From mboxrd@z Thu Jan 1 00:00:00 1970 From: =?ISO-8859-1?Q?Roger_Pau_Monn=E9?= Subject: Re: [PATCH] qemu-stubdom: prevent useless medium change Date: Mon, 3 Dec 2012 17:27:34 +0100 Message-ID: <50BCD2F6.8060106@citrix.com> References: <1352822539-56930-1-git-send-email-roger.pau@citrix.com> <20646.26443.690228.204535@mariner.uk.xensource.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20646.26443.690228.204535@mariner.uk.xensource.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xen.org Errors-To: xen-devel-bounces@lists.xen.org To: Ian Jackson Cc: "xen-devel@lists.xen.org" List-Id: xen-devel@lists.xenproject.org On 16/11/12 17:18, Ian Jackson wrote: > Roger Pau Monne writes ("[Xen-devel] [PATCH] qemu-stubdom: prevent useless medium change"): >> qemu-stubdom was stripping the prefix from the "params" xenstore >> key in xenstore_parse_domain_config, which was then saved stripped in >> a variable. In xenstore_process_event we compare the "param" from >> xenstore (not stripped) with the stripped "param" saved in the >> variable, which leads to a medium change (even if there isn't any), >> since we are comparing something like aio:/path/to/file with >> /path/to/file. This only happens one time, since >> xenstore_parse_domain_config is the only place where we strip the >> prefix. The result of this bug is the following: > > I have been thinking about this. > > The reason I'm reluctant to apply this patch is that I'm worried it > might cause some non-stubdom-related breakage. I know it feels EBW, > but perhaps the answer is _more_ #ifdef STUBDOM rather than less ? > > Or do you think I should just read the code closely enough to > understand it and your patch ? I suspect it's a can of worms... Yes, it's a can of worms indeed. The non-stubdom path is not modified, and the code changes (1st block of the patch) are contained inside a #ifdef STUBDOM (which is not seen on the patch itself, because it's already there). Thanks, Roger.