From patchwork Wed Mar 25 12:52:43 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Igor Russkikh X-Patchwork-Id: 221878 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-9.8 required=3.0 tests=DKIM_SIGNED,DKIM_VALID, DKIM_VALID_AU, HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_PATCH, MAILING_LIST_MULTI, SIGNED_OFF_BY,SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED,USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 8F0F1C54FD3 for ; Wed, 25 Mar 2020 12:53:30 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 536452076A for ; Wed, 25 Mar 2020 12:53:30 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (2048-bit key) header.d=marvell.com header.i=@marvell.com header.b="lRx2Qmdg" Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1727578AbgCYMx3 (ORCPT ); Wed, 25 Mar 2020 08:53:29 -0400 Received: from mx0a-0016f401.pphosted.com ([67.231.148.174]:36892 "EHLO mx0b-0016f401.pphosted.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1727046AbgCYMx2 (ORCPT ); Wed, 25 Mar 2020 08:53:28 -0400 Received: from pps.filterd (m0045849.ppops.net [127.0.0.1]) by mx0a-0016f401.pphosted.com (8.16.0.42/8.16.0.42) with SMTP id 02PCoHTg009160; Wed, 25 Mar 2020 05:53:25 -0700 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=marvell.com; h=from : to : cc : subject : date : message-id : in-reply-to : references : mime-version : content-type; s=pfpt0818; bh=NXK837NxwBCL0CGQnUZmq+nR/VKuIuc49LtckQd6ryU=; b=lRx2QmdgEF/cJ1Eyfs140Cn4fANbhqdGWYMWZVoYvwMyFYQg245+gM5OSbq8ctz2U72a feacfxxPaXNRiel/26IrResFwNMJ+8Xsb1gnNZhVlXUgRGCZSraru5jgnmU8QG47Tdqd opf4QKKKmNhQupNRyUKV/IaENqIx2aFHbc4hB8jusva/CDfSwgfmTb75IWP6t2t6zsAn 4kjXOjs6nqbg/cwEQdxHRzSqCTCJ1PP6KWKMigKkuQAgA1T8NTTsLintfeuN9toQ/L6j k1bghD7PCx7/XVQDWW0AuNU+GGOrdLnIo1AWzJEgJxfVNecHmmK0ghBtpFJBoBw2/au2 5Q== Received: from sc-exch01.marvell.com ([199.233.58.181]) by mx0a-0016f401.pphosted.com with ESMTP id 2ywg9nrgje-1 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-SHA384 bits=256 verify=NOT); Wed, 25 Mar 2020 05:53:24 -0700 Received: from SC-EXCH01.marvell.com (10.93.176.81) by SC-EXCH01.marvell.com (10.93.176.81) with Microsoft SMTP Server (TLS) id 15.0.1497.2; Wed, 25 Mar 2020 05:53:22 -0700 Received: from maili.marvell.com (10.93.176.43) by SC-EXCH01.marvell.com (10.93.176.81) with Microsoft SMTP Server id 15.0.1497.2 via Frontend Transport; Wed, 25 Mar 2020 05:53:22 -0700 Received: from localhost.localdomain (unknown [10.9.16.55]) by maili.marvell.com (Postfix) with ESMTP id C4A9C3F7040; Wed, 25 Mar 2020 05:53:21 -0700 (PDT) From: Igor Russkikh To: CC: Mark Starovoytov , Sabrina Dubroca , Antoine Tenart , "Igor Russkikh" Subject: [PATCH v2 net-next 14/17] net: atlantic: MACSec ingress offload implementation Date: Wed, 25 Mar 2020 15:52:43 +0300 Message-ID: <20200325125246.987-15-irusskikh@marvell.com> X-Mailer: git-send-email 2.17.1 In-Reply-To: <20200325125246.987-1-irusskikh@marvell.com> References: <20200325125246.987-1-irusskikh@marvell.com> MIME-Version: 