Re: [PATCH]x86-tsc.c : fix compile warning

[Subrata Modak <subrata@linux.vnet.ibm.com>]

On Thu, 2009-06-18 at 12:57 +0530, Balbir Singh wrote:
> * Ingo Molnar <mingo@elte.hu> [2009-06-17 18:23:03]:
> 
> > 
> > * Frans Pop <elendil@planet.nl> wrote:
> > 
> > > On Wednesday 17 June 2009, Ingo Molnar wrote:
> > > > Ah, and you fixed in what a superior way: you improved the code in
> > > > the process :-) This is how warnings should be fixed really.
> > > 
> > > Hmm. Did you also see Pavel's reply to that patch [1]:
> > > ! But that's a bug to be fixed, I'd say? ... actually I believe you are
> > > ! introducing a bug here. Yes, old code would put random numbers in
> > > ! loops_per_jiffy_ref for !CPUFREQ_CONST_LOOPS, but you are introducing
> > > ! oops there.
> > > 
> > > Was his comment incorrect?
> > > 
> > > Cheers,
> > > FJP
> > > 
> > > [1] http://lkml.org/lkml/2009/5/24/159
> > 
> > hm, Pavel seems right - i missed that. Subrata, mind sending an 
> > updated patch?
> >
> 
> Hi, Ingo,
> 
> Yes, it does seem that we'll oops at *lpj, but to be honest the code
> is badly written, ideally the CONFIG_SMP part should be abstracted
> out, having that in a if loop makes reading it time consuming and
> kills a few neuro cells each time.

Balbir/Ingo,

I find this fixed in todayś Linus´s git tree. dummy has been removed,
and, so does lpj initialization taking place properly.

631 static int time_cpufreq_notifier(struct notifier_block *nb, unsigned
long val,
632                                 void *data)
633 {
634         struct cpufreq_freqs *freq = data;
635         unsigned long *lpj;
636 
637         if (cpu_has(&cpu_data(freq->cpu), X86_FEATURE_CONSTANT_TSC))
638                 return 0;
639 
640         lpj = &boot_cpu_data.loops_per_jiffy;

this gets initialized here for the first time on any system.

641 #ifdef CONFIG_SMP
642         if (!(freq->flags & CPUFREQ_CONST_LOOPS))
643                 lpj = &cpu_data(freq->cpu).loops_per_jiffy;
644 #endif

And gets re-assigned if CONFIG_SMP.

Regards--
Subrata

> 
>