Re: [PATCH 1/7] x86: Limit the number of processor bootup messages

[Mike Travis]

Ingo Molnar wrote:
> * Mike Travis <travis@xxxxxxx> wrote:
>
> > Ingo Molnar wrote:
> > > * Mike Travis <travis@xxxxxxx> wrote:
> > >
> > > > --- linux.orig/arch/x86/kernel/smpboot.c
> > > > +++ linux/arch/x86/kernel/smpboot.c
> > > > @@ -442,6 +442,84 @@
> > > > 		return c->llc_shared_map;
> > > > }
> > > > +/* Summarize Processor Information */
> > > > +static void __init summarize_cpu_info(void)
> > > > +{
> > > > +	cpumask_var_t cpulist, cpusdone;
> > > > +	int cpu;
> > > > +	int err = 0;
> > > > +
> > > > +	if (!alloc_cpumask_var(&cpulist, GFP_KERNEL))
> > > > +		err = 1;
> > > > +	else if (!alloc_cpumask_var(&cpusdone, GFP_KERNEL)) {
> > > > +		free_cpumask_var(cpulist);
> > > > +		err = 1;
> > > > +	}
> > > > +	if (err) {
> > > > +		printk(KERN_INFO "Can't print processor summaries\n");
> > > > +		return;
> > > > +	}
> > > > +
> > > > +	cpumask_clear(cpusdone);
> > > > +	for (cpu = 0; cpu < nr_cpu_ids; cpu++) {
> > > > +		struct cpuinfo_x86 *c;
> > > > +		char buf[128];
> > > > +		int ncpu, len;
> > > > +		unsigned long minlpj, maxlpj, avglpj = 0;
> > > > +
> > > > +		/* skip if cpu has already been displayed */
> > > > +		if (cpumask_test_cpu(cpu, cpusdone))
> > > > +			continue;
> > > > +
> > > > +		c = &cpu_data(cpu);
> > > > +		minlpj = ULONG_MAX;
> > > > +		maxlpj = 0;
> > > > +
> > > > +		cpumask_clear(cpulist);
> > > > +
> > > > +		/* collate all cpus with same specifics */
> > > > +		for (ncpu = cpu; ncpu < nr_cpu_ids; ncpu++) {
> > > > +			struct cpuinfo_x86 *n = &cpu_data(ncpu);
> > > > +
> > > > +			if (c->x86 != n->x86 ||
> > > > +			    c->x86_vendor != n->x86_vendor ||
> > > > +			    c->x86_model  != n->x86_model  ||
> > > > +			    c->x86_mask   != n->x86_mask   ||
> > > > +			    strcmp(c->x86_model_id, n->x86_model_id))
> > > > +				continue;
> > > > +
> > > > +			cpumask_set_cpu(ncpu, cpulist);
> > > > +			cpumask_set_cpu(ncpu, cpusdone);
> > > > +
> > > > +			if (cpu_data(ncpu).loops_per_jiffy < minlpj)
> > > > +				minlpj = cpu_data(ncpu).loops_per_jiffy;
> > > > +
> > > > +			if (cpu_data(ncpu).loops_per_jiffy > maxlpj)
> > > > +				maxlpj = cpu_data(ncpu).loops_per_jiffy;
> > > > +
> > > > +			avglpj += cpu_data(ncpu).loops_per_jiffy;
> > > > +		}
> > > > +
> > > > +		len = cpulist_scnprintf(buf, sizeof(buf), cpulist);
> > > > +		printk(KERN_INFO
> > > > +			"Processor Information for CPUS: %s%s\n",
> > > > +				buf, (len == sizeof(buf)-1) ? "..." : "");
> > > > +
> > > > +		printk(KERN_INFO);
> > > > +		print_cpu_info(c);
> > > > +
> > > > +		avglpj /= cpumask_weight(cpulist);
> > > > +		printk(KERN_INFO "BogoMIPS: MIN %lu.%02lu (%lu) "
> > > > +			"AVG %lu.%02lu (%lu) MAX %lu.%02lu (%lu)\n",
> > > > +			minlpj/(500000/HZ), (minlpj/(5000/HZ)) % 100, minlpj,
> > > > +			avglpj/(500000/HZ), (avglpj/(5000/HZ)) % 100, avglpj,
> > > > +			maxlpj/(500000/HZ), (maxlpj/(5000/HZ)) % 100, maxlpj);
> > > > +	}
> > > > +
> > > > +	free_cpumask_var(cpusdone);
> > > > +	free_cpumask_var(cpulist);
> > > > +}
> > > > +
> > > Sigh, that's _way_ too complex.
> > >
> > > If you cannot print it in a summarized way without carrying over
> > > stupid state like bitmaps then please do the simple and obvious,
> > > and print:
> > >
> > > booting CPUs: #0 #1 #2 #3 #4 #5 #6 ...
> > That is almost exactly what it does. [...]
>
> You are missing my point i think, which is that we dont want the 76 
> lines long summarize_cpu_info() complexity during bootup. Lets keep it 
> _very_ simple and zap excessive messages - like DaveJ's patch did.
>
> 	Ingo

Ok, I'll zap the summary as I'm not sure how much easier to make a function
that scans the cpus to find ones that are common, and print their specific
information.  If I can include the model id and stepping in some /proc
interface, then a user app could do the same thing (though I'm not sure if
BogoMIPs is available or not.)

Btw, I don't actually see the difference between the two functions, both
scan a bit map list (apics vs. cpus) to find unprocessed items, then
collates similar items comparing (nodeid vs. cpu specifics) and then
prints a summary using the scnlistprintf to gather progressive item #'s.
But maybe I'm missing something.

Thanks,
Mike
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/