[PULL,1/2] xen/blkif: Avoid double access to src->nr_segments

Commit Message

Stefano Stabellini Dec. 18, 2015, 3:17 p.m. UTC

src is stored in shared memory and src->nr_segments is dereferenced
twice at the end of the function.  If a compiler decides to compile this
into two separate memory accesses then the size limitation could be
bypassed.

Fix it by removing the double access to src->nr_segments.

This is part of XSA-155.

Signed-off-by: Stefano Stabellini <[e-mail]>
---
 hw/block/xen_blkif.h |   12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

Patch

diff --git a/hw/block/xen_blkif.h b/hw/block/xen_blkif.h
index 711b692..c68487cb 100644
--- a/hw/block/xen_blkif.h
+++ b/hw/block/xen_blkif.h
@@ -85,8 +85,10 @@  static inline void blkif_get_x86_32_req(blkif_request_t *dst, blkif_x86_32_reque
 		d->nr_sectors = s->nr_sectors;
 		return;
 	}
-	if (n > src->nr_segments)
-		n = src->nr_segments;
+	/* prevent the compiler from optimizing the code and using src->nr_segments instead */
+	barrier();
+	if (n > dst->nr_segments)
+		n = dst->nr_segments;
 	for (i = 0; i < n; i++)
 		dst->seg[i] = src->seg[i];
 }
@@ -106,8 +108,10 @@  static inline void blkif_get_x86_64_req(blkif_request_t *dst, blkif_x86_64_reque
 		d->nr_sectors = s->nr_sectors;
 		return;
 	}
-	if (n > src->nr_segments)
-		n = src->nr_segments;
+	/* prevent the compiler from optimizing the code and using src->nr_segments instead */
+	barrier();
+	if (n > dst->nr_segments)
+		n = dst->nr_segments;
 	for (i = 0; i < n; i++)
 		dst->seg[i] = src->seg[i];
 }