Re: Trying to get 3c59x to work with two 3c900bs ...
Update: I found that my machine reacts similarly when I have two PCI
Bus LinkSys cards installed. A single LinkSys worked, but two of them
did not (using the pci-scan and tulip drivers off of Becker's site). My
next attempt then seems to have accomplished something...however it's
not fully tested yet. When I plug in one 3c900b and one LinkSys NIC and
then load >A different driver for each of my four NICs< then the 3c900b
responds okay. I now need to go through and try all the others and get
things together. The main thing of interest here seems to be that
having two of the same kind of PCI NIC failed for me for two kinds of
NICs, albeit not exactly in the same way. When I get the rest of this
put together and working, I'd like to revisit this if I can allocate the
time. It sounds to me like some test time is called for in these recent
NIC driver versions. Thanks for any feedback or further suggestions.
Xeno Campanoli wrote:
>
> according to Becker's suggestion to fix it, I needed to build his source
> file, as the version he built worked fine with multiple 3c900bs. I got
> the 3c59x.o file built with the 2.2.17 kernel, whose sources I
> downloaded in terms of .deb files (including the sources) from a debian
> mirror. I built the 3c59x.c from Donald Becker's netdrivers.tgz, but it
> gets the following diagnostic messages from an attempt to insmod:
>
> 3c59x.o: unresolved symbol acpi_wake
> 3c59x.o: unresolved symbol acpi_set_pwr_state
> 3c59x.o: unresolved symbol pci_drv_unregister
> 3c59x.o: unresolved symbol
> pci_drv_register
>
> I've never had to accompany this driver with any other module before, so
> presumably the version I've got isn't building right, either because it
> doesn't work with my kernel, or I'm just leaving something out in my
> novices attempt to hack the build.
>
> Has anyone else seen this problem or especially seen a solution to it?
> I really want to eventually stick in three 3c900b cards into this
> machine, and my testing shows with this kernel from Debian it won't work
> with two, but it will with one. Just in case you didn't see my other
> posting, Donald Becker sent me a response saying the 3c59x module I was
> using (from the latest Debian install of the time, early last week) was
> modified from the one he maintained, as his did not have this problem.
> Presumably this should be considered a bug with the Debian install for
> 2.2.17 kernel, or whatever the install set is called.
>
> --
> To UNSUBSCRIBE, email to debian-qa-request@lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: