From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andrea Arcangeli Subject: Re: swapping with MMU Notifiers V2 Date: Thu, 31 Jan 2008 13:58:42 +0100 Message-ID: <20080131125842.GL7185@v2.random> References: <20080129145021.GJ7233@v2.random> <20080130185735.GS7233@v2.random> <47A16F99.8060502@qumranet.com> <20080131101519.GG7185@v2.random> <47A1A43D.6020203@qumranet.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Cc: kvm-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org To: Avi Kivity Return-path: Content-Disposition: inline In-Reply-To: <47A1A43D.6020203-atKUWr5tajBWk0Htik3J/w@public.gmane.org> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: kvm-devel-bounces-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org Errors-To: kvm-devel-bounces-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org List-Id: kvm.vger.kernel.org On Thu, Jan 31, 2008 at 12:34:37PM +0200, Avi Kivity wrote: > I see. Will merge that patch, thanks. Thanks. BTW, with this fix I finally got KVM swapping 100% stable on my test system. However I had to rollback everything: I'm using my last mmu notifier patch (not Christoph's ones), my mmu locking patch to browse memslots with only kvm->mmu_lock, my last kvm patch against my last mmu notifier patch (only difference is that mmu notifier registers in common code instead of x86), the pagefault_disable to avoid the deadlock lock inversion, kvm.git at tag 3d12af8d03b98644df72c9ac59416d37ea000303 and kvm-userland at tag 2439b8e5ab4e733d7793b54315098567ce5deff3. Host kernel was rolled back from 2.6.24 to 2.6.23-something too. The last bit I had to rollback was kvm-userland, reconfigure, make clean, though it'd be surprising if kvm-userland updates are leading to a host crash, but they may be in the kernel/* directory (I'm using the external module). It might also be something stale in the buildsystem (perhaps a distcc of ccache glitch?), I also cleared 1G of ccache just to be sure in case it was a ccache glitch, after "make clean". Or I may not have run make clean after changing --kerneldir or after git pull, thinking the buildsystem would figure out the change itself (it should have in theory ;). Anyway because things are rock solid again now in both my workstation and in the test system, I will be able to find quickly if something's really wrong in kvm/kvm-userland or if it was a buildsystem issue of some sort and I'll move to the latest code again for all pieces. ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/