On Fri, 2006-08-04 at 15:05 -0700, Dave Hansen wrote: > On Fri, 2006-08-04 at 15:01 -0700, Paul Menage wrote: > > So is there an argument against moving to reporting things like > > membership (and maybe arbitrary stats) via sysfs rather than configfs? > > Is there a reason to do it, other than keeping configfs simple?
Yes. Can userspace do a mkdir in sysfs? rmdir? My understanding is it cannot because sysfs directories represent groups of attributes the represent pre-existing kernel objects. In order to create a directory we'd likely have to invent some roundabout method for creating these objects -- perhaps by echo'ing to a file: echo 'classes/foo/bar/baz' > /sys/ckrm/mkdir My rough understanding is: sysfs is for "objects" created by the kernel which userspace may wish to manipulate. configfs is for "objects" created by userspace which describe/affect kernel behavior. This means the "objects" have different lifetimes (as I think Joel mentioned) -- sysfs objects can appear and disappear without userspace interaction whereas configfs objects generally can only appear or disappear in response to userspace interaction. Because userspace creates and destroys resource groups configfs seems the appropriate choice. Cheers, -Matt Helsley ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys -- and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV _______________________________________________ ckrm-tech mailing list https://lists.sourceforge.net/lists/listinfo/ckrm-tech