View previous topic :: View next topic |
Author |
Message |
steveL Watchman
Joined: 13 Sep 2006 Posts: 5153 Location: The Peanut Gallery
|
Posted: Wed Dec 11, 2013 12:08 am Post subject: |
|
|
tld wrote: | While writing this I just realized that that frontend machine doesn't have any 70-persistent-net.rules file (and apparently never has), while this machine (and my others) do. I'm guessing that may be related, but why would reverting back to udev not correct it?
EDIT: OK...I'm sure this is related:
https://bugs.gentoo.org/show_bug.cgi?id=453494
...for the love of all that is holy...reading that made me want to cry. I see that that specifically relates to drivers compiled directly into the kernel, which is what I have. However I only have one device. Can someone explain how the above bug affects me in something less than a million words? For one thing, how is it that even with no 70-persistent-net.rules and udev at version 208, I've been reliably getting a name of eth0...that bug doesn't seem to imply that that would work. I don't actually even understand if 70-persistent-net.rules even works any more..seriously confused. |
I don't think it does, and if it does then you have to rename to a non-eth* or wlan* name. I went the route of:
Code: | echo '# dummy empty file to keep eth* and wlan*' > /etc/udev/rules.d/80-net-name-slot.rules |
which you didn't mention; that should get udev to leave network names as the kernel assigned them, and like you I only have the one ethernet card. (I find the whole thing really messed up, considering that it's supposed to be for desktop end-users, and I've never seen one with more than one ethernet, and one wifi card, so I'm left wondering who this really helps.) |
|
Back to top |
|
|
tld Veteran
Joined: 09 Dec 2003 Posts: 1816
|
Posted: Wed Dec 11, 2013 12:38 am Post subject: |
|
|
steveL wrote: |
I don't think it does, and if it does then you have to rename to a non-eth* or wlan* name. I went the route of:
Code: | echo '# dummy empty file to keep eth* and wlan*' > /etc/udev/rules.d/80-net-name-slot.rules |
which you didn't mention; that should get udev to leave network names as the kernel assigned them, and like you I only have the one ethernet card. (I find the whole thing really messed up, considering that it's supposed to be for desktop end-users, and I've never seen one with more than one ethernet, and one wifi card, so I'm left wondering who this really helps.) |
Thanks for the reply! I'll probably go that route myself, but as per my last reply here I'm still going nuts trying to understand what's been keeping my machines from using the new "predictable" names in the first place!...it has me absolutely stumped, and I'm leery of trying to "fix' this when, based on everything I've read, all my machines should have switched to the new names...so there's clearly something I'm not understanding here. The wiki says that the new names will be used unless one of these is met:
/etc/udev/rules.d/80-net-name-slot.rules is an empty file with only comments inside of it
/etc/udev/rules.d/80-net-name-slot.rules is a symlink to /dev/null
the kernel commandline contains net.ifnames=0
NONE of the above are true for ANY of my machines. The only difference I can see is that the one I had issues with when moving to eudev is the only one I have that didn't have 70-persistent-net.rules. The others have 70-persistent-net.rules specifically set up to use "eth0" which is NOT supposed to work. Just now out of curiosity I tried rebooting this machine without 70-persistent-net.rules. To make me even more confused, 70-persistent-net.rules got automatically created, all set up to use eth0! I think this is being caused by the fact that eudev apparently has installed /lib/udev/rules.d/75-persistent-net-generator.rules, which is not in the current version of udev...This all has me ready to f****** cry. The more I dig the less sense any of it makes.
Yea, this whole thing is really messed up. I wish someone could clear me up on this one...
Tom |
|
Back to top |
|
|
steveL Watchman
Joined: 13 Sep 2006 Posts: 5153 Location: The Peanut Gallery
|
Posted: Wed Dec 11, 2013 3:16 pm Post subject: |
|
|
tld wrote: | I wish someone could clear me up on this one... |
Try #gentoo-udev on IRC: chat.freenode.net for help with eudev. |
|
Back to top |
|
|
tld Veteran
Joined: 09 Dec 2003 Posts: 1816
|
Posted: Thu Dec 12, 2013 9:43 am Post subject: |
|
|
Thanks! I may do that...especially if I have any issue with the 80-net-name-slot.rules approach.
