[PATCH 3/3] scsi: sg: recheck MMAP_IO request length with lock held

[Tyler Hicks]

From: Todd Poynor <toddpoynor at google.com>

Commit 1bc0eb044615 ("scsi: sg: protect accesses to 'reserved' page
array") adds needed concurrency protection for the "reserve" buffer.
Some checks that are initially made outside the lock are replicated once
the lock is taken to ensure the checks and resulting decisions are made
using consistent state.

The check that a request with flag SG_FLAG_MMAP_IO set fits in the
reserve buffer also needs to be performed again under the lock to ensure
the reserve buffer length compared against matches the value in effect
when the request is linked to the reserve buffer.  An -ENOMEM should be
returned in this case, instead of switching over to an indirect buffer
as for non-MMAP_IO requests.

Signed-off-by: Todd Poynor <toddpoynor at google.com>
Acked-by: Douglas Gilbert <dgilbert at interlog.com>
Signed-off-by: Martin K. Petersen <martin.petersen at oracle.com>

CVE-2017-0794

(cherry picked from commit 8d26f491116feaa0b16de370b6a7ba40a40fa0b4)
Signed-off-by: Tyler Hicks <tyhicks at canonical.com>
---
 drivers/scsi/sg.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/drivers/scsi/sg.c b/drivers/scsi/sg.c
index 42e5a140ada5..7b99026238eb 100644
--- a/drivers/scsi/sg.c
+++ b/drivers/scsi/sg.c
@@ -1703,9 +1703,12 @@ static int sg_start_req(Sg_request *srp, unsigned char *cmd)
 		    !sfp->res_in_use) {
 			sfp->res_in_use = 1;
 			sg_link_reserve(sfp, srp, dxfer_len);
-		} else if ((hp->flags & SG_FLAG_MMAP_IO) && sfp->res_in_use) {
+		} else if (hp->flags & SG_FLAG_MMAP_IO) {
+			res = -EBUSY; /* sfp->res_in_use == 1 */
+			if (dxfer_len > rsv_schp->bufflen)
+				res = -ENOMEM;
 			mutex_unlock(&sfp->f_mutex);
-			return -EBUSY;
+			return res;
 		} else {
 			res = sg_build_indirect(req_schp, sfp, dxfer_len);
 			if (res) {
-- 
2.7.4