Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 13192 - Routes not established correctly for x86 stage 1 build of 1.4_rc2
Summary: Routes not established correctly for x86 stage 1 build of 1.4_rc2
Status: RESOLVED FIXED
Alias: None
Product: Gentoo Release Media
Classification: Unclassified
Component: Everything (show other bugs)
Hardware: x86 Linux
: High minor (vote)
Assignee: Seth Chandler
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2003-01-03 17:41 UTC by Andrew Bevitt
Modified: 2005-03-25 11:24 UTC (History)
1 user (show)

See Also:
Package list:
Runtime testing required: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Andrew Bevitt 2003-01-03 17:41:14 UTC
When setting up 1.4_rc2 earlier this morning I could not get networking to work.

After issuing 
ifconfig eth0 192.168.1.11 broadcast 192.168.1.255 netmask 255.255.255.0
route add default gw 192.168.1.10 netmask 0.0.0.0 metric 1

It appeared that my network was up and running (ifconfig output was correct).
But all traffic (eg ping requests to my LAN gateway or domains) did not appear
to find a valid path through eth0. Instead the actaul traffic values seemed to
be registering on lo (the localhost device).

I now have rc2 installed but I had to boot from an old 1.2 CD and then use the
1.4_rc2 stage 1 tar ball instead of the 1.2 tar ball.

Networking worked perfectly from the 1.2 CD 

I am using the 8139too driver for a SMC Ultra EX NIC which is working (as I am
writting this from rc2). This problem also exists on 1.4_rc1 LiveCD's (for i686)
 as I first went back to it before the 1.2 CD.

I am waiting for a few things to finish emerging and then I will be testing this
with a few other cards (tulip driver's) and see if the problem exists there aswell.
Comment 1 Daniel Robbins (RETIRED) gentoo-dev 2003-01-03 22:44:01 UTC
OK, we'll be awaiting the results of your additional testing.
Comment 2 Andrew Bevitt 2003-01-09 00:33:57 UTC
Sorry for the delay, my  machine became LAN mission critical for a while but now
I have had some time to do some futher tests and have found the same problem
using the following drivers...

8139too (original)
tulip
8139c
smc-ultra
3c501

All cards work under installed systems with the above drivers and still function
at maximum speed after being moved between test machines.
Comment 3 Tobias Eichert 2003-01-09 07:09:36 UTC
Hmm..that's strange as it happens with a wide range of drivers..
maybe have a look at bug #8495..it's a similar situation.

Which kernels do you run on your installed systems? 
Comment 4 Andrew Bevitt 2003-01-09 10:13:17 UTC
Installed systems (I have three, two gentoo's and one debian).

The first Gentoo is running 2.4.18 (vanilla-sources) no patches whatsoever
contains the 3Com card.

The second Gentoo is 1.4_rc2 running the latest gentoo-source from the 2.4 tree
(19-r10) and that machine (this one) runs the 8139too card. I also tried a
different Realtek card on this machine for the 8139c driver.

The debian box, which I tried booting with the x86 stage1 liveCD has two Netgear
cards (tulip drivers) and a SMC Ultra card -- LAN gateway for two subnets.
The debian box runs a kernel.org download of the 2.4.20 kernel.

That bug #8495 seems fairly similar, am going to reboot now and see if the
noapci options work, will post reply in another few minutes to resolve that issue.
Comment 5 Andrew Bevitt 2003-01-09 10:44:01 UTC
OK ive just been playing around passing noapic or acpi=no to the kernel, on both
my 1.4_rc2 machine and the Debian machine and in both cases the results were
still the same.

I did some log reading and found that there was this message reported on both
machines aswell, only after I tried pinging a network machine.

NETDEV WATCHDOG: eth0: Transmit Timeout.

Possible thoughts on the issue are 
http://www.uwsg.iu.edu/hypermail/linux/kernel/0112.0/1613.html

Comment 6 Seth Chandler 2003-02-06 05:12:47 UTC
have you tried the newest stages?
Comment 7 Andrew Bevitt 2003-02-13 00:39:02 UTC
Finally found the problem and the solution.

The 1.4_rc2 candidate CD actually needs acpi=off passed to it, not acpi=no this disables acpi during the install, and the networking actually works.

Bug can be CLOSED now I think, it works with all the previousl;y listed cards and drivers with acpi=off passed to the kernel. 

NOTE: In the future, the acpi=X option may change, if off does NOT work then maybe you should try no :)
Comment 8 John Davis (zhen) (RETIRED) gentoo-dev 2003-04-04 01:20:11 UTC
db fix
Comment 9 John Davis (zhen) (RETIRED) gentoo-dev 2003-04-04 01:25:05 UTC
db fix
Comment 10 Chris Gianelloni (RETIRED) gentoo-dev 2005-03-25 11:24:54 UTC
Moving these so we can remove the "Install CD" component from "Gentoo Linux".

I apologize to everyone for this spam, but according to the bugzilla developers,
this is the only reasonable way to do this.