Hi!
I'm getting `subj` (with different numbers, 1020002, 1020004,
sometimes 102006). ifconfig eth0 up fails, network is completely
broken.
Devices are 14e4:1639... Any hints/ideas? (It is possible this is
broken hardware).
Pavel
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
On Thu, 2008-12-04 at 02:55 -0800, Pavel Machek wrote:
> Hi!
>
> I'm getting `subj` (with different numbers, 1020002, 1020004,
> sometimes 102006). ifconfig eth0 up fails, network is completely
> broken.
Which kernel are you using? If it used to work before 2.6.26 or so, it
may be because of the PCI VPD access going beyond the size limit. We
should have a PCI quirk to prevent this, but may be it is not complete.
>
> Devices are 14e4:1639... Any hints/ideas? (It is possible this is
> broken hardware).
> Pavel
>
> On Thu, 2008-12-04 at 02:55 -0800, Pavel Machek wrote:
> > Hi!
> >
> > I'm getting `subj` (with different numbers, 1020002, 1020004,
> > sometimes 102006). ifconfig eth0 up fails, network is completely
> > broken.
>
> Which kernel are you using? If it used to work before 2.6.26 or so,
> it
I'm using 2.6.28-rc7. I did not try any previous kernels.
> may be because of the PCI VPD access going beyond the size limit. We
> should have a PCI quirk to prevent this, but may be it is not
> complete.
I could not find quirk handling, so I enabled the debug and workaround
options I could find...
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
On Mon, 2008-12-08 at 06:54 -0800, Pavel Machek wrote:
> >
> > On Thu, 2008-12-04 at 02:55 -0800, Pavel Machek wrote:
> > > Hi!
> > >
> > > I'm getting `subj` (with different numbers, 1020002, 1020004,
> > > sometimes 102006). ifconfig eth0 up fails, network is completely
> > > broken.
> >
> > Which kernel are you using? If it used to work before 2.6.26 or so,
> > it
>
> I'm using 2.6.28-rc7. I did not try any previous kernels.
Please provide ethtool -i eth0
and lspci -xxx on the Broadcom device.
Thanks.
>
> > may be because of the PCI VPD access going beyond the size limit. We
> > should have a PCI quirk to prevent this, but may be it is not
> > complete.
>
> I could not find quirk handling, so I enabled the debug and workaround
> options I could find...
>
On Mon 2008-12-08 09:12:25, Michael Chan wrote:
>
> On Mon, 2008-12-08 at 06:54 -0800, Pavel Machek wrote:
> > >
> > > On Thu, 2008-12-04 at 02:55 -0800, Pavel Machek wrote:
> > > > Hi!
> > > >
> > > > I'm getting `subj` (with different numbers, 1020002, 1020004,
> > > > sometimes 102006). ifconfig eth0 up fails, network is completely
> > > > broken.
> > >
> > > Which kernel are you using? If it used to work before 2.6.26 or so,
> > > it
> >
> > I'm using 2.6.28-rc7. I did not try any previous kernels.
>
> Please provide ethtool -i eth0
>
> and lspci -xxx on the Broadcom device.
Attached. It actually has eth0 & eth1, both seem to be same broadcom...
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
On Tue, 2008-12-09 at 23:56 -0800, Pavel Machek wrote:
> On Mon 2008-12-08 09:12:25, Michael Chan wrote:
> >
> > Please provide ethtool -i eth0
> >
> > and lspci -xxx on the Broadcom device.
>
> Attached. It actually has eth0 & eth1, both seem to be same broadcom...
>
It looks like the chip is bad or the firmware is bad. The PCI revision
ID and the serial number are all zeros. Perhaps we can try to reprogram
the NVRAM or do further diagnostics tests. I'll have someone contact
you. Thanks.