diff --git a/drivers/net/ethernet/mellanox/mlx5/core/accel/ipsec.c b/drivers/net/ethernet/mellanox/mlx5/core/accel/ipsec.c index 53e69edaedde..9f1b1939716a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/accel/ipsec.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/accel/ipsec.c @@ -37,24 +37,11 @@ #include "mlx5_core.h" #include "fpga/ipsec.h" -void *mlx5_accel_ipsec_sa_cmd_exec(struct mlx5_core_dev *mdev, - struct mlx5_accel_ipsec_sa *cmd) -{ - if (!MLX5_IPSEC_DEV(mdev)) - return ERR_PTR(-EOPNOTSUPP); - - return mlx5_fpga_ipsec_sa_cmd_exec(mdev, cmd); -} - -int mlx5_accel_ipsec_sa_cmd_wait(void *ctx) -{ - return mlx5_fpga_ipsec_sa_cmd_wait(ctx); -} - u32 mlx5_accel_ipsec_device_caps(struct mlx5_core_dev *mdev) { return mlx5_fpga_ipsec_device_caps(mdev); } +EXPORT_SYMBOL_GPL(mlx5_accel_ipsec_device_caps); unsigned int mlx5_accel_ipsec_counters_count(struct mlx5_core_dev *mdev) { @@ -67,6 +54,21 @@ int mlx5_accel_ipsec_counters_read(struct mlx5_core_dev *mdev, u64 *counters, return mlx5_fpga_ipsec_counters_read(mdev, counters, count); } +void *mlx5_accel_esp_create_hw_context(struct mlx5_core_dev *mdev, + struct mlx5_accel_esp_xfrm *xfrm, + const __be32 saddr[4], + const __be32 daddr[4], + const __be32 spi, bool is_ipv6) +{ + return mlx5_fpga_ipsec_create_sa_ctx(mdev, xfrm, saddr, daddr, + spi, is_ipv6); +} + +void mlx5_accel_esp_free_hw_context(void *context) +{ + mlx5_fpga_ipsec_delete_sa_ctx(context); +} + int mlx5_accel_ipsec_init(struct mlx5_core_dev *mdev) { return mlx5_fpga_ipsec_init(mdev); @@ -76,3 +78,32 @@ void mlx5_accel_ipsec_cleanup(struct mlx5_core_dev *mdev) { mlx5_fpga_ipsec_cleanup(mdev); } + +struct mlx5_accel_esp_xfrm * +mlx5_accel_esp_create_xfrm(struct mlx5_core_dev *mdev, + const struct mlx5_accel_esp_xfrm_attrs *attrs, + u32 flags) +{ + struct mlx5_accel_esp_xfrm *xfrm; + + xfrm = mlx5_fpga_esp_create_xfrm(mdev, attrs, flags); + if (IS_ERR(xfrm)) + return xfrm; + + xfrm->mdev = mdev; + return xfrm; +} +EXPORT_SYMBOL_GPL(mlx5_accel_esp_create_xfrm); + +void mlx5_accel_esp_destroy_xfrm(struct mlx5_accel_esp_xfrm *xfrm) +{ + mlx5_fpga_esp_destroy_xfrm(xfrm); +} +EXPORT_SYMBOL_GPL(mlx5_accel_esp_destroy_xfrm); + +int mlx5_accel_esp_modify_xfrm(struct mlx5_accel_esp_xfrm *xfrm, + const struct mlx5_accel_esp_xfrm_attrs *attrs) +{ + return mlx5_fpga_esp_modify_xfrm(xfrm, attrs); +} +EXPORT_SYMBOL_GPL(mlx5_accel_esp_modify_xfrm); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/accel/ipsec.h b/drivers/net/ethernet/mellanox/mlx5/core/accel/ipsec.h index 67cda8871f5a..024dbd22a89b 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/accel/ipsec.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/accel/ipsec.h @@ -35,88 +35,24 @@ #define __MLX5_ACCEL_IPSEC_H__ #include +#include #ifdef CONFIG_MLX5_ACCEL -enum { - MLX5_ACCEL_IPSEC_DEVICE = BIT(1), - MLX5_ACCEL_IPSEC_IPV6 = BIT(2), - MLX5_ACCEL_IPSEC_ESP = BIT(3), - MLX5_ACCEL_IPSEC_LSO = BIT(4), -}; - -#define MLX5_IPSEC_SADB_IP_AH BIT(7) -#define MLX5_IPSEC_SADB_IP_ESP BIT(6) -#define MLX5_IPSEC_SADB_SA_VALID BIT(5) -#define MLX5_IPSEC_SADB_SPI_EN BIT(4) -#define MLX5_IPSEC_SADB_DIR_SX BIT(3) -#define MLX5_IPSEC_SADB_IPV6 BIT(2) - -enum { - MLX5_IPSEC_CMD_ADD_SA = 0, - MLX5_IPSEC_CMD_DEL_SA = 1, -}; - -enum mlx5_accel_ipsec_enc_mode { - MLX5_IPSEC_SADB_MODE_NONE = 0, - MLX5_IPSEC_SADB_MODE_AES_GCM_128_AUTH_128 = 1, - MLX5_IPSEC_SADB_MODE_AES_GCM_256_AUTH_128 = 3, -}; - #define MLX5_IPSEC_DEV(mdev) (mlx5_accel_ipsec_device_caps(mdev) & \ - MLX5_ACCEL_IPSEC_DEVICE) - -struct mlx5_accel_ipsec_sa { - __be32 cmd; - u8 key_enc[32]; - u8 key_auth[32]; - __be32 sip[4]; - __be32 dip[4]; - union { - struct { - __be32 reserved; - u8 salt_iv[8]; - __be32 salt; - } __packed gcm; - struct { - u8 salt[16]; - } __packed cbc; - }; - __be32 spi; - __be32 sw_sa_handle; - __be16 tfclen; - u8 enc_mode; - u8 sip_masklen; - u8 dip_masklen; - u8 flags; - u8 reserved[2]; -} __packed; - -/** - * mlx5_accel_ipsec_sa_cmd_exec - Execute an IPSec SADB command - * @mdev: mlx5 device - * @cmd: command to execute - * May be called from atomic context. Returns context pointer, or error - * Caller must eventually call mlx5_accel_ipsec_sa_cmd_wait from non-atomic - * context, to cleanup the context pointer - */ -void *mlx5_accel_ipsec_sa_cmd_exec(struct mlx5_core_dev *mdev, - struct mlx5_accel_ipsec_sa *cmd); - -/** - * mlx5_accel_ipsec_sa_cmd_wait - Wait for command execution completion - * @context: Context pointer returned from call to mlx5_accel_ipsec_sa_cmd_exec - * Sleeps (killable) until command execution is complete. - * Returns the command result, or -EINTR if killed - */ -int mlx5_accel_ipsec_sa_cmd_wait(void *context); - -u32 mlx5_accel_ipsec_device_caps(struct mlx5_core_dev *mdev); + MLX5_ACCEL_IPSEC_CAP_DEVICE) unsigned int mlx5_accel_ipsec_counters_count(struct mlx5_core_dev *mdev); int mlx5_accel_ipsec_counters_read(struct mlx5_core_dev *mdev, u64 *counters, unsigned int count); +void *mlx5_accel_esp_create_hw_context(struct mlx5_core_dev *mdev, + struct mlx5_accel_esp_xfrm *xfrm, + const __be32 saddr[4], + const __be32 daddr[4], + const __be32 spi, bool is_ipv6); +void mlx5_accel_esp_free_hw_context(void *context); + int mlx5_accel_ipsec_init(struct mlx5_core_dev *mdev); void mlx5_accel_ipsec_cleanup(struct mlx5_core_dev *mdev); @@ -124,6 +60,20 @@ void mlx5_accel_ipsec_cleanup(struct mlx5_core_dev *mdev); #define MLX5_IPSEC_DEV(mdev) false +static inline void * +mlx5_accel_esp_create_hw_context(struct mlx5_core_dev *mdev, + struct mlx5_accel_esp_xfrm *xfrm, + const __be32 saddr[4], + const __be32 daddr[4], + const __be32 spi, bool is_ipv6) +{ + return NULL; +} + +static inline void mlx5_accel_esp_free_hw_context(void *context) +{ +} + static inline int mlx5_accel_ipsec_init(struct mlx5_core_dev *mdev) { return 0; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/cq.c b/drivers/net/ethernet/mellanox/mlx5/core/cq.c index 669ed16938b3..a4179122a279 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/cq.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/cq.c @@ -109,8 +109,7 @@ int mlx5_core_create_cq(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq, cq->cons_index = 0; cq->arm_sn = 0; cq->eq = eq; - refcount_set(&cq->refcount, 0); - mlx5_cq_hold(cq); + refcount_set(&cq->refcount, 1); init_completion(&cq->free); if (!cq->comp) cq->comp = mlx5_add_cq_to_tasklet; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c index 1b49afca65c0..cf58c9637904 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.c @@ -38,17 +38,24 @@ #include #include "en.h" -#include "accel/ipsec.h" #include "en_accel/ipsec.h" #include "en_accel/ipsec_rxtx.h" -struct mlx5e_ipsec_sa_entry { - struct hlist_node hlist; /* Item in SADB_RX hashtable */ - unsigned int handle; /* Handle in SADB_RX */ - struct xfrm_state *x; - struct mlx5e_ipsec *ipsec; - void *context; -}; + +static struct mlx5e_ipsec_sa_entry *to_ipsec_sa_entry(struct xfrm_state *x) +{ + struct mlx5e_ipsec_sa_entry *sa; + + if (!x) + return NULL; + + sa = (struct mlx5e_ipsec_sa_entry *)x->xso.offload_handle; + if (!sa) + return NULL; + + WARN_ON(sa->x != x); + return sa; +} struct xfrm_state *mlx5e_ipsec_sadb_rx_lookup(struct mlx5e_ipsec *ipsec, unsigned int handle) @@ -105,78 +112,93 @@ static void mlx5e_ipsec_sadb_rx_free(struct mlx5e_ipsec_sa_entry *sa_entry) ida_simple_remove(&ipsec->halloc, sa_entry->handle); } -static enum mlx5_accel_ipsec_enc_mode mlx5e_ipsec_enc_mode(struct xfrm_state *x) +static bool mlx5e_ipsec_update_esn_state(struct mlx5e_ipsec_sa_entry *sa_entry) { - unsigned int key_len = (x->aead->alg_key_len + 7) / 8 - 4; + struct xfrm_replay_state_esn *replay_esn; + u32 seq_bottom; + u8 overlap; + u32 *esn; - switch (key_len) { - case 16: - return MLX5_IPSEC_SADB_MODE_AES_GCM_128_AUTH_128; - case 32: - return MLX5_IPSEC_SADB_MODE_AES_GCM_256_AUTH_128; - default: - netdev_warn(x->xso.dev, "Bad key len: %d for alg %s\n", - key_len, x->aead->alg_name); - return -1; + if (!