Re: cpufreq_stats NULL deref on second system suspend

["Srivatsa S. Bhat" <srivatsa.bhat@linux.vnet.ibm.com>]

On 09/12/2013 12:11 PM, Viresh Kumar wrote:
> On 12 September 2013 11:56, Srivatsa S. Bhat
> <srivatsa.bhat@linux.vnet.ibm.com> wrote:
>> I had the same thought when solving this bug.. We have had similar issues with
>> CPU hotplug notifiers too: why are they invoked in the same order during both
>> CPU down and up, instead of reversing the order? I even had a patchset to perform
>> reverse-invocation of notifiers.. http://lwn.net/Articles/508072/
>> ... but people didn't find that very compelling to have.
>>
> 
>> It does to me too, but I think the reason nobody really bothered is because perhaps
>> not many other subsystems care about the order in which CPUs are torn down or
>> brought up; they just need the total number to match.. cpufreq is one exception
>> as we saw with this bug.
> 
> Probably its time to re-spin that series and make CPUFreq as one of the users
> of that patchset.. Resume should be just opposite of suspend and so
> that patchset
> would make sense even if not many people care about it :)
> 
> Over that there is one more problem that I see, don't know if it is really a big
> issue..
> 
> After a suspend/resume value of policy->cpu may get changed... And so the
> hierarchy of sysfs cpufreq files too.. Folder that had links to other
> CPUs folder
> can now be actual folders instead of links and vice versa..
> 
> Don't know if this can break something ??
> 

Interesting observation :-) But we just managed to retain sysfs file permissions
across suspend/resume with a lot of trouble and regressions. That's probably
good enough for some time to come ;-) We can retain folder/links when somebody
really finds a need to do that ;-)

Of course, if we change the suspend/resume sequence and that fixes this, that
would be like getting it for free, nobody would say no to it ;-)

Regards,
Srivatsa S. Bhat