Re: [gentoo-user] Another Install Issue

[Mick]

On Sunday 16 Oct 2011 15:52:51 CJoeB wrote:
> On 10/16/11 06:43, Mick wrote:
> > On Sunday 16 Oct 2011 00:05:01 CJoeB wrote:
> >> Hi everyone,
> >>
> >>
> >> However, now I have a networking issue.
> >>
> >> In past, when I booted to the install CD and my ethernet connection was
> >> not active, I typed net-setup eth0 and was able to set it up.  This
> >> time, when I booted to the install CD and typed net-setup eth0, the
> >> network card was not recognized.
> >
> > This means that your kernel is not configured with the corresponding
> > modules for your network card, or that there is some other configuration
> > problem with e.g. firmware loading (if such a thing is necessary for
> > your card).
>
> I use genkernel.  Granted, rather than just running genkernel all, I run
> genkernel --menuconfig all because I remove stuff that I know I won't
> need.  However, in the network section, everything related to broadcom
> and tg3 is built as a module.  I even tried building tg3 directly into
> the kernel and it didn't help.  I had to do the 'modprobe -r broadcom'
> and 'modprobe -r tg3' and then, 'modprobe broadcom' and 'modprobe tg3'
> prior to running net-setup even when booted to the install CD.

From what your dmesg shows tg3 fails to load.  So there should not be a need 
to remove it.


> >> I googled and found a post where
> >> someone said that they had to 'modprobe -r broadcom' and 'modprobe -r
> >> tg3' and then 'modprobe broadcom' and 'modprobe tg3' and then, run
> >> net-setup.  I did this and then ifconfig returned my eth0 connection.
> >>
> >> Of course, later you have to do the cp -L /etc/resolv.conf
> >> /mnt/gentoo/etc/ .... which I did and dhcpcd has been added to my
> >> default runlevel.
> >>
> >> However, when I boot, eth0 does not start.
> >
> > What does dmesg show?
>
> On line states:
> tg3 0000:03:00.0:  Problem fetching invariants of chip, aborting

It fails to load.  Something is blocking it from accessing the NIC chipset.


> > What does cat /var/log/syslog | grep -i tg3 show
>
> Don't have a syslog file .... however,
>
> 'cat /var/log/rc.log | grep -i tg3' shows:
> *Loading modul tg3 ...
>
> The line is repeated a total of 7 times.

It should only be there once, unless you tried to load it 7 times?


> >> I can start it manually by
> >> doing 'modprobe -r broadcom' and 'modprobe -r tg3' and then 'modprobe
> >> broadcom' and 'modprobe tg3'
> >
> > So the question is why when you remove and install the module manually
> > your NIC driver loads, but at boot time it does not ...
>
> Good question and I have no idea!  :-)

I looked into it and the answer seems to be a bug in the kernel (there are 
patches out which fix this) that does not make sure that broadcom is loaded 
before tg3.


> >   ifconfig  and ifconfig -a  (before and after you modprobe the driver)
>
> Before loading:
>
> ifconfig just lists:
> lo    Link encap:Local Loopback
>        inet addr:127.0.0.1 Mask:255.0.0.0
>        UP LOOPBACK RUNNING MTU:16436 Metric:1
>        RX packets:11 errors:0 dropped:0 overruns:0 frame:0
>        TX packets:11 errors:0 dropped:0 overruns:0 carrier:0
>        collisions:0 txqueuelen:0
>        RX bytes:198 (198.0 B)  TX bytes:198 (198.0 B)
>
> ifconfig -a lists:
> lo    Link encap:Local Loopback
>        inet addr:127.0.0.1 Mask:255.0.0.0
>        UP LOOPBACK RUNNING MTU:16436 Metric:1
>        RX packets:18 errors:0 dropped:0 overruns:0 frame:0
>        TX packets:18 errors:0 dropped:0 overruns:0 carrier:0
>        collisions:0 txqueuelen:0
>        RX bytes:324 (324.0 B)  TX bytes:324 (324.0 B)

The module is not loaded and the kernel has not seen the NIC.


