Re: [PATCH v3 1/3] mm/swapfile: unuse_pte can map random data if swap read fails

[HORIGUCHI NAOYA(堀口 直也)]

On Mon, Apr 25, 2022 at 04:47:41PM +0800, Miaohe Lin wrote:
> On 2022/4/25 15:45, David Hildenbrand wrote:
> > On 25.04.22 03:08, HORIGUCHI NAOYA(堀口 直也) wrote:
> >> On Sun, Apr 24, 2022 at 05:11:03PM +0800, Miaohe Lin wrote:
> >>> There is a bug in unuse_pte(): when swap page happens to be unreadable,
> >>> page filled with random data is mapped into user address space.  In case
> >>> of error, a special swap entry indicating swap read fails is set to the
> >>> page table.  So the swapcache page can be freed and the user won't end up
> >>> with a permanently mounted swap because a sector is bad.  And if the page
> >>> is accessed later, the user process will be killed so that corrupted data
> >>> is never consumed.  On the other hand, if the page is never accessed, the
> >>> user won't even notice it.
> >>>
> >>> Signed-off-by: Miaohe Lin <linmiaohe@xxxxxxxxxx>
> >>> Acked-by: David Hildenbrand <david@xxxxxxxxxx>
> >>
> >> Hi Miaohe,
> >>
> >> This bug sounds relatively serious to me, and it seems old, so is it worth
> >> sending to -stable?
> > 
> > I'm not sure if this is worth -stable, but no strong opinion.
> 
> I have no strong opinion too. I'm just afraid someone might run into it. But swapoff is
> expected to be a rare operation anyway...
> 
> > 
> > The do_swap_page() part was added in 2005:
> > 
> > commit b81074800b98ac50b64d4c8d34e8abf0fda5e3d1
> > Author: Kirill Korotaev <dev@xxxxx>
> > Date:   Mon May 16 21:53:50 2005 -0700
> > 
> >     [PATCH] do_swap_page() can map random data if swap read fails
> >     
> >     There is a bug in do_swap_page(): when swap page happens to be unreadable,
> >     page filled with random data is mapped into user address space.  The fix is
> >     to check for PageUptodate and send SIGBUS in case of error.
> >     
> >     Signed-Off-By: Kirill Korotaev <dev@xxxxx>
> >     Signed-Off-By: Alexey Kuznetsov <kuznet@xxxxxxxxxxxxx>
> >     Acked-by: Hugh Dickins <hugh@xxxxxxxxxxx>
> >     Signed-off-by: Andrew Morton <akpm@xxxxxxxx>
> >     Signed-off-by: Linus Torvalds <torvalds@xxxxxxxx>
> > 
> > So the do_swap_page() part has been fixed for quite a while already.
> 
> Does this mean only do_swap_page maps random data if swap read fails is observed from that time on?
> So this might not be worth -stable as it's never seen more than a decade?

OK, both choices seems possible, so not sending to -stable is fine to me.
It's finally up to you.

Thanks,
Naoya Horiguchi