[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] Kernel printk timestamps and walltime drift

  • To: Roger Cruz <rcruz@xxxxxxxxxxxxxxxxxxxxxxxx>, xen-devel <xen-devel@xxxxxxxxxxxxxxxxxxx>
  • From: Keir Fraser <keir.fraser@xxxxxxxxxxxxx>
  • Date: Fri, 13 Jun 2008 21:31:18 +0100
  • Delivery-date: Fri, 13 Jun 2008 13:30:52 -0700
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>
  • Thread-index: AcjM0Rhtsp0FqzEJQH+pTArFmFSulQACI1i0ACgZ8YAAApu8UAACeOCZAABf+DAAASPDdw==
  • Thread-topic: [Xen-devel] Kernel printk timestamps and walltime drift

On 13/6/08 21:10, "Roger Cruz" <rcruz@xxxxxxxxxxxxxxxxxxxxxxxx> wrote:

> The walltime appears to be generated from periodic timer ticks that
> update "xtime".  These values have always been correct.  Only the ones
> that get generated with the vprintk appear to slip when compared to this
> wall time.

Yes, sched_clock() can drift relative to wall time. It's an internal
monotonic timesource for the kernel, and if it drifts it doesn't usually
matter. It's not visible to user space. So, that explains your vprintk
timestamp drift. I hope it doesn't bother you too much. :-)

 -- Keir

Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.