On 02-06-16, 19:02, Javi Merino wrote:
On Thu, Jun 02, 2016 at 09:06:26PM +0530, Viresh Kumar wrote:
On 2 June 2016 at 20:29, Javi Merino javi.merino@arm.com wrote:
In 5a31d594a973 ("cpufreq: Allow freq_table to be obtained for offline CPUs") you did the opposite: don't use cpufreq_cpu_get_raw() because it won't give you the policy of a cpu that is offline. Now you are arguing that we should go back to cpufreq_cpu_get() which implicitly calls cpufreq_cpu_get_raw(). Won't we hit the same issue that 5a31d594a973 was trying to prevent: that we can't get a freq_table for a cpu that is offline?
Yes, that should be fixed. Thanks for letting me know about it :)
Ok, that was my only nit. Other than that, it looks good to me. For cpu_cooling.c
Acked-by: Javi Merino javi.merino@arm.com
Thanks, I will be adding this to the original patch.
diff --git a/drivers/thermal/cpu_cooling.c b/drivers/thermal/cpu_cooling.c index 63f760869651..4d678cfc81b1 100644 --- a/drivers/thermal/cpu_cooling.c +++ b/drivers/thermal/cpu_cooling.c @@ -792,10 +792,12 @@ __cpufreq_cooling_register(struct device_node *np, struct cpufreq_cooling_device *cpufreq_dev; char dev_name[THERMAL_NAME_LENGTH]; struct cpufreq_frequency_table *pos, *table; + struct cpumask temp_mask; unsigned int freq, i, num_cpus; int ret;
- policy = cpufreq_cpu_get(cpumask_first(clip_cpus)); + cpumask_and(&temp_mask, clip_cpus, cpu_online_mask); + policy = cpufreq_cpu_get(cpumask_first(&temp_mask)); if (!policy) { pr_debug("%s: CPUFreq policy not found\n", __func__); return ERR_PTR(-EPROBE_DEFER);
This will also make this problem clear, otherwise it was just hidden in the function call which was really easy to miss, as I missed it as well :(