Tom |
|
Back to top |
|
|
dmpogo Advocate
Joined: 02 Sep 2004 Posts: 3267 Location: Canada
|
Posted: Thu Dec 12, 2013 10:59 am Post subject: |
|
|
I also did
/etc/udev/rules.d/80-net-name-slot.rules is a symlink to /dev/null
immediately I heard about this persistent name adventure ( totally unneeded for my setup) |
|
Back to top |
|
|
krinn Watchman
Joined: 02 May 2003 Posts: 7470
|
Posted: Thu Dec 12, 2013 11:22 am Post subject: |
|
|
Fixed.
dmpogo wrote: | immediately I heard about this persistent name adventure ( totally unneeded for anyone) |
|
|
Back to top |
|
|
Checko55 n00b
Joined: 02 Feb 2014 Posts: 28
|
Posted: Sat Mar 15, 2014 12:59 pm Post subject: |
|
|
Hi,
I followed your steps and everything compiled without issues.
Unfortunately when I start KDE (kdm) no keyboard or mouse is present and I cannot login.
Could you please give me a hint what may be missing?
thx |
|
Back to top |
|
|
steveL Watchman
Joined: 13 Sep 2006 Posts: 5153 Location: The Peanut Gallery
|
Posted: Sat Mar 15, 2014 9:38 pm Post subject: |
|
|
Checko55: That's hard to know: it may be evdev, it may be x11 drivers. Your best bet is to ask in #gentoo-udev on IRC: chat.freenode.net; you can do that from a minimal livedisk(irssi), a livedvd for GUI, or a livedisk from any distro, if you don't have another install. I'd actually ask #gentoo first (or start another thread), as this sounds like a basic installation issue.
I have this in /etc/portage/make.conf:
Code: | VIDEO_CARDS="vesa nvidia"
INPUT_DEVICES="keyboard mouse evdev" |
AFAIK you just need evdev, but if it doesn't work for w/e reason the old ones come in handy.
Also try logging in without X. When xdm starts, hit ctrl-alt-F1 to get to console, and then (as root) run /etc/init.d/xdm stop and rc-update del xdm default (so subsequent boot goes straight to console.)
If you can't use the keyboard at all, when you boot, edit the grub line and add init=/bin/bash. That will get you to a rescue shell (ignore warnings about terminal), but you'll need to remount root read-write in order to run rc-update del xdm default and then shutdown -r now. If shutdown doesn't reboot the machine, then remount it read-only before you use the reset switch. |
|
Back to top |
|
|
Checko55 n00b
Joined: 02 Feb 2014 Posts: 28
|
Posted: Sun Mar 16, 2014 7:57 am Post subject: |
|
|
Hi,
thanks for your answer.
I have evdev activated in make.conf and system compiled without any problem. Actually tried to
switch back to "legacy" udev and it worked.
Do I have to create any rules or something maybe?
regards |
|
Back to top |
|
|
Tractor Girl Apprentice
Joined: 16 May 2013 Posts: 159
|
Posted: Sun Mar 16, 2014 12:54 pm Post subject: |
|
|
@Checko55
This is probably irrelevant but, did you add udev-postmount to default runlevel? |
|
Back to top |
|
|
Checko55 n00b
Joined: 02 Feb 2014 Posts: 28
|
Posted: Sun Mar 16, 2014 7:53 pm Post subject: |
|
|
Hi,
Yes it is started by default. |
|
Back to top |
|
|
steveL Watchman
Joined: 13 Sep 2006 Posts: 5153 Location: The Peanut Gallery
|
Posted: Mon Mar 17, 2014 5:10 am Post subject: |
|
|
Checko: #gentoo-udev; sorry I'm not on eudev yet, will make switch after I get kernel and --toolchain updated. That's where the developers of the actual software are, as it's an in-house project. ryao is very capable and so's blueness. (not sure who else is involved right now.) They're also very helpful, from what I've seen. They'd probably appreciate being able to fix your issue, as others may hit the same thing.