(sa_entry->x->props.flags & XFRM_STATE_ESN)) { + sa_entry->esn_state.trigger = 0; + return false; } + + replay_esn = sa_entry->x->replay_esn; + seq_bottom = replay_esn->seq - replay_esn->replay_window + 1; + overlap = sa_entry->esn_state.overlap; + + sa_entry->esn_state.esn = xfrm_replay_seqhi(sa_entry->x, + htonl(seq_bottom)); + esn = &sa_entry->esn_state.esn; + + sa_entry->esn_state.trigger = 1; + if (unlikely(overlap && seq_bottom < MLX5E_IPSEC_ESN_SCOPE_MID)) { + ++(*esn); + sa_entry->esn_state.overlap = 0; + return true; + } else if (unlikely(!overlap && + (seq_bottom >= MLX5E_IPSEC_ESN_SCOPE_MID))) { + sa_entry->esn_state.overlap = 1; + return true; + } + + return false; } -static void mlx5e_ipsec_build_hw_sa(u32 op, struct mlx5e_ipsec_sa_entry *sa_entry, - struct mlx5_accel_ipsec_sa *hw_sa) +static void +mlx5e_ipsec_build_accel_xfrm_attrs(struct mlx5e_ipsec_sa_entry *sa_entry, + struct mlx5_accel_esp_xfrm_attrs *attrs) { struct xfrm_state *x = sa_entry->x; + struct aes_gcm_keymat *aes_gcm = &attrs->keymat.aes_gcm; struct aead_geniv_ctx *geniv_ctx; - unsigned int crypto_data_len; struct crypto_aead *aead; - unsigned int key_len; + unsigned int crypto_data_len, key_len; int ivsize; - memset(hw_sa, 0, sizeof(*hw_sa)); + memset(attrs, 0, sizeof(*attrs)); - if (op == MLX5_IPSEC_CMD_ADD_SA) { - crypto_data_len = (x->aead->alg_key_len + 7) / 8; - key_len = crypto_data_len - 4; /* 4 bytes salt at end */ - aead = x->data; - geniv_ctx = crypto_aead_ctx(aead); - ivsize = crypto_aead_ivsize(aead); + /* key */ + crypto_data_len = (x->aead->alg_key_len + 7) / 8; + key_len = crypto_data_len - 4; /* 4 bytes salt at end */ - memcpy(&hw_sa->key_enc, x->aead->alg_key, key_len); - /* Duplicate 128 bit key twice according to HW layout */ - if (key_len == 16) - memcpy(&hw_sa->key_enc[16], x->aead->alg_key, key_len); - memcpy(&hw_sa->gcm.salt_iv, geniv_ctx->salt, ivsize); - hw_sa->gcm.salt = *((__be32 *)(x->aead->alg_key + key_len)); + memcpy(aes_gcm->aes_key, x->aead->alg_key, key_len); + aes_gcm->key_len = key_len * 8; + + /* salt and seq_iv */ + aead = x->data; + geniv_ctx = crypto_aead_ctx(aead); + ivsize = crypto_aead_ivsize(aead); + memcpy(&aes_gcm->seq_iv, &geniv_ctx->salt, ivsize); + memcpy(&aes_gcm->salt, x->aead->alg_key + key_len, + sizeof(aes_gcm->salt)); + + /* iv len */ + aes_gcm->icv_len = x->aead->alg_icv_len; + + /* esn */ + if (sa_entry->esn_state.trigger) { + attrs->flags |= MLX5_ACCEL_ESP_FLAGS_ESN_TRIGGERED; + attrs->esn = sa_entry->esn_state.esn; + if (sa_entry->esn_state.overlap) + attrs->flags |= MLX5_ACCEL_ESP_FLAGS_ESN_STATE_OVERLAP; } - hw_sa->cmd = htonl(op); - hw_sa->flags |= MLX5_IPSEC_SADB_SA_VALID | MLX5_IPSEC_SADB_SPI_EN; - if (x->props.family == AF_INET) { - hw_sa->sip[3] = x->props.saddr.a4; - hw_sa->dip[3] = x->id.daddr.a4; - hw_sa->sip_masklen = 32; - hw_sa->dip_masklen = 32; - } else { - memcpy(hw_sa->sip, x->props.saddr.a6, sizeof(hw_sa->sip)); - memcpy(hw_sa->dip, x->id.daddr.a6, sizeof(hw_sa->dip)); - hw_sa->sip_masklen = 128; - hw_sa->dip_masklen = 128; - hw_sa->flags |= MLX5_IPSEC_SADB_IPV6; - } - hw_sa->spi = x->id.spi; - hw_sa->sw_sa_handle = htonl(sa_entry->handle); - switch (x->id.proto) { - case IPPROTO_ESP: - hw_sa->flags |= MLX5_IPSEC_SADB_IP_ESP; - break; - case IPPROTO_AH: - hw_sa->flags |= MLX5_IPSEC_SADB_IP_AH; - break; - default: - break; - } - hw_sa->enc_mode = mlx5e_ipsec_enc_mode(x); - if (!(x->xso.flags & XFRM_OFFLOAD_INBOUND)) - hw_sa->flags |= MLX5_IPSEC_SADB_DIR_SX; + /* rx handle */ + attrs->sa_handle = sa_entry->handle; + + /* algo type */ + attrs->keymat_type = MLX5_ACCEL_ESP_KEYMAT_AES_GCM; + + /* action */ + attrs->action = (!(x->xso.flags & XFRM_OFFLOAD_INBOUND)) ? + MLX5_ACCEL_ESP_ACTION_ENCRYPT : + MLX5_ACCEL_ESP_ACTION_DECRYPT; + /* flags */ + attrs->flags |= (x->props.mode == XFRM_MODE_TRANSPORT) ? + MLX5_ACCEL_ESP_FLAGS_TRANSPORT : + MLX5_ACCEL_ESP_FLAGS_TUNNEL; } static inline int mlx5e_xfrm_validate_state(struct xfrm_state *x) @@ -198,7 +220,9 @@ static inline int mlx5e_xfrm_validate_state(struct xfrm_state *x) netdev_info(netdev, "Cannot offload compressed xfrm states\n"); return -EINVAL; } - if (x->props.flags & XFRM_STATE_ESN) { + if (x->props.flags & XFRM_STATE_ESN && + !(mlx5_accel_ipsec_device_caps(priv->mdev) & + MLX5_ACCEL_IPSEC_CAP_ESN)) { netdev_info(netdev, "Cannot offload ESN xfrm states\n"); return -EINVAL; } @@ -246,7 +270,8 @@ static inline int mlx5e_xfrm_validate_state(struct xfrm_state *x) return -EINVAL; } if (x->props.family == AF_INET6 && - !(mlx5_accel_ipsec_device_caps(priv->mdev) & MLX5_ACCEL_IPSEC_IPV6)) { + !(mlx5_accel_ipsec_device_caps(priv->mdev) & + MLX5_ACCEL_IPSEC_CAP_IPV6)) { netdev_info(netdev, "IPv6 xfrm state offload is not supported by this device\n"); return -EINVAL; } @@ -257,9 +282,10 @@ static int mlx5e_xfrm_add_state(struct xfrm_state *x) { struct mlx5e_ipsec_sa_entry *sa_entry = NULL; struct net_device *netdev = x->xso.dev; - struct mlx5_accel_ipsec_sa hw_sa; + struct mlx5_accel_esp_xfrm_attrs attrs; struct mlx5e_priv *priv; - void *context; + __be32 saddr[4] = {0}, daddr[4] = {0}, spi; + bool is_ipv6 = false; int err; priv = netdev_priv(netdev); @@ -286,22 +312,49 @@ static int mlx5e_xfrm_add_state(struct xfrm_state *x) netdev_info(netdev, "Failed adding to SADB_RX: %d\n", err); goto err_entry; } + } else { + sa_entry->set_iv_op = (x->props.flags & XFRM_STATE_ESN) ? + mlx5e_ipsec_set_iv_esn : mlx5e_ipsec_set_iv; } - mlx5e_ipsec_build_hw_sa(MLX5_IPSEC_CMD_ADD_SA, sa_entry, &hw_sa); - context = mlx5_accel_ipsec_sa_cmd_exec(sa_entry->ipsec->en_priv->mdev, &hw_sa); - if (IS_ERR(context)) { - err = PTR_ERR(context); + /* check esn */ + mlx5e_ipsec_update_esn_state(sa_entry); + + /* create xfrm */ + mlx5e_ipsec_build_accel_xfrm_attrs(sa_entry, &attrs); + sa_entry->xfrm = + mlx5_accel_esp_create_xfrm(priv->mdev, &attrs, + MLX5_ACCEL_XFRM_FLAG_REQUIRE_METADATA); + if (IS_ERR(sa_entry->xfrm)) { + err = PTR_ERR(sa_entry->xfrm); goto err_sadb_rx; } - err = mlx5_accel_ipsec_sa_cmd_wait(context); - if (err) - goto err_sadb_rx; + /* create hw context */ + if (x->props.family == AF_INET) { + saddr[3] = x->props.saddr.a4; + daddr[3] = x->id.daddr.a4; + } else { + memcpy(saddr, x->props.saddr.a6, sizeof(saddr)); + memcpy(daddr, x->id.daddr.a6, sizeof(daddr)); + is_ipv6 = true; + } + spi = x->id.spi; + sa_entry->hw_context = + mlx5_accel_esp_create_hw_context(priv->mdev, + sa_entry->xfrm, + saddr, daddr, spi, + is_ipv6); + if (IS_ERR(sa_entry->hw_context)) { + err = PTR_ERR(sa_entry->hw_context); + goto err_xfrm; + } x->xso.offload_handle = (unsigned long)sa_entry; goto out; +err_xfrm: + mlx5_accel_esp_destroy_xfrm(sa_entry->xfrm); err_sadb_rx: if (x->xso.flags & XFRM_OFFLOAD_INBOUND) { mlx5e_ipsec_sadb_rx_del(sa_entry); @@ -315,43 +368,26 @@ out: static void mlx5e_xfrm_del_state(struct xfrm_state *x) { - struct mlx5e_ipsec_sa_entry *sa_entry; - struct mlx5_accel_ipsec_sa hw_sa; - void *context; + struct mlx5e_ipsec_sa_entry *sa_entry = to_ipsec_sa_entry(x); - if (!x->xso.offload_handle) + if (!sa_entry) return; - sa_entry = (struct mlx5e_ipsec_sa_entry *)x->xso.offload_handle; - WARN_ON(sa_entry->x != x); - if (x->xso.flags & XFRM_OFFLOAD_INBOUND) mlx5e_ipsec_sadb_rx_del(sa_entry); - - mlx5e_ipsec_build_hw_sa(MLX5_IPSEC_CMD_DEL_SA, sa_entry, &hw_sa); - context = mlx5_accel_ipsec_sa_cmd_exec(sa_entry->ipsec->en_priv->mdev, &hw_sa); - if (IS_ERR(context)) - return; - - sa_entry->context = context; } static void mlx5e_xfrm_free_state(struct xfrm_state *x) { - struct mlx5e_ipsec_sa_entry *sa_entry; - int res; + struct mlx5e_ipsec_sa_entry *sa_entry = to_ipsec_sa_entry(x); - if (!x->xso.offload_handle) + if (!sa_entry) return; - sa_entry = (struct mlx5e_ipsec_sa_entry *)x->xso.offload_handle; - WARN_ON(sa_entry->x != x); - - res = mlx5_accel_ipsec_sa_cmd_wait(sa_entry->context); - sa_entry->context = NULL; - if (res) { - /* Leftover object will leak */ - return; + if (sa_entry->hw_context) { + flush_workqueue(sa_entry->ipsec->wq); + mlx5_accel_esp_free_hw_context(sa_entry->hw_context); + mlx5_accel_esp_destroy_xfrm(sa_entry->xfrm); } if (x->xso.flags & XFRM_OFFLOAD_INBOUND) @@ -378,6 +414,14 @@ int mlx5e_ipsec_init(struct mlx5e_priv *priv) ida_init(&ipsec->halloc); ipsec->en_priv = priv; ipsec->en_priv->ipsec = ipsec; + ipsec->no_trailer = !!