Re: Further on P4R800-VM and 3c59x
Andrej Prsa <[email protected]> Mon, 12 Jan 2004 02:08:49 +0100
| Newsgroups | gmane.linux.drivers.vortex.devel |
|---|---|
| Organization | Faculty of Mathematics and Physics |
| Message-ID | <[email protected]> |
Hi!
Sorry for the late reply; I didn't have Windows installed on my desktop
and after your suggestion to try to warm-boot from windows to linux to see
if functionality is preserved, I tried installing them. Sadly, the CD was
bad and it somehow managed to destroy hard disk's superblock. %-((( I
spent two days re-installing Debian, getting new Windows CD, installing
that, ... Phew, I just remembered why I didn't like windows in the first
place! ;)
So, to my testing; I didn't yet thoroughly test windows->linux transition
+ disabling all ACPI functionality, so I'll get back on this shortly. And
now the results of your newest suggestion; since I'm not very skilled with
networking, I'll document each step I made, so please correct me if I did
anything foolish that could compromise the results of the tests! :)
> - no TxReset and no RxReset (this is a bit crazy, but if the
> card comes up in a working state and the driver is destroying this)
1) Commented out 3c59x entry /etc/modules
2) Commented out alias and options entries in /etc/modutils/eth0
3) Ran update-modules
4) Checked /etc/modules.conf to make sure it's clean
5) Cold-rebooted desktop
6) Issued ifconfig to make sure that eth0 wasn't configured
7) NEdited 3c59x.c and made changes according to instructions
8) Compiled both pci-scan.c and 3c59x.c
9) Insmod'd "pci-scan.o" and "3c59x debug=31"
10) ifconfig'd eth0
11) Tried pinging my laptop, unfortunately to no avail. I continuously
ran tcpdump on my laptop to see whether anything comes through and
it doesn't.
> - TxReset with a mask: 0, 2, 4 and 6. For this the line 1303 should look
> like:
>
> outw(TxReset | 4, ioaddr + EL3_CMD);
1) Removed 3c59x.o and pci-scan.o from /lib/modules/2.4.18/kernel/net
just to be absolutely sure
2) Added printk ("*** My change: ***\n"); to the source to recognize the
modified module insertion in dmesg output
3) Compiled, ifconfig eth0; rmmod 3c59x; rmmod pci-scan; insmod pci-scan;
insmod 3c59x; ifconfig eth0 [options] up; ping 192.168.0.2. While
doing this, I monitored the connection with tcpdump on the laptop and
with dmesg on the desktop. For all combinations (TxReset 0 through 8
and RxReset 0 through 8) no improvement was noted.
Just to be absolutely sure: 4 is equivalent to 0x04, right? The original
source contained 0x04 and I was using simply | 4.
Finally, under Windows I checked how the Device manager sees this card:
3Com 3C920B-EMB-WNM Lan
Driver: el90xbc5.sys
I/O Range: DC00-DC7F
Memory Range: FDE00000-FDE0007F
IRQ 18
I don't believe this is relevant, but I don't want to miss anything that
may give you guys a clue! O:)
Best wishes,
Andrej
_______________________________________________
vortex mailing list
[email protected]
http://www.scyld.com/mailman/listinfo/vortex