On Wed, Sep 16, 2020 at 4:21 AM Juraj Linkeš <juraj.lin...@pantheon.tech> wrote:
>
> From: Ruifeng Wang <ruifeng.w...@arm.com>
>
> Expand vector PMD support to aarch32.
>
> Signed-off-by: Ruifeng Wang <ruifeng.w...@arm.com>
> Acked-by: Ajit Khaparde <ajit.khapa...@broadcom.com>
> ---
> drivers/net/bnxt/bnxt_rxq.h | 2 +-
> drivers/net/bnxt/bnxt_rxr.h | 2 +-
> drivers/net/bnxt/bnxt_txr.h | 2 +-
> 3 files changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/net/bnxt/bnxt_rxq.h b/drivers/net/bnxt/bnxt_rxq.h
> index d5ce3b6d5..1c4027711 100644
> --- a/drivers/net/bnxt/bnxt_rxq.h
> +++ b/drivers/net/bnxt/bnxt_rxq.h
> @@ -22,7 +22,7 @@ struct bnxt_rx_queue {
> uint16_t nb_rx_hold; /* num held free RX desc */
> uint16_t rx_free_thresh; /* max free RX desc to hold */
> uint16_t queue_id; /* RX queue index */
> -#if defined(RTE_ARCH_X86) || defined(RTE_ARCH_ARM64)
> +#if defined(RTE_ARCH_X86) || defined(RTE_ARCH_ARM64) || defined(RTE_ARCH_ARM)
> uint16_t rxrearm_nb; /* number of descs to reinit. */
> uint16_t rxrearm_start; /* next desc index to reinit.
> */
> #endif
> diff --git a/drivers/net/bnxt/bnxt_rxr.h b/drivers/net/bnxt/bnxt_rxr.h
> index 2bf46cd91..e2fba1647 100644
> --- a/drivers/net/bnxt/bnxt_rxr.h
> +++ b/drivers/net/bnxt/bnxt_rxr.h
> @@ -221,7 +221,7 @@ int bnxt_init_one_rx_ring(struct bnxt_rx_queue *rxq);
> int bnxt_rx_queue_start(struct rte_eth_dev *dev, uint16_t rx_queue_id);
> int bnxt_rx_queue_stop(struct rte_eth_dev *dev, uint16_t rx_queue_id);
>
> -#if defined(RTE_ARCH_X86) || defined(RTE_ARCH_ARM64)
> +#if defined(RTE_ARCH_X86) || defined(RTE_ARCH_ARM64) || defined(RTE_ARCH_ARM)
> uint16_t bnxt_recv_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
> uint16_t nb_pkts);
> int bnxt_rxq_vec_setup(struct bnxt_rx_queue *rxq);
> diff --git a/drivers/net/bnxt/bnxt_txr.h b/drivers/net/bnxt/bnxt_txr.h
> index 7715c11b8..38e5ac9df 100644
> --- a/drivers/net/bnxt/bnxt_txr.h
> +++ b/drivers/net/bnxt/bnxt_txr.h
> @@ -59,7 +59,7 @@ uint16_t bnxt_xmit_pkts(void *tx_queue, struct rte_mbuf
> **tx_pkts,
> uint16_t nb_pkts);
> uint16_t bnxt_dummy_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
> uint16_t nb_pkts);
> -#if defined(RTE_ARCH_X86) || defined(RTE_ARCH_ARM64)
> +#if defined(RTE_ARCH_X86) || defined(RTE_ARCH_ARM64) || defined(RTE_ARCH_ARM)
> uint16_t bnxt_xmit_pkts_vec(void *tx_queue, struct rte_mbuf **tx_pkts,
> uint16_t nb_pkts);
> #endif
> --
> 2.20.1
>
Hi Juraj,
I might be missing something, but I don't
believe these changes are sufficient to
enable vector mode for ARM32. For one
thing, bnxt_receive_function() in bnxt_ethdev.c
would need to be changed to enable the
selection of the vector receive function.
Also, meson.build has this condition
for building bnxt_rxtx_vec_neon.c:
if arch_subdir == 'x86'
sources += files('bnxt_rxtx_vec_sse.c')
elif arch_subdir == 'arm' and host_machine.cpu_family().startswith('aarch64')
sources += files('bnxt_rxtx_vec_neon.c')
endif
Were you able to build with these changes
and confirm that the vector mode functions
are selected and used? (There is a log at INFO
level to indicate which function is selected,
as should "show rxq info ..." from the
testpmd CLI.
Thanks,
Lance
