From: Bruce Richardson <bruce.richardson@intel.com>
To: Wang Xiao W <xiao.w.wang@intel.com>
Cc: jing.d.chen@intel.com, dev@dpdk.org
Subject: Re: [dpdk-dev] [PATCH] net/fm10k: fix Rx descriptor read timing
Date: Tue, 5 Jul 2016 10:43:42 +0100 [thread overview]
Message-ID: <20160705094342.GD10232@bricha3-MOBL3> (raw)
In-Reply-To: <1467618668-28611-1-git-send-email-xiao.w.wang@intel.com>
On Mon, Jul 04, 2016 at 03:51:08PM +0800, Wang Xiao W wrote:
> We find that when traffic is light, a few amount of packets will be
> wrongly parsed (e.g. packet type), however this issue will not happen
> when traffic is heavy.
>
> The root cause is some fields in fm10k_rx_desc are read at wrong timing.
> When the input speed is slower than software's capability, fm10k scalar
> Rx function accesses descriptors closely to HW, so there's potential
> sequence: some fields like pkt_info in fm10k_rx_desc are read before HW
> writeback but some fields like DD bit are read after HW writeback, this
> will lead to the later packet parsing function using incorrect value.
>
> This patch fixes this issue by reading and parsing Rx descriptor after
> DD bit is set.
>
> Fixes: 4b61d3bfa941 ("fm10k: add receive and tranmit")
> Fixes: c82dd0a7bfa5 ("fm10k: add scatter receive")
>
> Signed-off-by: Wang Xiao W <xiao.w.wang@intel.com>
Fix looks correct to me. The key part I think is that the read of the
descriptor at line "desc = q->hw_ring[next_dd];" is not atomic as it's
a read of 16B, allowing part of the descriptor to be read either side of
a HW write-back.
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
> ---
> drivers/net/fm10k/fm10k_rxtx.c | 8 ++++----
> 1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/net/fm10k/fm10k_rxtx.c b/drivers/net/fm10k/fm10k_rxtx.c
> index dd92a91..5b2d04b 100644
> --- a/drivers/net/fm10k/fm10k_rxtx.c
> +++ b/drivers/net/fm10k/fm10k_rxtx.c
> @@ -114,10 +114,10 @@ fm10k_recv_pkts(void *rx_queue, struct rte_mbuf **rx_pkts,
>
> nb_pkts = RTE_MIN(nb_pkts, q->alloc_thresh);
> for (count = 0; count < nb_pkts; ++count) {
> + if (!(q->hw_ring[next_dd].d.staterr & FM10K_RXD_STATUS_DD))
> + break;
> mbuf = q->sw_ring[next_dd];
> desc = q->hw_ring[next_dd];
> - if (!(desc.d.staterr & FM10K_RXD_STATUS_DD))
> - break;
> #ifdef RTE_LIBRTE_FM10K_DEBUG_RX
> dump_rxd(&desc);
> #endif
> @@ -228,10 +228,10 @@ fm10k_recv_scattered_pkts(void *rx_queue, struct rte_mbuf **rx_pkts,
>
> nb_seg = RTE_MIN(nb_pkts, q->alloc_thresh);
> for (count = 0; count < nb_seg; count++) {
> + if (!(q->hw_ring[next_dd].d.staterr & FM10K_RXD_STATUS_DD))
> + break;
> mbuf = q->sw_ring[next_dd];
> desc = q->hw_ring[next_dd];
> - if (!(desc.d.staterr & FM10K_RXD_STATUS_DD))
> - break;
> #ifdef RTE_LIBRTE_FM10K_DEBUG_RX
> dump_rxd(&desc);
> #endif
> --
> 1.9.3
>
next prev parent reply other threads:[~2016-07-05 9:43 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-07-04 7:51 Wang Xiao W
2016-07-05 9:43 ` Bruce Richardson [this message]
2016-07-05 9:47 ` Bruce Richardson
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20160705094342.GD10232@bricha3-MOBL3 \
--to=bruce.richardson@intel.com \
--cc=dev@dpdk.org \
--cc=jing.d.chen@intel.com \
--cc=xiao.w.wang@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).