You are not logged in.
A recent update removed module-init-tools and replaced it with kmod.
The wiki has a number of methods to prevent loading of ipv6 that all seem to rely on /etc/modprobe.d/modprobe.conf. Seeing that this file is no longer used by kmod, how do I prevent ipv6 module being loaded?
Last edited by skunktrader (2012-01-21 03:27:57)
Offline
kmod is a dropin replacement. it honors exactly the same config files. The package doesn't ship a modprobe.conf because it was shipped empty with m-i-t. I didn't see a reason to continue this bizzare practice.
Rename the file to ditch the pacsave.
edit for posterity: http://mailman.archlinux.org/pipermail/ … 22324.html
Last edited by falconindy (2012-01-21 03:28:47)
Offline
Thanks
# mv /etc/modprobe.d/modprobe.conf.pacsave /etc/modprobe.d/modprobe.conf
Offline
Shouldn't this be worth an announcement post?
This is not immediately clear.
Regards
+1
foggy shades, bright outside
Offline
Also, shouldn't the replacement of module-init-tools with kmod initiate a rebuild of the initramfs or at least show a message that you have to do it manualy?
Offline
Same here, I expected an announcement that would explain this but there is none.
Offline
But isn't this just merging pacnew files? What's so new in this?
Offline
Same here, I expected an announcement that would explain this but there is none.
+1
Offline
But isn't this just merging pacnew files? What's so new in this?
true, I guess the flock of new users never learned they need to merge .pacnew files yet. Takes a deep cut to learn
foggy shades, bright outside
Offline
Rolinh wrote:Same here, I expected an announcement that would explain this but there is none.
+1
+1
http://euterpe.blog.br/
Blog de música clássica (pt-BR)
Offline
Instead of worthless '+1' it would be interesting to read *why* do you think this update was so special that it deserved an announcement?
Offline
First of, it's not a pacnew file, it's a pacsave making things quiet intriguing at first sight.
Offline
Because someone has already stated why it might have been a good idea, however the people are quoting the wrong post.
Shouldn't this be worth an announcement post?
This is not immediately clear.
Regards
Fortunately for me, after looking for announcements, there were enough forum threads (2) discussing the change for me to understand what was happening.
Cheers.
Last edited by Earnestly (2012-01-21 12:56:35)
Offline
First of, it's not a pacnew file, it's a pacsave making things quiet intriguing at first sight.
Why is it intriguing? It's a standard procedure: a package has been removed, files from the backup array that had been modified have been saved as .pacsave files.
At least my pacman output said clearly:
[2012-01-13 01:48] warning: /etc/modprobe.d/modprobe.conf saved as /etc/modprobe.d/modprobe.conf.pacsave
Offline
Of course it does. But this does not tell us if modprobe.conf can still be used (which is the case btw). So we learn that modprobe.conf has been moved to a pacsave file... and then? As there is no more modprobe.conf after that, one might think that it cannot be used anymore... (and this was my thoughts actually). So we have to search in the forums to find what we are supposed to do instead of reading a 2 lines announcement.
Offline
See https://bbs.archlinux.org/viewtopic.php?id=134045
I'm not against adding some info to pacman's output so the average user doesn't have to search through the ML trying to understand the discussion etc. but '+1's don't help.
If you want to get something fixed you should be much more precise:
1. Why do you think an announcement was in order?
2. Why not add some info to pacman's output instead?
3. Give an example of what said announcement / other info should look like.
Offline
Also, shouldn't the replacement of module-init-tools with kmod initiate a rebuild of the initramfs or at least show a message that you have to do it manualy?
Why should I manually rebuild initramfs after upgrading to kmod? Is this really necessary?
Offline
twouters wrote:Also, shouldn't the replacement of module-init-tools with kmod initiate a rebuild of the initramfs or at least show a message that you have to do it manualy?
Why should I manually rebuild initramfs after upgrading to kmod? Is this really necessary?
Rebuilding the initramfs has always been something you're encouraged to do after a udev upgrade. Mismatches in versions on the initramfs and on the real root can cause problems when upstream decides to play with the udev DB format (it's entirely internal to udev so there's no API guarantee). In this case, there was no udev DB format change, so it's not strictly necessary, but it's still a good idea.
Offline
Sorry for OT, but in the current version (kmod 4-1) the output of 'kmod -h' is lacking. Your hint http://mailman.archlinux.org/pipermail/ … 22324.html
One other thing you might notice is that kmod doesn't currently include
man pages. I don't consider this a loss -- the m-i-t manpages did not
provide full coverage, nor did the command line help. kmod's binaries
all currently have full coverage of options via -h, --help.
saved me from committing some horrible acts I might later regret.
kmod [options] command [command_options]
doesn't make much sense.
Upstream http://git.profusion.mobi/cgit.cgi/kmod.git is dead atm. How should I proceed if I want this fixed?
Offline
Evilandi666 wrote:twouters wrote:Also, shouldn't the replacement of module-init-tools with kmod initiate a rebuild of the initramfs or at least show a message that you have to do it manualy?
