Re: [PATCH 1/1] net/ixgbe: add a proper memory barrier for LoongArch

[zhoumin <zhoumin@loongson.cn>]

On Fri, Apr 7, 2023 at 4:50PM, Min Zhou wrote:
> Segmentation fault has been observed while running the
> ixgbe_recv_pkts_lro() function to receive packets on the Loongson
> 3C5000 processor which has 64 cores and 4 NUMA nodes.
>
> Reason is the read ordering of the status and the rest of the
> descriptor fields in this function may not be correct on the
> LoongArch processor. We should add rte_rmb() to ensure the read
> ordering be correct.
>
> We also did the same thing in the ixgbe_recv_pkts() function.
>
> Signed-off-by: Min Zhou <zhoumin@loongson.cn>
> ---
>   drivers/net/ixgbe/ixgbe_rxtx.c | 6 ++++++
>   1 file changed, 6 insertions(+)
>
> diff --git a/drivers/net/ixgbe/ixgbe_rxtx.c b/drivers/net/ixgbe/ixgbe_rxtx.c
> index c9d6ca9efe..16391a42f9 100644
> --- a/drivers/net/ixgbe/ixgbe_rxtx.c
> +++ b/drivers/net/ixgbe/ixgbe_rxtx.c
> @@ -1823,6 +1823,9 @@ ixgbe_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts,
>   		staterr = rxdp->wb.upper.status_error;
>   		if (!(staterr & rte_cpu_to_le_32(IXGBE_RXDADV_STAT_DD)))
>   			break;
> +#if defined(RTE_ARCH_LOONGARCH)
> +		rte_rmb();
> +#endif
>   		rxd = *rxdp;
>   
>   		/*
> @@ -2122,6 +2125,9 @@ ixgbe_recv_pkts_lro(void *rx_queue, struct rte_mbuf **rx_pkts, uint16_t nb_pkts,
>   		if (!(staterr & IXGBE_RXDADV_STAT_DD))
>   			break;
>   
> +#if defined(RTE_ARCH_LOONGARCH)
> +		rte_rmb();
> +#endif
>   		rxd = *rxdp;
>   
>   		PMD_RX_LOG(DEBUG, "port_id=%u queue_id=%u rx_id=%u "

Kindly ping.

Any comments or suggestions will be appreciated.


Min