kernel: spurious 8259A interrupt: IRQ7
by Brian Millett
Hello,
I have a toshiba laptop with a miniPCI EL-2511MP PLUS 802.11/b card in it.
I am running kernel 2.6.5-1.358 with all of the updates as of today.
I am a bit confused (understatement)
I just lost ethernet connectivity and had to power cycle the laptop. I
was getting so many messages written to /var/log/messages that the system
was unresponsive. I was using XMMS listening to radio station.
Messages start with:
May 11 10:10:31 mktg6nt kernel: spurious 8259A interrupt: IRQ7.
May 11 10:41:27 mktg6nt kernel: eth1: error -110 reading info frame. Frame
dropped.
May 11 10:41:27 mktg6nt kernel: eth1: Error -110 writing Tx descriptor to BAP
May 11 10:41:29 mktg6nt last message repeated 349 times
May 11 10:41:29 mktg6nt kernel: hermes @ MEM 0x21ea1000: Timeout waiting
for command completion.
May 11 10:41:29 mktg6nt kernel: hermes @ MEM 0x21ea1000: Error -16 issuing
command.
May 11 10:41:29 mktg6nt kernel: eth1: Error -110 writing Tx descriptor to BAP
May 11 10:41:30 mktg6nt last message repeated 63 times
May 11 10:41:30 mktg6nt kernel: eth1: Ereth1: Error -110 writing Tx
descriptor to BAP
May 11 10:41:30 mktg6nt kernel: eth1: Error -110 writing Tx descriptor to BAP
May 11 10:41:31 mktg6nt last message repeated 81 times
May 11 10:41:31 mktg6nt kernel: eth1: Error -110 writing Tx eth1: Error
-110 writing Tx descriptor to BAP
May 11 10:41:31 mktg6nt kernel: eth1: Error -110 writing Tx descriptor to BAP
May 11 10:41:31 mktg6nt last message repeated 81 times
May 11 10:41:31 mktg6nt kernel: eth1: Error -110 writing Tx eth1: Error
-110 writing Tx descriptor to BAP
What I am really confused about are the following output from dmesg.
eth0: Station identity 001f:0009:0001:0004
eth0: Looks like an Intersil firmware version 1.4.9
eth0: Ad-hoc demo mode supported
eth0: IEEE standard IBSS ad-hoc mode supported
eth0: WEP supported, 104-bit key
eth0: MAC address 00:02:6F:04:64:6A
eth0: Station name "Prism I"
eth0: ready
8139too Fast Ethernet driver 0.9.27
divert: allocating divert_blk for eth1
eth1: RealTek RTL8139 at 0x3000, 00:02:3f:8f:19:0e, IRQ 10
eth1: Identified 8139 chip type 'RTL-8100B/8139D'
BUT I have this from ifconfig:
mktg6nt: ifconfig -a
eth1 Link encap:Ethernet HWaddr 00:02:6F:04:64:6A
inet addr:192.9.200.42 Bcast:192.9.200.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:4129 errors:0 dropped:0 overruns:0 frame:0
TX packets:1028 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:826960 (807.5 Kb) TX bytes:177917 (173.7 Kb)
Interrupt:11 Base address:0x8000 Memory:e0500000-e0500fff
and in my modprobe.conf:
alias eth0 8139too
alias eth1 orinoco_pci
I am writting this with my laptop, so it came back up clean. I just do
not understand the differences in the eth0/eth1 being reversed in the
dmesg output from the ifconfig output.
Should I change the modprobe.conf to have the orinoco_pci aliased to eth0
and the builtin 8139too aliased to eth0?
What is the 8259A interupt? It seems like the builtin sound card.
Thanks.
--
Brian Millett
Enterprise Consulting Group "Shifts in paradigms
(314) 205-9030 often cause nose bleeds."
bpmATec-groupDOTcom Greg Glenn
20 years
Latest tree broke rpm?
by Matthias Saou
Hi,
I just finished upgrading to today's development tree, and was surprised to
see yum bailing out when I tried to clean up the download archives :
[... complete yum upgrade ...]
Transaction(s) Complete
[dude@python2 dude]$ yum clean all
Traceback (most recent call last):
File "/usr/bin/yum", line 22, in ?
import yummain
File "/usr/share/yum/yummain.py", line 22, in ?
import clientStuff
File "/usr/share/yum/clientStuff.py", line 18, in ?
import rpm
ImportError: /lib/librt.so.1: symbol __librt_multiple_threads, version GLIBC_PRIVATE not defined in file libc.so.6 with link time reference
[dude@python2 dude]$
And now I get this :
[dude@python2 dude]$ rpm -qa
rpm: relocation error: /lib/librt.so.1: symbol __librt_multiple_threads, version GLIBC_PRIVATE not defined in file libc.so.6 with link time reference
[dude@python2 dude]$
I've just noticed the problem, now time to investigate some more...