Why should I manually rebuild initramfs after upgrading to kmod? Is this really necessary?
Rebuilding the initramfs has always been something you're encouraged to do after a udev upgrade. Mismatches in versions on the initramfs and on the real root can cause problems when upstream decides to play with the udev DB format (it's entirely internal to udev so there's no API guarantee). In this case, there was no udev DB format change, so it's not strictly necessary, but it's still a good idea.
Thx, done.
But I really ask myself, why doesn't udev tell me that while updating via pacman's output. Or why doesn't if start a initramfs rebuild itself. Afaik not even the udev wiki page tells me to do so.
Offline
moderator comment: Posts consisting of only '+1' add nothing to the discussion. Please refrain from making them. Forum Etiquette: No Power-Posting/Empty Posts
aur S & M :: forum rules :: Community Ethos
Resources for Women, POC, LGBT*, and allies
Offline
Sorry for OT, but in the current version (kmod 4-1) the output of 'kmod -h' is lacking. Your hint http://mailman.archlinux.org/pipermail/ … 22324.html
One other thing you might notice is that kmod doesn't currently include
man pages. I don't consider this a loss -- the m-i-t manpages did not
provide full coverage, nor did the command line help. kmod's binaries
all currently have full coverage of options via -h, --help.saved me from committing some horrible acts I might later regret.
kmod [options] command [command_options]
doesn't make much sense.
Upstream http://git.profusion.mobi/cgit.cgi/kmod.git is dead atm. How should I proceed if I want this fixed?
You can look harder before saying things like "kmod isn't feature complete" when it's indeed feature complete (and beyond) compared to m-i-t.
man pages exist with kmod-4, and all the symlinks to kmod have complete help output... a manpage for the multi-call kmod itself is on the TODO list.
Offline
stefanwilkens said "kmod --help is feature complete", I disagreed https://bbs.archlinux.org/viewtopic.php … 5#p1045115
I was talking about the man page.
'Usage' part mentions options, command and command options, but command options are not listed. Instead we get a list of compat commands / symlinks.
How should I use the symlinks? What's the usage syntax for them? I thought I can use a symlink (e.g. rmmod) in place of 'kmod' but that's not the case.
I can do 'kmod help' but
[karol@black ~]$ lsmod help
Usage: lsmod
[karol@black ~]$ rmmod help
Error: Module help is not currently loaded
[karol@black ~]$ modprobe help
FATAL: Module help not found.
Does the man page tell you that if an option is present, the command will not be executed?
If I need that much trial and error to get it going, either the man page leaves something to be desired or the underlying software is too baroque for such a simple manual.
Or I'm too dumb to comprehend this ingenius ingenuous simplicity.
Offline
You can't possibly be serious. kmod uses GNU's getopt_long(3) like almost every other C program out there does (including module-init-tools).
$ rmmod --help
Usage:
rmmod [options] modulename ...
Options:
-f, --force forces a module unload and may crash your
machine. This requires Forced Module Removal
option in your kernel. DANGEROUS
-s, --syslog print to syslog, not stderr
-v, --verbose enables more messages
-V, --version show version
-w, --wait begins module removal even if it is used and
will stop new users from accessing it.
-h, --help show this help
$ modprobe --help
Usage:
modprobe [options] [-i] [-b] modulename
modprobe [options] -a [-i] [-b] modulename [modulename...]
modprobe [options] -r [-i] modulename
modprobe [options] -r -a [-i] modulename [modulename...]
modprobe [options] -l [-t dirname] [wildcard]
modprobe [options] -c
modprobe [options] --dump-modversions filename
Management Options:
-a, --all Consider every non-argument to
be a module name to be inserted
or removed (-r)
-r, --remove Remove modules instead of inserting
--remove-dependencies Also remove modules depending on it
-R, --resolve-alias Only lookup and print alias and exit
--first-time Fail if module already inserted or removed
-i, --ignore-install Ignore install commands
-i, --ignore-remove Ignore remove commands
-b, --use-blacklist Apply blacklist to resolved alias.
-f, --force Force module insertion or removal.
implies --force-modversions and
--force-vermagic
--force-modversion Ignore module's version
--force-vermagic Ignore module's version magic
Query Options:
-t, --type=DIR Limit type used by --list
-D, --show-depends Only print module dependencies and exit
-c, --showconfig Print out known configuration and exit
-c, --show-config Same as --showconfig
--show-modversions Dump module symbol version and exit
--dump-modversions Same as --show-modversions
General Options:
-n, --dry-run Do not execute operations, just print out
-n, --show Same as --dry-run
-C, --config=FILE Use FILE instead of default search paths
-d, --dirname=DIR Use DIR as filesystem root for /lib/modules
-S, --set-version=VERSION Use VERSION instead of `uname -r`
-s, --syslog print to syslog, not stderr
-q, --quiet disable messages
-v, --verbose enables more messages
-V, --version show version
-h, --help show this help
Offline