Instructions here
I could login in again, get in touch with the channel and ask them to check in: but then you'd be missing out on the much wider world of the community on IRC: over 50,000 channels last time I checked with about 100,000 people. Yes there's many that are small, dead or even private, but there are an awful lot more where you can ask questions live, and get live support much better than any commercial setup. #gentoo always has over 900 people in (pure support), and #gentoo-chat about 150 (off-topic relaxation, still polite. Sometimes a lot firmer than other gentoo channels, but fun :-)
Just show you've tried things as you have, don't talk too much til you know your way round, and you'll be fine. I'm in #friendly-coders when I'm online: /join #channel and have fun! If you want to find channels about something use, eg:
Code: | /msg alis list gentoo-*
/msg alis list *apache*
/msg alis list * -topic *something* | -- I defy you not to find something you're interested in (besides Gentoo!)
Regards
igli.
PS Don't go into a channel that logs publically; half the point of IRC is that it's a safe space to explore in, and the flipside of that is privacy, at least not to release logs unless necessary, eg for a problem that needs to be resolved with staff; don't worry, rare, though you do get some unfriendly people. If someone is treating you disrespectfully or offensively, just type:
Code: | /ignore fool
fool: *plunk*
| So they know you can't hear them any more (I wish real-life were so easy..:) /unignore fool, to hear them again, but leave it for at least a few days if you've had to ignore someone, for things to cool down a bit, and life to get in the way. ('fool' being the irc nick you wish no longer to hear from, as I am 'igli'.)
Have a read of the catalyst page which is what operators, and indeed all of us, are supposed to be doing. Good luck! |
|
Back to top |
|
|
Checko55 n00b
Joined: 02 Feb 2014 Posts: 28
|
Posted: Mon Mar 17, 2014 3:25 pm Post subject: |
|
|
Hi SteveL,
thanks for your help and the instructions. Will definitely try that.
Best regards
Checko55 |
|
Back to top |
|
|
ExecutorElassus Veteran
Joined: 11 Mar 2004 Posts: 1435 Location: Berlin, Germany
|
Posted: Thu Aug 28, 2014 4:23 pm Post subject: |
|
|
incidentally, having just attempted this -- and due to util-linux-2.25 getting masked in between and thus requiring a downgrade -- I realized that it works just fine now just to run "emerge eudev," which handles installing eudev (first, to cover dependencines) as well as unmerging sys-fs/udev afterwards. YMMV, etc. |
|
Back to top |
|
|
maxmat n00b
Joined: 30 Jun 2011 Posts: 12
|
Posted: Fri Oct 10, 2014 1:16 pm Post subject: |
|
|
Regarding the switch udev -> eudev:
I would strongly advise to "emerge @preserved-rebuild" then to"emerge -euva world" after that (in fact I do that on system then on world). It's time, it's electricity and it's could be consider as ugly. But it is efficient.
After a complex update (complex from system point of view, not my own point of view) the rebuild of the whole system shows system's instabilities.
Regarding lvm2-activation-generator, the man of this tool says "generator for systemd units to activate LVM2 volumes on boot". This tool is not part of LVM2 but part of systemd.
Regarding systemd on Gentoo, with LVM used for root FS (aka /) and also a separated /usr (to keep fun in building the system): not the right place to speak about it ^^
PS: thanks a lot to libertytrek to have taken time to explain his solution. Now this solution is also ours : )
PPS: I use emerge -e world mainly for two things:
- rebuild the whole system: this must be done without human help, then your system would be consistent from emerge point of view (which few but huge).
- after --depclean, before reboot: ok you finished updating world, you did a gentle revdep-rebuild and you followed by emerge -va --depclean. Nothing which looked important was removed, you verified.
And now you reboot.
And now you have no computer working.
This happened to me, most certainly because I made an error of course, I'm not blaming Gentoo team ^^
So to avoid that kind of surprise (most of them, kernel/initramfs, grub and some others issues are nor managed by emerge) I launch a final emerge -evau world before rebooting after by update or --depclean, to check if nothing would be updated or added...
