Re: [PATCH] mm/rmap: fix BUG at rmap_walk

[Andrew Morton <akpm@linux-foundation.org>]

On Thu, 19 Dec 2013 09:58:05 +0900 Joonsoo Kim <iamjoonsoo.kim@lge.com> wrote:

> On Wed, Dec 18, 2013 at 04:28:58PM -0800, Andrew Morton wrote:
> > On Thu, 19 Dec 2013 08:16:35 +0800 Wanpeng Li <liwanp@linux.vnet.ibm.com> wrote:
> > 
> > > page_get_anon_vma() called in page_referenced_anon() will lock and 
> > > increase the refcount of anon_vma, page won't be locked for anonymous 
> > > page. This patch fix it by skip check anonymous page locked.
> > > 
> > > [  588.698828] kernel BUG at mm/rmap.c:1663!
> > 
> > Why is all this suddenly happening.  Did we change something, or did a
> > new test get added to trinity?
> 
> It is my fault.
> I should remove this VM_BUG_ON() since rmap_walk() can be called
> without holding PageLock() in this case.
> 
> I think that adding VM_BUG_ON() to each rmap_walk calllers is better
> than this patch, because, now, rmap_walk() is called by many places and
> each places has different contexts.

I don't think that putting the assertion into the caller makes a lot of
sense, particularly if that code just did a lock_page()!  If a *callee*
needs PageLocked() then that callee should assert that the page is
locked.  So

	VM_BUG_ON(!PageLocked(page));

means "this code requires that the page be locked".  And if that code
requires PageLocked(), there must be reasons for this.  Let's also
include an explanation of those reasons.