From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jeremy Fitzhardinge Subject: [patch 0/2] softlockup watchdog improvements Date: Mon, 26 Mar 2007 22:38:16 -0700 Message-ID: <20070327053816.881735237@goop.org> Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Return-path: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: virtualization-bounces@lists.linux-foundation.org Errors-To: virtualization-bounces@lists.linux-foundation.org To: Andrew Morton Cc: virtualization@lists.osdl.org, Ingo Molnar , Thomas Gleixner , Linux Kernel , john stultz List-Id: virtualization@lists.linuxfoundation.org Here's couple of patches to improve the softlockup watchdog. The first changes the softlockup timer from using jiffies to sched_clock() as a timebase. Xen and VMI implement sched_clock() as counting unstolen time, so time stolen by the hypervisor won't cause the watchdog to bite. The second adds per-cpu enable flags for the watchdog timer. This allows the timer to be disabled when the CPU goes into a (potentially unbounded) tickless sleep. I know this conflicts with fix-bogus-softlockup-warning-with-sysrq-t.patch in -mm2. I think that patch incorrectly changes the behaviour of the softlockup watchdog, and a better solution is to temporarily disable the watchdog while doing something known to be cpu-consuming, like a long sysreq output. J -- =