View previous topic :: View next topic |
Author |
Message |
Lebkoungcity Apprentice
Joined: 16 Nov 2008 Posts: 212 Location: near Lebkoungcity (='Gingerbreadcity' =Nuremberg)
|
Posted: Mon Jun 29, 2015 10:25 pm Post subject: strange boot message about rc-cgroup.sh |
|
|
Hi,
last week I got my old graphics card, a Nvidia GeForce210, replaced by a GeForce GTX 750 and emerged the latest nvidia-drivers. Since then I noticed a line at boot time saying:
Code: |
/lib64/rc/sh/rc-cgroup.sh: Zeile 89: printf: Schreibfehler: Kein passendes Gerät gefunden.
|
Which I think might be in english:
Code: |
/lib64/rc/sh/rc-cgroup.sh: line 89: printf: write error: No matching device found.
|
I don't know if this line was there before the replacement of the graphics card...
OK, I understand that it might be something related with rc-cgroups - but I don't even know why it appears since I never had something to do with rc-cgroups. Or do I but didn't know it?
Is there a way to get rid of this line and the problem it is pointing at?
Thanks a lot for your help!
Andy
EDIT: typo in subject corrected _________________ "The most dangerous world view is the world view of the ones who haven't viewed the world."
Alexander von Humboldt (1769-1859) |
|
Back to top |
|
|
Keruskerfuerst Advocate
Joined: 01 Feb 2006 Posts: 2289 Location: near Augsburg, Germany
|
Posted: Fri Jul 03, 2015 6:56 pm Post subject: |
|
|
Then analyze the rc-cgroup.sh. |
|
Back to top |
|
|
Lebkoungcity Apprentice
Joined: 16 Nov 2008 Posts: 212 Location: near Lebkoungcity (='Gingerbreadcity' =Nuremberg)
|
Posted: Fri Jul 03, 2015 7:58 pm Post subject: |
|
|
Hey Arminius
Thanks a lot for your response!
Unfortunately, I do not know anything about cgroups... What I found on the gentoo-wiki was as helpful as the article on wikipedia - now I know that it exists but I don't understand for what I would need it (and why it is enabled in the kernel since an upgrade to 4.0.5). And I don't know why it only gives errors on the machine with the changed graphics card...
Today I re-compiled the kernel (without altering the .config) and now the original line is gone but this came up instead:
Code: |
/lib64/rc/sh/rc-cgroup.sh: Zeile 89: /sys/fs/cgroup/openrc/keymaps/tasks: Datei oder Verzeichnis nicht gefunden
|
Which I would translate into:
Code: |
/lib64/rc/sh/rc-cgroup.sh: line 89: /sys/fs/cgroup/openrc/keymaps/tasks: file or directory not found
|
OK, this parts of /lib64/rc/sh/rc-cgroup.sh:
Code: |
(...)
cgroup_add_service()
{
# relocate starting process to the top of the cgroup
# it prevents from unwanted inheriting of the user
# cgroups. But may lead to a problems where that inheriting
# is needed.
for d in /sys/fs/cgroup/* ; do
[ -f "${d}"/tasks ] && printf "%d" 0 > "${d}"/tasks
done
openrc_cgroup=/sys/fs/cgroup/openrc
if [ -d "$openrc_cgroup" ]; then
cgroup="$openrc_cgroup/$RC_SVCNAME"
mkdir -p "$cgroup"
[ -f "$cgroup/tasks" ] && printf "%d" 0 > "$cgroup/tasks"
fi
}
(...)
|
with line 89 being:
Code: |
[ -f "$cgroup/tasks" ] && printf "%d" 0 > "$cgroup/tasks"
|
Now I'm really sorry but I don't know what to do next _________________ "The most dangerous world view is the world view of the ones who haven't viewed the world."
Alexander von Humboldt (1769-1859) |
|
Back to top |
|
|
steveL Watchman
Joined: 13 Sep 2006 Posts: 5153 Location: The Peanut Gallery
|
Posted: Sat Jul 04, 2015 7:54 pm Post subject: |
|
|
Lebkoungcity wrote: | Code: | [ -f "$cgroup/tasks" ] && printf "%d" 0 > "$cgroup/tasks" |
|
Ugh that's ugly. Apparently there was a bug in the kernel, but the documentation on cgroups, makes it clear that a newline is required. cf: less /usr/src/linux/Documentation/cgroups/cgroups.txt
Odd; ime Hubbs has always refused to "work around" bugs, or indeed add any portability mechanisms, on the grounds that whatever is at fault should correct its bug (even where there isn't one, which is helpful.)
File a bug. At least let's get a reference as to what bug in the kernel they're supposedly "working-round," though I imagine it'll just get reverted, likely in a way that doesn't look like a reversion.
Shame, he could have patched it to be more robust; certainly that was the intent when qnikst and I wrote it (that we'd come back and add robustness based on user feedback, but in the meantime we'd leave errors visible to get that feedback as to what mattered and what could be ignored.) In this case, I'd use an 'if', which was what I originally wanted to do.
OFC this could just be a race, but I don't think so; it's a shame the tests were modded to be less efficient as well.
No accounting for people. |
|
Back to top |
|
|
Lebkoungcity Apprentice
Joined: 16 Nov 2008 Posts: 212 Location: near Lebkoungcity (='Gingerbreadcity' =Nuremberg)
|
Posted: Sun Jul 05, 2015 11:42 am Post subject: |
|
|
Hi steveL,
thanks for reading and answering my post!
I filed a bug:
https://bugs.gentoo.org/show_bug.cgi?id=553978
Hope I did it the right way!
Greetings,
Andy _________________ "The most dangerous world view is the world view of the ones who haven't viewed the world."
Alexander von Humboldt (1769-1859) |
|
Back to top |
|
|
steveL Watchman
Joined: 13 Sep 2006 Posts: 5153 Location: The Peanut Gallery
|
Posted: Sun Jul 05, 2015 7:38 pm Post subject: |
|
|
Yeah that's fine Andy, thanks.
If you want to make it quicker, you can tell them this is related to https://github.com/OpenRC/openrc/issues/33 or "issue 33 upstream" (though it's an in-house project) as the commit which introduced the changes, is in relation to that.
You can put the URL in the bug's URL field if you want. |
|
Back to top |
|
|
Lebkoungcity Apprentice
Joined: 16 Nov 2008 Posts: 212 Location: near Lebkoungcity (='Gingerbreadcity' =Nuremberg)
|
Posted: Sun Jul 05, 2015 8:01 pm Post subject: |
|
|
OK, added it. Thanks for the advice!
Now, let's see what'll happen next _________________ "The most dangerous world view is the world view of the ones who haven't viewed the world."
Alexander von Humboldt (1769-1859) |
|
Back to top |
|
|
NTU Apprentice
Joined: 17 Jul 2015 Posts: 187
|
Posted: Fri Jul 17, 2015 7:22 pm Post subject: |
|
|
I'm not sure how common this bug is but I have it too. I commented in the bug report with more information.
grep "printf" /var/log/rc.log:
loopback |/lib64/rc/sh/rc-cgroup.sh: line 89: printf: write error: No such device
loopback |/lib64/rc/sh/rc-cgroup.sh: line 89: printf: write error: No such device
termencoding |/lib64/rc/sh/rc-cgroup.sh: line 89: printf: write error: No such device
termencoding |/lib64/rc/sh/rc-cgroup.sh: line 89: printf: write error: No such device |
|
Back to top |
|
|
|