[PATCH v7 2/2] ARM hibernation / suspend-to-disk

[Alexander Holler]

Am 26.03.2014 00:36, schrieb Alexander Holler:
> Am 25.03.2014 19:38, schrieb Alexander Holler:
>
>> reboot too. Another solution would be to unregister the reboot_notifier
>> in the reboot_nofifier itself. I've just seen one watchdog driver
>> (drivers/rtc/rtc-m41t80.c) which does that. But I still think such
>
> That, btw. is broken. ;)
>
> Right after having send the mail, I've became that intuition, had a look
> and ... notifier.h does state the following:
>
>   * atomic_notifier_chain_unregister(),
> blocking_notifier_chain_unregister(),
>   * and srcu_notifier_chain_unregister() _must not_ be called from within
>   * the call chain.
>
> (The reboot-notifier chain is of type blocking_notifier_chain)
>
> So I've picked up one user of reboot_notifier by random and had the luck
> to choose a broken one. ;)
>
> I will make a patch and will have a look if the same failure can be
> found elsewhere. There aren't that much users of the reboot-notifier, so
> it shouldn't cost me that much time.

Hmm, and either I was confused, or have looked at some other user of the 
reboot_notifier, but rtc-m41t80.c doesn't call unregister from it's 
notifier. And unfortunately there are a bit more users of the 
reboot_notifier than I first thought. :/

I will check if I can find out at least at which driver I had a look at 
which did call unregister_notifier from the notifier itself.

Regards,

Alexander Holler