Messages in this thread | | | Date | Fri, 13 Sep 2013 09:56:18 +0530 | Subject | Re: cpufreq_stats NULL deref on second system suspend | From | Viresh Kumar <> |
| |
On 12 September 2013 22:56, Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com> wrote: > On 09/12/2013 09:25 PM, Stephen Warren wrote: > Anyway, nevermind, as of now, subsystems do work around this suitably, so > there is no known bug as such at the present. Just that we could have probably > done it a better way, that's all.
Yeah, there is no bug as of now due to the number of hacks adopted by different framework.. I believe we can still have a cleanup series to take care of this stuff.. That would be some improvement and would be better for future.. Otherwise this kind of problems would keep coming again and again..
> You're absolutely right! Regular CPU hotplug is more demanding than > suspend/resume in the context we are discussing, since any CPU can be > hotplugged at any time and put back in any order. So code like cpufreq should > be prepared to work with any ordering.
And that part is well implemented and tested as far as I know..
| |