[net-next,1/2] packet: rework packet_pick_tx_queue() to use common code selection

Currently packet_pick_tx_queue() is the only caller of
ndo_select_queue() using a fallback argument other than
__netdev_pick_tx.

Leveraging rx queue, we can obtain a similar queue selection
behavior using core helpers. After this change, ndo_select_queue()
is always invoked with __netdev_pick_tx() as fallback.
We can change ndo_select_queue() signature in a followup patch,
dropping an indirect call per transmitted packet in some scenarios
(e.g. TCP syn and XDP generic xmit)

This changes slightly how af packet queue selection happens when
PACKET_QDISC_BYPASS is set. It's now more similar to plan dev_queue_xmit()
tacking in account both XPS and TC mapping.

 RFC -> v1:
  - initialize sender_cpu to the expected value

Signed-off-by: Paolo Abeni <pabeni@redhat.com>
---
Note: the main goal here is getting rid of the fallback() indirect
call in the device drivers implementing ndo_select_queue(). We can
obtain the same result with the INDIRECT_CALL() harness. Both ways
we need to export __netdev_pick_tx() but here we avoid the need
for additional branching.
---
 include/linux/netdevice.h |  2 ++
 net/core/dev.c            |  5 +++--
 net/packet/af_packet.c    | 15 +++++++--------
 3 files changed, 12 insertions(+), 10 deletions(-)

On 03/19/2019 06:25 AM, Paolo Abeni wrote:

> +u16 __netdev_pick_tx(struct net_device *dev, struct sk_buff *skb,
> +		     struct net_device *sb_dev)
>  {
>  	struct sock *sk = skb->sk;
>  	int queue_index = sk_tx_queue_get(sk);
> @@ -3729,6 +3729,7 @@ static u16 __netdev_pick_tx(struct net_device *dev, struct sk_buff *skb,
>  
>  	return queue_index;
>  }
> +EXPORT_SYMBOL(__netdev_pick_tx);
>
I would rather rename this function, now it is going to be used from many drivers.

Maybe the internal core-networking should be renamed as core_netdev_pick_tx()
and the 'fallback' would be named netdev_pick_tx()

From: Paolo Abeni <pabeni@redhat.com>
Date: Tue, 19 Mar 2019 14:25:01 +0100

> Currently packet_pick_tx_queue() is the only caller of
> ndo_select_queue() using a fallback argument other than
> __netdev_pick_tx.
> 
> Leveraging rx queue, we can obtain a similar queue selection
> behavior using core helpers. After this change, ndo_select_queue()
> is always invoked with __netdev_pick_tx() as fallback.
> We can change ndo_select_queue() signature in a followup patch,
> dropping an indirect call per transmitted packet in some scenarios
> (e.g. TCP syn and XDP generic xmit)
> 
> This changes slightly how af packet queue selection happens when
> PACKET_QDISC_BYPASS is set. It's now more similar to plan dev_queue_xmit()
> tacking in account both XPS and TC mapping.
> 
>  RFC -> v1:
>   - initialize sender_cpu to the expected value
> 
> Signed-off-by: Paolo Abeni <pabeni@redhat.com>

To reply indirectly to Eric's feedback.  I agree with him that maybe these
should have different names.

But having a global symbol for networking be core_*() doesn't sit
right with me.

netdev_core_*() might be better.