From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ulrich Obergfell Subject: Re: [PATCH 2/3] watchdog: control hard lockup detection default Date: Thu, 24 Jul 2014 08:02:22 -0400 (EDT) Message-ID: <300049235.17260559.1406203342695.JavaMail.zimbra@redhat.com> References: <1406196811-5384-1-git-send-email-drjones@redhat.com> <1406196811-5384-3-git-send-email-drjones@redhat.com> <53D0E3F3.6050206@redhat.com> <1140033921.17233579.1406200685037.JavaMail.zimbra@redhat.com> <53D0ED70.4070800@redhat.com> <146661093.17248776.1406202294303.JavaMail.zimbra@redhat.com> <53D0F1EB.9080504@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit Cc: Andrew Jones , linux-kernel@vger.kernel.org, kvm@vger.kernel.org, dzickus@redhat.com, akpm@linux-foundation.org, mingo@redhat.com To: Paolo Bonzini Return-path: In-Reply-To: <53D0F1EB.9080504@redhat.com> Sender: linux-kernel-owner@vger.kernel.org List-Id: kvm.vger.kernel.org >----- Original Message ----- >From: "Paolo Bonzini" >To: "Ulrich Obergfell" >Cc: "Andrew Jones" , linux-kernel@vger.kernel.org, kvm@vger.kernel.org, dzickus@redhat.com, akpm@linux-foundation.org, >mingo@redhat.com >Sent: Thursday, July 24, 2014 1:45:47 PM >Subject: Re: [PATCH 2/3] watchdog: control hard lockup detection default > >Il 24/07/2014 13:44, Ulrich Obergfell ha scritto: >> > But this means that it is not possible to re-enable softlockup detection >> > only. I think that should be the effect of echo 0 + echo 1, if >> > hardlockup detection was disabled by either the command line or patch 3. >> >> The idea was to give the user two options to override the effect of patch 3/3. >> Either via the kernel command line ('nmi_watchdog=') at boot time or via /proc >> ('echo 0' + 'echo 1') when the system is up and running. > > I think the kernel command line is enough; another alternative is to > split the nmi_watchdog /proc entry in two. > > Paolo The current behaviour (without the patch) already allows a user to disable NMI watchdog at boot time ('nmi_watchdog=0') and enable it explicitly when the system is up and running ('echo 0' + 'echo 1'). I think it would be more consistent with this behaviour and more intuitive if we would give the user the option to override the effect of patch 3/3 via /proc. By 'intuitive' I mean that the user says: 'I _want_ this to be enabled'. Regards, Uli