In rpc_t we declare data to be a uint8_t of size 2048, for a final size of 2048. We also however declare the reply part of the union to have a uint32_t data field of NFS_READ_SIZE (1024) for a final size of 4096+24=4120 and an overrun. Expand the comment above the struct to note that if NFS_READ_SIZE is increased then the data buf must also be increased and correct the declaration to be uint8_t.
Reported-by: Coverity (CID: 152888) Cc: Joe Hershberger joe.hershberger@ni.com Signed-off-by: Tom Rini trini@konsulko.com --- net/nfs.h | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/net/nfs.h b/net/nfs.h index 1aa06e8fb90f..b23b4088d825 100644 --- a/net/nfs.h +++ b/net/nfs.h @@ -39,8 +39,9 @@ /* * Block size used for NFS read accesses. A RPC reply packet (including all * headers) must fit within a single Ethernet frame to avoid fragmentation. - * However, if CONFIG_IP_DEFRAG is set, a bigger value could be used. In any - * case, most NFS servers are optimized for a power of 2. + * However, if CONFIG_IP_DEFRAG is set, a bigger value could be used, so long + * as rpc_t->u->data is incrased to match. In any case, most NFS servers are + * optimized for a power of 2. */ #define NFS_READ_SIZE 1024 /* biggest power of two that fits Ether frame */
@@ -73,7 +74,7 @@ struct rpc_t { uint32_t verifier; uint32_t v2; uint32_t astatus; - uint32_t data[NFS_READ_SIZE]; + uint8_t data[NFS_READ_SIZE]; } reply; } u; } __attribute__((packed));