viric | JanC: I found that upstart deadlock to be an issue in readdir() | 08:51 |
---|---|---|
viric | JanC: http://www.linux-mips.org/archives/linux-mips/2012-11/msg00028.html | 08:51 |
jodh | viric: I haven't seen that issue, but please also see: http://upstart.ubuntu.com/cookbook/#symbolic-links-don-t-work-in-etc-init | 09:50 |
viric | meh | 09:50 |
viric | jodh: all links can be followed :) we use that since early upstart, and still, in all platforms other than mips. Well, also in mips, but it stopped working at linux 3.5. | 09:51 |
viric | (stopped working only for mips) | 09:51 |
viric | nevertheless, thank you for that link! | 09:52 |
viric | I didn't know there was a problem about that. | 09:52 |
jodh | viric: it's also documented in init(5). | 09:58 |
viric | well, we take it as "our job" for symlinks to resolve. We don't expect upstart to do anything good in case they don't resolve. | 09:59 |
viric | no blame for upstart. :) | 09:59 |
viric | And in the mips issue.... clearly I tracked down the problem inside libnih, in its call to readdir(). And I could reproduce it out of libnih and out of upstart. | 10:00 |
jodh | yes, thank you for investigating this issue! | 10:08 |
viric | ah, you are welcome :) Let's say that I'd like my system to boot. hehe | 10:10 |
viric | But I don't see any response in linuxmips. | 10:10 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!