From: "Wang, Haiyue" <haiyue.wang@intel.com>
To: "Yu, DapengX" <dapengx.yu@intel.com>,
"Yang, Qiming" <qiming.yang@intel.com>,
"Zhang, Qi Z" <qi.z.zhang@intel.com>
Cc: "dev@dpdk.org" <dev@dpdk.org>, "stable@dpdk.org" <stable@dpdk.org>
Subject: Re: [dpdk-stable] [PATCH v2] net/ice: fix function pointer in multi-process
Date: Fri, 22 Oct 2021 11:08:15 +0000 [thread overview]
Message-ID: <DM4PR11MB55182FD67DA1C984363FA045F7809@DM4PR11MB5518.namprd11.prod.outlook.com> (raw)
In-Reply-To: <20211022085734.712382-1-dapengx.yu@intel.com>
> -----Original Message-----
> From: Yu, DapengX <dapengx.yu@intel.com>
> Sent: Friday, October 22, 2021 16:58
> To: Yang, Qiming <qiming.yang@intel.com>; Zhang, Qi Z <qi.z.zhang@intel.com>
> Cc: dev@dpdk.org; Wang, Haiyue <haiyue.wang@intel.com>; Yu, DapengX <dapengx.yu@intel.com>;
> stable@dpdk.org
> Subject: [PATCH v2] net/ice: fix function pointer in multi-process
>
> From: Dapeng Yu <dapengx.yu@intel.com>
>
> The sharing of function pointer may cause crash of secondary process.
>
> This patch removes the shared function pointer: "rxd_to_pkt_fields" in
> the instance of "struct ice_rx_queue" which is shared between primary
> and secondary process, and uses an index of function pointer array to
> replace it.
I think we can simplify it something as:
Use the index value to call the function, instead of the function pointer
assignment to save the selection of Receive Flex Descriptor profile ID.
Otherwise the secondary process will run with wrong function address from
primary process.
>
> Fixes: 7a340b0b4e03 ("net/ice: refactor Rx FlexiMD handling")
> Cc: stable@dpdk.org
>
> Signed-off-by: Dapeng Yu <dapengx.yu@intel.com>
> ---
> V2:
> * Remove redundant code
> ---
> drivers/net/ice/ice_rxtx.c | 35 +++++++++++++++++------------------
> drivers/net/ice/ice_rxtx.h | 2 +-
> 2 files changed, 18 insertions(+), 19 deletions(-)
>
> diff --git a/drivers/net/ice/ice_rxtx.c b/drivers/net/ice/ice_rxtx.c
> index ff362c21d9..667eae9f6d 100644
> --- a/drivers/net/ice/ice_rxtx.c
> +++ b/drivers/net/ice/ice_rxtx.c
> @@ -205,51 +205,50 @@ ice_rxd_to_pkt_fields_by_comms_aux_v2(struct ice_rx_queue *rxq,
> #endif
> }
>
> +static const ice_rxd_to_pkt_fields_t rxd_to_pkt_fields_ops[] = {
> + [ICE_RXDID_COMMS_AUX_VLAN] = ice_rxd_to_pkt_fields_by_comms_aux_v1,
> + [ICE_RXDID_COMMS_AUX_IPV4] = ice_rxd_to_pkt_fields_by_comms_aux_v1,
> + [ICE_RXDID_COMMS_AUX_IPV6] = ice_rxd_to_pkt_fields_by_comms_aux_v1,
> + [ICE_RXDID_COMMS_AUX_IPV6_FLOW] = ice_rxd_to_pkt_fields_by_comms_aux_v1,
