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

[Andrew Morton]

On Wed, 18 Dec 2013 19:41:44 -0500 Sasha Levin <sasha.levin@xxxxxxxxxx> wrote:

> On 12/18/2013 07:28 PM, Andrew Morton wrote:
> > On Thu, 19 Dec 2013 08:16:35 +0800 Wanpeng Li <liwanp@xxxxxxxxxxxxxxxxxx> 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?
> 
> Dave has improved mmap testing in trinity, maybe it's related?

Dave, can you please summarise recent trinity changes for us?

> > Or if there is no reason why the page must be locked for
> > rmap_walk_ksm() and rmap_walk_file(), let's just remove rmap_walk()'s
> > VM_BUG_ON()?  And rmap_walk_ksm()'s as well - it's duplicative anyway.
> 
> IMO, removing all these VM_BUG_ON()s (which is happening quite often recently) will
> lead to having bugs sneak by causing obscure undetected corruption instead of
> being very obvious through a BUG.
> 

Well.  a) My patch was functionally the same as the one Wanpeng
proposed, only better ;) and b) we shouldn't just assert X because we
observed that the existing code does X.  If a particular function
*needs* PageLocked(page) then sure, it can and should assert that the
page is locked.  Preferably with a comment explaining *why*
PageLocked() is needed.  That way we don't end up with years-old
assertions which nobody understands any more, which is what we have
now.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/