On Fri, 2016-03-25 at 09:21 -0700, David Decotigny wrote:

> From: David Decotigny <decot@googlers.com>
> 
> To benefit from this, kernel commit 025c68186e07 ("netlink: add support
> for NIC driver ioctls") is needed.
> 
> 
> Signed-off-by: David Decotigny <decot@googlers.com>
> ---
>  configure.ac | 2 +-
>  ethtool.c    | 7 +++++++
>  2 files changed, 8 insertions(+), 1 deletion(-)
> 
> diff --git a/configure.ac b/configure.ac
> index 3105415..47d2a0f 100644
> --- a/configure.ac
> +++ b/configure.ac
> @@ -15,7 +15,7 @@ AM_PROG_CC_C_O
>  dnl Checks for libraries.
>  
>  dnl Checks for header files.
> -AC_CHECK_HEADERS(sys/ioctl.h)
> +AC_CHECK_HEADERS(sys/ioctl.h linux/netlink.h)

This doesn't make a lot of sense.  That header has been around since
Linux 2.1.  But it didn't define NETLINK_GENERIC then.

(The test for <sys/ioctl.h> doesn't make sense either, as we include it
unconditionally!)

>  dnl Checks for typedefs, structures, and compiler characteristics.
>  AC_MSG_CHECKING([whether  defines big-endian types])
> diff --git a/ethtool.c b/ethtool.c
> index cb3d971..314b1b8 100644
> --- a/ethtool.c
> +++ b/ethtool.c
> @@ -42,6 +42,9 @@
>  #include 
>  
>  #include 
> +#ifdef HAVE_LINUX_NETLINK_H
> +# include 
> +#endif
> 

So what I've committed instead makes this unconditional, but adds a
conditional definition of NETLINK_GENERIC.

Ben.

>  
>  #ifndef MAX_ADDR_LEN
>  #define MAX_ADDR_LEN	32
> @@ -4645,6 +4648,10 @@ opt_found:
>  
>  		/* Open control socket. */
>  		ctx.fd = socket(AF_INET, SOCK_DGRAM, 0);
> +#ifdef HAVE_LINUX_NETLINK_H
> +		if (ctx.fd < 0)
> +			ctx.fd = socket(AF_NETLINK, SOCK_RAW, NETLINK_GENERIC);
> +#endif
>  		if (ctx.fd < 0) {
>  			perror("Cannot get control socket");
>  			return 70;
-- 

Ben Hutchings
compatible: Gracefully accepts erroneous data from any source