Re: [PATCH v3] mm: fix race between MADV_FREE reclaim and blkdev direct IO read

[Yu Zhao <yuzhao@google.com>]

On Wed, Feb 16, 2022 at 02:48:19PM +0800, Huang, Ying wrote:
> Yu Zhao <yuzhao@google.com> writes:
> 
> > On Wed, Feb 02, 2022 at 06:27:47PM -0300, Mauricio Faria de Oliveira wrote:
> >> On Wed, Feb 2, 2022 at 4:56 PM Yu Zhao <yuzhao@google.com> wrote:
> >> >
> >> > On Mon, Jan 31, 2022 at 08:02:55PM -0300, Mauricio Faria de Oliveira wrote:
> >> > > Problem:
> >> > > =======
> >> >
> >> > Thanks for the update. A couple of quick questions:
> >> >
> >> > > Userspace might read the zero-page instead of actual data from a
> >> > > direct IO read on a block device if the buffers have been called
> >> > > madvise(MADV_FREE) on earlier (this is discussed below) due to a
> >> > > race between page reclaim on MADV_FREE and blkdev direct IO read.
> >> >
> >> > 1) would page migration be affected as well?
> >> 
> >> Could you please elaborate on the potential problem you considered?
> >> 
> >> I checked migrate_pages() -> try_to_migrate() holds the page lock,
> >> thus shouldn't race with shrink_page_list() -> with try_to_unmap()
> >> (where the issue with MADV_FREE is), but maybe I didn't get you
> >> correctly.
> >
> > Could the race exist between DIO and migration? While DIO is writing
> > to a page, could migration unmap it and copy the data from this page
> > to a new page?
> 
> Check the migrate_pages() code,
> 
>   migrate_pages
>     unmap_and_move
>       __unmap_and_move
>         try_to_migrate // set PTE to swap entry with PTL
>         move_to_new_page
>           migrate_page
>             folio_migrate_mapping
>               folio_ref_count(folio) != expected_count // check page ref count
>             folio_migrate_copy
> 
> The page ref count is checked after unmapping and before copying.  This
> is good, but it appears that we need a memory barrier between checking
> page ref count and copying page.

I didn't look into this but, off the top of head, this should be
similar if not identical to the DIO case. Therefore, it requires two
barriers -- before and after the refcnt check (which may or may not
exist).