powerpc: ppc64 address space capped at 32TB, mmap randomisation disabled
authorAnton Blanchard <anton@samba.org>
Mon, 18 Nov 2013 03:55:28 +0000 (14:55 +1100)
committerBenjamin Herrenschmidt <benh@kernel.crashing.org>
Wed, 20 Nov 2013 23:33:41 +0000 (10:33 +1100)
Commit fba2369e6ceb (mm: use vm_unmapped_area() on powerpc architecture)
has a bug in slice_scan_available() where we compare an unsigned long
(high_slices) against a shifted int. As a result, comparisons against
the top 32 bits of high_slices (representing the top 32TB) always
returns 0 and the top of our mmap region is clamped at 32TB

This also breaks mmap randomisation since the randomised address is
always up near the top of the address space and it gets clamped down
to 32TB.

Cc: stable@vger.kernel.org # v3.10+
Signed-off-by: Anton Blanchard <anton@samba.org>
Acked-by: Michel Lespinasse <walken@google.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
arch/powerpc/mm/slice.c

index 3e99c149271aa0d4f454c95a9d23eaffa7debec6..7ce9cf3b698835c0dd2b2644d137ff7549c68e72 100644 (file)
@@ -258,7 +258,7 @@ static bool slice_scan_available(unsigned long addr,
                slice = GET_HIGH_SLICE_INDEX(addr);
                *boundary_addr = (slice + end) ?
                        ((slice + end) << SLICE_HIGH_SHIFT) : SLICE_LOW_TOP;
-               return !!(available.high_slices & (1u << slice));
+               return !!(available.high_slices & (1ul << slice));
        }
 }