(mlx5_accel_ipsec_device_caps(priv->mdev) & + MLX5_ACCEL_IPSEC_CAP_RX_NO_TRAILER); + ipsec->wq = alloc_ordered_workqueue("mlx5e_ipsec: %s", 0, + priv->netdev->name); + if (!ipsec->wq) { + kfree(ipsec); + return -ENOMEM; + } netdev_dbg(priv->netdev, "IPSec attached to netdevice\n"); return 0; } @@ -389,6 +433,9 @@ void mlx5e_ipsec_cleanup(struct mlx5e_priv *priv) if (!ipsec) return; + drain_workqueue(ipsec->wq); + destroy_workqueue(ipsec->wq); + ida_destroy(&ipsec->halloc); kfree(ipsec); priv->ipsec = NULL; @@ -409,11 +456,58 @@ static bool mlx5e_ipsec_offload_ok(struct sk_buff *skb, struct xfrm_state *x) return true; } +struct mlx5e_ipsec_modify_state_work { + struct work_struct work; + struct mlx5_accel_esp_xfrm_attrs attrs; + struct mlx5e_ipsec_sa_entry *sa_entry; +}; + +static void _update_xfrm_state(struct work_struct *work) +{ + int ret; + struct mlx5e_ipsec_modify_state_work *modify_work = + container_of(work, struct mlx5e_ipsec_modify_state_work, work); + struct mlx5e_ipsec_sa_entry *sa_entry = modify_work->sa_entry; + + ret = mlx5_accel_esp_modify_xfrm(sa_entry->xfrm, + &modify_work->attrs); + if (ret) + netdev_warn(sa_entry->ipsec->en_priv->netdev, + "Not an IPSec offload device\n"); + + kfree(modify_work); +} + +static void mlx5e_xfrm_advance_esn_state(struct xfrm_state *x) +{ + struct mlx5e_ipsec_sa_entry *sa_entry = to_ipsec_sa_entry(x); + struct mlx5e_ipsec_modify_state_work *modify_work; + bool need_update; + + if (!sa_entry) + return; + + need_update = mlx5e_ipsec_update_esn_state(sa_entry); + if (!need_update) + return; + + modify_work = kzalloc(sizeof(*modify_work), GFP_ATOMIC); + if (!modify_work) + return; + + mlx5e_ipsec_build_accel_xfrm_attrs(sa_entry, &modify_work->attrs); + modify_work->sa_entry = sa_entry; + + INIT_WORK(&modify_work->work, _update_xfrm_state); + WARN_ON(!queue_work(sa_entry->ipsec->wq, &modify_work->work)); +} + static const struct xfrmdev_ops mlx5e_ipsec_xfrmdev_ops = { .xdo_dev_state_add = mlx5e_xfrm_add_state, .xdo_dev_state_delete = mlx5e_xfrm_del_state, .xdo_dev_state_free = mlx5e_xfrm_free_state, .xdo_dev_offload_ok = mlx5e_ipsec_offload_ok, + .xdo_dev_state_advance_esn = mlx5e_xfrm_advance_esn_state, }; void mlx5e_ipsec_build_netdev(struct mlx5e_priv *priv) @@ -424,7 +518,7 @@ void mlx5e_ipsec_build_netdev(struct mlx5e_priv *priv) if (!priv->ipsec) return; - if (!(mlx5_accel_ipsec_device_caps(mdev) & MLX5_ACCEL_IPSEC_ESP) || + if (!(mlx5_accel_ipsec_device_caps(mdev) & MLX5_ACCEL_IPSEC_CAP_ESP) || !MLX5_CAP_ETH(mdev, swp)) { mlx5_core_dbg(mdev, "mlx5e: ESP and SWP offload not supported\n"); return; @@ -443,7 +537,7 @@ void mlx5e_ipsec_build_netdev(struct mlx5e_priv *priv) netdev->features |= NETIF_F_HW_ESP_TX_CSUM; netdev->hw_enc_features |= NETIF_F_HW_ESP_TX_CSUM; - if (!(mlx5_accel_ipsec_device_caps(mdev) & MLX5_ACCEL_IPSEC_LSO) || + if (!(mlx5_accel_ipsec_device_caps(mdev) & MLX5_ACCEL_IPSEC_CAP_LSO) || !MLX5_CAP_ETH(mdev, swp_lso)) { mlx5_core_dbg(mdev, "mlx5e: ESP LSO not supported\n"); return; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h index 56e00baf16cc..1198fc1eba4c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h @@ -40,7 +40,11 @@ #include #include +#include "accel/ipsec.h" + #define MLX5E_IPSEC_SADB_RX_BITS 10 +#define MLX5E_IPSEC_ESN_SCOPE_MID 0x80000000L + #define MLX5E_METADATA_ETHER_TYPE (0x8CE4) #define MLX5E_METADATA_ETHER_LEN 8 @@ -77,10 +81,30 @@ struct mlx5e_ipsec_stats { struct mlx5e_ipsec { struct mlx5e_priv *en_priv; DECLARE_HASHTABLE(sadb_rx, MLX5E_IPSEC_SADB_RX_BITS); + bool no_trailer; spinlock_t sadb_rx_lock; /* Protects sadb_rx and halloc */ struct ida halloc; struct mlx5e_ipsec_sw_stats sw_stats; struct mlx5e_ipsec_stats stats; + struct workqueue_struct *wq; +}; + +struct mlx5e_ipsec_esn_state { + u32 esn; + u8 trigger: 1; + u8 overlap: 1; +}; + +struct mlx5e_ipsec_sa_entry { + struct hlist_node hlist; /* Item in SADB_RX hashtable */ + struct mlx5e_ipsec_esn_state esn_state; + unsigned int handle; /* Handle in SADB_RX */ + struct xfrm_state *x; + struct mlx5e_ipsec *ipsec; + struct mlx5_accel_esp_xfrm *xfrm; + void *hw_context; + void (*set_iv_op)(struct sk_buff *skb, struct xfrm_state *x, + struct xfrm_offload *xo); }; void mlx5e_ipsec_build_inverse_table(void); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c index 6a7c8b04447e..c245d8e78509 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c @@ -42,10 +42,11 @@ enum { MLX5E_IPSEC_RX_SYNDROME_DECRYPTED = 0x11, MLX5E_IPSEC_RX_SYNDROME_AUTH_FAILED = 0x12, + MLX5E_IPSEC_RX_SYNDROME_BAD_PROTO = 0x17, }; struct mlx5e_ipsec_rx_metadata { - unsigned char reserved; + unsigned char nexthdr; __be32 sa_handle; } __packed; @@ -175,7 +176,30 @@ static void mlx5e_ipsec_set_swp(struct sk_buff *skb, } } -static void mlx5e_ipsec_set_iv(struct sk_buff *skb, struct xfrm_offload *xo) +void mlx5e_ipsec_set_iv_esn(struct sk_buff *skb, struct xfrm_state *x, + struct xfrm_offload *xo) +{ + struct xfrm_replay_state_esn *replay_esn = x->replay_esn; + __u32 oseq = replay_esn->oseq; + int iv_offset; + __be64 seqno; + u32 seq_hi; + + if (unlikely(skb_is_gso(skb) && oseq < MLX5E_IPSEC_ESN_SCOPE_MID && + MLX5E_IPSEC_ESN_SCOPE_MID < (oseq - skb_shinfo(skb)->gso_segs))) { + seq_hi = xo->seq.hi - 1; + } else { + seq_hi = xo->seq.hi; + } + + /* Place the SN in the IV field */ + seqno = cpu_to_be64(xo->seq.low + ((u64)seq_hi << 32)); + iv_offset = skb_transport_offset(skb) + sizeof(struct ip_esp_hdr); + skb_store_bits(skb, iv_offset, &seqno, 8); +} + +void mlx5e_ipsec_set_iv(struct sk_buff *skb, struct xfrm_state *x, + struct xfrm_offload *xo) { int iv_offset; __be64 seqno; @@ -227,6 +251,7 @@ struct sk_buff *mlx5e_ipsec_handle_tx_skb(struct net_device *netdev, struct mlx5e_priv *priv = netdev_priv(netdev); struct xfrm_offload *xo = xfrm_offload(skb); struct mlx5e_ipsec_metadata *mdata; + struct mlx5e_ipsec_sa_entry *sa_entry; struct xfrm_state *x; if (!xo) @@ -261,7 +286,8 @@ struct sk_buff *mlx5e_ipsec_handle_tx_skb(struct net_device *netdev, goto drop; } mlx5e_ipsec_set_swp(skb, &wqe->eth, x->props.mode, xo); - mlx5e_ipsec_set_iv(skb, xo); + sa_entry = (struct mlx5e_ipsec_sa_entry *)x->xso.offload_handle; + sa_entry->set_iv_op(skb, x, xo); mlx5e_ipsec_set_metadata(skb, mdata, xo); return skb; @@ -301,10 +327,17 @@ mlx5e_ipsec_build_sp(struct net_device *netdev, struct sk_buff *skb, switch (mdata->syndrome) { case MLX5E_IPSEC_RX_SYNDROME_DECRYPTED: xo->status = CRYPTO_SUCCESS; + if (likely(priv->ipsec->no_trailer)) { + xo->flags |= XFRM_ESP_NO_TRAILER; + xo->proto = mdata->content.rx.nexthdr; + } break; case MLX5E_IPSEC_RX_SYNDROME_AUTH_FAILED: xo->status = CRYPTO_TUNNEL_ESP_AUTH_FAILED; break; + case MLX5E_IPSEC_RX_SYNDROME_BAD_PROTO: + xo->status = CRYPTO_INVALID_PROTOCOL; + break; default: atomic64_inc(&priv->ipsec->sw_stats.ipsec_rx_drop_syndrome); return NULL; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.h index e37ae2598dbb..2bfbbef1b054 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.h @@ -37,6 +37,7 @@ #ifdef CONFIG_MLX5_EN_IPSEC #include +#include #include "en.h" struct sk_buff *mlx5e_ipsec_handle_rx_skb(struct net_device *netdev, @@ -46,6 +47,10 @@ void mlx5e_ipsec_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe); void mlx5e_ipsec_inverse_table_init(void); bool mlx5e_ipsec_feature_check(struct sk_buff *skb, struct net_device *netdev, netdev_features_t features); +void mlx5e_ipsec_set_iv_esn(struct sk_buff *skb, struct xfrm_state *x, + struct xfrm_offload *xo); +void mlx5e_ipsec_set_iv(struct sk_buff *skb, struct xfrm_state *x, + struct xfrm_offload *xo); struct sk_buff *mlx5e_ipsec_handle_tx_skb(struct net_device *netdev, struct mlx5e_tx_wqe *wqe, struct sk_buff *skb); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/ipsec.c b/drivers/net/ethernet/mellanox/mlx5/core/fpga/ipsec.c index 95f9c5a8619b..4f1568528738 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fpga/ipsec.