1.0 X-Proofpoint-Virus-Version: vendor=fsecure engine=2.50.10434:6.0.138, 18.0.645 definitions=2020-03-25_05:2020-03-24,2020-03-25 signatures=0 Sender: netdev-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: netdev@vger.kernel.org From: Mark Starovoytov This patch adds support for MACSec ingress HW offloading on Atlantic network cards. Signed-off-by: Mark Starovoytov Signed-off-by: Igor Russkikh --- .../ethernet/aquantia/atlantic/aq_macsec.c | 421 +++++++++++++++++- .../ethernet/aquantia/atlantic/aq_macsec.h | 5 + 2 files changed, 420 insertions(+), 6 deletions(-) diff --git a/drivers/net/ethernet/aquantia/atlantic/aq_macsec.c b/drivers/net/ethernet/aquantia/atlantic/aq_macsec.c index 6ca5361387e6..53e533f8a53d 100644 --- a/drivers/net/ethernet/aquantia/atlantic/aq_macsec.c +++ b/drivers/net/ethernet/aquantia/atlantic/aq_macsec.c @@ -25,6 +25,10 @@ static int aq_clear_txsc(struct aq_nic_s *nic, const int txsc_idx, enum aq_clear_type clear_type); static int aq_clear_txsa(struct aq_nic_s *nic, struct aq_macsec_txsc *aq_txsc, const int sa_num, enum aq_clear_type clear_type); +static int aq_clear_rxsc(struct aq_nic_s *nic, const int rxsc_idx, + enum aq_clear_type clear_type); +static int aq_clear_rxsa(struct aq_nic_s *nic, struct aq_macsec_rxsc *aq_rxsc, + const int sa_num, enum aq_clear_type clear_type); static int aq_clear_secy(struct aq_nic_s *nic, const struct macsec_secy *secy, enum aq_clear_type clear_type); static int aq_apply_macsec_cfg(struct aq_nic_s *nic); @@ -57,6 +61,22 @@ static int aq_get_txsc_idx_from_secy(struct aq_macsec_cfg *macsec_cfg, return -1; } +static int aq_get_rxsc_idx_from_rxsc(struct aq_macsec_cfg *macsec_cfg, + const struct macsec_rx_sc *rxsc) +{ + int i; + + if (unlikely(!rxsc)) + return -1; + + for (i = 0; i < AQ_MACSEC_MAX_SC; i++) { + if (macsec_cfg->aq_rxsc[i].sw_rxsc == rxsc) + return i; + } + + return -1; +} + static int aq_get_txsc_idx_from_sc_idx(const enum aq_macsec_sc_sa sc_sa, const int sc_idx) { @@ -506,34 +526,351 @@ static int aq_mdo_del_txsa(struct macsec_context *ctx) return ret; } +static int aq_rxsc_validate_frames(const enum macsec_validation_type validate) +{ + switch (validate) { + case MACSEC_VALIDATE_DISABLED: + return 2; + case MACSEC_VALIDATE_CHECK: + return 1; + case MACSEC_VALIDATE_STRICT: + return 0; + default: + WARN_ONCE(true, "Invalid validation type"); + } + + return 0; +} + +static int aq_set_rxsc(struct aq_nic_s *nic, const u32 rxsc_idx) +{ + const struct aq_macsec_rxsc *aq_rxsc = + &nic->macsec_cfg->aq_rxsc[rxsc_idx]; + struct aq_mss_ingress_preclass_record pre_class_record; + const struct macsec_rx_sc *rx_sc = aq_rxsc->sw_rxsc; + const struct macsec_secy *secy = aq_rxsc->sw_secy; + const u32 hw_sc_idx = aq_rxsc->hw_sc_idx; + struct aq_mss_ingress_sc_record sc_record; + struct aq_hw_s *hw = nic->aq_hw; + int ret = 0; + + memset(&pre_class_record, 0, sizeof(pre_class_record)); + put_unaligned_be64((__force