Hi,
Environment: Kernel 2.5.24, SuSE Linux 7.2, Intel Chipset 440FX
NIC: RTL1839 connected to ADSL Modem
When trying to connect to ADSL, I get the following error msgs (I guess the
"ioctl(SIOCGIFHWADDR): No such device" is the important part):
pppd[823]: pppd 2.4.1 started by root, uid 0
pppd[823]: Using interface ppp0
pppd[823]: Connect: ppp0 <--> /dev/pts/0
pppoe[824]: ioctl(SIOCGIFHWADDR): No such device
pppd[823]: Modem hangup
pppd[823]: Connection terminated.
pppd[823]: Exit.
Have not tried 2.5.23 but 2.5.22 works fine. Since there have been changes to
the 8139too driver I guess thats it. Unfortunately I do not know where to fix
this.
Regards,
Axel
Axel Siebenwirth wrote:
> Have not tried 2.5.23 but 2.5.22 works fine. Since there have been changes to
> the 8139too driver I guess thats it. Unfortunately I do not know where to fix
> this.
None that should affect this, however.
Can you please copy your 2.5.22 drivers/net/8139too.c into 2.5.24, and
test 2.5.24 with the older driver?
Jeff
Hi Jeff!
On Mon, 01 Jul 2002, Jeff Garzik wrote:
> >Have not tried 2.5.23 but 2.5.22 works fine. Since there have been changes
> >to the 8139too driver I guess thats it. Unfortunately I do not know where
> >to fix
> >this.
>
> None that should affect this, however.
>
> Can you please copy your 2.5.22 drivers/net/8139too.c into 2.5.24, and
> test 2.5.24 with the older driver?
Ok. I copied 8139too driver of 2.5.22 to 2.5.24 and it it was still the same.
I attached output of "lspci -vvv" and "dmesg".
"ifconfig eth1 up" returns with something like "device not found".
Regards,
Axel
Hi Jeff!
On Mon, 01 Jul 2002, Jeff Garzik wrote:
> Axel Siebenwirth wrote:
> >Have not tried 2.5.23 but 2.5.22 works fine. Since there have been changes
> >to the 8139too driver I guess thats it. Unfortunately I do not know where
> >to fix
> >this.
Kai made some changes about net_dev_init ...
<[email protected]>
Make net_dev_init() an __initcall
If you were looking to find where net_dev_init() is called, you
wouldn'
t
have guessed it's in drivers/block/genhd.c, would you?
Nothing should break if the __initcall net_dev_init is called too
late
now, since register_netdevice() will call it for us in that case.
With my poor knowledge I guess this could be it?! If you are sure it's not
the 8139too driver.
Regards,
Axel
Hi!
Kai Germaschewski schrieb am Donnerstag, den 04. Juli 2002:
> On Tue, 2 Jul 2002, Axel Siebenwirth wrote:
>
> > Since I cannot initialize the network device eth1 for the RTL8139 card, I
> > thought your changes about net_dev_init
> >
> > <[email protected]>
> > Make net_dev_init() an __initcall
> >
> > may have caused this.
>
> Yes, that seems quite possible. I'll submit a patch which fixes this
> shortly. In the mean time, you could try to find the __initcall line
> in net/core/dev.c, and replace __initcall by subsys_initcall.
>
> If you do so, please let me know if it fixes the problem for you.
>
Still the same problem in 2.5.25. Replacing __initcall with subsys_initcall
fixes it.
Regards,
Axel