[dpdk-dev,v2,1/2] net/i40e: eliminate mbuf write on rearm

  With the mbuf rework, we now have 8 contiguous bytes to be rearmed in the
mbuf just before the 8-bytes of olflags. If we don't do the rearm write
inside the descriptor ring replenishment function, and delay it to
receiving the packet, we can do a single 16B write inside the RX function
to set both the rearm data, and the flags together.

Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
---
V2: fix a checkpatch warning. One warning remains, which is being left
    as-is as the code line in question is being removed by patch 2.
---
 drivers/net/i40e/i40e_rxtx_vec_sse.c | 46 +++++++++++++++++++++---------------
 1 file changed, 27 insertions(+), 19 deletions(-)
  

Hi Bruce:

> -----Original Message-----
> From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Bruce Richardson
> Sent: Thursday, April 6, 2017 7:32 PM
> To: Zhang, Helin <helin.zhang@intel.com>
> Cc: dev@dpdk.org; Yigit, Ferruh <ferruh.yigit@intel.com>; Richardson, Bruce
> <bruce.richardson@intel.com>
> Subject: [dpdk-dev] [PATCH v2 1/2] net/i40e: eliminate mbuf write on rearm
> 
> With the mbuf rework, we now have 8 contiguous bytes to be rearmed in the
> mbuf just before the 8-bytes of olflags. If we don't do the rearm write inside
> the descriptor ring replenishment function, and delay it to receiving the
> packet, we can do a single 16B write inside the RX function to set both the
> rearm data, and the flags together.
> 
> Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
> ---
> V2: fix a checkpatch warning. One warning remains, which is being left
>     as-is as the code line in question is being removed by patch 2.
> ---
>  drivers/net/i40e/i40e_rxtx_vec_sse.c | 46
> +++++++++++++++++++++---------------
>  1 file changed, 27 insertions(+), 19 deletions(-)
> 
> diff --git a/drivers/net/i40e/i40e_rxtx_vec_sse.c
> b/drivers/net/i40e/i40e_rxtx_vec_sse.c
> index fdd4a34..c43d1c3 100644
> --- a/drivers/net/i40e/i40e_rxtx_vec_sse.c
> +++ b/drivers/net/i40e/i40e_rxtx_vec_sse.c
> @@ -82,19 +82,10 @@ i40e_rxq_rearm(struct i40e_rx_queue *rxq)
>  	/* Initialize the mbufs in vector, process 2 mbufs in one loop */
>  	for (i = 0; i < RTE_I40E_RXQ_REARM_THRESH; i += 2, rxep += 2) {
>  		__m128i vaddr0, vaddr1;
> -		uintptr_t p0, p1;
> 
>  		mb0 = rxep[0].mbuf;
>  		mb1 = rxep[1].mbuf;
> 
> -		/* Flush mbuf with pkt template.
> -		 * Data to be rearmed is 6 bytes long.
> -		 */
> -		p0 = (uintptr_t)&mb0->rearm_data;
> -		*(uint64_t *)p0 = rxq->mbuf_initializer;
> -		p1 = (uintptr_t)&mb1->rearm_data;
> -		*(uint64_t *)p1 = rxq->mbuf_initializer;
> -
>  		/* load buf_addr(lo 64bit) and buf_physaddr(hi 64bit) */
>  		vaddr0 = _mm_loadu_si128((__m128i *)&mb0->buf_addr);
>  		vaddr1 = _mm_loadu_si128((__m128i *)&mb1->buf_addr); @@
> -125,6 +116,13 @@ i40e_rxq_rearm(struct i40e_rx_queue *rxq)
>  	I40E_PCI_REG_WRITE(rxq->qrx_tail, rx_id);  }
> 
> +static inline void
> +desc_to_olflags_v(struct i40e_rx_queue *rxq, __m128i descs[4]
> __rte_unused,
> +	struct rte_mbuf **rx_pkts)


Should we change the function name? since its scope is changed.
I'm Ok with all the other part.