u64)rx_sc->sci, pre_class_record.sci); + pre_class_record.sci_mask = 0xff; + /* match all MACSEC ethertype packets */ + pre_class_record.eth_type = ETH_P_MACSEC; + pre_class_record.eth_type_mask = 0x3; + + aq_ether_addr_to_mac(pre_class_record.mac_sa, (char *)&rx_sc->sci); + pre_class_record.sa_mask = 0x3f; + + pre_class_record.an_mask = nic->macsec_cfg->sc_sa; + pre_class_record.sc_idx = hw_sc_idx; + /* strip SecTAG & forward for decryption */ + pre_class_record.action = 0x0; + pre_class_record.valid = 1; + + ret = aq_mss_set_ingress_preclass_record(hw, &pre_class_record, + 2 * rxsc_idx + 1); + if (ret) + return ret; + + /* If SCI is absent, then match by SA alone */ + pre_class_record.sci_mask = 0; + pre_class_record.sci_from_table = 1; + + ret = aq_mss_set_ingress_preclass_record(hw, &pre_class_record, + 2 * rxsc_idx); + if (ret) + return ret; + + memset(&sc_record, 0, sizeof(sc_record)); + sc_record.validate_frames = + aq_rxsc_validate_frames(secy->validate_frames); + if (secy->replay_protect) { + sc_record.replay_protect = 1; + sc_record.anti_replay_window = secy->replay_window; + } + sc_record.valid = 1; + sc_record.fresh = 1; + + ret = aq_mss_set_ingress_sc_record(hw, &sc_record, hw_sc_idx); + if (ret) + return ret; + + return ret; +} + static int aq_mdo_add_rxsc(struct macsec_context *ctx) { - return -EOPNOTSUPP; + struct aq_nic_s *nic = netdev_priv(ctx->netdev); + struct aq_macsec_cfg *cfg = nic->macsec_cfg; + const u32 rxsc_idx_max = aq_sc_idx_max(cfg->sc_sa); + u32 rxsc_idx; + int ret = 0; + + if (hweight32(cfg->rxsc_idx_busy) >= rxsc_idx_max) + return -ENOSPC; + + rxsc_idx = ffz(cfg->rxsc_idx_busy); + if (rxsc_idx >= rxsc_idx_max) + return -ENOSPC; + + if (ctx->prepare) + return 0; + + cfg->aq_rxsc[rxsc_idx].hw_sc_idx = aq_to_hw_sc_idx(rxsc_idx, + cfg->sc_sa); + cfg->aq_rxsc[rxsc_idx].sw_secy = ctx->secy; + cfg->aq_rxsc[rxsc_idx].sw_rxsc = ctx->rx_sc; + + if (netif_carrier_ok(nic->ndev) && netif_running(ctx->secy->netdev)) + ret = aq_set_rxsc(nic, rxsc_idx); + + if (ret < 0) + return ret; + + set_bit(rxsc_idx, &cfg->rxsc_idx_busy); + + return 0; } static int aq_mdo_upd_rxsc(struct macsec_context *ctx) { - return -EOPNOTSUPP; + struct aq_nic_s *nic = netdev_priv(ctx->netdev); + int rxsc_idx; + int ret = 0; + + rxsc_idx = aq_get_rxsc_idx_from_rxsc(nic->macsec_cfg, ctx->rx_sc); + if (rxsc_idx < 0) + return -ENOENT; + + if (ctx->prepare) + return 0; + + if (netif_carrier_ok(nic->ndev) && netif_running(ctx->secy->netdev)) + ret = aq_set_rxsc(nic, rxsc_idx); + + return ret; +} + +static int aq_clear_rxsc(struct aq_nic_s *nic, const int rxsc_idx, + enum aq_clear_type clear_type) +{ + struct aq_macsec_rxsc *rx_sc = &nic->macsec_cfg->aq_rxsc[rxsc_idx]; + struct aq_hw_s *hw = nic->aq_hw; + int ret = 0; + int sa_num; + + for_each_set_bit (sa_num, &rx_sc->rx_sa_idx_busy, AQ_MACSEC_MAX_SA) { + ret = aq_clear_rxsa(nic, rx_sc, sa_num, clear_type); + if (ret) + return ret; + } + + if (clear_type & AQ_CLEAR_HW) { + struct aq_mss_ingress_preclass_record pre_class_record; + struct aq_mss_ingress_sc_record sc_record; + + memset(&pre_class_record, 0, sizeof(pre_class_record)); + memset(&sc_record, 0, sizeof(sc_record)); + + ret = aq_mss_set_ingress_preclass_record(hw, &pre_class_record, + 2 * rxsc_idx); + if (ret) + return ret; + + ret = aq_mss_set_ingress_preclass_record(hw, &pre_class_record, + 2 * rxsc_idx + 1); + if (ret) + return ret; + + sc_record.fresh = 1; + ret = aq_mss_set_ingress_sc_record(hw, &sc_record, + rx_sc->hw_sc_idx); + if (ret) + return ret; + } + + if (clear_type & AQ_CLEAR_SW) { + clear_bit(rxsc_idx, &nic->macsec_cfg->rxsc_idx_busy); + rx_sc->sw_secy = NULL; + rx_sc->sw_rxsc = NULL; + } + + return ret; } static int aq_mdo_del_rxsc(struct macsec_context *ctx) { - return -EOPNOTSUPP; + struct aq_nic_s *nic = netdev_priv(ctx->netdev); + enum aq_clear_type clear_type = AQ_CLEAR_SW; + int rxsc_idx; + int ret = 0; + + rxsc_idx = aq_get_rxsc_idx_from_rxsc(nic->macsec_cfg, ctx->rx_sc); + if (rxsc_idx < 0) + return -ENOENT; + + if (ctx->prepare) + return 0; + + if (netif_carrier_ok(nic->ndev)) + clear_type = AQ_CLEAR_ALL; + + ret = aq_clear_rxsc(nic, rxsc_idx, clear_type); + + return ret; +} + +static int aq_update_rxsa(struct aq_nic_s *nic, const unsigned int sc_idx, + const struct macsec_secy *secy, + const struct macsec_rx_sa *rx_sa, + const unsigned char *key, const unsigned char an) +{ + struct aq_mss_ingress_sakey_record sa_key_record; + struct aq_mss_ingress_sa_record sa_record; + struct aq_hw_s *hw = nic->aq_hw; + const int sa_idx = sc_idx | an; + int ret = 0; + + memset(&sa_record, 0, sizeof(sa_record)); + sa_record.valid = rx_sa->active; + sa_record.fresh = 1; + sa_record.next_pn = rx_sa->next_pn; + + ret = aq_mss_set_ingress_sa_record(hw, &sa_record, sa_idx); + if (ret) + return ret; + + if (!key) + return ret; + + memset(&sa_key_record, 0, sizeof(sa_key_record)); + memcpy(&sa_key_record.key, key, secy->key_len); + + switch (secy->key_len) { + case AQ_MACSEC_KEY_LEN_128_BIT: + sa_key_record.key_len = 0; + break; + case AQ_MACSEC_KEY_LEN_192_BIT: + sa_key_record.key_len = 1; + break; + case AQ_MACSEC_KEY_LEN_256_BIT: + sa_key_record.key_len = 2; + break; + default: + return -1; + } + + aq_rotate_keys(&sa_key_record.key, secy->key_len); + + ret = aq_mss_set_ingress_sakey_record(hw, &sa_key_record, sa_idx); + + return ret; } static int aq_mdo_add_rxsa(struct macsec_context *ctx) { - return -EOPNOTSUPP; + const struct macsec_rx_sc *rx_sc = ctx->sa.rx_sa->sc; + struct aq_nic_s *nic = netdev_priv(ctx->netdev); + const struct macsec_secy *secy = ctx->secy; + struct aq_macsec_rxsc *aq_rxsc; + int rxsc_idx; + int ret = 0; + + rxsc_idx = aq_get_rxsc_idx_from_rxsc(nic->macsec_cfg, rx_sc); + if (rxsc_idx < 0) + return -EINVAL; + + if (ctx->prepare) + return 0; + + aq_rxsc = &nic->macsec_cfg->aq_rxsc[rxsc_idx]; + set_bit(ctx->sa.assoc_num, &aq_rxsc->rx_sa_idx_busy); + + memcpy(aq_rxsc->rx_sa_key[ctx->sa.assoc_num], ctx->sa.key, + secy->key_len); + + if (netif_carrier_ok(nic->ndev) && netif_running(secy->netdev)) + ret = aq_update_rxsa(nic, aq_rxsc->hw_sc_idx, secy, + ctx->sa.rx_sa, ctx->sa.key, + ctx->sa.assoc_num); + + return ret; } static int aq_mdo_upd_rxsa(struct macsec_context *ctx) { - return -EOPNOTSUPP; + const struct macsec_rx_sc *rx_sc = ctx->sa.rx_sa->sc; + struct aq_nic_s *nic = netdev_priv(ctx->netdev); + struct aq_macsec_cfg *cfg = nic->macsec_cfg; + const struct macsec_secy *secy = ctx->secy; + int rxsc_idx; + int ret = 0; + + rxsc_idx = aq_get_rxsc_idx_from_rxsc(cfg, rx_sc); + if (rxsc_idx < 0) + return -EINVAL; + + if (ctx->prepare) + return 0; + + if (netif_carrier_ok(nic->ndev) && netif_running(secy->netdev)) + ret = aq_update_rxsa(nic, cfg->aq_rxsc[rxsc_idx].hw_sc_idx, + secy, ctx->sa.rx_sa, NULL, + ctx->sa.assoc_num); + + return ret; +} + +static int aq_clear_rxsa(struct aq_nic_s *nic, struct aq_macsec_rxsc *aq_rxsc, + const int sa_num, enum aq_clear_type clear_type) +{ + int sa_idx = aq_rxsc->hw_sc_idx | sa_num; + struct aq_hw_s *hw = nic->aq_hw; + int ret = 0; + + if (clear_type & AQ_CLEAR_SW) + clear_bit(sa_num, &aq_rxsc->rx_sa_idx_busy); + + if ((clear_type & AQ_CLEAR_HW) && netif_carrier_ok(nic->ndev)) { + struct aq_mss_ingress_sakey_record sa_key_record; + struct aq_mss_ingress_sa_record sa_record; + + memset(&sa_key_record, 0, sizeof(sa_key_record)); + memset(&sa_record, 0, sizeof(sa_record)); + sa_record.fresh = 1; + ret = aq_mss_set_ingress_sa_record(hw, &sa_record, sa_idx); + if (ret) + return ret; + + return aq_mss_set_ingress_sakey_record(hw, &sa_key_record, + sa_idx); + } + + return ret; } static int aq_mdo_del_rxsa(struct macsec_context *ctx) { - return -EOPNOTSUPP; + const struct macsec_rx_sc *rx_sc = ctx->sa.rx_sa->sc; + struct aq_nic_s *nic = netdev_priv(ctx->netdev); + struct aq_macsec_cfg *cfg = nic->macsec_cfg; + int rxsc_idx; + int ret = 0; + + rxsc_idx = aq_get_rxsc_idx_from_rxsc(cfg, rx_sc); + if (rxsc_idx < 0) + return -EINVAL; + + if (ctx->prepare) + return 0; + + ret = aq_clear_rxsa(nic, &cfg->aq_rxsc[rxsc_idx], ctx->sa.assoc_num, + AQ_CLEAR_ALL); + + return ret; } static int apply_txsc_cfg(struct aq_nic_s *nic, const int txsc_idx) @@ -564,10 +901,40 @@ static int apply_txsc_cfg(struct aq_nic_s *nic, const int txsc_idx) return ret; } +static int apply_rxsc_cfg(struct aq_nic_s *nic, const int rxsc_idx) +{ + struct aq_macsec_rxsc *aq_rxsc = &nic->macsec_cfg->aq_rxsc[rxsc_idx]; + const struct macsec_secy *secy = aq_rxsc->sw_secy; + struct macsec_rx_sa *rx_sa; + int ret = 0; + int i; + + if (!netif_running(secy->netdev)) + return ret; + + ret = aq_set_rxsc(nic, rxsc_idx); + if (ret) + return ret; + + for (i = 0; i < MACSEC_NUM_AN; i++) { + rx_sa = rcu_dereference_bh(aq_rxsc->sw_rxsc->sa[i]); + if (rx_sa) { + ret = aq_update_rxsa(nic, aq_rxsc->hw_sc_idx, secy, + rx_sa, aq_rxsc->rx_sa_key[i], i); + if (ret) + return ret; + } + } + + return