[dpdk-dev] [PATCH v2 13/17] net/virtio: add checks for max SIMD bitwidth

Xia, Chenbo chenbo.xia at intel.com
Mon Aug 31 04:39:00 CEST 2020


Hi Ciara,

Sorry for late response and thanks for working on this! Currently virtio
driver puts all vector-related conditions in dev_configure (virtio_dev_configure).
Do you think it's ok to put all below code logic to dev_configure?

Thanks!
Chenbo

> -----Original Message-----
> From: Power, Ciara <ciara.power at intel.com>
> Sent: Friday, August 28, 2020 12:13 AM
> To: dev at dpdk.org
> Cc: Power, Ciara <ciara.power at intel.com>; Maxime Coquelin
> <maxime.coquelin at redhat.com>; Xia, Chenbo <chenbo.xia at intel.com>; Wang,
> Zhihong <zhihong.wang at intel.com>
> Subject: [PATCH v2 13/17] net/virtio: add checks for max SIMD bitwidth
> 
> When choosing a vector path to take, an extra condition must be
> satisfied to ensure the max SIMD bitwidth allows for the CPU enabled
> path.
> 
> Cc: Maxime Coquelin <maxime.coquelin at redhat.com>
> Cc: Chenbo Xia <chenbo.xia at intel.com>
> Cc: Zhihong Wang <zhihong.wang at intel.com>
> 
> Signed-off-by: Ciara Power <ciara.power at intel.com>
> ---
>  drivers/net/virtio/virtio_ethdev.c | 12 ++++++++----
>  1 file changed, 8 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/net/virtio/virtio_ethdev.c
> b/drivers/net/virtio/virtio_ethdev.c
> index dc0093bdf0..f779ce8396 100644
> --- a/drivers/net/virtio/virtio_ethdev.c
> +++ b/drivers/net/virtio/virtio_ethdev.c
> @@ -1517,9 +1517,11 @@ set_rxtx_funcs(struct rte_eth_dev *eth_dev)
>  	if (vtpci_packed_queue(hw)) {
>  		PMD_INIT_LOG(INFO,
>  			"virtio: using packed ring %s Tx path on port %u",
> -			hw->use_vec_tx ? "vectorized" : "standard",
> +			(hw->use_vec_tx && rte_get_max_simd_bitwidth()
> +			> RTE_MAX_256_SIMD) ? "vectorized" : "standard",
>  			eth_dev->data->port_id);
> -		if (hw->use_vec_tx)
> +		if (hw->use_vec_tx && rte_get_max_simd_bitwidth()
> +				> RTE_MAX_256_SIMD)
>  			eth_dev->tx_pkt_burst = virtio_xmit_pkts_packed_vec;
>  		else
>  			eth_dev->tx_pkt_burst = virtio_xmit_pkts_packed;
> @@ -1536,7 +1538,8 @@ set_rxtx_funcs(struct rte_eth_dev *eth_dev)
>  	}
> 
>  	if (vtpci_packed_queue(hw)) {
> -		if (hw->use_vec_rx) {
> +		if (hw->use_vec_rx && rte_get_max_simd_bitwidth()
> +				> RTE_MAX_256_SIMD) {
>  			PMD_INIT_LOG(INFO,
>  				"virtio: using packed ring vectorized Rx path on
> port %u",
>  				eth_dev->data->port_id);
> @@ -1555,7 +1558,8 @@ set_rxtx_funcs(struct rte_eth_dev *eth_dev)
>  			eth_dev->rx_pkt_burst = &virtio_recv_pkts_packed;
>  		}
>  	} else {
> -		if (hw->use_vec_rx) {
> +		if (hw->use_vec_rx && rte_get_max_simd_bitwidth()
> +				>= RTE_MAX_128_SIMD) {
>  			PMD_INIT_LOG(INFO, "virtio: using vectorized Rx path on
> port %u",
>  				eth_dev->data->port_id);
>  			eth_dev->rx_pkt_burst = virtio_recv_pkts_vec;
> --
> 2.17.1



More information about the dev mailing list