On 2 February 2015 at 09:08, ethan zhao ethan.zhao@oracle.com wrote:
We take cpufreq_driver_lock() here, and so this will block thread B.
No, there is no cpufreq_driver_lock acquired between
cpufreq_cpu_get() and cpufreq_cpu_put()
I am not saying that the lock is taken between them. But within cpufreq_cpu_get() to make sure policy doesn't get freed while we are doing kobject_get().
beginning the deference of policy Thread B: ... ... __cpufreq_remove_dev_finish()cpufreq_policy_free(policy);
Perhaps move policy->rwsem out side the policy structure is a way to avoid it completely. and you could stopping the PPC thread stepping forward as my patch as temporary workaround.
I couldn't understand your problem completely. Apart from giving a detailed look of what's going on both threads, always specify where the BUG actually is..
The problem is you are using a rwsem inside policy structure to protect its assessment, that is bad design.
What is the current bug you are facing right now, I haven't understood it well. Also a lock within the structure isn't new. Its all over the kernel. I don't understand why you say its a bad design.
-- viresh