Re: [PATCH 3/3] led-triggers: create a trigger for CPU activity

[Bryan Wu]

On Wed, Mar 14, 2012 at 7:11 AM, Andrew Morton
<akpm@xxxxxxxxxxxxxxxxxxxx> wrote:
> On Thu,  8 Mar 2012 17:11:03 +0800
> Bryan Wu <bryan.wu@xxxxxxxxxxxxx> wrote:
>
>> Attempting to consolidate the ARM LED code, this removes the
>> custom RealView LED trigger code to turn LEDs on and off in
>> response to CPU activity and replace it with a standard trigger.
>>
>> (bryan.wu@xxxxxxxxxxxxx:
>> It moves arch/arm/kernel/leds.c syscore stubs into this trigger.
>> It also provides ledtrig_cpu trigger event stub in <linux/leds.h>.
>> Although it was inspired by ARM work, it can be used in other arch.)
>
> The patch doesn't "remove" or "move" anything.  It wholly consists of
> additions.  Confused.
>

I'm working on rebase my previous ARM leds consolidation patches to
the linux-next. So this patch will come with those patches soon.
Thanks for pointing out this.

>> Cc: Richard Purdie <rpurdie@xxxxxxxxx>
>> Signed-off-by: Linus Walleij <linus.walleij@xxxxxxxxxx>
>> Signed-off-by: Bryan Wu <bryan.wu@xxxxxxxxxxxxx>
>>
>> Reviewed-by: Jamie Iles <jamie@xxxxxxxxxxxxx>
>> Tested-by: Jochen Friedrich <jochen@xxxxxxxx>
>
> The authorship is a bit unclear.  You're the primary author, yes?
>

My work is based on Linus Walleij's patch, so I keep his signed off
here. Basically I rewrote most of things.

>> --- /dev/null
>> +++ b/drivers/leds/ledtrig-cpu.c
>> @@ -0,0 +1,119 @@
>> +/*
>> + * ledtrig-cpu.c - LED trigger based on CPU activity
>> + *
>> + * Copyright 2011 Linus Walleij <linus.walleij@xxxxxxxxxx>
>> + * Copyright 2011 Bryan Wu <bryan.wu@xxxxxxxxxxxxx>
>> + *
>> + * This program is free software; you can redistribute it and/or modify
>> + * it under the terms of the GNU General Public License version 2 as
>> + * published by the Free Software Foundation.
>> + *
>> + */
>> +
>> +#include <linux/module.h>
>> +#include <linux/kernel.h>
>> +#include <linux/init.h>
>> +#include <linux/slab.h>
>> +#include <linux/percpu.h>
>> +#include <linux/syscore_ops.h>
>> +#include "leds.h"
>> +
>> +static DEFINE_PER_CPU(struct led_trigger *, cpu_trig);
>> +static DEFINE_PER_CPU(char [8], trig_name);
>> +
>> +void ledtrig_cpu(enum cpu_led_event ledevt)
>> +{
>> +     struct led_trigger *trig = __get_cpu_var(cpu_trig);
>> +
>> +     if (!trig)
>> +             return;
>> +
>> +     /* Locate the correct CPU LED */
>> +     switch (ledevt) {
>> +     case CPU_LED_IDLE_END:
>> +     case CPU_LED_START:
>> +             /* Will turn the LED on, max brightness */
>> +             led_trigger_event(trig, LED_FULL);
>> +             break;
>> +
>> +     case CPU_LED_IDLE_START:
>> +     case CPU_LED_STOP:
>> +     case CPU_LED_HALTED:
>> +             /* Will turn the LED off */
>> +             led_trigger_event(trig, LED_OFF);
>> +             break;
>> +
>> +     default:
>> +             /* Will leave the LED as it is */
>> +             break;
>> +     }
>> +}
>> +EXPORT_SYMBOL(ledtrig_cpu);
>
> This is a global, exported-to-modules API function.  It should be
> documented.
>
> It should especially be documented if it has tricky unobvious calling
> conditions, as this function does.  AFAICT it must be called by code
> which is running on the target CPU and which is pinned to that CPU via
> either preempt_disable() or set_cpus_allowed().
>
> Or something else - I don't have a clue, and the author didn't tell me :(
>

OK, I will sort out the documents about this API.

