Re: Cpufreq shutdown patch (kernel 2.6.35.3)

From: Dave Jones
Date: Mon Feb 13 2012 - 11:29:28 EST


On Thu, Jan 19, 2012 at 10:18:10PM +0100, N. Coesel wrote:

> I've found a problem in drivers/cpufreq/cpufreq.c. The driver does
> not execute the exit member of the low level driver when shutting
> down for a reboot (reset). This potentially leaves the power supply
> at a too low voltage to boot the system properly. The patch below
> adds a shutdown function which executes the exit member of the low
> level driver which allows a system to be properly prepared for a reset.

I'm curious what hardware you saw this problem on ?

I've just seen a report on an x86 system which looks like it might
be solved by this.
https://bugzilla.redhat.com/show_bug.cgi?id=789964

> --- drivers/cpufreq/cpufreq.c.orig 2010-08-20 20:55:55.000000000 +0200
> +++ drivers/cpufreq/cpufreq.c 2012-01-19 21:50:46.000000000 +0100
> @@ -1431,11 +1431,34 @@ fail:
> return ret;
> }
>
> +
> +static int cpufreq_shutdown(struct sys_device *sysdev)
> +{
> + struct cpufreq_policy *cpu_policy;
> + int cpu = sysdev->id;
> + int ret =0;
> +
> + if (!cpu_online(cpu))
> + return 0;
> +
> + cpu_policy = cpufreq_cpu_get(cpu);
> + if (!cpu_policy)
> + return -EINVAL;
> +
> +// printk("cpufreq_shutdown %d \n", cpu);
> + if (cpufreq_driver->exit)
> + ret = cpufreq_driver->exit(cpu_policy);
> +
> + return ret;
> +}
> +
> +
> static struct sysdev_driver cpufreq_sysdev_driver = {
> .add = cpufreq_add_dev,
> .remove = cpufreq_remove_dev,
> .suspend = cpufreq_suspend,
> .resume = cpufreq_resume,
> + .shutdown = cpufreq_shutdown,
> };

I'll queue this up, with the whitespace fixed, and commented out line removed.

Dave

--
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/