Re: [PATCH 4/4] uprobes: __replace_page() needs munlock_vma_page()

[Oleg Nesterov <oleg@redhat.com>]

On 07/26, Srikar Dronamraju wrote:
>
> * Oleg Nesterov <oleg@redhat.com> [2012-07-12 19:10:25]:
>
> > Like do_wp_page(), __replace_page() should do munlock_vma_page()
> > for the case when the old page still has other !VM_LOCKED mappings.
> > Unfortunately this needs mm/internal.h.
> >
> > Also, move put_page() outside of ptl lock. This doesn't really
> > matter but looks a bit better.
> >
> > Signed-off-by: Oleg Nesterov <oleg@redhat.com>
>
> One thing I wanted to check is, should we mlock the new page, i.e the
> replacing page.

Yes, currently page_add_new_anon_rmap() makes it Mlocked/Unevictable.

> It may not a good idea to mlock the new page
> because then we can end up adding too many pages to the unevictable
> list.

Perhaps.

But. I think this is not really important. What is more important,
uprobe_register() should simply not add "too many pages", iow we
should re-use the same page if possible. This was another reason
for (buggy) http://marc.info/?l=linux-kernel&m=134013566617717
I sent before. I'll try to return to this later.

> Acked-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>

Thanks!

Oleg.