I am considering applying some variation of Greg's systemd.debug patch in our systemd ebuild. For reference: https://www.mail-archive.com/systemd-devel%40lists.freedesktop.org/msg18094.html There is currently no way to enable kernel debug output without also enabling systemd debug output -- and that really sucks. This is probably never getting applied upstream. I am willing to update/maintain the patch moving forward. Are there any strong objections to this idea?
I don't see the final (working) version of the patch there.
Concept seems like a good one to me - a bit annoying that it won't be upstreamed. What are other distros doing? The only refinement I'd suggest is that if everybody but Fedora/RHEL/CentOS is doing it a slightly different way that we do it the same way. But, like I said I agree in concept. I don't want one debug flag that makes every log on the system 100x larger.
(In reply to Michał Górny from comment #1) > I don't see the final (working) version of the patch there. I still need to track that down, or more likely just implement the missing pieces myself. If you would like to review what I come up with, that's not a problem. I'm looking for feedback on the idea at this point.
(In reply to Richard Freeman from comment #2) > What are other distros doing? The only refinement I'd suggest is that if > everybody but Fedora/RHEL/CentOS is doing it a slightly different way that > we do it the same way. OK. I will look around a bit to see if I can find anything.
(In reply to Michał Górny from comment #1) > I don't see the final (working) version of the patch there. Here is the latest version of the patch I see on the systemd-devel list. http://thread.gmane.org/gmane.comp.sysutils.systemd.devel/18251/focus=18309
Created attachment 382160 [details, diff] Ignore 'debug' kernel parameter Seems that part of the effected code has been removed, so the ML patches are no longer valid. Here's a patch that works.
(In reply to Richard Freeman from comment #2) > What are other distros doing? The only refinement I'd suggest is that if > everybody but Fedora/RHEL/CentOS is doing it a slightly different way that > we do it the same way. I checked the Debian, Arch, and Fedora git repos, and did not find any indication that they are doing something similar.