ret; +} + static int aq_clear_secy(struct aq_nic_s *nic, const struct macsec_secy *secy, enum aq_clear_type clear_type) { + struct macsec_rx_sc *rx_sc; int txsc_idx; + int rxsc_idx; int ret = 0; txsc_idx = aq_get_txsc_idx_from_secy(nic->macsec_cfg, secy); @@ -577,19 +944,43 @@ static int aq_clear_secy(struct aq_nic_s *nic, const struct macsec_secy *secy, return ret; } + for (rx_sc = rcu_dereference_bh(secy->rx_sc); rx_sc; + rx_sc = rcu_dereference_bh(rx_sc->next)) { + rxsc_idx = aq_get_rxsc_idx_from_rxsc(nic->macsec_cfg, rx_sc); + if (rxsc_idx < 0) + continue; + + ret = aq_clear_rxsc(nic, rxsc_idx, clear_type); + if (ret) + return ret; + } + return ret; } static int aq_apply_secy_cfg(struct aq_nic_s *nic, const struct macsec_secy *secy) { + struct macsec_rx_sc *rx_sc; int txsc_idx; + int rxsc_idx; int ret = 0; txsc_idx = aq_get_txsc_idx_from_secy(nic->macsec_cfg, secy); if (txsc_idx >= 0) apply_txsc_cfg(nic, txsc_idx); + for (rx_sc = rcu_dereference_bh(secy->rx_sc); rx_sc && rx_sc->active; + rx_sc = rcu_dereference_bh(rx_sc->next)) { + rxsc_idx = aq_get_rxsc_idx_from_rxsc(nic->macsec_cfg, rx_sc); + if (unlikely(rxsc_idx < 0)) + continue; + + ret = apply_rxsc_cfg(nic, rxsc_idx); + if (ret) + return ret; + } + return ret; } @@ -606,6 +997,14 @@ static int aq_apply_macsec_cfg(struct aq_nic_s *nic) } } + for (i = 0; i < AQ_MACSEC_MAX_SC; i++) { + if (nic->macsec_cfg->rxsc_idx_busy & BIT(i)) { + ret = apply_rxsc_cfg(nic, i); + if (ret) + return ret; + } + } + return ret; } @@ -781,6 +1180,7 @@ int aq_macsec_enable(struct aq_nic_s *nic) /* Init Ethertype bypass filters */ for (index = 0; index < ARRAY_SIZE(ctl_ether_types); index++) { + struct aq_mss_ingress_prectlf_record rx_prectlf_rec; struct aq_mss_egress_ctlf_record tx_ctlf_rec; if (ctl_ether_types[index] == 0) @@ -794,6 +1194,15 @@ int aq_macsec_enable(struct aq_nic_s *nic) tbl_idx = NUMROWS_EGRESSCTLFRECORD - num_ctl_ether_types - 1; aq_mss_set_egress_ctlf_record(hw, &tx_ctlf_rec, tbl_idx); + memset(&rx_prectlf_rec, 0, sizeof(rx_prectlf_rec)); + rx_prectlf_rec.eth_type = ctl_ether_types[index]; + rx_prectlf_rec.match_type = 4; /* Match eth_type only */ + rx_prectlf_rec.match_mask = 0xf; /* match for eth_type */ + rx_prectlf_rec.action = 0; /* Bypass MACSEC modules */ + tbl_idx = + NUMROWS_INGRESSPRECTLFRECORD - num_ctl_ether_types - 1; + aq_mss_set_ingress_prectlf_record(hw, &rx_prectlf_rec, tbl_idx); + num_ctl_ether_types++; } diff --git a/drivers/net/ethernet/aquantia/atlantic/aq_macsec.h b/drivers/net/ethernet/aquantia/atlantic/aq_macsec.h index 5ab0ee4bea73..b8485c1cb667 100644 --- a/drivers/net/ethernet/aquantia/atlantic/aq_macsec.h +++ b/drivers/net/ethernet/aquantia/atlantic/aq_macsec.h @@ -31,6 +31,11 @@ struct aq_macsec_txsc { }; struct aq_macsec_rxsc { + u32 hw_sc_idx; + unsigned long rx_sa_idx_busy; + const struct macsec_secy *sw_secy; + const struct macsec_rx_sc *sw_rxsc; + u8 rx_sa_key[MACSEC_NUM_AN][MACSEC_KEYID_LEN]; }; struct aq_macsec_cfg {