PPS: I love that system |
|
Back to top |
|
|
ArneBab Guru
Joined: 24 Jan 2006 Posts: 429 Location: Graben-Neudorf, Germany
|
Posted: Thu Oct 23, 2014 7:14 am Post subject: |
|
|
I just tried installing eudev and it bailed out with blockers:
Code: | [ebuild NS ] app-text/docbook-xml-dtd-4.5-r1 [4.1.2-r6, 4.2-r2, 4.3-r1, 4.4-r2]
[ebuild N ] sys-fs/eudev-1.10-r2 USE="gudev hwdb introspection keymap kmod modutils openrc rule-generator -doc (-selinux) -static-libs {-test}" ABI_X86="(64) (-32) (-x32)"
[blocks B ] sys-apps/gentoo-systemd-integration ("sys-apps/gentoo-systemd-integration" is blocking sys-fs/eudev-1.10-r2)
[blocks B ] sys-apps/systemd ("sys-apps/systemd" is blocking sys-fs/eudev-1.10-r2)
* Error: The above package list contains packages which cannot be
* installed at the same time on the same system.
(sys-fs/eudev-1.10-r2:0/0::gentoo, ebuild scheduled for merge) pulled in by
eudev
(sys-apps/systemd-212-r5:0/2::gentoo, installed) pulled in by
>=sys-apps/systemd-207 required by (sys-apps/gentoo-systemd-integration-4:0/0::gentoo, installed)
>=sys-apps/systemd-208:0/2[abi_x86_64(-),gudev,introspection] required by (virtual/libgudev-208:0/0::gentoo, installed)
>=sys-apps/systemd-44-r1[pam] required by (sys-auth/pambase-20120417-r3:0/0::gentoo, installed)
>=sys-apps/systemd-208:0/2[abi_x86_64(-)] required by (virtual/libudev-208:0/1::gentoo, installed)
>=sys-apps/systemd-208:0 required by (virtual/udev-208-r2:0/0::gentoo, installed)
|
I had thought that portage would take care of these itself, but I don’t have the time to dig right now so I just wanted to report it. _________________ Being unpolitical means being political without realizing it. - Arne Babenhauserheide ( http://draketo.de )
pkgcore: So fast that it feels unreal - by doing only what is needed. |
|
Back to top |
|
|
steveL Watchman
Joined: 13 Sep 2006 Posts: 5153 Location: The Peanut Gallery
|
Posted: Thu Oct 23, 2014 3:02 pm Post subject: |
|
|
ArneBab wrote: | I had thought that portage would take care of these itself, but I don’t have the time to dig right now so I just wanted to report it. |
You already have systemd installed, which is a whole other ball game from converting an openrc/udev system to an openrc/eudev one.
If you want to stay with systemd, then eudev is not for you, since systemd incorporates udev (and recent versions of the latter appear to hang for a bit, if not under systemd.) I doubt it would work at all with eudev.
If not then I'd switch back to openrc first; though make sure to get rid of all systemd-supplied udev rules if you do that. It likely involves a profile change. |
|
Back to top |
|
|
ArneBab Guru
Joined: 24 Jan 2006 Posts: 429 Location: Graben-Neudorf, Germany
|
Posted: Mon Oct 27, 2014 10:46 am Post subject: |
|
|
steveL wrote: | ArneBab wrote: | I had thought that portage would take care of these itself, but I don’t have the time to dig right now so I just wanted to report it. |
You already have systemd installed, which is a whole other ball game from converting an openrc/udev system to an openrc/eudev one.
If you want to stay with systemd, then eudev is not for you, since systemd incorporates udev (and recent versions of the latter appear to hang for a bit, if not under systemd.) I doubt it would work at all with eudev.
If not then I'd switch back to openrc first; though make sure to get rid of all systemd-supplied udev rules if you do that. It likely involves a profile change. |
I use profile default/linux/amd64/13.0/desktop/kde (no systemd there), and I do not want systemd. My system starts with openrc, but something seems to have pulled in systemd.
