Messages in this thread Patch in this message |  | | | Date | Sun, 1 Jun 2008 11:22:04 +0300 (EEST) | | From | Pekka J Enberg <> | | Subject | Re: [PATCH] nommu: fix kobjsize() for SLOB and SLUB | |
On Sun, 1 Jun 2008, Pekka Enberg wrote:
> > Not until the page->index bits are killed, otherwise you aren't fixing
> > anything. SLOB on nommu with those page->index tests will automatically
> > oops today, before or after your patches. Until that's resolved, there's
> > no point in pretending like kobjsize() has been "fixed". As no one has
> > come up with a valid reason for those tests existing in the first place,
> > simply having your patches and killing the BUG_ON()'s seems ok.
>
> Sorry if I'm starting to sound like a broken record, but can you explain why
> removing the ->index bits are safe? I mean, if removing them is really okay,
> that means we don't hit that code path with SLAB at all?
Paul, so with something like this, the WARN_ON never triggers?
diff --git a/mm/nommu.c b/mm/nommu.c
index dca93fc..38eec2e 100644
--- a/mm/nommu.c
+++ b/mm/nommu.c
@@ -109,16 +109,23 @@ unsigned int kobjsize(const void *objp)
* If the object we have should not have ksize performed on it,
* return size of 0
*/
- if (!objp || (unsigned long)objp >= memory_end || !((page = virt_to_page(objp))))
+ if (!objp)
+ return 0;
+
+ if ((unsigned long)objp >= memory_end)
+ return 0;
+
+ page = virt_to_page(objp);
+ if (!page)
return 0;
if (PageSlab(page))
return ksize(objp);
- BUG_ON(page->index < 0);
- BUG_ON(page->index >= MAX_ORDER);
+ if (WARN_ON(!PageCompound(page)))
+ return 0;
- return (PAGE_SIZE << page->index);
+ return PAGE_SIZE << compound_order(page);
}
/*
|  |