Matthias
--
Clean custom Red Hat Linux rpm packages : http://freshrpms.net/
Fedora Core release 1.92 (FC2 Test 3) - Linux kernel 2.6.5-1.349
Load : 0.06 0.36 0.45
20 years
re: Help Needed: Firefox 0.8 and Thunderbird 0.6
by PLinnell
<snip>
>For example, we have made this kind of very successful partnership
>with
>three upstream gaim developers. They are automatically added to all
>Fedora gaim bug reports, and have helped us immensely to be very
>responsive to user bugzilla reports and package fixing. (Currently
>there are ZERO open gaim reports in FC!)
>
>Warren Togami
>wtogami(a)redhat.com
This is one of the main reasons I began to become involved in Fedora.
Speaking only of Scribus, I want to know of *any* issues so that: a. we
can determine if it is a Fedora specific or packaging problem. b. make
Fedora aware of updates and new releases. c. provide a work around if
possible.
Once CVS is opened, this is something IMO which needs to be formalized.
Every project should have at its option the ability to be cc'd on all
bug reports relating to their package.
Our experience with Debian and Scribus is instructive. There were
unknown to us for a long time, many stale bug reports in Debian's
tracker for Scribus. We also received a disproportionate amount of
complaints and issues from Debian users.
Scribus now has an outstanding Debian maintainer, who has worked very
closely with us to resolve most all the issues. Results: happy users,
fewer issues.
Cheers,
Peter
20 years
IPSEC NETLINK errors
by Nathanael D. Noblet
Hello,
I'm a little unsure of where to post this problem, but google turned
up some results relating to it on this list I figured I might at least
get a pointer of where to go.
I am attempting to setup an IPSEC VPN in a net-to-net configuration.
I've done it with freeswan/openswan and openvpn, so do know a bit about
the stuff going on. I recently learned that the RH supplied kernels
contain the 2.6 IPSEC stack backported, and the package ipsec-tools can
be used to set up these tunnels. I started to learn the setkey to
manually set one up. As I did that I found out that the
redhat-config-network contains a tab for IPSEC stuff. Made me happy.
Unfortunately I can't get it to work. The command ifup ipsec0 returns
with NETLINK answers: Network is unreachable.
here is my ifcfg-ipsec0 file
# COMP A ifcfg-ipsec0
DSTGW=192.168.0.1
SRCGW=10.0.0.1
DSTNET=192.168.0.0/24
SRCNET=10.0.0.0/24
DST=24.72.x.x
TYPE=IPSEC
ONBOOT=no
-------------- ---------------
10.0.0.0/24---| COMP A | 24.68.x.x --- internet --- 24.72.x.x | COMP B
| --- 192.168.0.0/24
--------------- ---------------
I've tried 2 different configuration setups with the compA's
ifcfg-ipsec0 file.
this is the other one
# COMP A ifcfg-ipsec0
DSTGW=24.72.x.x
SRCGW=24.68.x.x
DSTNET=192.168.0.0/24
SRCNET=10.0.0.0/24
DST=24.72.x.x
TYPE=IPSEC
ONBOOT=no
my iptables contain on both sides...
iptables -t udp -p udp --dport 500 -j ACCEPT
iptables -p 50 -j ACCEPT
iptables -p 51 -j ACCEPT
So my two questions are:
1) What am I doing wrong?
1a) How can I get greater debug info if that is what is needed?
2) If here isn't a good place to ask the above question, where do I go?
Thanks for any help you can provide.
--
Nathanael D. Noblet
Gnat Solutions
412 - 135 Gorge Road E
Victoria, BC V9A 1L1
T/F 250.385.4613
http://www.gnat.ca/
20 years
Re: systematic Kerberization
by "Nils O. Selåsdal"
On Tue, 2004-05-11 at 15:30, Havoc Pennington wrote:
> > Yeah, I'm not quite sure what's going on here. At the same time, it's
> > definitely not an unsolvable problem. And since this is Havoc's
> > wishlist thread, we should make sure that fixing this ends up in
> > there ;)
>
> This isn't the first strong customer request for disconnected operation.
> I have no idea what's involved though (it seems like there would be some
> tricky security issues?). I could ask Nalin, but public lists beat
> hallway conversations. ;-)
It's twofold. One might need disconnected operations, in that you log
in as a user found in an LDAP directory. It can probably be discussed
wether you should be allowed to log in as that user if you pull your
network plug ;), or that everything should continue to work if you pull
it after you already logged in.
