From mboxrd@z Thu Jan 1 00:00:00 1970 From: Oren Laadan Subject: Re: liblxc and cgroups Date: Thu, 20 Nov 2008 10:40:11 -0500 Message-ID: <492584DB.8030604@cs.columbia.edu> References: <20081108235107.GA15761@us.ibm.com> <491849DA.7060204@fr.ibm.com> <49228C12.3060802@fr.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <49228C12.3060802-NmTC/0ZBporQT0dZR+AlfA@public.gmane.org> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: containers-bounces-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org Errors-To: containers-bounces-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org To: Daniel Lezcano Cc: Linux Containers List-Id: containers.vger.kernel.org Daniel Lezcano wrote: > Daniel Lezcano wrote: >> Serge E. Hallyn wrote: >>> Hi Daniel, >>> >>> I'm playing with liblxc containers and the device whitelist cgroup. >>> One thing which makes the devices cgroup unique from the others is >>> that there can be many entries to the devices.allow (and in theory >>> also to devices.deny) file. liblxc doesn't support that right now. >>> This needs to be fixed in two places. >>> First, lxc_conf.c:write_info needs to write multiple entries >>> from the .conf file into the cgroups/devices.allow file. I just >>> changed the creat(f, 0755) to open(f, O_CREAT|O_WRONLY|O_APPEND, 0755) >>> which seemed to work for me, but I'm not sure if that might adversely >>> affect other code which counted on the truncation implicit in creat()? >>> Secondly, the lxc_cgroup_copy needs to do a loop and write the >>> entries one by one into the cgroup file. I'm just doing a dumb >>> unsafe fgets loop, but I actually don't have that working yet, >>> (which is why I'm not sending a patch - I figure you can whip >>> something robust up in 2 seconds) >> >> Serge, thanks for investigating this bug. >> I will look how to fix that without breaking previous container >> configuration. > > Fixed and commited to CVS. > > I will do a new release as soon as I finish the man pages. > > Oren, > > is there any change I have to care about before releasing a new version ? not yet. I'm running a bit behind, and hope to post new version around the weekend. Oren.