Re: [PATCH v1 2/3] mm: process_mrelease: skip LRU movement for exclusive file folios

["David Hildenbrand (Arm)" <david@kernel.org>]

>>
>> But we'd have to learn from Liam :)
> 
> Yes, it will be a zero entry tree now.
> 
> I left the flag to indicate that it's an unstable mm, not for faulting
> in but to be skipped in OOM events and process_mrelease since neither
> should bother doing anything in the window between the dup_mm() failure
> and the exit_mmap() window where the write lock was dropped.
> 
> We can safely drop the flag now if you want to, because everything has
> to deal with an empty vma tree anyways - a race can occur between a call
> to unmap everything and the task seg faulting.

Thanks for confirming!

> 
>>
>>
>>>
>>> The bottom line is the same. Make sure PF fails rather than silently
>>> provide potentially corrupted data.
>>>
>>>
>>> Ohh, you mean to add a new flag, right?
>>
>> We could do that as well, if it's of any help.
> 
> I really think this goes back to the life cycle of the mm being somewhat
> difficult to figure out.  

Agreed.

> I'm fine with another flag.

Right, alternatively we could just turn the unstable flag into a "OOM is hiding
in the bushes to reap this MM".

-- 
Cheers,

David