Re: [rfc patch 3/3] mm: munlock COW pages on truncation unmap

[KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>]

> Hi,
> 
> On Sat, Oct 03, 2009 at 10:56:55PM +0900, KOSAKI Motohiro wrote:
> > >> Umm..
> > >> I haven't understand this.
> > >>
> > >> (1) unmap_mapping_range() is called twice.
> > >>
> > >>       unmap_mapping_range(mapping, new + PAGE_SIZE - 1, 0, 1);
> > >>       truncate_inode_pages(mapping, new);
> > >>       unmap_mapping_range(mapping, new + PAGE_SIZE - 1, 0, 1);
> > >>
> > >> (2) PG_mlock is turned on from mlock() and vmscan.
> > >> (3) vmscan grab anon_vma, but mlock don't grab anon_vma.
> > >
> > > You are right, I was so focused on the LRU side that I missed an
> > > obvious window here: an _explicit_ mlock can still happen between the
> > > PG_mlocked clearing section and releasing the page.
> 
> Okay, so what are the opinions on this?  Would you consider my patches
> to fix the most likely issues?  Dropping them in favor of looking for
> a complete fix?  Revert the warning on freeing PG_mlocked pages?

Honestly, I don't have any good idea. but luckly, we have enough time.
the false-positve warning is not so big problem. then, I prefer looking for
complete solusion.

Thanks.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>