Code: | equery d systemd
… (removed those which only depend on systemd with systemd useflag)…
sys-apps/gentoo-systemd-integration-4 (>=sys-apps/systemd-207)
virtual/libgudev-208 (>=sys-apps/systemd-208:0[abi_x86_64(-),gudev,introspection])
virtual/libudev-208 (>=sys-apps/systemd-208:0[abi_x86_64(-)])
virtual/service-manager-0 (kernel_linux ? sys-apps/systemd)
virtual/udev-208-r2 (>=sys-apps/systemd-208:0)
|
Code: | equery d gentoo-systemd-integration
sys-apps/systemd-212-r5 (!vanilla ? sys-apps/gentoo-systemd-integration)
|
^ is that a circular dependency?
and what is virtual/service-manager - why do I need it? _________________ Being unpolitical means being political without realizing it. - Arne Babenhauserheide ( http://draketo.de )
pkgcore: So fast that it feels unreal - by doing only what is needed. |
|
Back to top |
|
|
steveL Watchman
Joined: 13 Sep 2006 Posts: 5153 Location: The Peanut Gallery
|
Posted: Mon Oct 27, 2014 12:40 pm Post subject: |
|
|
ArneBab wrote: | Code: | equery d systemd
… (removed those which only depend on systemd with systemd useflag)…
sys-apps/gentoo-systemd-integration-4 (>=sys-apps/systemd-207)
virtual/libgudev-208 (>=sys-apps/systemd-208:0[abi_x86_64(-),gudev,introspection])
virtual/libudev-208 (>=sys-apps/systemd-208:0[abi_x86_64(-)])
virtual/service-manager-0 (kernel_linux ? sys-apps/systemd)
virtual/udev-208-r2 (>=sys-apps/systemd-208:0)
|
Code: | equery d gentoo-systemd-integration
sys-apps/systemd-212-r5 (!vanilla ? sys-apps/gentoo-systemd-integration)
|
^ is that a circular dependency?
and what is virtual/service-manager - why do I need it? |
That (kernel_linux ? sys-apps/systemd) tripped me for a second til I checked the ebuild:
Code: | DESCRIPTION="Virtual for various service managers"
..
RDEPEND="
prefix? ( >=sys-apps/baselayout-prefix-2.2 )
!prefix? (
|| (
sys-apps/openrc
kernel_linux? ( || (
sys-apps/systemd
sys-process/runit
virtual/daemontools
) ) ) )"
DEPEND="" |
I think it's what you pointed out about sys-apps/gentoo-systemd-integration, which is being pulled in by systemd, but I have no idea why openrc is not satisfying that virtual itself, or was rather. Now systemd is on the machine, it's being selected. It's not in world, is it?
I'd just emerge -Cq it first, and if that doesn't work use -cq and go a lot more carefully. If as you said you have openrc, and it's starting your machine, there shouldn't be any problem removing systemd.
You may need to remove the virtual, though in the first instance I'd just emerge -cq systemd myself, and then see what the tree is saying. It should then take openrc for the virtual, though you will likely want to remove other stuff and do a depclean after upgrade too.
Oh and check those udev rules like I said, as I wouldn't trust it not to "backup essential files" somewhere or the other.
Hmm virtual/udev-208-r2 (>=sys-apps/systemd-208:0) might be another issue, as you appear not to have the normal udev any more (if that's just what it's going to do, then no bother.)
Just be careful if systemd is in-place as udev. |
|
Back to top |
|
|
eflothmeier n00b
Joined: 20 Sep 2014 Posts: 58 Location: Tucson, AZ
|
Posted: Tue Nov 25, 2014 3:37 am Post subject: |
|
|
On this machine I have:
mmom ~ # emerge -pv eudev
These are the packages that would be merged, in order:
Calculating dependencies... done!
[ebuild NS ] app-text/docbook-xml-dtd-4.5-r1:4.5 [4.1.2-r6:4.1.2, 4.2-r2:4.2, 4.3-r1:4.3, 4.4-r2:4.4] 97 kB
[ebuild N ] sys-fs/eudev-1.10-r2 USE="gudev hwdb introspection keymap kmod modutils openrc rule-generator -doc (-selinux) -static-libs {-test}" 1,736 kB
[uninstall ] sys-fs/udev-216 USE="acl firmware-loader gudev introspection kmod -doc (-selinux) -static-libs"
[blocks b ] sys-fs/udev ("sys-fs/udev" is blocking sys-fs/eudev-1.10-r2)
Total: 2 packages (1 new, 1 in new slot, 1 uninstall), Size of downloads: 1,832 kB
Conflict: 1 block
I looked at the docimentation and I figure that I don't have to do anything more that this emerge
because:
1)
mmom ~ # udevadm test-builtin net_id /sys/class/net/eth0 2> /dev/null
returns nothing
2)
mmom ~ # ls -la /sys/class/net
total 0
drwxr-xr-x 2 root root 0 Nov 24 20:16 .