> + [ICE_RXDID_COMMS_AUX_TCP] = ice_rxd_to_pkt_fields_by_comms_aux_v1,
> + [ICE_RXDID_COMMS_AUX_IP_OFFSET] = ice_rxd_to_pkt_fields_by_comms_aux_v2,
> + [ICE_RXDID_COMMS_GENERIC] = ice_rxd_to_pkt_fields_by_comms_generic,
> + [ICE_RXDID_COMMS_OVS] = ice_rxd_to_pkt_fields_by_comms_ovs,
> +};
> +
> void
> ice_select_rxd_to_pkt_fields_handler(struct ice_rx_queue *rxq, uint32_t rxdid)
> {
> + rxq->rxdid = rxdid;
> +
> switch (rxdid) {
> case ICE_RXDID_COMMS_AUX_VLAN:
> rxq->xtr_ol_flag = rte_net_ice_dynflag_proto_xtr_vlan_mask;
> - rxq->rxd_to_pkt_fields = ice_rxd_to_pkt_fields_by_comms_aux_v1;
> break;
>
> case ICE_RXDID_COMMS_AUX_IPV4:
> rxq->xtr_ol_flag = rte_net_ice_dynflag_proto_xtr_ipv4_mask;
> - rxq->rxd_to_pkt_fields = ice_rxd_to_pkt_fields_by_comms_aux_v1;
> break;
>
> case ICE_RXDID_COMMS_AUX_IPV6:
> rxq->xtr_ol_flag = rte_net_ice_dynflag_proto_xtr_ipv6_mask;
> - rxq->rxd_to_pkt_fields = ice_rxd_to_pkt_fields_by_comms_aux_v1;
> break;
>
> case ICE_RXDID_COMMS_AUX_IPV6_FLOW:
> rxq->xtr_ol_flag = rte_net_ice_dynflag_proto_xtr_ipv6_flow_mask;
> - rxq->rxd_to_pkt_fields = ice_rxd_to_pkt_fields_by_comms_aux_v1;
> break;
>
> case ICE_RXDID_COMMS_AUX_TCP:
> rxq->xtr_ol_flag = rte_net_ice_dynflag_proto_xtr_tcp_mask;
> - rxq->rxd_to_pkt_fields = ice_rxd_to_pkt_fields_by_comms_aux_v1;
> break;
>
> case ICE_RXDID_COMMS_AUX_IP_OFFSET:
> rxq->xtr_ol_flag = rte_net_ice_dynflag_proto_xtr_ip_offset_mask;
> - rxq->rxd_to_pkt_fields = ice_rxd_to_pkt_fields_by_comms_aux_v2;
> - break;
> -
> - case ICE_RXDID_COMMS_GENERIC:
> - rxq->rxd_to_pkt_fields = ice_rxd_to_pkt_fields_by_comms_generic;
> - break;
> -
> - case ICE_RXDID_COMMS_OVS:
> - rxq->rxd_to_pkt_fields = ice_rxd_to_pkt_fields_by_comms_ovs;
> break;
My fault, we need to keep the above case explicitly, otherwise it will be
set to default ICE_RXDID_COMMS_OVS. ;-)
case ICE_RXDID_COMMS_GENERIC:
/* fallthrough */
case ICE_RXDID_COMMS_OVS:
break;
>
> default:
> /* update this according to the RXDID for PROTO_XTR_NONE */
> - rxq->rxd_to_pkt_fields = ice_rxd_to_pkt_fields_by_comms_ovs;
> + rxq->rxdid = ICE_RXDID_COMMS_OVS;
> break;
> }
>
> @@ -1622,7 +1621,7 @@ ice_rx_scan_hw_ring(struct ice_rx_queue *rxq)
> mb->packet_type = ptype_tbl[ICE_RX_FLEX_DESC_PTYPE_M &
> rte_le_to_cpu_16(rxdp[j].wb.ptype_flex_flags0)];
> ice_rxd_to_vlan_tci(mb, &rxdp[j]);
> - rxq->rxd_to_pkt_fields(rxq, mb, &rxdp[j]);
> + rxd_to_pkt_fields_ops[rxq->rxdid](rxq, mb, &rxdp[j]);
> #ifndef RTE_LIBRTE_ICE_16BYTE_RX_DESC
> if (rxq->offloads & DEV_RX_OFFLOAD_TIMESTAMP) {
> ts_ns = ice_tstamp_convert_32b_64b(hw,
