| View previous topic :: View next topic |
| Author |
Message |
mdeininger
Veteran
Joined: 15 Jun 2005
Posts: 1740
Location: Emerald Isles, observing Dublin's docklands
|
 Posted: Tue Mar 27, 2007 11:02 am Post subject: [einit] 0.21.0 -- now with gentoo-compat-0.99.0 (superseded) |
 |
|
oi everyone, einit 0.20.0 is out, so here we go with a new thread 
just like last time there's been lots of cleanups, bugfixes, worked out bugtracker entries... this time i also revamped most of the build system -- einit even seems to build (and work, in sandbox mode) fine on x86 freebsd 
Also, Ryan made a lot of improvements to the ebuild, so that now you can select which modules to build and I kept that for the new 0.20.0 ebuilds. Go Ryan!
linky:
http://einit.sourceforge.net/
Documentation:
http://einit.sourceforge.net/documentation/users/ (needs updating, included in the main distribution with the +doc flag)
http://wiki.jyujin.de
IRC:
irc.freenode.net#einit
PS.: I added mailing lists on sf.net... not sure if anyone likes 'em, but it can't hurt to have them around 
EDIT: don't forget about layman now, having the overlay around is quite a nice thing -- instructions on the wiki
EDIT.2: *cough* the BIOS bug is reported to be fixed, thus the fast new 0.21.0 release. don't forget to update -- it never occured on my box, but it's reported not to occur with the latest svn, thus the 0.21.0 release. yeah, i'm faster and off university work for at least 3 weeks from now on .
EDIT.3: baselayout-2 is in portage (go baselayout team! go UberLord! *cheers*), so now the gentoo-compat module works quite nicely. emerge einit-modules-gentoo off the overlay and get amazing features like /etc/init.d/X (start|stop|status) and native support for these scripts through einit (well, okay, maybe not that amazing, but it's definitely handy ).
EDIT.4: superseded, see https://forums.gentoo.org/viewtopic-t-555649.html
_________________
"Confident, lazy, cocky, dead." -- Felix Jongleur, Otherland
( Twitter | Blog | GitHub )
Last edited by mdeininger on Thu Apr 26, 2007 7:00 pm; edited 3 times in total |
|
| Back to top |
|
 |
jvale
n00b
Joined: 23 Jan 2006
Posts: 74
Location: Porto, Portugal
|
| Posted: Tue Mar 27, 2007 11:29 am Post subject: |
|
|
I emerged the SVN version and when I run 'einit --wtf' I get this error:
| Code: |
serenity jvale # einit --wtf
eINIT 0.21.0-svn-984: Initialising: Linux
>> initialising in-core modules:
* module "module-so"
* module "module-logic-v3"
* module "bootstrap-configuration-xml-expat"
* module "bootstrap-configuration-stree"
* module "log"
* module "linux-sysconf"
* module "linux-mount"
* module "linux-process"
>> parsing "/lib/einit/einit.xml".
>> parsing "/etc/einit/local.xml".
** updating modules (module-so)
>> updating modules in "/lib/einit/modules/".
>> :253(read_metadata_linux): opening file failed. (System Error #123 [No medium found])
* module: shell-unicode: undefined variable: configuration-system-setfont/ttys
* module: shell-net-ath0: undefined node: configuration-kernel-module-ath0
* module: shell-net-ath0: undefined node: configuration-network-wireless-ssid
|
|
|
| Back to top |
|
|
mdeininger
Veteran
Joined: 15 Jun 2005
Posts: 1740
Location: Emerald Isles, observing Dublin's docklands
|
| Posted: Tue Mar 27, 2007 11:39 am Post subject: |
|
|
that's okay, nothing tragic in there...
