[PATCH] mm/pagewalk.c: report holes in hugetlb ranges

[PATCH] mm/pagewalk.c: report holes in hugetlb ranges

[Jann Horn]

commit 373c4557d2aa362702c4c2d41288fb1e54990b7c upstream.

This matters at least for the mincore syscall, which will otherwise copy
uninitialized memory from the page allocator to userspace.  It is
probably also a correctness error for /proc/$pid/pagemap, but I haven't
tested that.

Removing the `walk->hugetlb_entry` condition in walk_hugetlb_range() has
no effect because the caller already checks for that.

This only reports holes in hugetlb ranges to callers who have specified
a hugetlb_entry callback.

This issue was found using an AFL-based fuzzer.

v2:
 - don't crash on ->pte_hole==NULL (Andrew Morton)
 - add Cc stable (Andrew Morton)

Changed for 4.4/4.9 stable backport:
 - fix up conflict in the huge_pte_offset() call

Fixes: 1e25a271c8ac ("mincore: apply page table walker on do_mincore()")
Signed-off-by: Jann Horn <jannh@google.com>
---
Please apply this patch to <=4.9 stable trees instead of the
original patch.

 mm/pagewalk.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/mm/pagewalk.c b/mm/pagewalk.c
index 29f2f8b853ae..c2cbd2620169 100644
--- a/mm/pagewalk.c
+++ b/mm/pagewalk.c
@@ -142,8 +142,12 @@ static int walk_hugetlb_range(unsigned long addr, unsigned long end,
 	do {
 		next = hugetlb_entry_end(h, addr, end);
 		pte = huge_pte_offset(walk->mm, addr & hmask);
-		if (pte && walk->hugetlb_entry)
+
+		if (pte)
 			err = walk->hugetlb_entry(pte, hmask, addr, next, walk);
+		else if (walk->pte_hole)
+			err = walk->pte_hole(addr, next, walk);
+
 		if (err)
 			break;
 	} while (addr = next, addr != end);
-- 
2.15.0.448.gf294e3d99a-goog

Re: [PATCH] mm/pagewalk.c: report holes in hugetlb ranges

[Greg KH]

On Wed, Nov 22, 2017 at 02:58:59PM +0100, Jann Horn wrote:
> commit 373c4557d2aa362702c4c2d41288fb1e54990b7c upstream.
> 
> This matters at least for the mincore syscall, which will otherwise copy
> uninitialized memory from the page allocator to userspace.  It is
> probably also a correctness error for /proc/$pid/pagemap, but I haven't
> tested that.
> 
> Removing the `walk->hugetlb_entry` condition in walk_hugetlb_range() has
> no effect because the caller already checks for that.
> 
> This only reports holes in hugetlb ranges to callers who have specified
> a hugetlb_entry callback.
> 
> This issue was found using an AFL-based fuzzer.
> 
> v2:
>  - don't crash on ->pte_hole==NULL (Andrew Morton)
>  - add Cc stable (Andrew Morton)
> 
> Changed for 4.4/4.9 stable backport:
>  - fix up conflict in the huge_pte_offset() call
> 
> Fixes: 1e25a271c8ac ("mincore: apply page table walker on do_mincore()")
> Signed-off-by: Jann Horn <jannh@google.com>
> ---
> Please apply this patch to <=4.9 stable trees instead of the
> original patch.

Thanks for the backport!

greg k-h