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/ipsec.c @@ -31,9 +31,14 @@ * */ +#include #include +#include +#include +#include #include "mlx5_core.h" +#include "fs_cmd.h" #include "fpga/ipsec.h" #include "fpga/sdk.h" #include "fpga/core.h" @@ -41,40 +46,76 @@ #define SBU_QP_QUEUE_SIZE 8 #define MLX5_FPGA_IPSEC_CMD_TIMEOUT_MSEC (60 * 1000) -enum mlx5_ipsec_response_syndrome { - MLX5_IPSEC_RESPONSE_SUCCESS = 0, - MLX5_IPSEC_RESPONSE_ILLEGAL_REQUEST = 1, - MLX5_IPSEC_RESPONSE_SADB_ISSUE = 2, - MLX5_IPSEC_RESPONSE_WRITE_RESPONSE_ISSUE = 3, +enum mlx5_fpga_ipsec_cmd_status { + MLX5_FPGA_IPSEC_CMD_PENDING, + MLX5_FPGA_IPSEC_CMD_SEND_FAIL, + MLX5_FPGA_IPSEC_CMD_COMPLETE, }; -enum mlx5_fpga_ipsec_sacmd_status { - MLX5_FPGA_IPSEC_SACMD_PENDING, - MLX5_FPGA_IPSEC_SACMD_SEND_FAIL, - MLX5_FPGA_IPSEC_SACMD_COMPLETE, -}; - -struct mlx5_ipsec_command_context { +struct mlx5_fpga_ipsec_cmd_context { struct mlx5_fpga_dma_buf buf; - struct mlx5_accel_ipsec_sa sa; - enum mlx5_fpga_ipsec_sacmd_status status; + enum mlx5_fpga_ipsec_cmd_status status; + struct mlx5_ifc_fpga_ipsec_cmd_resp resp; int status_code; struct completion complete; struct mlx5_fpga_device *dev; struct list_head list; /* Item in pending_cmds */ + u8 command[0]; }; -struct mlx5_ipsec_sadb_resp { - __be32 syndrome; - __be32 sw_sa_handle; - u8 reserved[24]; -} __packed; +struct mlx5_fpga_esp_xfrm; + +struct mlx5_fpga_ipsec_sa_ctx { + struct rhash_head hash; + struct mlx5_ifc_fpga_ipsec_sa hw_sa; + struct mlx5_core_dev *dev; + struct mlx5_fpga_esp_xfrm *fpga_xfrm; +}; + +struct mlx5_fpga_esp_xfrm { + unsigned int num_rules; + struct mlx5_fpga_ipsec_sa_ctx *sa_ctx; + struct mutex lock; /* xfrm lock */ + struct mlx5_accel_esp_xfrm accel_xfrm; +}; + +struct mlx5_fpga_ipsec_rule { + struct rb_node node; + struct fs_fte *fte; + struct mlx5_fpga_ipsec_sa_ctx *ctx; +}; + +static const struct rhashtable_params rhash_sa = { + .key_len = FIELD_SIZEOF(struct mlx5_fpga_ipsec_sa_ctx, hw_sa), + .key_offset = offsetof(struct mlx5_fpga_ipsec_sa_ctx, hw_sa), + .head_offset = offsetof(struct mlx5_fpga_ipsec_sa_ctx, hash), + .automatic_shrinking = true, + .min_size = 1, +}; struct mlx5_fpga_ipsec { + struct mlx5_fpga_device *fdev; struct list_head pending_cmds; spinlock_t pending_cmds_lock; /* Protects pending_cmds */ u32 caps[MLX5_ST_SZ_DW(ipsec_extended_cap)]; struct mlx5_fpga_conn *conn; + + struct notifier_block fs_notifier_ingress_bypass; + struct notifier_block fs_notifier_egress; + + /* Map hardware SA --> SA context + * (mlx5_fpga_ipsec_sa) (mlx5_fpga_ipsec_sa_ctx) + * We will use this hash to avoid SAs duplication in fpga which + * aren't allowed + */ + struct rhashtable sa_hash; /* hw_sa -> mlx5_fpga_ipsec_sa_ctx */ + struct mutex sa_hash_lock; + + /* Tree holding all rules for this fpga device + * Key for searching a rule (mlx5_fpga_ipsec_rule) is (ft, id) + */ + struct rb_root rules_rb; + struct mutex rules_rb_lock; /* rules lock */ }; static bool mlx5_fpga_is_ipsec_device(struct mlx5_core_dev *mdev) @@ -98,28 +139,29 @@ static void mlx5_fpga_ipsec_send_complete(struct mlx5_fpga_conn *conn, struct mlx5_fpga_dma_buf *buf, u8 status) { - struct mlx5_ipsec_command_context *context; + struct mlx5_fpga_ipsec_cmd_context *context; if (status) { - context = container_of(buf, struct mlx5_ipsec_command_context, + context = container_of(buf, struct mlx5_fpga_ipsec_cmd_context, buf); mlx5_fpga_warn(fdev, "IPSec command send failed with status %u\n", status); - context->status = MLX5_FPGA_IPSEC_SACMD_SEND_FAIL; + context->status = MLX5_FPGA_IPSEC_CMD_SEND_FAIL; complete(&context->complete); } } -static inline int syndrome_to_errno(enum mlx5_ipsec_response_syndrome syndrome) +static inline +int syndrome_to_errno(enum mlx5_ifc_fpga_ipsec_response_syndrome syndrome) { switch (syndrome) { - case MLX5_IPSEC_RESPONSE_SUCCESS: + case MLX5_FPGA_IPSEC_RESPONSE_SUCCESS: return 0; - case MLX5_IPSEC_RESPONSE_SADB_ISSUE: + case MLX5_FPGA_IPSEC_RESPONSE_SADB_ISSUE: return -EEXIST; - case MLX5_IPSEC_RESPONSE_ILLEGAL_REQUEST: + case MLX5_FPGA_IPSEC_RESPONSE_ILLEGAL_REQUEST: return -EINVAL; - case MLX5_IPSEC_RESPONSE_WRITE_RESPONSE_ISSUE: + case MLX5_FPGA_IPSEC_RESPONSE_WRITE_RESPONSE_ISSUE: return -EIO; } return -EIO; @@ -127,9 +169,9 @@ static inline int syndrome_to_errno(enum mlx5_ipsec_response_syndrome syndrome) static void mlx5_fpga_ipsec_recv(void *cb_arg, struct mlx5_fpga_dma_buf *buf) { - struct mlx5_ipsec_sadb_resp *resp = buf->sg[0].data; - struct mlx5_ipsec_command_context *context; - enum mlx5_ipsec_response_syndrome syndrome; + struct mlx5_ifc_fpga_ipsec_cmd_resp *resp = buf->sg[0].data; + struct mlx5_fpga_ipsec_cmd_context *context; + enum mlx5_ifc_fpga_ipsec_response_syndrome syndrome; struct mlx5_fpga_device *fdev = cb_arg; unsigned long flags; @@ -139,12 +181,12 @@ static void mlx5_fpga_ipsec_recv(void *cb_arg, struct mlx5_fpga_dma_buf *buf) return; } - mlx5_fpga_dbg(fdev, "mlx5_ipsec recv_cb syndrome %08x sa_id %x\n", - ntohl(resp->syndrome), ntohl(resp->sw_sa_handle)); + mlx5_fpga_dbg(fdev, "mlx5_ipsec recv_cb syndrome %08x\n", + ntohl(resp->syndrome)); spin_lock_irqsave(&fdev->ipsec->pending_cmds_lock, flags); context = list_first_entry_or_null(&fdev->ipsec->pending_cmds, - struct mlx5_ipsec_command_context, + struct mlx5_fpga_ipsec_cmd_context, list); if (context) list_del(&context->list); @@ -156,51 +198,48 @@ static void mlx5_fpga_ipsec_recv(void *cb_arg, struct mlx5_fpga_dma_buf *buf) } mlx5_fpga_dbg(fdev, "Handling response for %p\n", context); - if (context->sa.sw_sa_handle != resp->sw_sa_handle) { - mlx5_fpga_err(fdev, "mismatch SA handle. cmd 0x%08x vs resp 0x%08x\n", - ntohl(context->sa.sw_sa_handle), - ntohl(resp->sw_sa_handle)); - return; - } - syndrome = ntohl(resp->syndrome); context->status_code = syndrome_to_errno(syndrome); - context->status = MLX5_FPGA_IPSEC_SACMD_COMPLETE; + context->status = MLX5_FPGA_IPSEC_CMD_COMPLETE; + memcpy(&context->resp, resp, sizeof(*resp)); if (context->status_code) - mlx5_fpga_warn(fdev, "IPSec SADB command failed with syndrome %08x\n", + mlx5_fpga_warn(fdev, "IPSec command failed with syndrome %08x\n", syndrome); + complete(&context->complete); } -void *mlx5_fpga_ipsec_sa_cmd_exec(struct mlx5_core_dev *mdev, - struct mlx5_accel_ipsec_sa *cmd) +static void *mlx5_fpga_ipsec_cmd_exec(struct mlx5_core_dev *mdev, + const void *cmd, int cmd_size) { - struct mlx5_ipsec_command_context *context; + struct mlx5_fpga_ipsec_cmd_context *context; struct mlx5_fpga_device *fdev = mdev->fpga; unsigned long flags; - int res = 0; + int