> After doing ''modprobe -r broadcom' and 'modprobe -r tg3' and then,
> 'modprobe broadcom' and 'modprobe tg3'
>
> ifconfig lists:
>
> eth0    Link encap:Ethernet  HWaddr 78:2b:cb:ad:9e:47
>            inet addr:192.168.2.15  Bcast:192:168.2.255  Mask:255.255.255.0
>            UP BROADCAST RUNNING MULTICAST  MTU:1500 Metric:1
>            RX packets:1727 errors:0 dropped:0 overruns:0 frame:0
>            TX packets:908 errors:0 dropped:0 overruns:0 carrier:0
>             collisions:0 txqueuelen:1000
>             RX bytes:1826702 (1.7 MiB)  TX bytes:67525 (65.9 KiB)
>
>
> lo         Link encap:Local Loopback
>             inet addr:127.0.0.1 Mask:255.0.0.0
>             UP LOOPBACK RUNNING MTU:16436 Metric:1
>             RX packets:46 errors:0 dropped:0 overruns:0 frame:0
>             TX packets:46 errors:0 dropped:0 overruns:0 carrier:0
>             collisions:0 txqueuelen:0
>             RX bytes:828 (828.0 B)  TX bytes:828 (828.0 B)
>
> ifconfig -a lists
>
> eth0    Link encap:Ethernet  HWaddr 78:2b:cb:ad:9e:47
>            inet addr:192.168.2.15  Bcast:192:168.2.255  Mask:255.255.255.0
>            UP BROADCAST RUNNING MULTICAST  MTU:1500 Metric:1
>            RX packets:2024 errors:0 dropped:0 overruns:0 frame:0
>            TX packets:958 errors:0 dropped:0 overruns:0 carrier:0
>             collisions:0 txqueuelen:1000
>             RX bytes:1846649 (1.7 MiB)  TX bytes:70725 (69.0 KiB)
>
>
> lo         Link encap:Local Loopback
>             inet addr:127.0.0.1 Mask:255.0.0.0
>             UP LOOPBACK RUNNING MTU:16436 Metric:1
>             RX packets:50 errors:0 dropped:0 overruns:0 frame:0
>             TX packets:50 errors:0 dropped:0 overruns:0 carrier:0
>             collisions:0 txqueuelen:0
>             RX bytes:1062 (1.0 KiB)  TX bytes:1062 (1.0 KiB)
>
> > It may also help to know what is your card (lshw and lspci -v).
>
> Okay, I had to install  pciutils before I could get the output so,
> obviously, this was done after the drivers loaded.
>
> lspci -v shows:
>
> 03:00.0    Ethernet controller:  Broadcom Corporation Netlink BCM57788
> Gigabit Ethernet PCIe (rev 01)
>                 Subsystem: Dell Device 04aa
>                 Flags:  bus master, fast devsel, latency 0, IRQ 47
>                 Memory at fe500000 (64-bit, non-prefetchable) [size=64K]
>                 Capabilities:  [48] Power Management version 3
>                 Capabilities:  [60] Vendor Specific Information: Len=6c <?>
>                 Capabilities:  [50] MSI: Enable+ Count=1/1 Maskable- 64bit+
>                 Capabilities:  [cc] Express Endpoint, MSI 00
>                 Capabilities:  [100] Advanced Error Reporting
>                 Capabilities:  [13c] Virtual Channel
>                 Capabilities:  [160] Device Serial Number
> 78-2b-cb-ff-fe-ad-9e-47
>                 Capabilities:  [16c] Power Budgeting <?>
>                 Kernel driver in use: tg3
>                 Kernel modules:  tg3

As I said above the kernel ought to manage the order in which the modules and 
dependencies are loaded.  This is a bug.  There are patches out to address 
this in future kernels.

Until then you can try to:

1. Configure your /etc/conf.d/modules as suggested here:

    http://www.gentoo.org/doc/en/openrc-migration.xml

and have broadcom before tg3.

2. Configure broadcom in the kernel and tg3 as module (in case it makes a 
difference, or swap and see if that helps).

3. Have a custom init script (re)loading them in the right order as Florian 
has already suggested.

HTH.
-- 
Regards,
Mick