The other thing is wether to log in to the box at all(to a local
account) if you configured ldap/kerberos/etc. in system-config-auth
Currently you cannot, which is very bad. Last I looked at it, it's just
a matter of changing some "required" to "suffcient" or similar in
/etc/pam.d/system-auth
20 years
rawhide report: 20040511 changes
by Build System
Updated Packages:
cyrus-imapd-2.2.3-11
--------------------
* Sun May 09 2004 Karsten Hopp <karsten(a)redhat.de> 2.2.3-11
- don't enable cyrus-imapd per default (#122615)
fedora-release-2-2
------------------
fedora-release-2-3
------------------
im-sdk-11.4-42
--------------
* Mon May 10 2004 Akira TAGOH <tagoh(a)redhat.com> 1:11.4-42
- im-sdk-11.4-initscript-lock.patch: to avoid the conflicts, which causes 100%
CPU eaten. (Warren Togami, FC2 BLOCKER #122631)
rpmdb-fedora-1.92-0.20040511
----------------------------
xinitrc-3.39-1
--------------
* Tue Mar 30 2004 Mike A. Harris <mharris(a)redhat.com> 3.39-1
- Fix problem in xinput script caused due to a bit of over-quoting in recent
changes, which result in logged erros (FC2 BLOCKER #119529)
* Mon Mar 29 2004 Mike A. Harris <mharris(a)redhat.com> 3.38-1
- xinput fixes:
- Fix missing ']' on elif test on line 25 (FC2 BLOCKER #119284)
- Change test for /etc/profile.d/lang.sh to test with -r instead of -f
- Change all non-null string tests to use -n instead of != ""
- Change broken test logic in KDE block to use -o instead of ||
- Fixes for iiimf getting turned on even for European languages where it
is not useful. (Jens Petersen, FC2 BLOCKER #119241)
- Xsession cleanups:
- Update Xsession for new switchdesk (Than Ngo, FC2 BLOCKER #116164)
- Change all file existance tests using -f, to -r since we care if it is
readable rather than if it exists or not.
- Use new style $() command substitution rather than ``
- Updated the copyright messages of all script files, and added missing
GPLv2 notices where appropriate. Updated License: tag of spec file to be
"GPLv2, MIT/X11"
* Tue Mar 16 2004 Mike A. Harris <mharris(a)redhat.com> 3.37-1
- Disable Requires: XFree86 dep, and add comment to spec file with reasoning,
so if anything breaks, we can fix it up in a better way later on
- Xclients cleanups:
- Check for existance of xclock and xterm binaries, both in /usr/bin, and
in /usr/X11R6/bin before trying to execute either in failsafe
- Check for executable existance of mozilla/fvwm/twm with -x, rather than
just checking for existance with -f, and then use the explicit path to
these binaries when executing them
- xinput cleanups:
- Change shebang to use /bin/bash instead of /bin/sh, as our OS is always
guaranteed to have /bin/bash installed on it because many other things
require it anyway. This makes it explicitly legal to use bashisms
everywhere, of which are usually already being used anyway without notice,
because /bin/sh is a link to /bin/bash by default.
- Replace old style usage of "test x$foo" with 'if [ "$foo"' as it is much
more readable.
- Test for the readability of /etc/sysconfig/i18n with -r, rather than
it's mere existance with -e
- Everywhere -e was used to test for the existance of a binary prior to
executing it, has been replaced with -x, which tests for executability
- Added "Requires: /usr/X11R6/bin/sessreg" for GiveConsole and TakeConsole
- Added "Requires: which /usr/X11R6/bin/RunWM" for Xclients
- Added "Requires: /sbin/pidof /usr/X11R6/bin/xsetroot" for Xsetup_0
20 years
Help Needed: Firefox 0.8 and Thunderbird 0.6
by Warren Togami
https://bugzilla.fedora.us/show_bug.cgi?id=1460
Please assist Fedora Extras in preparing revisions of the firefox-0.8
and thunderbird-0.6 packages before the release of FC2. These packages
are patched with Blizzard's xremote fixes which finally allow perfect
integration with both firefox & thunderbird running simultaneously. We
have also spent a great deal of time making sure they work within the
newly fixed Preferred Application framework of FC2.
Only minor changes are needed to both packages before publication. Also
it would help if Fedora Desktop volunteers can help to contact upstream
developers to either act as liasons, or work directly with Fedora Extras
development.
https://bugzilla.fedora.us/show_bug.cgi?id=1443
Even non-programmers can help by searching upstream mozilla.org bugzilla
and talking to firefox or mozilla developers about this issue. It is
very likely that there already exists a patch or workaround for this
issue, but the developers are too busy to search for it. You can help!
For example, we have made this kind of very successful partnership with
three upstream gaim developers. They are automatically added to all
Fedora gaim bug reports, and have helped us immensely to be very
responsive to user bugzilla reports and package fixing. (Currently
there are ZERO open gaim reports in FC!)
Warren Togami
wtogami(a)redhat.com
20 years
Re: systematic Kerberization
by "Nils O. Selåsdal"
On Tue, 2004-05-11 at 01:15, Enrico Scholz wrote:
> hp(a)redhat.com (Havoc Pennington) writes:
>
> > Any real-world experience/scenarios where Kerberos support was needed
> > and not available? (Which things should be Kerberized first?)
...
> * CVS gserver is not runnable as non-root (a small LD_PRELOAD hack
> which overloads getuid()/getgid() helps; but native functionality
> would be nice) --> perhaps the alternative version management systems
> (subversion, arch) need kerberos support also
>
There is afaik gssapi support in neon now, and gssapi modules for
apache.
Should probably do to make Subversion kerberized. Never tried it myself
thoug.
--
Nils Olav Selåsdal
System Engineer
w w w . u t e l s y s t e m s . c o m
20 years