res; - BUILD_BUG_ON((sizeof(struct mlx5_accel_ipsec_sa) & 3) != 0); if (!fdev || !fdev->ipsec) return ERR_PTR(-EOPNOTSUPP); - context = kzalloc(sizeof(*context), GFP_ATOMIC); + if (cmd_size & 3) + return ERR_PTR(-EINVAL); + + context = kzalloc(sizeof(*context) + cmd_size, GFP_ATOMIC); if (!context) return ERR_PTR(-ENOMEM); - memcpy(&context->sa, cmd, sizeof(*cmd)); - context->buf.complete = mlx5_fpga_ipsec_send_complete; - context->buf.sg[0].size = sizeof(context->sa); - context->buf.sg[0].data = &context->sa; - init_completion(&context->complete); + context->status = MLX5_FPGA_IPSEC_CMD_PENDING; context->dev = fdev; + context->buf.complete = mlx5_fpga_ipsec_send_complete; + init_completion(&context->complete); + memcpy(&context->command, cmd, cmd_size); + context->buf.sg[0].size = cmd_size; + context->buf.sg[0].data = &context->command; + spin_lock_irqsave(&fdev->ipsec->pending_cmds_lock, flags); list_add_tail(&context->list, &fdev->ipsec->pending_cmds); spin_unlock_irqrestore(&fdev->ipsec->pending_cmds_lock, flags); - context->status = MLX5_FPGA_IPSEC_SACMD_PENDING; - res = mlx5_fpga_sbu_conn_sendmsg(fdev->ipsec->conn, &context->buf); if (res) { mlx5_fpga_warn(fdev, "Failure sending IPSec command: %d\n", @@ -215,9 +254,9 @@ void *mlx5_fpga_ipsec_sa_cmd_exec(struct mlx5_core_dev *mdev, return context; } -int mlx5_fpga_ipsec_sa_cmd_wait(void *ctx) +static int mlx5_fpga_ipsec_cmd_wait(void *ctx) { - struct mlx5_ipsec_command_context *context = ctx; + struct mlx5_fpga_ipsec_cmd_context *context = ctx; unsigned long timeout = msecs_to_jiffies(MLX5_FPGA_IPSEC_CMD_TIMEOUT_MSEC); int res; @@ -228,36 +267,90 @@ int mlx5_fpga_ipsec_sa_cmd_wait(void *ctx) return -ETIMEDOUT; } - if (context->status == MLX5_FPGA_IPSEC_SACMD_COMPLETE) + if (context->status == MLX5_FPGA_IPSEC_CMD_COMPLETE) res = context->status_code; else res = -EIO; - kfree(context); return res; } +static inline bool is_v2_sadb_supported(struct mlx5_fpga_ipsec *fipsec) +{ + if (MLX5_GET(ipsec_extended_cap, fipsec->caps, v2_command)) + return true; + return false; +} + +static int mlx5_fpga_ipsec_update_hw_sa(struct mlx5_fpga_device *fdev, + struct mlx5_ifc_fpga_ipsec_sa *hw_sa, + int opcode) +{ + struct mlx5_core_dev *dev = fdev->mdev; + struct mlx5_ifc_fpga_ipsec_sa *sa; + struct mlx5_fpga_ipsec_cmd_context *cmd_context; + size_t sa_cmd_size; + int err; + + hw_sa->ipsec_sa_v1.cmd = htonl(opcode); + if (is_v2_sadb_supported(fdev->ipsec)) + sa_cmd_size = sizeof(*hw_sa); + else + sa_cmd_size = sizeof(hw_sa->ipsec_sa_v1); + + cmd_context = (struct mlx5_fpga_ipsec_cmd_context *) + mlx5_fpga_ipsec_cmd_exec(dev, hw_sa, sa_cmd_size); + if (IS_ERR(cmd_context)) + return PTR_ERR(cmd_context); + + err = mlx5_fpga_ipsec_cmd_wait(cmd_context); + if (err) + goto out; + + sa = (struct mlx5_ifc_fpga_ipsec_sa *)&cmd_context->command; + if (sa->ipsec_sa_v1.sw_sa_handle != cmd_context->resp.sw_sa_handle) { + mlx5_fpga_err(fdev, "mismatch SA handle. cmd 0x%08x vs resp 0x%08x\n", + ntohl(sa->ipsec_sa_v1.sw_sa_handle), + ntohl(cmd_context->resp.sw_sa_handle)); + err = -EIO; + } + +out: + kfree(cmd_context); + return err; +} + u32 mlx5_fpga_ipsec_device_caps(struct mlx5_core_dev *mdev) { struct mlx5_fpga_device *fdev = mdev->fpga; u32 ret = 0; - if (mlx5_fpga_is_ipsec_device(mdev)) - ret |= MLX5_ACCEL_IPSEC_DEVICE; - else + if (mlx5_fpga_is_ipsec_device(mdev)) { + ret |= MLX5_ACCEL_IPSEC_CAP_DEVICE; + ret |= MLX5_ACCEL_IPSEC_CAP_REQUIRED_METADATA; + } else { return ret; + } if (!fdev->ipsec) return ret; if (MLX5_GET(ipsec_extended_cap, fdev->ipsec->caps, esp)) - ret |= MLX5_ACCEL_IPSEC_ESP; + ret |= MLX5_ACCEL_IPSEC_CAP_ESP; if (MLX5_GET(ipsec_extended_cap, fdev->ipsec->caps, ipv6)) - ret |= MLX5_ACCEL_IPSEC_IPV6; + ret |= MLX5_ACCEL_IPSEC_CAP_IPV6; if (MLX5_GET(ipsec_extended_cap, fdev->ipsec->caps, lso)) - ret |= MLX5_ACCEL_IPSEC_LSO; + ret |= MLX5_ACCEL_IPSEC_CAP_LSO; + + if (MLX5_GET(ipsec_extended_cap, fdev->ipsec->caps, rx_no_trailer)) + ret |= MLX5_ACCEL_IPSEC_CAP_RX_NO_TRAILER; + + if (MLX5_GET(ipsec_extended_cap, fdev->ipsec->caps, esn)) { + ret |= MLX5_ACCEL_IPSEC_CAP_ESN; + ret |= MLX5_ACCEL_IPSEC_CAP_TX_IV_IS_ESN; + } return ret; } @@ -321,6 +414,828 @@ out: return ret; } +static int mlx5_fpga_ipsec_set_caps(struct mlx5_core_dev *mdev, u32 flags) +{ + struct mlx5_fpga_ipsec_cmd_context *context; + struct mlx5_ifc_fpga_ipsec_cmd_cap cmd = {0}; + int err; + + cmd.cmd = htonl(MLX5_FPGA_IPSEC_CMD_OP_SET_CAP); + cmd.flags = htonl(flags); + context = mlx5_fpga_ipsec_cmd_exec(mdev, &cmd, sizeof(cmd)); + if (IS_ERR(context)) { + err = PTR_ERR(context); + goto out; + } + + err = mlx5_fpga_ipsec_cmd_wait(context); + if (err) + goto out; + + if ((context->resp.flags & cmd.flags) != cmd.flags) { + mlx5_fpga_err(context->dev, "Failed to set capabilities. cmd 0x%08x vs resp 0x%08x\n", + cmd.flags, + context->resp.flags); + err = -EIO; + } + +out: + return err; +} + +static int mlx5_fpga_ipsec_enable_supported_caps(struct mlx5_core_dev *mdev) +{ + u32 dev_caps = mlx5_fpga_ipsec_device_caps(mdev); + u32 flags = 0; + + if (dev_caps & MLX5_ACCEL_IPSEC_CAP_RX_NO_TRAILER) + flags |= MLX5_FPGA_IPSEC_CAP_NO_TRAILER; + + return mlx5_fpga_ipsec_set_caps(mdev, flags); +} + +static void +mlx5_fpga_ipsec_build_hw_xfrm(struct mlx5_core_dev *mdev, + const struct mlx5_accel_esp_xfrm_attrs *xfrm_attrs, + struct mlx5_ifc_fpga_ipsec_sa *hw_sa) +{ + const struct aes_gcm_keymat *aes_gcm = &xfrm_attrs->keymat.aes_gcm; + + /* key */ + memcpy(&hw_sa->ipsec_sa_v1.key_enc, aes_gcm->aes_key, + aes_gcm->key_len / 8); + /* Duplicate 128 bit key twice according to HW layout */ + if (aes_gcm->key_len == 128) + memcpy(&hw_sa->ipsec_sa_v1.key_enc[16], + aes_gcm->aes_key, aes_gcm->key_len / 8); + + /* salt and seq_iv */ + memcpy(&hw_sa->ipsec_sa_v1.gcm.salt_iv, &aes_gcm->seq_iv, + sizeof(aes_gcm->seq_iv)); + memcpy(&hw_sa->ipsec_sa_v1.gcm.salt, &aes_gcm->salt, + sizeof(aes_gcm->salt)); + + /* esn */ + if (xfrm_attrs->flags & MLX5_ACCEL_ESP_FLAGS_ESN_TRIGGERED) { + hw_sa->ipsec_sa_v1.flags |= MLX5_FPGA_IPSEC_SA_ESN_EN; + hw_sa->ipsec_sa_v1.flags |= + (xfrm_attrs->flags & + MLX5_ACCEL_ESP_FLAGS_ESN_STATE_OVERLAP) ? + MLX5_FPGA_IPSEC_SA_ESN_OVERLAP : 0; + hw_sa->esn = htonl(xfrm_attrs->esn); + } else { + hw_sa->ipsec_sa_v1.flags &= ~MLX5_FPGA_IPSEC_SA_ESN_EN; + hw_sa->ipsec_sa_v1.flags &= + ~(xfrm_attrs->flags & + MLX5_ACCEL_ESP_FLAGS_ESN_STATE_OVERLAP) ? + MLX5_FPGA_IPSEC_SA_ESN_OVERLAP : 0; + hw_sa->esn = 0; + } + + /* rx handle */ + hw_sa->ipsec_sa_v1.sw_sa_handle = htonl(xfrm_attrs->sa_handle); + + /* enc mode */ + switch (aes_gcm->key_len) { + case 128: + hw_sa->ipsec_sa_v1.enc_mode = + MLX5_FPGA_IPSEC_SA_ENC_MODE_AES_GCM_128_AUTH_128; + break; + case 256: + hw_sa->ipsec_sa_v1.enc_mode = + MLX5_FPGA_IPSEC_SA_ENC_MODE_AES_GCM_256_AUTH_128; + break; + } + + /* flags */ + hw_sa->ipsec_sa_v1.flags |= MLX5_FPGA_IPSEC_SA_SA_VALID | + MLX5_FPGA_IPSEC_SA_SPI_EN | + MLX5_FPGA_IPSEC_SA_IP_ESP; + + if (xfrm_attrs->action & MLX5_ACCEL_ESP_ACTION_ENCRYPT) + hw_sa->ipsec_sa_v1.flags |= MLX5_FPGA_IPSEC_SA_DIR_SX; + else + hw_sa->ipsec_sa_v1.flags &= ~MLX5_FPGA_IPSEC_SA_DIR_SX; +} + +static void +mlx5_fpga_ipsec_build_hw_sa(struct mlx5_core_dev *mdev, + struct mlx5_accel_esp_xfrm_attrs *xfrm_attrs, + const __be32 saddr[4], + const __be32 daddr[4], + const __be32 spi, bool is_ipv6, + struct mlx5_ifc_fpga_ipsec_sa *hw_sa) +{ + mlx5_fpga_ipsec_build_hw_xfrm(mdev, xfrm_attrs, hw_sa); + + /* IPs */ + memcpy(hw_sa->ipsec_sa_v1.sip, saddr, sizeof(hw_sa->ipsec_sa_v1.sip)); + memcpy(hw_sa->ipsec_sa_v1.dip, daddr, sizeof(hw_sa->ipsec_sa_v1.dip)); + + /* SPI */ + hw_sa->ipsec_sa_v1.spi = spi; + + /* flags */ + if (is_ipv6) + hw_sa->ipsec_sa_v1.flags |= MLX5_FPGA_IPSEC_SA_IPV6; +} + +static bool is_full_mask(const void *p, size_t len) +{ + WARN_ON(len % 4); + + return !memchr_inv(p, 0xff, len); +} + +static bool validate_fpga_full_mask(struct mlx5_core_dev *dev, + const u32 *match_c, + const u32 *match_v) +{ + const void *misc_params_c = MLX5_ADDR_OF(fte_match_param, + match_c, + misc_parameters); + const void *headers_c = MLX5_ADDR_OF(fte_match_param, + match_c, + outer_headers); + const void *headers_v = MLX5_ADDR_OF(fte_match_param, + match_v, + outer_headers); + + if (mlx5_fs_is_outer_ipv4_flow(dev, headers_c, headers_v)) { + const void *s_ipv4_c = MLX5_ADDR_OF(fte_match_set_lyr_2_4, + headers_c, + src_ipv4_src_ipv6.ipv4_layout.ipv4); + const void *d_ipv4_c = MLX5_ADDR_OF(fte_match_set_lyr_2_4, + headers_c, + dst_ipv4_dst_ipv6.ipv4_layout.ipv4); + + if (!is_full_mask(s_ipv4_c, MLX5_FLD_SZ_BYTES(ipv4_layout, + ipv4)) || + !is_full_mask(d_ipv4_c, MLX5_FLD_SZ_BYTES(ipv4_layout, + ipv4))) + return false; + } else { + const void *s_ipv6_c = MLX5_ADDR_OF(fte_match_set_lyr_2_4, + headers_c, + src_ipv4_src_ipv6.ipv6_layout.ipv6); + const void *d_ipv6_c = MLX5_ADDR_OF(fte_match_set_lyr_2_4, + headers_c, + dst_ipv4_dst_ipv6.ipv6_layout.ipv6); + + if (!is_full_mask(s_ipv6_c, MLX5_FLD_SZ_BYTES(ipv6_layout, + ipv6)) || + !is_full_mask(d_ipv6_c, MLX5_FLD_SZ_BYTES(ipv6_layout, + ipv6))) + return false; + } + + if (!is_full_mask(MLX5_ADDR_OF(fte_match_set_misc, misc_params_c, + outer_esp_spi), + MLX5_FLD_SZ_BYTES(fte_match_set_misc, outer_esp_spi))) + return false; + + return true; +} + +static bool mlx5_is_fpga_ipsec_rule(struct mlx5_core_dev *dev, + u8 match_criteria_enable, + const u32 *match_c, + const u32 *match_v) +{ + u32 ipsec_dev_caps = mlx5_accel_ipsec_device_caps(dev); + bool ipv6_flow; + + ipv6_flow = mlx5_fs_is_outer_ipv6_flow(dev, match_c, match_v); + + if (!(match_criteria_enable & MLX5_MATCH_OUTER_HEADERS) || + mlx5_fs_is_outer_udp_flow(match_c, match_v) || + mlx5_fs_is_outer_tcp_flow(match_c, match_v) || + mlx5_fs_is_vxlan_flow(match_c) || + !(mlx5_fs_is_outer_ipv4_flow(dev, match_c, match_v) || + ipv6_flow)) + return false; + + if (!(ipsec_dev_caps & MLX5_ACCEL_IPSEC_CAP_DEVICE)) + return false; + + if (!(ipsec_dev_caps & MLX5_ACCEL_IPSEC_CAP_ESP) && + mlx5_fs_is_outer_ipsec_flow(match_c)) + return false; + + if (!(ipsec_dev_caps & MLX5_ACCEL_IPSEC_CAP_IPV6) && + ipv6_flow) + return false; + + if (!validate_fpga_full_mask(dev, match_c, match_v)) + return false; + + return true; +} + +static bool mlx5_is_fpga_egress_ipsec_rule(struct mlx5_core_dev *dev, + u8 match_criteria_enable, + const u32 *match_c, + const u32 *match_v, + struct mlx5_flow_act *flow_act) +{ + const void *outer_c = MLX5_ADDR_OF(fte_match_param, match_c, + outer_headers); + bool is_dmac = MLX5_GET(fte_match_set_lyr_2_4, outer_c, dmac_47_16) || + MLX5_GET(fte_match_set_lyr_2_4, outer_c, dmac_15_0); + bool is_smac = MLX5_GET(fte_match_set_lyr_2_4, outer_c, smac_47_16) || + MLX5_GET(fte_match_set_lyr_2_4, outer_c, smac_15_0); + int ret; + + ret = mlx5_is_fpga_ipsec_rule(dev, match_criteria_enable, match_c, + match_v); + if (!ret) + return ret; + + if (is_dmac || is_smac || + (match_criteria_enable & + ~(MLX5_MATCH_OUTER_HEADERS | MLX5_MATCH_MISC_PARAMETERS)) || + (flow_act->action & ~(MLX5_FLOW_CONTEXT_ACTION_ENCRYPT | MLX5_FLOW_CONTEXT_ACTION_ALLOW)) || + flow_act->has_flow_tag) + return false; + + return true; +} + +void *mlx5_fpga_ipsec_create_sa_ctx(struct mlx5_core_dev *mdev, + struct mlx5_accel_esp_xfrm *accel_xfrm, + const __be32 saddr[4], + const __be32 daddr[4], + const __be32 spi, bool is_ipv6) +{ + struct mlx5_fpga_ipsec_sa_ctx *sa_ctx; + struct mlx5_fpga_esp_xfrm *fpga_xfrm = + container_of(accel_xfrm, typeof(*fpga_xfrm), + accel_xfrm); + struct mlx5_fpga_device *fdev = mdev->fpga; + struct mlx5_fpga_ipsec *fipsec = fdev->ipsec; + int opcode, err; + void *context; + + /* alloc SA */ + sa_ctx = kzalloc(sizeof(*sa_ctx), GFP_KERNEL); + if (!sa_ctx) + return ERR_PTR(-ENOMEM); + + sa_ctx->dev = mdev; + + /* build candidate SA */ + mlx5_fpga_ipsec_build_hw_sa(mdev, &accel_xfrm->attrs, + saddr, daddr, spi, is_ipv6, + &sa_ctx->hw_sa); + + mutex_lock(&fpga_xfrm->lock); + + if (fpga_xfrm->sa_ctx) { /* multiple rules for same accel_xfrm */ + /* all rules must be with same IPs and SPI */ + if (memcmp(&sa_ctx->hw_sa, &fpga_xfrm->sa_ctx->hw_sa, + sizeof(sa_ctx->hw_sa))) { + context = ERR_PTR(-EINVAL); + goto exists; + } + + ++fpga_xfrm->num_rules; + context = fpga_xfrm->sa_ctx; + goto exists; + } + + /* This is unbounded fpga_xfrm, try to add to hash */ + mutex_lock(&fipsec->sa_hash_lock); + + err = rhashtable_lookup_insert_fast(&fipsec->sa_hash, &sa_ctx->hash, + rhash_sa); + if (err) { + /* Can't bound different accel_xfrm to already existing sa_ctx. + * This is because we can't support multiple ketmats for + * same IPs and SPI + */ + context = ERR_PTR(-EEXIST); + goto unlock_hash; + } + + /* Bound accel_xfrm to sa_ctx */ + opcode = is_v2_sadb_supported(fdev->ipsec) ? + MLX5_FPGA_IPSEC_CMD_OP_ADD_SA_V2 : + MLX5_FPGA_IPSEC_CMD_OP_ADD_SA; + err = mlx5_fpga_ipsec_update_hw_sa(fdev, &sa_ctx->hw_sa, opcode); + sa_ctx->hw_sa.ipsec_sa_v1.cmd = 0; + if (err) { + context = ERR_PTR(err); + goto delete_hash; + } + + mutex_unlock(&fipsec->sa_hash_lock); + + ++fpga_xfrm->num_rules; + fpga_xfrm->sa_ctx = sa_ctx; + sa_ctx->fpga_xfrm = fpga_xfrm; + + mutex_unlock(&fpga_xfrm->lock); + + return sa_ctx; + +delete_hash: + WARN_ON(rhashtable_remove_fast(&fipsec->sa_hash, &sa_ctx->hash, + rhash_sa)); +unlock_hash: + mutex_unlock(&fipsec->sa_hash_lock); + +exists: + mutex_unlock(&fpga_xfrm->lock); + kfree(sa_ctx); + return context; +} + +static void * +mlx5_fpga_ipsec_fs_create_sa_ctx(struct mlx5_core_dev *mdev, + struct fs_fte *fte, + bool is_egress) +{ + struct mlx5_accel_esp_xfrm *accel_xfrm; + __be32 saddr[4], daddr[4], spi; + struct mlx5_flow_group *fg; + bool is_ipv6 = false; + + fs_get_obj(fg, fte->node.parent); + /* validate */ + if (is_egress && + !mlx5_is_fpga_egress_ipsec_rule(mdev, + fg->mask.match_criteria_enable, + fg->mask.match_criteria, + fte->val, + &fte->action)) + return ERR_PTR(-EINVAL); + else if (!mlx5_is_fpga_ipsec_rule(mdev, + fg->mask.match_criteria_enable, + fg->mask.match_criteria, + fte->val)) + return ERR_PTR(-EINVAL); + + /* get xfrm context */ + accel_xfrm = + (struct mlx5_accel_esp_xfrm *)fte->action.esp_id; + + /* IPs */ + if (mlx5_fs_is_outer_ipv4_flow(mdev, fg->mask.match_criteria, + fte->val)) { + memcpy(&saddr[3], + MLX5_ADDR_OF(fte_match_set_lyr_2_4, + fte->val, + src_ipv4_src_ipv6.ipv4_layout.ipv4), + sizeof(saddr[3])); + memcpy(&daddr[3], + MLX5_ADDR_OF(fte_match_set_lyr_2_4, + fte->val, + dst_ipv4_dst_ipv6.ipv4_layout.ipv4), + sizeof(daddr[3])); + } else { + memcpy(saddr, + MLX5_ADDR_OF(fte_match_param, + fte->val, + outer_headers.src_ipv4_src_ipv6.ipv6_layout.ipv6), + sizeof(saddr)); + memcpy(daddr, + MLX5_ADDR_OF(fte_match_param, + fte->val, + outer_headers.dst_ipv4_dst_ipv6.ipv6_layout.ipv6), + sizeof(daddr)); + is_ipv6 = true; + } + + /* SPI */ + spi = MLX5_GET_BE(typeof(spi), + fte_match_param, fte->val, + misc_parameters.outer_esp_spi); + + /* create */ + return mlx5_fpga_ipsec_create_sa_ctx(mdev, accel_xfrm, + saddr, daddr, + spi, is_ipv6); +} + +static void +mlx5_fpga_ipsec_release_sa_ctx(struct mlx5_fpga_ipsec_sa_ctx *sa_ctx) +{ + struct mlx5_fpga_device *fdev = sa_ctx->dev->fpga; + struct mlx5_fpga_ipsec *fipsec = fdev->ipsec; + int opcode = is_v2_sadb_supported(fdev->ipsec) ? + MLX5_FPGA_IPSEC_CMD_OP_DEL_SA_V2 : + MLX5_FPGA_IPSEC_CMD_OP_DEL_SA; + int err; + + err = mlx5_fpga_ipsec_update_hw_sa(fdev, &sa_ctx->hw_sa, opcode); + sa_ctx->hw_sa.ipsec_sa_v1.cmd = 0; + if (err) { + WARN_ON(err); + return; + } + + mutex_lock(&fipsec->sa_hash_lock); + WARN_ON(rhashtable_remove_fast(&fipsec->sa_hash, &sa_ctx->hash, + rhash_sa)); + mutex_unlock(&fipsec->sa_hash_lock); +} + +void mlx5_fpga_ipsec_delete_sa_ctx(void *context) +{ + struct mlx5_fpga_esp_xfrm *fpga_xfrm = + ((struct mlx5_fpga_ipsec_sa_ctx *)context)->fpga_xfrm; + + mutex_lock(&fpga_xfrm->lock); + if (!