Re: [PATCH v4 1/3] mm: Don't pin ZERO_PAGE in pin_user_pages()

[David Howells <dhowells@redhat.com>]

David Hildenbrand <david@redhat.com> wrote:

> > Make pin_user_pages*() leave a ZERO_PAGE unpinned if it extracts a pointer
> > to it from the page tables and make unpin_user_page*() correspondingly
> > ignore a ZERO_PAGE when unpinning.  We don't want to risk overrunning a
> > zero page's refcount as we're only allowed ~2 million pins on it -
> > something that userspace can conceivably trigger.
> 
> 2 millions pins (FOLL_PIN, which increments the refcount by 1024) or 2 million
> references ?

Definitely pins.  It's tricky because we've been using "pinned" to mean held
by a refcount or held by a flag too.

2 million pins on the zero page is in the realms of possibility.  It only
takes 32768 64-page DIO writes.

> > @@ -3079,6 +3096,9 @@ EXPORT_SYMBOL_GPL(get_user_pages_fast);
> >    *
> >    * FOLL_PIN means that the pages must be released via unpin_user_page(). Please
> >    * see Documentation/core-api/pin_user_pages.rst for further details.
> > + *
> > + * Note that if a zero_page is amongst the returned pages, it will not have
> > + * pins in it and unpin_user_page() will not remove pins from it.
> >    */
> 
> "it will not have pins in it" sounds fairly weird to a non-native speaker.

Oh, I know.  The problem is that "pin" is now really ambiguous.  Can we change
"FOLL_PIN" to "FOLL_NAIL"?  Or maybe "FOLL_SCREW" - your pages are screwed if
you use DIO and fork at the same time.

> "Note that the refcount of any zero_pages returned among the pinned pages will
> not be incremented, and unpin_user_page() will similarly not decrement it."

That's not really right (although it happens to be true), because we're
talking primarily about the pin counter, not the refcount - and they may be
separate.

David