are pending! CSR5 is fc06c812.
(note that the other tulip 21142/43 works fine with both versions of the
tulip module)
What could one do about it?
Juhan Ernits
P.S. I posted this also to tulip-bug@scyld.com.
This archive was generated by hypermail 2b29 : Tue Apr 30 2002 - 22:00:19 EST