--fpga_xfrm->num_rules) { + mlx5_fpga_ipsec_release_sa_ctx(fpga_xfrm->sa_ctx); + fpga_xfrm->sa_ctx = NULL; + } + mutex_unlock(&fpga_xfrm->lock); +} + +static inline struct mlx5_fpga_ipsec_rule * +_rule_search(struct rb_root *root, struct fs_fte *fte) +{ + struct rb_node *node = root->rb_node; + + while (node) { + struct mlx5_fpga_ipsec_rule *rule = + container_of(node, struct mlx5_fpga_ipsec_rule, + node); + + if (rule->fte < fte) + node = node->rb_left; + else if (rule->fte > fte) + node = node->rb_right; + else + return rule; + } + return NULL; +} + +static struct mlx5_fpga_ipsec_rule * +rule_search(struct mlx5_fpga_ipsec *ipsec_dev, struct fs_fte *fte) +{ + struct mlx5_fpga_ipsec_rule *rule; + + mutex_lock(&ipsec_dev->rules_rb_lock); + rule = _rule_search(&ipsec_dev->rules_rb, fte); + mutex_unlock(&ipsec_dev->rules_rb_lock); + + return rule; +} + +static inline int _rule_insert(struct rb_root *root, + struct mlx5_fpga_ipsec_rule *rule) +{ + struct rb_node **new = &root->rb_node, *parent = NULL; + + /* Figure out where to put new node */ + while (*new) { + struct mlx5_fpga_ipsec_rule *this = + container_of(*new, struct mlx5_fpga_ipsec_rule, + node); + + parent = *new; + if (rule->fte < this->fte) + new = &((*new)->rb_left); + else if (rule->fte > this->fte) + new = &((*new)->rb_right); + else + return -EEXIST; + } + + /* Add new node and rebalance tree. */ + rb_link_node(&rule->node, parent, new); + rb_insert_color(&rule->node, root); + + return 0; +} + +static int rule_insert(struct mlx5_fpga_ipsec *ipsec_dev, + struct mlx5_fpga_ipsec_rule *rule) +{ + int ret; + + mutex_lock(&ipsec_dev->rules_rb_lock); + ret = _rule_insert(&ipsec_dev->rules_rb, rule); + mutex_unlock(&ipsec_dev->rules_rb_lock); + + return ret; +} + +static inline void _rule_delete(struct mlx5_fpga_ipsec *ipsec_dev, + struct mlx5_fpga_ipsec_rule *rule) +{ + struct rb_root *root = &ipsec_dev->rules_rb; + + mutex_lock(&ipsec_dev->rules_rb_lock); + rb_erase(&rule->node, root); + mutex_unlock(&ipsec_dev->rules_rb_lock); +} + +static void rule_delete(struct mlx5_fpga_ipsec *ipsec_dev, + struct mlx5_fpga_ipsec_rule *rule) +{ + _rule_delete(ipsec_dev, rule); + kfree(rule); +} + +struct mailbox_mod { + uintptr_t saved_esp_id; + u32 saved_action; + u32 saved_outer_esp_spi_value; +}; + +static void restore_spec_mailbox(struct fs_fte *fte, + struct mailbox_mod *mbox_mod) +{ + char *misc_params_v = MLX5_ADDR_OF(fte_match_param, + fte->val, + misc_parameters); + + MLX5_SET(fte_match_set_misc, misc_params_v, outer_esp_spi, + mbox_mod->saved_outer_esp_spi_value); + fte->action.action |= mbox_mod->saved_action; + fte->action.esp_id = (uintptr_t)mbox_mod->saved_esp_id; +} + +static void modify_spec_mailbox(struct mlx5_core_dev *mdev, + struct fs_fte *fte, + struct mailbox_mod *mbox_mod) +{ + char *misc_params_v = MLX5_ADDR_OF(fte_match_param, + fte->val, + misc_parameters); + + mbox_mod->saved_esp_id = fte->action.esp_id; + mbox_mod->saved_action = fte->action.action & + (MLX5_FLOW_CONTEXT_ACTION_ENCRYPT | + MLX5_FLOW_CONTEXT_ACTION_DECRYPT); + mbox_mod->saved_outer_esp_spi_value = + MLX5_GET(fte_match_set_misc, misc_params_v, + outer_esp_spi); + + fte->action.esp_id = 0; + fte->action.action &= ~(MLX5_FLOW_CONTEXT_ACTION_ENCRYPT | + MLX5_FLOW_CONTEXT_ACTION_DECRYPT); + if (!MLX5_CAP_FLOWTABLE(mdev, + flow_table_properties_nic_receive.ft_field_support.outer_esp_spi)) + MLX5_SET(fte_match_set_misc, misc_params_v, outer_esp_spi, 0); +} + +static enum fs_flow_table_type egress_to_fs_ft(bool egress) +{ + return egress ? FS_FT_NIC_TX : FS_FT_NIC_RX; +} + +static int fpga_ipsec_fs_create_flow_group(struct mlx5_core_dev *dev, + struct mlx5_flow_table *ft, + u32 *in, + unsigned int *group_id, + bool is_egress) +{ + int (*create_flow_group)(struct mlx5_core_dev *dev, + struct mlx5_flow_table *ft, u32 *in, + unsigned int *group_id) = + mlx5_fs_cmd_get_default(egress_to_fs_ft(is_egress))->create_flow_group; + char *misc_params_c = MLX5_ADDR_OF(create_flow_group_in, in, + match_criteria.misc_parameters); + u32 saved_outer_esp_spi_mask; + u8 match_criteria_enable; + int ret; + + if (MLX5_CAP_FLOWTABLE(dev, + flow_table_properties_nic_receive.ft_field_support.outer_esp_spi)) + return create_flow_group(dev, ft, in, group_id); + + match_criteria_enable = + MLX5_GET(create_flow_group_in, in, match_criteria_enable); + saved_outer_esp_spi_mask = + MLX5_GET(fte_match_set_misc, misc_params_c, outer_esp_spi); + if (!match_criteria_enable || !saved_outer_esp_spi_mask) + return create_flow_group(dev, ft, in, group_id); + + MLX5_SET(fte_match_set_misc, misc_params_c, outer_esp_spi, 0); + + if (!(*misc_params_c) && + !memcmp(misc_params_c, misc_params_c + 1, MLX5_ST_SZ_BYTES(fte_match_set_misc) - 1)) + MLX5_SET(create_flow_group_in, in, match_criteria_enable, + match_criteria_enable & ~MLX5_MATCH_MISC_PARAMETERS); + + ret = create_flow_group(dev, ft, in, group_id); + + MLX5_SET(fte_match_set_misc, misc_params_c, outer_esp_spi, saved_outer_esp_spi_mask); + MLX5_SET(create_flow_group_in, in, match_criteria_enable, match_criteria_enable); + + return ret; +} + +static int fpga_ipsec_fs_create_fte(struct mlx5_core_dev *dev, + struct mlx5_flow_table *ft, + struct mlx5_flow_group *fg, + struct fs_fte *fte, + bool is_egress) +{ + int (*create_fte)(struct mlx5_core_dev *dev, + struct mlx5_flow_table *ft, + struct mlx5_flow_group *fg, + struct fs_fte *fte) = + mlx5_fs_cmd_get_default(egress_to_fs_ft(is_egress))->create_fte; + struct mlx5_fpga_device *fdev = dev->fpga; + struct mlx5_fpga_ipsec *fipsec = fdev->ipsec; + struct mlx5_fpga_ipsec_rule *rule; + bool is_esp = fte->action.esp_id; + struct mailbox_mod mbox_mod; + int ret; + + if (!is_esp || + !(fte->action.action & + (MLX5_FLOW_CONTEXT_ACTION_ENCRYPT | + MLX5_FLOW_CONTEXT_ACTION_DECRYPT))) + return create_fte(dev, ft, fg, fte); + + rule = kzalloc(sizeof(*rule), GFP_KERNEL); + if (!rule) + return -ENOMEM; + + rule->ctx = mlx5_fpga_ipsec_fs_create_sa_ctx(dev, fte, is_egress); + if (IS_ERR(rule->ctx)) { + kfree(rule); + return PTR_ERR(rule->ctx); + } + + rule->fte = fte; + WARN_ON(rule_insert(fipsec, rule)); + + modify_spec_mailbox(dev, fte, &mbox_mod); + ret = create_fte(dev, ft, fg, fte); + restore_spec_mailbox(fte, &mbox_mod); + if (ret) { + _rule_delete(fipsec, rule); + mlx5_fpga_ipsec_delete_sa_ctx(rule->ctx); + kfree(rule); + } + + return ret; +} + +static int fpga_ipsec_fs_update_fte(struct mlx5_core_dev *dev, + struct mlx5_flow_table *ft, + unsigned int group_id, + int modify_mask, + struct fs_fte *fte, + bool is_egress) +{ + int (*update_fte)(struct mlx5_core_dev *dev, + struct mlx5_flow_table *ft, + unsigned int group_id, + int modify_mask, + struct fs_fte *fte) = + mlx5_fs_cmd_get_default(egress_to_fs_ft(is_egress))->update_fte; + bool is_esp = fte->action.esp_id; + struct mailbox_mod mbox_mod; + int ret; + + if (!is_esp || + !