net: fix rds_iovec page count overflow

commit 1b1f693d7ad6d193862dcb1118540a030c5e761f upstream.

As reported by Thomas Pollet, the rdma page counting can overflow.  We
get the rdma sizes in 64-bit unsigned entities, but then limit it to
UINT_MAX bytes and shift them down to pages (so with a possible "+1" for
an unaligned address).

So each individual page count fits comfortably in an 'unsigned int' (not
even close to overflowing into signed), but as they are added up, they
might end up resulting in a signed return value. Which would be wrong.

Catch the case of tot_pages turning negative, and return the appropriate
error code.

[PG: In 34, var names are slightly different, 1b1f6's tot_pages is
 34's nr_pages, and 1b1f6's nr_pages is 34's nr; so map accordingly.]

Reported-by: Thomas Pollet <thomas.pollet@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Andy Grover <andy.grover@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>

diff --git a/net/rds/rdma.c b/net/rds/rdma.c
index cf0dfa77fb7ef2e0d3accd11e327f0a9aa4d4ff3..34819316f93aadc8add1deb7652c352e23496d8e 100644 (file)
--- a/net/rds/rdma.c
+++ b/net/rds/rdma.c
@@ -498,6 +498,15 @@ static struct rds_rdma_op *rds_rdma_prepare(struct rds_sock *rs,
 
                max_pages = max(nr, max_pages);
                nr_pages += nr;
+
+               /*
+                * nr for one entry is limited to (UINT_MAX>>PAGE_SHIFT)+1,
+                * so nr_pages cannot overflow without first going negative.
+                */
+               if ((int)nr_pages < 0) {
+                       ret = -EINVAL;
+                       goto out;
+               }
        }
 
        pages = kcalloc(max_pages, sizeof(struct page *), GFP_KERNEL);