Re: [PATCH 2/3] network: recvmsg and sendmsg standard compliance (BZ#16919)

[Szabolcs Nagy <szabolcs dot nagy at arm dot com>]

On 28/03/16 15:34, Adhemerval Zanella wrote:
> POSIX specifies that both msghdr::msg_iovlen and msghdr::msg_controllen
> to be of size int and socklen_t respectively.  However Linux defines it as
> both size_t and for 64-bit it requires some adjustments to make the
> functions standard compliance.
> 
> This patch fixes it by creating a temporary header and zeroing the pad
> fields for 64-bits architecture where size of size_t exceeds the size of
> the int.

sendmsg is harder to fix because cmsghdr also needs fix ups:

>  /* Structure used for storage of ancillary data object information.  */
>  struct cmsghdr
>    {
> -    size_t cmsg_len;		/* Length of data in cmsg_data plus length
> -				   of cmsghdr structure.
> -				   !! The type should be socklen_t but the
> -				   definition of the kernel is incompatible
> -				   with this.  */
> +#if __BYTE_ORDER == __BIG_ENDIAN
> +    int __glibc_reserved1;	/* Pad toadjust Linux size to POSIX defined
> +				   size for cmsg_len.  */
> +    socklen_t cmsg_len;		/* Length of data in cmsg_data plus length
> +				   of cmsghdr structure.  */
> +#else
> +    socklen_t cmsg_len;
> +    int __glibc_reserved1;
> +#endif
>      int cmsg_level;		/* Originating protocol.  */
>      int cmsg_type;		/* Protocol specific type.  */
...

>  ssize_t
>  __libc_sendmsg (int fd, const struct msghdr *msg, int flags)
>  {
> +  /* POSIX specifies that both msghdr::msg_iovlen and msghdr::msg_controllen
> +     to be int and socklen_t respectively.  However Linux defines it as
> +     both size_t.  So for 64-bit it requires some adjustments by copying to
> +     temporary header and zeroing the pad fields.  */
> +#if __WORDSIZE == 64
> +  struct msghdr hdr;
> +  if (msg != NULL)
> +    {
> +      hdr = *msg;
> +      hdr.__glibc_reserved1 = 0;
> +      hdr.__glibc_reserved2 = 0;
> +      msg = &hdr;
> +    }
> +#endif

e.g. user supplied msg.msg_control might contain cmsghdr
with __glibc_reserved1 != 0 since user code might not
initialize the struct.

in musl this is fixed by copying the controls to a tmp
buf on the stack (which has fixed size so it can fail)
and fixing the paddings there.