Re: [PATCH net-next 6/6] net: ethtool: add the ability to run ethtool_[gs]et_rxnfc() without RTNL

[Willem de Bruijn <willemdebruijn.kernel@gmail.com>]

Eric Dumazet wrote:
> For better scalability, drivers can prefer to implement their own locking schem
> (for instance one mutex per port or queue) instead of relying on RTNL.
> 
> This patch adds a new boolean field in ethtool_ops : rxnfc_parallel
> 
> Drivers can opt-in to this new behavior.
> 
> Signed-off-by: Eric Dumazet <edumazet@google.com>
> ---
>  include/linux/ethtool.h |  2 ++
>  net/ethtool/ioctl.c     | 43 +++++++++++++++++++++++++++--------------
>  2 files changed, 31 insertions(+), 14 deletions(-)
> 
> diff --git a/include/linux/ethtool.h b/include/linux/ethtool.h
> index 6fd9107d3cc010dd2f1ecdb005c412145c461b6c..ee9b8054165361c9236186ff61f886e53cfa6b49 100644
> --- a/include/linux/ethtool.h
> +++ b/include/linux/ethtool.h
> @@ -748,6 +748,7 @@ struct ethtool_rxfh_param {
>   *	error code or zero.
>   * @set_rxnfc: Set RX flow classification rules.  Returns a negative
>   *	error code or zero.
> + * @rxnfc_parallel: true if @set_rxnfc, @get_rxnfc and @get_rxfh do not need RTNL.
>   * @flash_device: Write a firmware image to device's flash memory.
>   *	Returns a negative error code or zero.
>   * @reset: Reset (part of) the device, as specified by a bitmask of
> @@ -907,6 +908,7 @@ struct ethtool_ops {
>  	int	(*get_rxnfc)(struct net_device *,
>  			     struct ethtool_rxnfc *, u32 *rule_locs);
>  	int	(*set_rxnfc)(struct net_device *, struct ethtool_rxnfc *);
> +	bool	rxnfc_parallel;

Would it make sense to make this a bit, as there already are u32 bits
at the start of the struct, with a 29-bit gap?