drwxr-xr-x 48 root root 0 Nov 24 20:16 ..
lrwxrwxrwx 1 root root 0 Nov 24 20:02 enp2s8 -> ../../devices/pci0000:00/0000:00:1e.0/0000:02:08.0/net/enp2s8
lrwxrwxrwx 1 root root 0 Nov 24 13:02 lo -> ../../devices/virtual/net/lo
lrwxrwxrwx 1 root root 0 Nov 24 13:02 sit0 -> ../../devices/virtual/net/sit0
shows that sys-fs/udev-216 is sufficiently advanced to anticipate ipv6
(I already did this as per the Gentoo handbook what is listed here in the eudev instructions)
3)
when I run:
mmom ~ # grep -r 'eth0' /etc/* | grep -v '#'
/etc/conf.d/net:config_eth0="DHCP"
/etc/udhcpd.conf:interface eth0
But it has to be this anyway because I'm still in an ipv4 world
Looks like the emerge will autmatically remove the block too
I'm posting this because messing with udev is dangerous
Erich |
|
Back to top |
|
|
eflothmeier n00b
Joined: 20 Sep 2014 Posts: 58 Location: Tucson, AZ
|
Posted: Sun Nov 30, 2014 12:03 am Post subject: |
|
|
Succsess.
I ran
emerge eudev
After reboot, the result was:
mmom ~ # ls -la /sys/class/net
total 0
drwxr-xr-x 2 root root 0 Nov 29 16:24 .
drwxr-xr-x 48 root root 0 Nov 29 16:24 ..
lrwxrwxrwx 1 root root 0 Nov 29 12:28 eth0 -> ../../devices/pci0000:00/0000:00:1e.0/0000:02:08.0/net/eth0
lrwxrwxrwx 1 root root 0 Nov 29 05:28 lo -> ../../devices/virtual/net/lo
lrwxrwxrwx 1 root root 0 Nov 29 05:28 sit0 -> ../../devices/virtual/net/sit0
So the interface had to be changed back to eth0
Nonetheless:
mmom ~ # udevadm test-builtin net_id /sys/class/net/eth0 2>/dev/null
ID_NET_NAME_MAC=enx0002a5e1a232
ID_OUI_FROM_DATABASE=Hewlett-Packard Company
ID_NET_NAME_PATH=enp2s8
So enp2s8 is still around. It simply doesn't present a problem.
I had trouble getting DHCP to work because my /etc/config.d/net file
had this:
config_eth0="DHCP"
I never took thought of this because other of my boxes have it this
way, and it presented no problem.
config_eth0="dhcp"
Is what got it working. What goes in quotes is now case sensitive.
Some thoughts on systemd. I looks like blocks are eliminated and it
will install. It wont execute so long as the kernel configuration is set for
init:
Symbol: GENTOO_LINUX_INIT_SYSTEMD [=n]
Type : boolean
Prompt: systemd
Location:
-> Gentoo Linux
(1) -> Support for init systems, system and service managers
Defined at distro/Kconfig:68
Depends on: GENTOO_LINUX [=y] && GENTOO_LINUX_UDEV [=y]
Erich |
|
Back to top |
|
|
WWWW Tux's lil' helper
Joined: 30 Nov 2014 Posts: 143
|
Posted: Tue Dec 02, 2014 9:12 am Post subject: |
|
|
hello,
I finally switched to eudev thanks for making the process as easy as emerge -C udev and emerge eudev.
Now, I want eudev part of the install stages by default. Gentoo should be promoting its own solutions, openrc, eudev, etc.
Another question is whether it's possible to contribute to this project. Can I help with something? I don't know how to code but I can hunt for things, debug, test to some menial diff tasks etc. Is it possible to help with small things?
One more thing, the old ethX naming convention is eudev's default now? I just ask to confirm and not be worried about my net devices have their names flipped arbitrarily later on.
