From mboxrd@z Thu Jan 1 00:00:00 1970 From: Dave Hansen Subject: Re: [RFC v11][PATCH 05/13] Dump memory address space Date: Thu, 18 Dec 2008 07:05:20 -0800 Message-ID: <1229612720.17206.505.camel@nimitz> References: <1228498282-11804-1-git-send-email-orenl@cs.columbia.edu> <1228498282-11804-6-git-send-email-orenl@cs.columbia.edu> <4949B4ED.9060805@google.com> <494A2F94.2090800@cs.columbia.edu> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <494A2F94.2090800-eQaUEPhvms7ENvBUuze7eA@public.gmane.org> Sender: linux-api-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Oren Laadan Cc: Mike Waychison , jeremy-TSDbQ3PG+2Y@public.gmane.org, arnd-r2nGTMty4D4@public.gmane.org, linux-api-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, containers-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org, linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-mm-Bw31MaZKKs3YtjvyW6yDsg@public.gmane.org, Linux Torvalds , Alexander Viro , "H. Peter Anvin" , Thomas Gleixner , Ingo Molnar List-Id: linux-api@vger.kernel.org On Thu, 2008-12-18 at 06:10 -0500, Oren Laadan wrote: > >> + mutex_lock(&mm->context.lock); > >> + > >> + hh->ldt_entry_size = LDT_ENTRY_SIZE; > >> + hh->nldt = mm->context.size; > >> + > >> + cr_debug("nldt %d\n", hh->nldt); > >> + > >> + ret = cr_write_obj(ctx, &h, hh); > >> + cr_hbuf_put(ctx, sizeof(*hh)); > >> + if (ret < 0) > >> + goto out; > >> + > >> + ret = cr_kwrite(ctx, mm->context.ldt, > >> + mm->context.size * LDT_ENTRY_SIZE); > > > > Do we really want to emit anything under lock? I realize that this > > patch goes and does a ton of writes with mmap_sem held for read -- is > > this ok? > > Because all tasks in the container must be frozen during the checkpoint, > there is no performance penalty for keeping the locks. Although the object > should not change in the interim anyways, the locks protects us from, e.g. > the task unfreezing somehow, or being killed by the OOM killer, or any > other change incurred from the "outside world" (even future code). > > Put in other words - in the long run it is safer to assume that the > underlying object may otherwise change. > > (If we want to drop the lock here before cr_kwrite(), we need to copy the > data to a temporary buffer first. If we also want to drop mmap_sem(), we > need to be more careful with following the vma's.) > > Do you see a reason to not keeping the locks ? Mike, although we're doing writes of the checkpoint file here, the *mm* access is read-only. We only need really mmap_sem for write if we're creating new VMAs, which we only do on restore. Was there an action taken on the mm that would require a write that we missed? Oren, I never considered the locking overhead, either. The fact that the processes are frozen is very, very important here. The code is fine as it stands because this *is* a very simple way to do it. But, this probably deserves a comment. -- Dave -- To unsubscribe from this list: send the line "unsubscribe linux-api" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html