Thanks
Qi
  

On Thu, Apr 06, 2017 at 05:24:44PM +0100, Zhang, Qi Z wrote:
> Hi Bruce:
> 
> > -----Original Message-----
> > From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Bruce Richardson
> > Sent: Thursday, April 6, 2017 7:32 PM
> > To: Zhang, Helin <helin.zhang@intel.com>
> > Cc: dev@dpdk.org; Yigit, Ferruh <ferruh.yigit@intel.com>; Richardson, Bruce
> > <bruce.richardson@intel.com>
> > Subject: [dpdk-dev] [PATCH v2 1/2] net/i40e: eliminate mbuf write on rearm
> > 
> > With the mbuf rework, we now have 8 contiguous bytes to be rearmed in the
> > mbuf just before the 8-bytes of olflags. If we don't do the rearm write inside
> > the descriptor ring replenishment function, and delay it to receiving the
> > packet, we can do a single 16B write inside the RX function to set both the
> > rearm data, and the flags together.
> > 
> > Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
> > ---
> > V2: fix a checkpatch warning. One warning remains, which is being left
> >     as-is as the code line in question is being removed by patch 2.
> > ---
> >  drivers/net/i40e/i40e_rxtx_vec_sse.c | 46
> > +++++++++++++++++++++---------------
> >  1 file changed, 27 insertions(+), 19 deletions(-)
> > 
> > diff --git a/drivers/net/i40e/i40e_rxtx_vec_sse.c
> > b/drivers/net/i40e/i40e_rxtx_vec_sse.c
> > index fdd4a34..c43d1c3 100644
> > --- a/drivers/net/i40e/i40e_rxtx_vec_sse.c
> > +++ b/drivers/net/i40e/i40e_rxtx_vec_sse.c
> > @@ -82,19 +82,10 @@ i40e_rxq_rearm(struct i40e_rx_queue *rxq)
> >  	/* Initialize the mbufs in vector, process 2 mbufs in one loop */
> >  	for (i = 0; i < RTE_I40E_RXQ_REARM_THRESH; i += 2, rxep += 2) {
> >  		__m128i vaddr0, vaddr1;
> > -		uintptr_t p0, p1;
> > 
> >  		mb0 = rxep[0].mbuf;
> >  		mb1 = rxep[1].mbuf;
> > 
> > -		/* Flush mbuf with pkt template.
> > -		 * Data to be rearmed is 6 bytes long.
> > -		 */
> > -		p0 = (uintptr_t)&mb0->rearm_data;
> > -		*(uint64_t *)p0 = rxq->mbuf_initializer;
> > -		p1 = (uintptr_t)&mb1->rearm_data;
> > -		*(uint64_t *)p1 = rxq->mbuf_initializer;
> > -
> >  		/* load buf_addr(lo 64bit) and buf_physaddr(hi 64bit) */
> >  		vaddr0 = _mm_loadu_si128((__m128i *)&mb0->buf_addr);
> >  		vaddr1 = _mm_loadu_si128((__m128i *)&mb1->buf_addr); @@
> > -125,6 +116,13 @@ i40e_rxq_rearm(struct i40e_rx_queue *rxq)
> >  	I40E_PCI_REG_WRITE(rxq->qrx_tail, rx_id);  }
> > 
> > +static inline void
> > +desc_to_olflags_v(struct i40e_rx_queue *rxq, __m128i descs[4]
> > __rte_unused,
> > +	struct rte_mbuf **rx_pkts)
> 
> 
> Should we change the function name? since its scope is changed.
> I'm Ok with all the other part.
> 
> Thanks
> Qi
>
Yes, it perhaps should, though it's not a big deal IMHO. Alternatively,
the function should be changed to return the flags values as an output
and we do the writes themselves in the main rx function.

/Bruce