Messages in this thread | | | Date | Thu, 31 Jul 2014 16:48:00 -0700 | From | Saravana Kannan <> | Subject | Re: [PATCH v4 3/5] cpufreq: Don't destroy/realloc policy/sysfs on hotplug/suspend |
| |
On 07/31/2014 02:56 PM, Rafael J. Wysocki wrote: > On Thursday, July 24, 2014 06:07:26 PM Saravana Kannan wrote: >> This patch simplifies a lot of the hotplug/suspend code by not >> adding/removing/moving the policy/sysfs/kobj during hotplug and just leaves >> the cpufreq directory and policy in place irrespective of whether the CPUs >> are ONLINE/OFFLINE. > > I'm still quite unsure how this is going to work with the real CPU hot-remove > that makes the entire sysfs cpu directories go away. Can you please explain > that?
Sure. Not a problem. I just wanted to make sure you had a chance to look at the code first.
Physical hot-remove triggers a "remove" for all the registered subsys_interfaces for that CPU (after going through a couple of functions). So, when that happens, the cpufreq subsys_interface remove for that CPU gets called. At that point, I clean up that CPU's SW states as if it was never plugged in from the start. If that CPU was the owner of the sysfs directory, I move it over to a different CPU.
-Saravana
-- The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum, hosted by The Linux Foundation
| |