the only actual error is
| Code: | | >> :253(read_metadata_linux): opening file failed. (System Error #123 [No medium found]) |
but that just means that the mounter tried to identify the filesystem in your cd/dvd drive and it couldn't open that as there was no medium in there, no need to worry =)
these notices:
| Code: | * module: shell-unicode: undefined variable: configuration-system-setfont/ttys
* module: shell-net-ath0: undefined node: configuration-kernel-module-ath0
* module: shell-net-ath0: undefined node: configuration-network-wireless-ssid |
mean you may or may not have forgotten to add some new nodes to your local.xml (the ones mentioned) -- you can check your local.xml.example and copy the defaults from there
(none of these are critical, or even major though)
_________________
"Confident, lazy, cocky, dead." -- Felix Jongleur, Otherland
( Twitter | Blog | GitHub ) |
|
| Back to top |
|
|
jvale
n00b
Joined: 23 Jan 2006
Posts: 74
Location: Porto, Portugal
|
| Posted: Tue Mar 27, 2007 11:44 am Post subject: |
|
|
Hm. Pasted too much info. 
Actually I was just talking about the
| Code: | | >> :253(read_metadata_linux): opening file failed. (System Error #123 [No medium found]) |
But I'm glad it's just minor stuff.
Interesting thing: I managed to emerge einit at the same time you were updating the repository. I got a segfault while emerging it. 
Emerged it again and everything was fine. ^^ |
|
| Back to top |
|
|
mdeininger
Veteran
Joined: 15 Jun 2005
Posts: 1740
Location: Emerald Isles, observing Dublin's docklands
|
| Posted: Tue Mar 27, 2007 11:50 am Post subject: |
|
|
| jvale wrote: | Hm. Pasted too much info.
Actually I was just talking about the
| Code: | | >> :253(read_metadata_linux): opening file failed. (System Error #123 [No medium found]) |
But I'm glad it's just minor stuff.
Interesting thing: I managed to emerge einit at the same time you were updating the repository. I got a segfault while emerging it.
Emerged it again and everything was fine. ^^ |
 omg, i have no idea how *that* happened?
well, reporting this kind of errors was actually requested...
https://sourceforge.net/tracker/index.php?func=detail&aid=1658023&group_id=159321&atid=811366
apparently, completely dropping non-critical errors (even if they aren't errors) isn't acceptable... :-/
_________________
"Confident, lazy, cocky, dead." -- Felix Jongleur, Otherland
( Twitter | Blog | GitHub ) |
|
| Back to top |
|
|
jvale
n00b
Joined: 23 Jan 2006
Posts: 74
Location: Porto, Portugal
|
| Posted: Tue Mar 27, 2007 12:47 pm Post subject: |
|
|
I valgrind'ed 'einit --wtf' and it reported quite a few leaks. After browsing einit's source I'm not sure about the accuracy though, as valgrind's backtraces don't seem very complete. They're only showing where the lost memory was allocated.
Anyway, here it is: http://pastebin.ca/411990 |
|
| Back to top |
|
|
mdeininger
Veteran
Joined: 15 Jun 2005
Posts: 1740
Location: Emerald Isles, observing Dublin's docklands
|
| Posted: Tue Mar 27, 2007 1:00 pm Post subject: |
|
|
| jvale wrote: | I valgrind'ed 'einit --wtf' and it reported quite a few leaks. After browsing einit's source I'm not sure about the accuracy though, as valgrind's backtraces don't seem very complete. They're only showing where the lost memory was allocated.
Anyway, here it is: http://pastebin.ca/411990 |
i admit that einit does have a large number of memory leaks at the moment 
it doesn't hog on any resources that the system wouldn't be able to get back from einit after killing it, though, so that should be okay. glibc actually does the same thing (i.e. purposefully not cleaning up after itself), because the OS will deallocate all the used memory after a process dies.
then again, i think the cleanup() function of einit is actually not called at all, that should clean a lot of the leaked memory...
(i'm using valgrind to stress-test einit for overflows while developing, btw... great tool, i really miss it on darwin )
_________________
"Confident, lazy, cocky, dead." -- Felix Jongleur, Otherland
( Twitter | Blog | GitHub ) |
|
| Back to top |
|
|
jack_mort
Apprentice
Joined: 30 Jan 2003
Posts: 167
|
| Posted: Tue Mar 27, 2007 4:38 pm Post subject: |
|
|
Hi,
I wanted to try einit for a while and today I did it ! But it seems I've gone wrong somewhere or I didn't open my eyes but I can't seem to make it work with my lvm system which is booting via an initrd... Has someone made einit work with lvm partitions ?