> @@ -1939,7 +1938,7 @@ ice_recv_scattered_pkts(void *rx_queue,
> first_seg->packet_type = ptype_tbl[ICE_RX_FLEX_DESC_PTYPE_M &
> rte_le_to_cpu_16(rxd.wb.ptype_flex_flags0)];
> ice_rxd_to_vlan_tci(first_seg, &rxd);
> - rxq->rxd_to_pkt_fields(rxq, first_seg, &rxd);
> + rxd_to_pkt_fields_ops[rxq->rxdid](rxq, first_seg, &rxd);
> pkt_flags = ice_rxd_error_to_pkt_flags(rx_stat_err0);
> #ifndef RTE_LIBRTE_ICE_16BYTE_RX_DESC
> if (rxq->offloads & DEV_RX_OFFLOAD_TIMESTAMP) {
> @@ -2370,7 +2369,7 @@ ice_recv_pkts(void *rx_queue,
> rxm->packet_type = ptype_tbl[ICE_RX_FLEX_DESC_PTYPE_M &
> rte_le_to_cpu_16(rxd.wb.ptype_flex_flags0)];
> ice_rxd_to_vlan_tci(rxm, &rxd);
> - rxq->rxd_to_pkt_fields(rxq, rxm, &rxd);
> + rxd_to_pkt_fields_ops[rxq->rxdid](rxq, rxm, &rxd);
> pkt_flags = ice_rxd_error_to_pkt_flags(rx_stat_err0);
> #ifndef RTE_LIBRTE_ICE_16BYTE_RX_DESC
> if (rxq->offloads & DEV_RX_OFFLOAD_TIMESTAMP) {
> diff --git a/drivers/net/ice/ice_rxtx.h b/drivers/net/ice/ice_rxtx.h
> index e1c644fb63..146dc1f95d 100644
> --- a/drivers/net/ice/ice_rxtx.h
> +++ b/drivers/net/ice/ice_rxtx.h
> @@ -89,7 +89,7 @@ struct ice_rx_queue {
> bool rx_deferred_start; /* don't start this queue in dev start */
> uint8_t proto_xtr; /* Protocol extraction from flexible descriptor */
> uint64_t xtr_ol_flag; /* Protocol extraction offload flag */
> - ice_rxd_to_pkt_fields_t rxd_to_pkt_fields; /* handle FlexiMD by RXDID */
> + uint32_t rxdid; /* Receive Flex Descriptor profile ID */
> ice_rx_release_mbufs_t rx_rel_mbufs;
> uint64_t offloads;
> uint32_t time_high;
> --
> 2.27.0
next prev parent reply other threads:[~2021-10-22 11:08 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-10-21 3:35 [dpdk-stable] [PATCH] " dapengx.yu
2021-10-22 7:55 ` Wang, Haiyue
2021-10-22 8:57 ` [dpdk-stable] [PATCH v2] " dapengx.yu
2021-10-22 11:08 ` Wang, Haiyue [this message]
2021-10-25 3:40 ` [dpdk-stable] [PATCH v3] " dapengx.yu
2021-10-25 5:40 ` [dpdk-stable] [PATCH v4] " dapengx.yu
2021-10-25 5:47 ` Wang, Haiyue
2021-10-26 1:55 ` [dpdk-stable] [PATCH v5] " dapengx.yu
2021-10-27 3:28 ` Zhang, Qi Z
2022-02-24 18:00 ` [dpdk-dev] " Navin Srinivas
2022-02-25 1:34 ` Zhang, Qi Z
2022-02-25 10:24 ` Navin Srinivas
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=DM4PR11MB55182FD67DA1C984363FA045F7809@DM4PR11MB5518.namprd11.prod.outlook.com \
--to=haiyue.wang@intel.com \
--cc=dapengx.yu@intel.com \
--cc=dev@dpdk.org \
--cc=qi.z.zhang@intel.com \
--cc=qiming.yang@intel.com \
--cc=stable@dpdk.org \
/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).