From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jim Fehlig Subject: Re: That xenstored console leak... Date: Tue, 15 Jan 2008 08:59:49 -0700 Message-ID: <478CD875.1050303@novell.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Keir Fraser Cc: xen-devel@lists.xensource.com, John Levon List-Id: xen-devel@lists.xenproject.org Keir Fraser wrote: > On 15/1/08 14:38, "John Levon" wrote: > > >>>> Anyway, seems like we need a simple working fix for 3.1 series, even if >>>> there's larger surgery done. >>>> >>> I checked your patch in for 3.1, which is probably sufficient there. >>> >> The fix I was referring to was the one for the MAC-lossage issue. >> > > I'd consider taking Jim's patch to remove the path deletion, and swallow the > (rather theoretical) xenstore leak across large numbers of device hotplugs. > It's reality, not just theory :-/. xm network-[attach|detach] does leak. block does not as nothing is written to /vm/uuid/device for block devices. Not sure about other devices, e.g. vtpm, security labels, etc. A quick grep through the sources reveals netif as the only dev controller invoking _writeVm(). Perhaps we should look at handling network and block devices in a similar fashion - eliminating the need to write any device info to this path? Jim