[PATCH 3.13.y-ckt 077/108] xen-blkback: read from indirect descriptors only once

[Kamal Mostafa]

3.13.11-ckt33 -stable review patch.  If anyone has any objections, please let me know.

---8<------------------------------------------------------------

From: =?UTF-8?q?Roger=20Pau=20Monn=C3=A9?= <roger.pau at citrix.com>

commit 18779149101c0dd43ded43669ae2a92d21b6f9cb upstream.

Since indirect descriptors are in memory shared with the frontend, the
frontend could alter the first_sect and last_sect values after they have
been validated but before they are recorded in the request.  This may
result in I/O requests that overflow the foreign page, possibly
overwriting local pages when the I/O request is executed.

When parsing indirect descriptors, only read first_sect and last_sect
once.

This is part of XSA155.

Signed-off-by: Roger Pau Monné <roger.pau at citrix.com>
Signed-off-by: David Vrabel <david.vrabel at citrix.com>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk at oracle.com>
[ luis: backported to 3.16:
  - Use ACCESS_ONCE instead of READ_ONCE
  - Use PAGE_SIZE instead of XEN_PAGE_SIZE ]
Signed-off-by: Luis Henriques <luis.henriques at canonical.com>
Signed-off-by: Kamal Mostafa <kamal at canonical.com>
---
 drivers/block/xen-blkback/blkback.c | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)

diff --git a/drivers/block/xen-blkback/blkback.c b/drivers/block/xen-blkback/blkback.c
index 08c8185..0219bf2 100644
--- a/drivers/block/xen-blkback/blkback.c
+++ b/drivers/block/xen-blkback/blkback.c
@@ -853,6 +853,8 @@ static int xen_blkbk_parse_indirect(struct blkif_request *req,
 		goto unmap;
 
 	for (n = 0, i = 0; n < nseg; n++) {
+		uint8_t first_sect, last_sect;
+
 		if ((n % SEGS_PER_INDIRECT_FRAME) == 0) {
 			/* Map indirect segments */
 			if (segments)
@@ -860,15 +862,18 @@ static int xen_blkbk_parse_indirect(struct blkif_request *req,
 			segments = kmap_atomic(pages[n/SEGS_PER_INDIRECT_FRAME]->page);
 		}
 		i = n % SEGS_PER_INDIRECT_FRAME;
+
 		pending_req->segments[n]->gref = segments[i].gref;
-		seg[n].nsec = segments[i].last_sect -
-			segments[i].first_sect + 1;
-		seg[n].offset = (segments[i].first_sect << 9);
-		if ((segments[i].last_sect >= (PAGE_SIZE >> 9)) ||
-		    (segments[i].last_sect < segments[i].first_sect)) {
+
+		first_sect = ACCESS_ONCE(segments[i].first_sect);
+		last_sect = ACCESS_ONCE(segments[i].last_sect);
+		if (last_sect >= (PAGE_SIZE >> 9) || last_sect < first_sect) {
 			rc = -EINVAL;
 			goto unmap;
 		}
+
+		seg[n].nsec = last_sect - first_sect + 1;
+		seg[n].offset = first_sect << 9;
 		preq->nr_sects += seg[n].nsec;
 	}
 
-- 
1.9.1