Thanks and anyway, keep up the good work ! |
|
| Back to top |
|
|
mdeininger
Veteran
Joined: 15 Jun 2005
Posts: 1740
Location: Emerald Isles, observing Dublin's docklands
|
| Posted: Tue Mar 27, 2007 6:19 pm Post subject: |
|
|
| jack_mort wrote: | Hi,
I wanted to try einit for a while and today I did it ! But it seems I've gone wrong somewhere or I didn't open my eyes but I can't seem to make it work with my lvm system which is booting via an initrd... Has someone made einit work with lvm partitions ?
Thanks and anyway, keep up the good work ! |
aw my, i'm not sure... there's an example line for a mountpoint in einit.xml that you could use as a template for this kind of operation... haven't tested this yet though :/
i'll try to set up a qemu box to play with lvm and udev a bit... can't promise anything right now though 
_________________
"Confident, lazy, cocky, dead." -- Felix Jongleur, Otherland
( Twitter | Blog | GitHub ) |
|
| Back to top |
|
|
UberPinguin
Guru
Joined: 20 Nov 2005
Posts: 510
Location: 2416.94 Miles From Home
|
| Posted: Tue Mar 27, 2007 8:31 pm Post subject: |
|
|
| mdeininger wrote: | | i'll try to set up a qemu box to play with lvm and udev a bit... can't promise anything right now though |
FWIW, I've found VirtualBox to be much faster/more stable than qemu or kqemu.
_________________
| aidanjt wrote: | | You see, instead of arguing from ignorance, and fear, there is only one way to verify a theory. And that's not by clutching a black book and begging the sky fairy for deliverance from the mad scientists and their big machines. |
|
|
| Back to top |
|
|
jack_mort
Apprentice
Joined: 30 Jan 2003
Posts: 167
|
| Posted: Wed Mar 28, 2007 9:37 am Post subject: |
|
|
Well... I found how to use einit with an initrd image : you must pass | Code: | | real_init=/sbin/einit |
to your bootloader and not
Sorry for the buzz |
|
| Back to top |
|
|
mdeininger
Veteran
Joined: 15 Jun 2005
Posts: 1740
Location: Emerald Isles, observing Dublin's docklands
|
| Posted: Wed Mar 28, 2007 9:41 am Post subject: |
|
|
| UberPinguin wrote: | | mdeininger wrote: | | i'll try to set up a qemu box to play with lvm and udev a bit... can't promise anything right now though |
FWIW, I've found VirtualBox to be much faster/more stable than qemu or kqemu. |
can't say I'd have had stability issues with qemu yet... yeah, it's slow, but it's interesting how easy it is to test different architectures with it...
| jack_mort wrote: | Well... I found how to use einit with an initrd image : you must pass | Code: | | real_init=/sbin/einit |
to your bootloader and not
Sorry for the buzz |
wait... it works like that?
if so, that really needs to get into the wiki
_________________
"Confident, lazy, cocky, dead." -- Felix Jongleur, Otherland
( Twitter | Blog | GitHub ) |
|
| Back to top |
|
|
UberPinguin
Guru
Joined: 20 Nov 2005
Posts: 510
Location: 2416.94 Miles From Home
|
| Posted: Wed Mar 28, 2007 12:55 pm Post subject: |
|
|
| mdeininger wrote: | | UberPinguin wrote: | | mdeininger wrote: | | i'll try to set up a qemu box to play with lvm and udev a bit... can't promise anything right now though |
FWIW, I've found VirtualBox to be much faster/more stable than qemu or kqemu. |
can't say I'd have had stability issues with qemu yet... yeah, it's slow, but it's interesting how easy it is to test different architectures with it... |
Yeah, the multiple architecture support is neat. I just think that it's absurd that the thing is practically unusable on a 2.0Ghz/1GiB machine. I was also disappointed to find that I couldn't actually emulate a Mac with it - I had fantasies of running OSX for educational purposes :/
_________________
| aidanjt wrote: | | You see, instead of arguing from ignorance, and fear, there is only one way to verify a theory. And that's not by clutching a black book and begging the sky fairy for deliverance from the mad scientists and their big machines. |
|
|
| Back to top |
|
|
jack_mort
Apprentice
Joined: 30 Jan 2003
Posts: 167
|
| Posted: Wed Mar 28, 2007 1:18 pm Post subject: |
|
|
| mdeininger wrote: |
wait... it works like that?
