Francois Romieu writes:
sebastian.ionita@focomunicatii.ro [off-list ref] :
[...]
quoted
Later i will restart the server with the new kernel and i will tell you if
something gets bad or if it will work ... :)
If it does not work as expected, please describe the minimal level of
operation which is ok (i.e. networking without vlan for instance).
The range of possible errors should be rather limited: mostly gross
typos or endianness issues when mucking with vlan related registers
(I am not too sure about the layout of the VID for Rx/Tx).
Without vlan's it works but with vlan the behavier is just like with the old
drivers. Shoud I give some param's to the module if want to have vlans ?
--
Ueimor
____________________________________________________________
SC. FO Comunicatii SRL.
Sebastian Ionita
Administrator Sistem
mobil: 0724 212408
tel fix: 0264 450456
sebastian.ionita@focomunicatii.ro [off-list ref] :
[...]
Without vlan's it works but with vlan the behavier is just like with the
old drivers. Shoud I give some param's to the module if want to have vlans ?
I have modified the original patch. Does the network appliance allow you
to configure some specific VID, say 68, 1028 or 1092 ?
If yes, pick one and test the behavior of the three VID on the network adapter,
ideally within their own insmod, config up, ping (from host and remote),
config down and rmmod session.
Provided the issue is only a matter of ordering of the VID, something should
appear in the log/packet statistics when the VID overlap.
dmesg/kernel log messages as well as the serie of commands issued would be
welcome. bzip2 is your friend.
If the VID on the lan side is fixed for you, please send it so I'll see which
values on the network adapter could match (the values above are not sensitive
to endianness for instance).
A patch against vanilla 2.4.18-rc2 is available at:
http://www.fr.zoreil.com/people/francois/misc/20041112-2.4.28-rc2-via-velocity-backport.patch
--
Ueimor