Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1156979

Re: [PATCH v1 2/9]powerpc/powernv: nest pmu init function with cpumask attr

Show all headers | View raw

[Multipart message — attachments visible in raw view] - view raw

On Tue, 2015-06-02 at 21:29 +0530, Madhavan Srinivasan wrote:
> Patch creates a file "nest-pmu-c" to contain nest pmu related functions.
"nest-pmu.c"
> Patch adds nest pmu init function and cpumask function since Nest pmu units
> are per-chip. First online cpu for a given node is picked as
> designated thread to read the counter data.
> 
> Subsequent patch adds the hotplug support.
> 
> Cc: Michael Ellerman <mpe@ellerman.id.au>
> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
> Cc: Paul Mackerras <paulus@samba.org>
> Cc: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
> Cc: Anshuman Khandual <khandual@linux.vnet.ibm.com>
> Cc: Stephane Eranian <eranian@google.com>
> Cc: Preeti U Murthy <preeti@linux.vnet.ibm.com>
> Cc: Ingo Molnar <mingo@kernel.org>
> Cc: Peter Zijlstra <peterz@infradead.org>
> Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
> ---
>  arch/powerpc/perf/nest-pmu.c | 70 ++++++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 70 insertions(+)
>  create mode 100644 arch/powerpc/perf/nest-pmu.c
> 
> diff --git a/arch/powerpc/perf/nest-pmu.c b/arch/powerpc/perf/nest-pmu.c
> new file mode 100644
> index 0000000..d4413bb
> --- /dev/null
> +++ b/arch/powerpc/perf/nest-pmu.c
> @@ -0,0 +1,70 @@
> +/*
> + * Nest Performance Monitor counter support for POWER8 processors.
> + *
> + * Copyright 2015 Madhavan Srinivasan, IBM Corporation.
> + *
> + * This program is free software; you can redistribute it and/or
> + * modify it under the terms of the GNU General Public License
> + * as published by the Free Software Foundation; either version
> + * 2 of the License, or (at your option) any later version.
> + */
> +
Again, I think this is supposed to be v2 only.

> +#include "nest-pmu.h"
> +
> +static cpumask_t cpu_mask_nest_pmu;
> +
> +static ssize_t cpumask_nest_pmu_get_attr(struct device *dev,
> +				struct device_attribute *attr, char *buf)
> +{
> +	return cpumap_print_to_pagebuf(true, buf, &cpu_mask_nest_pmu);
> +}
> +
> +static DEVICE_ATTR(cpumask, S_IRUGO, cpumask_nest_pmu_get_attr, NULL);
> +
> +static struct attribute *cpumask_nest_pmu_attrs[] = {
> +	&dev_attr_cpumask.attr,
> +	NULL,
> +};
> +
> +static struct attribute_group cpumask_nest_pmu_attr_group = {
> +	.attrs = cpumask_nest_pmu_attrs,
> +};
> +
> +void cpumask_chip(void)
> +{
> +	const struct cpumask *l_cpumask;
> +	int cpu, nid;
> +
> +	if (!cpumask_empty(&cpu_mask_nest_pmu)) {
> +		printk(KERN_INFO "cpumask not empty\n");
> +		return;
> +	}
> +
> +	cpu_notifier_register_begin();
> +	for_each_online_node(nid) {
> +		l_cpumask = cpumask_of_node(nid);
> +		cpu = cpumask_first(l_cpumask);
> +		cpumask_set_cpu(cpu, &cpu_mask_nest_pmu);
> +	}
> +
> +	cpu_notifier_register_done();
> +}
It's not clear from the name of this function what it does. I don't
think I actually understand what it does: it appears to register a
notifier on the first cpu of each node; maybe that should be reflected
in the name.

> +static int __init nest_pmu_init(void)
> +{
> +	int ret = 0;
> +
> +	/*
> +	 * Lets do this only if we are hypervisor
> +	 */
> +	if (!cur_cpu_spec->oprofile_cpu_type ||
> +	   strcmp(cur_cpu_spec->oprofile_cpu_type, "ppc64/power8") ||
> +	   !cpu_has_feature(CPU_FTR_HVMODE))
> +		return ret;
> +
> +	cpumask_chip();
> +
> +	return 0;
> +}
 - Where is ret set? I can only see it set when it's defined: the if
statment doesn't change the value of ret as far as I can see...

 - Would it be clearer if you said
        !(strcmp(cur_cpu_spec->oprofile_cpu_type, "ppc64/power8") == 0)
   That would make it clearer that you're trying to get a list of
possible failure conditions.
 
 - Is there really no better way to check if a CPU is a power 8 than an
string comparison?

> +device_initcall(nest_pmu_init);

Regards,
Daniel Axtens

Back to linux.kernel | Previous | Next — Next in thread | Find similar | Unroll thread

Thread

Re: [PATCH v1 2/9]powerpc/powernv: nest pmu init function with  cpumask attr Daniel Axtens <dja@axtens.net> - 2015-06-03 01:20 +0200
  Re: [PATCH v1 2/9]powerpc/powernv: nest pmu init function with cpumask  attr Madhavan Srinivasan <maddy@linux.vnet.ibm.com> - 2015-06-04 10:10 +0200

csiph-web