if so, that really needs to get into the wiki |
Yeah it works like that ^^
Well in fact I just made it boot with einit Now I need to configure einit to suite my needs, but einit is now able to boot my computer ! Btw, I saw some things about kdm or gdm in config file but I'm using Slim, is there a way to tell einit which session manager is used ? (It tells me it could not start gdm or kdm).
Sorry if my question has already been asked |
|
| Back to top |
|
|
rmh3093
Advocate
Joined: 06 Aug 2003
Posts: 2138
Location: Albany, NY
|
| Posted: Wed Mar 28, 2007 1:23 pm Post subject: |
|
|
| UberPinguin wrote: | | mdeininger wrote: | | UberPinguin wrote: | | mdeininger wrote: | | i'll try to set up a qemu box to play with lvm and udev a bit... can't promise anything right now though |
FWIW, I've found VirtualBox to be much faster/more stable than qemu or kqemu. |
can't say I'd have had stability issues with qemu yet... yeah, it's slow, but it's interesting how easy it is to test different architectures with it... |
Yeah, the multiple architecture support is neat. I just think that it's absurd that the thing is practically unusable on a 2.0Ghz/1GiB machine. I was also disappointed to find that I couldn't actually emulate a Mac with it - I had fantasies of running OSX for educational purposes :/ |
yeah can run OSX natively though... with the right version... leopard will definitely run on x86 when its out you just need a cpu with sse3
... kvm is only useful on a cpu that supports it, if not you are better to use the kqemu accelerator. that might be why your 2ghz machine seems slow
_________________
Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
| Back to top |
|
|
UberPinguin
Guru
Joined: 20 Nov 2005
Posts: 510
Location: 2416.94 Miles From Home
|
| Posted: Wed Mar 28, 2007 1:31 pm Post subject: |
|
|
| jack_mort wrote: | | mdeininger wrote: |
wait... it works like that?
if so, that really needs to get into the wiki |
Yeah it works like that ^^
Well in fact I just made it boot with einit Now I need to configure einit to suite my needs, but einit is now able to boot my computer ! Btw, I saw some things about kdm or gdm in config file but I'm using Slim, is there a way to tell einit which session manager is used ? (It tells me it could not start gdm or kdm).
Sorry if my question has already been asked |
Sure. Assuming that slim is a daemon that starts its own X server (á la entranced): | /etc/einit/local.xml in <services><virtual><module>: | <daemon id="daemon-slim"
name="XDM (Slim)"
provides="xdm:x11"
requires="net-lo:video:mount/critical:xfs"
command="/path/to/slim -args"
restart="yes"
DISPLAY=":0.0"
HOME="/root" /> |
| /etc/einit/local.xml in <services><prefer>: | | <xdm s="daemon-entranced" /> |
If Slim is just a login manager that needs an existing X session (á la entrance): | /etc/einit/local.xml in <services><virtual><module>: | <daemon id="daemon-slim"
name="XDM (Slim)"
provides="xdm"
requires="x11"
command="/path/to/slim -args"
restart="yes"
DISPLAY=":0.0"
HOME="/root" /> |
| /etc/einit/local.xml in <services><prefer>: | | <xdm s="daemon-entranced" /> |
_________________
| aidanjt wrote: | | You see, instead of arguing from ignorance, and fear, there is only one way to verify a theory. And that's not by clutching a black book and begging the sky fairy for deliverance from the mad scientists and their big machines. |
|
|
| Back to top |
|
|
UberPinguin
Guru
Joined: 20 Nov 2005
Posts: 510
Location: 2416.94 Miles From Home
|
| Posted: Wed Mar 28, 2007 1:35 pm Post subject: |
|
|
| rmh3093 wrote: | yeah can run OSX natively though... with the right version... leopard will definitely run on x86 when its out you just need a cpu with sse3
... kvm is only useful on a cpu that supports it, if not you are better to use the kqemu accelerator. that might be why your 2ghz machine seems slow |
Not using kvm ('cause my proc doesn't support it) - the slowness was in kqemu. I'm using VirtualBox right now, which is worlds faster than kqemu. Perhaps I'll pick up a copy of Leopard, but I still have to support some Panther users, so it'd be nice to be able to run that too. I may hunt around for an x86 version of Tiger, but I'd rather not buy it as I already have the ppc version of Tiger and OSX Server.
