private_data_len is defined as a u8.  If the user specifies a large
private_data size (> 220 bytes), we will calculate a total length that
exceeds 255, resulting in private_data_len wrapping back to 0.  This
can lead to overwriting random kernel memory.  Avoid this by verifying
that the resulting size fits into a u8.

Problem reported by B. Thery, OFA bug 2335.

Signed-off-by: Sean Hefty <[email protected]>
---
Please consider queuing for 3.2 if it's not too late.  This bug has been
around for a while though, so hitting 3.3 is fine.

 drivers/infiniband/core/cma.c |    6 ++++++
 1 files changed, 6 insertions(+), 0 deletions(-)

diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c
index c3a0c62..fdd8ef8 100644
--- a/drivers/infiniband/core/cma.c
+++ b/drivers/infiniband/core/cma.c
@@ -2513,6 +2513,9 @@ static int cma_resolve_ib_udp(struct rdma_id_private 
*id_priv,
 
        req.private_data_len = sizeof(struct cma_hdr) +
                               conn_param->private_data_len;
+       if (req.private_data_len < conn_param->private_data_len)
+               return -EINVAL;
+
        req.private_data = kzalloc(req.private_data_len, GFP_ATOMIC);
        if (!req.private_data)
                return -ENOMEM;
@@ -2562,6 +2565,9 @@ static int cma_connect_ib(struct rdma_id_private *id_priv,
        memset(&req, 0, sizeof req);
        offset = cma_user_data_offset(id_priv->id.ps);
        req.private_data_len = offset + conn_param->private_data_len;
+       if (req.private_data_len < conn_param->private_data_len)
+               return -EINVAL;
+
        private_data = kzalloc(req.private_data_len, GFP_ATOMIC);
        if (!private_data)
                return -ENOMEM;


--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to