On Sun, Aug 20, 2017 at 9:40 PM, Tom Rini <tr...@konsulko.com> wrote:
> 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.hershber...@ni.com>
> Signed-off-by: Tom Rini <tr...@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];
All of the pointer math would also need to be updated. Didn't notice
that at first so,
Nacked-by: Joe Hershberger <joe.hershber...@ni.com>
> } reply;
> } u;
> } __attribute__((packed));
> --
> 1.9.1
>
> _______________________________________________
> U-Boot mailing list
> U-Boot@lists.denx.de
> https://lists.denx.de/listinfo/u-boot
_______________________________________________
U-Boot mailing list
U-Boot@lists.denx.de
https://lists.denx.de/listinfo/u-boot