_________________
| aidanjt wrote: | | You see, instead of arguing from ignorance, and fear, there is only one way to verify a theory. And that's not by clutching a black book and begging the sky fairy for deliverance from the mad scientists and their big machines. |
|
|
| Back to top |
|
|
mdeininger
Veteran
Joined: 15 Jun 2005
Posts: 1740
Location: Emerald Isles, observing Dublin's docklands
|
| Posted: Wed Mar 28, 2007 2:47 pm Post subject: |
|
|
jack_mort: that's pretty good news, i'll be sure to update the wiki over that
what uberpinguin posted should get you going... basically, you'd need to add a new module that starts your DM/SM, then select that by "preferring" it. i think uberpinguin summed that up nicely
EDIT: can't say I'd have much trouble with qemu+kqemu on my x2... x86 freebsd runs fast enough to actually install stuff through ports
on the other hand, you really have to make sure it's actually *using* kqemu -- took me ages to figure out that i had to call a different binary for it to use kqemu, and i kept wondring why the hell it ran so slow :-/
_________________
"Confident, lazy, cocky, dead." -- Felix Jongleur, Otherland
( Twitter | Blog | GitHub ) |
|
| Back to top |
|
|
jack_mort
Apprentice
Joined: 30 Jan 2003
Posts: 167
|
| Posted: Wed Mar 28, 2007 3:53 pm Post subject: |
|
|
| Thanks for your answers and your explanations, I'll now have take some time to make my computer nicely boot with einit |
|
| Back to top |
|
|
UberPinguin
Guru
Joined: 20 Nov 2005
Posts: 510
Location: 2416.94 Miles From Home
|
| Posted: Wed Mar 28, 2007 4:00 pm Post subject: |
|
|
| mdeininger wrote: | jack_mort: that's pretty good news, i'll be sure to update the wiki over that
what uberpinguin posted should get you going... basically, you'd need to add a new module that starts your DM/SM, then select that by "preferring" it. i think uberpinguin summed that up nicely
EDIT: can't say I'd have much trouble with qemu+kqemu on my x2... x86 freebsd runs fast enough to actually install stuff through ports
on the other hand, you really have to make sure it's actually *using* kqemu -- took me ages to figure out that i had to call a different binary for it to use kqemu, and i kept wondring why the hell it ran so slow :-/ |
Hmmmm. What binary was that? I thought that by just having kqemu modprobed it would work.
