From mboxrd@z Thu Jan 1 00:00:00 1970 From: Anthony Liguori Subject: Re: [PATCH] [RFC] sysfs support for xen linux Date: Wed, 11 Jan 2006 10:50:56 -0600 Message-ID: <43C53770.1090403@us.ibm.com> References: <43C53237.2020802@us.ibm.com> <200601111628.49811.mark.williamson@cl.cam.ac.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <200601111628.49811.mark.williamson@cl.cam.ac.uk> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Mark Williamson Cc: "Mike D. Day" , xen-devel@lists.xensource.com List-Id: xen-devel@lists.xenproject.org Mark Williamson wrote: >>Yes, true. I think privcmd might be more appropriate for /proc, but >>would like to see how it works as a binary file under /sys. ioctl's are >>not supported by sysfs that I can see, so privcmd would have to be a >>read/write interface. >> >>Regardless of privcmd there are numerous simple attributes related to >>Xen that are consistent with "zen" of sysfs. :-) >> >> > >Agreed. It's an ideal place for many of the things that are in proc, and many >things we'd want to add in the future. > >You could possibly abuse sysfs into supporting ioctl but I'm pretty sure the >kernel people would find that quite distressing ;-) I had the impression >binary files under sysfs were also a no-no (all meant to be cat-able - >addressing one of the other problems of /proc), but I could be mistaken. > > There was a discussion a while back about this. http://permalink.gmane.org/gmane.comp.emulators.xen.devel/14273 I think the consensus was to move any ioctl interface to a char device and everything else to sysfs. You'll probably want a proper kobject hierarchy too which means working with GKH et al to figure out where out stuff should be (/sys/hypervisor/xen??) Regards, Anthony Liguori >Cheers, >Mark > > >