>> +static int ledtrig_cpu_syscore_suspend(void)
>> +{
>> +     ledtrig_cpu(CPU_LED_STOP);
>> +     return 0;
>> +}
>> +
>> +static void ledtrig_cpu_syscore_resume(void)
>> +{
>> +     ledtrig_cpu(CPU_LED_START);
>> +}
>> +
>> +static void ledtrig_cpu_syscore_shutdown(void)
>> +{
>> +     ledtrig_cpu(CPU_LED_HALTED);
>> +}
>> +
>> +static struct syscore_ops ledtrig_cpu_syscore_ops = {
>> +     .shutdown       = ledtrig_cpu_syscore_shutdown,
>> +     .suspend        = ledtrig_cpu_syscore_suspend,
>> +     .resume         = ledtrig_cpu_syscore_resume,
>> +};
>> +
>> +static void __init ledtrig_cpu_register(void *info)
>> +{
>> +     int cpuid = smp_processor_id();
>> +     struct led_trigger *trig;
>> +     char *name = __get_cpu_var(trig_name);
>> +
>> +     snprintf(name, 8, "cpu%d", cpuid);
>
> This explodes at 10,000 CPUS ;)
>

I think it's enough for a while -:). or need I use snprintf(name,
4+NR_CPUS, "cpu%d", cpuid);

>> +     led_trigger_register_simple(name, &trig);
>> +
>> +     pr_info("LED trigger %s indicate activity on CPU %d\n",
>> +             trig->name, cpuid);
>> +
>> +     __get_cpu_var(cpu_trig) = trig;
>> +}
>> +
>> +static void __exit ledtrig_cpu_unregister(void *info)
>> +{
>> +     struct led_trigger *trig = __get_cpu_var(cpu_trig);
>> +     char *name = __get_cpu_var(trig_name);
>> +
>> +     led_trigger_unregister_simple(trig);
>> +     __get_cpu_var(cpu_trig) = NULL;
>> +     memset(name, 0, 8);
>> +}
>> +
>> +static int __init ledtrig_cpu_init(void)
>> +{
>> +     on_each_cpu(ledtrig_cpu_register, NULL, 1);
>
> on_each_cpu() only calls the function on presently-onlined CPUs.  How
> does this code interact with CPU hotplug?
>
> And as you've already noted, led_trigger_register_simple() is totally
> not callable from interrupt context.  And we're not just talking about
> under local_irq_disable() here: on_each_cpu() will raise cross-cpu hard
> interrupts to call the function on remote CPUs.  We end up trying to do
> down_write() and who knows what else from within a hard interrupt
> handler.
>

Right, I will change that to for_each_cpu() and resubmmit the patch.

>> --- a/include/linux/leds.h
>> +++ b/include/linux/leds.h
>> @@ -210,4 +210,19 @@ struct gpio_led_platform_data {
>>  struct platform_device *gpio_led_register_device(
>>               int id, const struct gpio_led_platform_data *pdata);
>>
>> +#if defined(CONFIG_LEDS_TRIGGER_CPU) || defined(CONFIG_LEDS_TRIGGER_CPU_MODULE)
>> +enum cpu_led_event {
>> +     CPU_LED_IDLE_START,     /* CPU enters idle */
>> +     CPU_LED_IDLE_END,       /* CPU idle ends */
>> +     CPU_LED_START,          /* Machine starts, especially resume */
>> +     CPU_LED_STOP,           /* Machine stops, especially suspend */
>> +     CPU_LED_HALTED,         /* Machine shutdown */
>> +};
>> +
>> +/* Use this routine to handle LEDs */
>> +extern void ledtrig_cpu(enum cpu_led_event evt);
>> +#else
>> +#define ledtrig_cpu(evt) do {} while (0)
>> +#endif
>
> a) the stub function should be written in C, not CPP please
>

Actually, I'm only familiar with C and forget to write CPP for a
while. is this looks like CPP?

> b) it makes no sense to provide the stub function when the
>   definition of its argument type (enum cpu_led_event) is not
>   also available.
>

OK, l will move out the enum cpu_led_event out of this ifdef.

Thanks,
-- 
Bryan Wu <bryan.wu@xxxxxxxxxxxxx>
Kernel Developer    +86.138-1617-6545 Mobile
Canonical Ltd.      www.canonical.com
Ubuntu - Linux for human beings | www.ubuntu.com
--
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/