From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Serge E. Hallyn" Subject: Re: [RFC PATCH 0/2] cr: Introduce s390x checkpoint/restart code Date: Thu, 15 Jan 2009 16:25:39 -0600 Message-ID: <20090115222539.GA24296@us.ibm.com> References: <20090115050523.GA10415@us.ibm.com> <1232012395.30152.13.camel@localhost> <496F082C.3020008@cs.columbia.edu> <1232014244.30152.19.camel@localhost> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Content-Disposition: inline In-Reply-To: <1232014244.30152.19.camel@localhost> 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: Martin Schwidefsky Cc: Linux Containers , Arnd Bergmann List-Id: containers.vger.kernel.org Quoting Martin Schwidefsky (schwidefsky-tA70FqPdS9bQT0dZR+AlfA@public.gmane.org): > What is left are the pgstes tables. After you forked the new process > that is used to restart a KVM enabled process you need to call > s390_enable_sie(), preferably before you recreate the VMAs. So, let's say we're checkpointing 3 tasks in a s390-kvm. Now we restart them outside of kvm. Some process will fork(), exec() some restart program, which will in turn fork() twice, then each of those programs will call sys_restart(), read the info pertaining to the checkpoint task they are to re-create, and set themselves up. Should the s390_enable_sie() then not be correctly set by the system automatically? So whether or not I'm restarting in kvm, the kvm_arch_create_vm() for s390 will have been called? Or do I misunderstand? thanks, -serge