Some time ago there was a convoluted way of switching between ethX and enpXsX. I want to know if eudev stays with ethX from now on.
thanks. |
|
Back to top |
|
|
Fitzcarraldo Advocate
Joined: 30 Aug 2008 Posts: 2034 Location: United Kingdom
|
Posted: Tue Dec 02, 2014 10:29 am Post subject: |
|
|
WWWW wrote: | One more thing, the old ethX naming convention is eudev's default now? I just ask to confirm and not be worried about my net devices have their names flipped arbitrarily later on.
Some time ago there was a convoluted way of switching between ethX and enpXsX. I want to know if eudev stays with ethX from now on. |
No, it isn't eudev's default now. I found that sys-fs/eudev-2.1.1 behaves in the same way as sys-fs/udev-217. I found that, whether using sys-fs/udev-217 or sys-fs/eudev-2.1.1, I needed to add the kernel boot parameter "net.ifnames=0" in order to ensure the interface names "eth0" and "wlan0" are used instead of "enp4s0" and "wlp3s0" respectively, because even the configuration listed below on my laptop (as per Ref. 1) did not reliably stop the freedesktop.org so-called 'predicable network interface names':
Code: | # ls -la /etc/udev/rules.d/80-net-setup-link.rules
lrwxrwxrwx 1 root root 9 Nov 30 15:25 /etc/udev/rules.d/80-net-setup-link.rules -> /dev/null
# ls -la /etc/systemd/network/99-default.link
-rw-r--r-- 1 root root 0 Oct 31 07:41 /etc/systemd/network/99-default.link |
The problems and hassle I've suffered as a result of the various changes to udev over the last couple of years have left me disliking it (and freedesktop.org) with a vengeance.
Reference 1:
eselect news wrote: | 2014-02-25-udev-upgrade
Title Upgrade to >=sys-fs/udev-210
Author Samuli Suominen <ssuominen@gentoo.org>
Posted 2014-02-25
Revision 1
The options CONFIG_FHANDLE and CONFIG_NET are now required in the kernel.
You will be warned of them if they are missing while you upgrade to
>=sys-fs/udev-210 by the package manager.
See the package's README at /usr/share/doc/udev-210/ for more optional
kernel options.
The most reliable way of disabling the new network interface scheme is still
the kernel parameter "net.ifnames=0" since overriding the
80-net-name-slot.rules in /etc/udev/rules.d/ no longer works since upstream
renamed the file to /lib/udev/rules.d/80-net-setup-link.rules
The actual configuration is at /lib/systemd/network/99-default.link, which
you can override in /etc/systemd/network/
So, to clarify, you can override the new .rules file or the .link file in /etc
but using the kernel parameter is the most consistent way.
Since both the systemd-udevd executable and the network configuration is stored
at /lib/systemd, using a too wide INSTALL_MASK would be a mistake.
[1] https://wiki.gentoo.org/wiki/Udev/upgrade#udev_208_to_210
[2] http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames |
_________________ Clevo W230SS: amd64, VIDEO_CARDS="intel modesetting nvidia".
Compal NBLB2: ~amd64, xf86-video-ati. Dual boot Win 7 Pro 64-bit.
OpenRC udev elogind & KDE on both.
Fitzcarraldo's blog |
|
Back to top |
|
|
Anon-E-moose Watchman
Joined: 23 May 2008 Posts: 6098 Location: Dallas area
|
Posted: Tue Dec 02, 2014 11:59 am Post subject: |
|
|
Quote: | but using the kernel parameter is the most consistent way. |
Next up on the udev change list....
Personally, I masked eudev-2.* and above
One of my fears about eudev was that they were going to follow the upstream changes to udev too closely. _________________ PRIME x570-pro, 3700x, 6.1 zen kernel
gcc 13, profile 17.0 (custom bare multilib), openrc, wayland |
|
Back to top |
|
|
Tony0945 Watchman
Joined: 25 Jul 2006 Posts: 5127 Location: Illinois, USA
|
Posted: Tue Dec 02, 2014 1:29 pm Post subject: |
|
|
Masking it myself, to keep eth0 and also that they are changing the way firmware is handled. |
|
Back to top |
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
|