Re: nohz: Warn if the machine can not perform nohz_full

[Frederic Weisbecker <fweisbec@gmail.com>]

On Mon, Jul 15, 2013 at 03:30:20PM -0400, Steven Rostedt wrote:
> On Mon, 2013-07-15 at 14:56 -0400, Dave Jones wrote:
> > On Mon, Jul 15, 2013 at 02:49:42PM -0400, Steven Rostedt wrote:
> >  
> > Is a printk not enough for that purpose ? Tainting the kernel is kinda anti-social.
> > 
> 
> printk's don't usually get peoples attention. Taints and warnings do. If
> the hardware doesn't allow for nohz-full, and the only way to enable it
> is via kernel command line, how do you scream to the user that it isn't
> going to work.
> 
> We could do a large banner saying:
> 
> ***************************************************************
> ***************************************************************
> ***************************************************************
> ***************************************************************
> ***                    UNSTABLE TSC                         ***
> ***                NO_HZ_FULL disabled                      ***
> ***************************************************************
> ***************************************************************
> ***************************************************************
> ***************************************************************

May be we can try that. In fact I have a pending patch that converts the
stacktrace warning to a one line printk message, as Ingo reported that issue
to me. But the risk is that it can be indeed lost in the flow.

> 
> 
> And not do the warning. Maybe that will get peoples attentions?
> 
> But then again, this could be lost in the boot up if the box prints a
> lot of data.

Well the above example is unlikely to be missed. If it is, then I believe a
traditional warning would be lost as well.

> 
> -- Steve
> 
>