_________________
| aidanjt wrote: | | You see, instead of arguing from ignorance, and fear, there is only one way to verify a theory. And that's not by clutching a black book and begging the sky fairy for deliverance from the mad scientists and their big machines. |
|
|
| Back to top |
|
|
mdeininger
Veteran
Joined: 15 Jun 2005
Posts: 1740
Location: Emerald Isles, observing Dublin's docklands
|
| Posted: Wed Mar 28, 2007 4:12 pm Post subject: |
|
|
| UberPinguin wrote: | | mdeininger wrote: | jack_mort: that's pretty good news, i'll be sure to update the wiki over that
what uberpinguin posted should get you going... basically, you'd need to add a new module that starts your DM/SM, then select that by "preferring" it. i think uberpinguin summed that up nicely
EDIT: can't say I'd have much trouble with qemu+kqemu on my x2... x86 freebsd runs fast enough to actually install stuff through ports
on the other hand, you really have to make sure it's actually *using* kqemu -- took me ages to figure out that i had to call a different binary for it to use kqemu, and i kept wondring why the hell it ran so slow :-/ |
Hmmmm. What binary was that? I thought that by just having kqemu modprobed it would work. |
well, on my box i had to use 'qemu-system-x86_64' to get it working (since i'm on an x2 in 64bit mode), simply 'qemu' didn't work... i had originally thought it'd just use it, but that doesn't seem to be the case. when your qemu is up, focus it and hit ctrl+alt+2, then type "info kqemu", that'll show you if it's actually using kqemu =)
_________________
"Confident, lazy, cocky, dead." -- Felix Jongleur, Otherland
( Twitter | Blog | GitHub ) |
|
| Back to top |
|
|
rmh3093
Advocate
Joined: 06 Aug 2003
Posts: 2138
Location: Albany, NY
|
| Posted: Wed Mar 28, 2007 9:23 pm Post subject: |
|
|
Magnus: is there really any circumstance where stdin!=stdout for feedback-visual? could these be combined into 1 variable? and, can you make broadcast-constraints ignore the stdin variable instead of hardcoding the removal of tty9
_________________
Do not meddle in the affairs of wizards, for they are subtle and quick to anger. |
|
| Back to top |
|
|
mdeininger
Veteran
Joined: 15 Jun 2005
Posts: 1740
Location: Emerald Isles, observing Dublin's docklands
|
| Posted: Thu Mar 29, 2007 3:19 pm Post subject: |
|
|
| rmh3093 wrote: | | Magnus: is there really any circumstance where stdin!=stdout for feedback-visual? could these be combined into 1 variable? and, can you make broadcast-constraints ignore the stdin variable instead of hardcoding the removal of tty9 |
you're perfectly right, it'd probably better to replace stdin= + stdout= with something like stdio="".
the thing with the broadcast-constraints is just temporary and more of a quick fix; i'll need to rewrite that code to not scan through all of /dev and send messages to the ttys, but to instead use utmp or something like that to find all the active sessions and then use that instead (it's actually a little dangerous like that... just guess what would happen if someone set that var to .* -- it'd fry all your partitions )
_________________
"Confident, lazy, cocky, dead." -- Felix Jongleur, Otherland
( Twitter | Blog | GitHub ) |
|
| Back to top |
|
|
deac
n00b
Joined: 19 Jun 2004
Posts: 3
Location: Colonia Claudia Ara Agrippinensis, Germany
|
| Posted: Thu Mar 29, 2007 10:59 pm Post subject: |
|
|
after einit-reset (reboot via ctrl-alt-del or einit-control power reset) my cmos (bios) was destroied: "cmos checksum error: defaults loaded".
i tried it 3 times and i tried it again with sysvinit, all times after einit-reset the same, but not after sysvinit.
my system:
board: via epia en12000 (on board cpu via eden 1.2ghz)
bios: phoenix (if somebody needs, i'll post the version later)
kernel: linux 2.6.20.3
einit: 0.20.0#982
CONFIG_EDD and CONFIG_RTC_DRV_CMOS aren't set.
now i tried a simple pressing reset, no einit-reset, no error.
after shutdown (einit-control power down) there's no error.
only after reboot. |
|
| Back to top |
|
|
mdeininger
Veteran
Joined: 15 Jun 2005
Posts: 1740
Location: Emerald Isles, observing Dublin's docklands
|
| Posted: Fri Mar 30, 2007 12:08 am Post subject: |
|
|
uhm, to be perfectly honest, i have no idea how that could have happened? the only code where einit is writing anything at all is the broadcast code... unless of course the broadcast constraints would match a node below /dev that was associated with your cmos -- could you post a
when logged into einit?
(i assume you haven't played with any of the nodes containing regular expressions in einit.xml, like the broadcast-constraints?)
i'm using the reboot() function to tell the kernel to reset, which is described in the manpage reboot(2).
einit works fine on my 1ghz epia-mII downstairs...
_________________
"Confident, lazy, cocky, dead." -- Felix Jongleur, Otherland
( Twitter | Blog | GitHub ) |
|
| 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
|
|
Unsupported Software