(fte->action.action & + (MLX5_FLOW_CONTEXT_ACTION_ENCRYPT | + MLX5_FLOW_CONTEXT_ACTION_DECRYPT))) + return update_fte(dev, ft, group_id, modify_mask, fte); + + modify_spec_mailbox(dev, fte, &mbox_mod); + ret = update_fte(dev, ft, group_id, modify_mask, fte); + restore_spec_mailbox(fte, &mbox_mod); + + return ret; +} + +static int fpga_ipsec_fs_delete_fte(struct mlx5_core_dev *dev, + struct mlx5_flow_table *ft, + struct fs_fte *fte, + bool is_egress) +{ + int (*delete_fte)(struct mlx5_core_dev *dev, + struct mlx5_flow_table *ft, + struct fs_fte *fte) = + mlx5_fs_cmd_get_default(egress_to_fs_ft(is_egress))->delete_fte; + struct mlx5_fpga_device *fdev = dev->fpga; + struct mlx5_fpga_ipsec *fipsec = fdev->ipsec; + struct mlx5_fpga_ipsec_rule *rule; + bool is_esp = fte->action.esp_id; + struct mailbox_mod mbox_mod; + int ret; + + if (!is_esp || + !(fte->action.action & + (MLX5_FLOW_CONTEXT_ACTION_ENCRYPT | + MLX5_FLOW_CONTEXT_ACTION_DECRYPT))) + return delete_fte(dev, ft, fte); + + rule = rule_search(fipsec, fte); + if (!rule) + return -ENOENT; + + mlx5_fpga_ipsec_delete_sa_ctx(rule->ctx); + rule_delete(fipsec, rule); + + modify_spec_mailbox(dev, fte, &mbox_mod); + ret = delete_fte(dev, ft, fte); + restore_spec_mailbox(fte, &mbox_mod); + + return ret; +} + +static int +mlx5_fpga_ipsec_fs_create_flow_group_egress(struct mlx5_core_dev *dev, + struct mlx5_flow_table *ft, + u32 *in, + unsigned int *group_id) +{ + return fpga_ipsec_fs_create_flow_group(dev, ft, in, group_id, true); +} + +static int +mlx5_fpga_ipsec_fs_create_fte_egress(struct mlx5_core_dev *dev, + struct mlx5_flow_table *ft, + struct mlx5_flow_group *fg, + struct fs_fte *fte) +{ + return fpga_ipsec_fs_create_fte(dev, ft, fg, fte, true); +} + +static int +mlx5_fpga_ipsec_fs_update_fte_egress(struct mlx5_core_dev *dev, + struct mlx5_flow_table *ft, + unsigned int group_id, + int modify_mask, + struct fs_fte *fte) +{ + return fpga_ipsec_fs_update_fte(dev, ft, group_id, modify_mask, fte, + true); +} + +static int +mlx5_fpga_ipsec_fs_delete_fte_egress(struct mlx5_core_dev *dev, + struct mlx5_flow_table *ft, + struct fs_fte *fte) +{ + return fpga_ipsec_fs_delete_fte(dev, ft, fte, true); +} + +static int +mlx5_fpga_ipsec_fs_create_flow_group_ingress(struct mlx5_core_dev *dev, + struct mlx5_flow_table *ft, + u32 *in, + unsigned int *group_id) +{ + return fpga_ipsec_fs_create_flow_group(dev, ft, in, group_id, false); +} + +static int +mlx5_fpga_ipsec_fs_create_fte_ingress(struct mlx5_core_dev *dev, + struct mlx5_flow_table *ft, + struct mlx5_flow_group *fg, + struct fs_fte *fte) +{ + return fpga_ipsec_fs_create_fte(dev, ft, fg, fte, false); +} + +static int +mlx5_fpga_ipsec_fs_update_fte_ingress(struct mlx5_core_dev *dev, + struct mlx5_flow_table *ft, + unsigned int group_id, + int modify_mask, + struct fs_fte *fte) +{ + return fpga_ipsec_fs_update_fte(dev, ft, group_id, modify_mask, fte, + false); +} + +static int +mlx5_fpga_ipsec_fs_delete_fte_ingress(struct mlx5_core_dev *dev, + struct mlx5_flow_table *ft, + struct fs_fte *fte) +{ + return fpga_ipsec_fs_delete_fte(dev, ft, fte, false); +} + +static struct mlx5_flow_cmds fpga_ipsec_ingress; +static struct mlx5_flow_cmds fpga_ipsec_egress; + +const struct mlx5_flow_cmds *mlx5_fs_cmd_get_default_ipsec_fpga_cmds(enum fs_flow_table_type type) +{ + switch (type) { + case FS_FT_NIC_RX: + return &fpga_ipsec_ingress; + case FS_FT_NIC_TX: + return &fpga_ipsec_egress; + default: + WARN_ON(true); + return NULL; + } +} + int mlx5_fpga_ipsec_init(struct mlx5_core_dev *mdev) { struct mlx5_fpga_conn_attr init_attr = {0}; @@ -335,6 +1250,8 @@ int mlx5_fpga_ipsec_init(struct mlx5_core_dev *mdev) if (!fdev->ipsec) return -ENOMEM; + fdev->ipsec->fdev = fdev; + err = mlx5_fpga_get_sbu_caps(fdev, sizeof(fdev->ipsec->caps), fdev->ipsec->caps); if (err) { @@ -358,14 +1275,47 @@ int mlx5_fpga_ipsec_init(struct mlx5_core_dev *mdev) goto error; } fdev->ipsec->conn = conn; + + err = rhashtable_init(&fdev->ipsec->sa_hash, &rhash_sa); + if (err) + goto err_destroy_conn; + mutex_init(&fdev->ipsec->sa_hash_lock); + + fdev->ipsec->rules_rb = RB_ROOT; + mutex_init(&fdev->ipsec->rules_rb_lock); + + err = mlx5_fpga_ipsec_enable_supported_caps(mdev); + if (err) { + mlx5_fpga_err(fdev, "Failed to enable IPSec extended capabilities: %d\n", + err); + goto err_destroy_hash; + } + return 0; +err_destroy_hash: + rhashtable_destroy(&fdev->ipsec->sa_hash); + +err_destroy_conn: + mlx5_fpga_sbu_conn_destroy(conn); + error: kfree(fdev->ipsec); fdev->ipsec = NULL; return err; } +static void destroy_rules_rb(struct rb_root *root) +{ + struct mlx5_fpga_ipsec_rule *r, *tmp; + + rbtree_postorder_for_each_entry_safe(r, tmp, root, node) { + rb_erase(&r->node, root); + mlx5_fpga_ipsec_delete_sa_ctx(r->ctx); + kfree(r); + } +} + void mlx5_fpga_ipsec_cleanup(struct mlx5_core_dev *mdev) { struct mlx5_fpga_device *fdev = mdev->fpga; @@ -373,7 +1323,209 @@ void mlx5_fpga_ipsec_cleanup(struct mlx5_core_dev *mdev) if (!mlx5_fpga_is_ipsec_device(mdev)) return; + destroy_rules_rb(&fdev->ipsec->rules_rb); + rhashtable_destroy(&fdev->ipsec->sa_hash); + mlx5_fpga_sbu_conn_destroy(fdev->ipsec->conn); kfree(fdev->ipsec); fdev->ipsec = NULL; } + +void mlx5_fpga_ipsec_build_fs_cmds(void) +{ + /* ingress */ + fpga_ipsec_ingress.create_flow_table = + mlx5_fs_cmd_get_default(egress_to_fs_ft(false))->create_flow_table; + fpga_ipsec_ingress.destroy_flow_table = + mlx5_fs_cmd_get_default(egress_to_fs_ft(false))->destroy_flow_table; + fpga_ipsec_ingress.modify_flow_table = + mlx5_fs_cmd_get_default(egress_to_fs_ft(false))->modify_flow_table; + fpga_ipsec_ingress.create_flow_group = + mlx5_fpga_ipsec_fs_create_flow_group_ingress; + fpga_ipsec_ingress.destroy_flow_group = + mlx5_fs_cmd_get_default(egress_to_fs_ft(false))->destroy_flow_group; + fpga_ipsec_ingress.create_fte = + mlx5_fpga_ipsec_fs_create_fte_ingress; + fpga_ipsec_ingress.update_fte = + mlx5_fpga_ipsec_fs_update_fte_ingress; + fpga_ipsec_ingress.delete_fte = + mlx5_fpga_ipsec_fs_delete_fte_ingress; + fpga_ipsec_ingress.update_root_ft = + mlx5_fs_cmd_get_default(egress_to_fs_ft(false))->update_root_ft; + + /* egress */ + fpga_ipsec_egress.create_flow_table = + mlx5_fs_cmd_get_default(egress_to_fs_ft(true))->create_flow_table; + fpga_ipsec_egress.destroy_flow_table = + mlx5_fs_cmd_get_default(egress_to_fs_ft(true))->destroy_flow_table; + fpga_ipsec_egress.modify_flow_table = + mlx5_fs_cmd_get_default(egress_to_fs_ft(true))->modify_flow_table; + fpga_ipsec_egress.create_flow_group = + mlx5_fpga_ipsec_fs_create_flow_group_egress; + fpga_ipsec_egress.destroy_flow_group = + mlx5_fs_cmd_get_default(egress_to_fs_ft(true))->destroy_flow_group; + fpga_ipsec_egress.create_fte = + mlx5_fpga_ipsec_fs_create_fte_egress; + fpga_ipsec_egress.update_fte = + mlx5_fpga_ipsec_fs_update_fte_egress; + fpga_ipsec_egress.delete_fte = + mlx5_fpga_ipsec_fs_delete_fte_egress; + fpga_ipsec_egress.update_root_ft = + mlx5_fs_cmd_get_default(egress_to_fs_ft(true))->update_root_ft; +} + +static int +mlx5_fpga_esp_validate_xfrm_attrs(struct mlx5_core_dev *mdev, + const struct mlx5_accel_esp_xfrm_attrs *attrs) +{ + if (attrs->tfc_pad) { + mlx5_core_err(mdev, "Cannot offload xfrm states with tfc padding\n"); + return -EOPNOTSUPP; + } + + if (attrs->replay_type != MLX5_ACCEL_ESP_REPLAY_NONE) { + mlx5_core_err(mdev, "Cannot offload xfrm states with anti replay\n"); + return -EOPNOTSUPP; + } + + if (attrs->keymat_type != MLX5_ACCEL_ESP_KEYMAT_AES_GCM) { + mlx5_core_err(mdev, "Only aes gcm keymat is supported\n"); + return -EOPNOTSUPP; + } + + if (attrs->keymat.aes_gcm.iv_algo != + MLX5_ACCEL_ESP_AES_GCM_IV_ALGO_SEQ) { + mlx5_core_err(mdev, "Only iv sequence algo is supported\n"); + return -EOPNOTSUPP; + } + + if (attrs->keymat.aes_gcm.icv_len != 128) { + mlx5_core_err(mdev, "Cannot offload xfrm states with AEAD ICV length other than 128bit\n"); + return -EOPNOTSUPP; + } + + if (attrs->keymat.aes_gcm.key_len != 128 && + attrs->keymat.aes_gcm.key_len != 256) { + mlx5_core_err(mdev, "Cannot offload xfrm states with AEAD key length other than 128/256 bit\n"); + return -EOPNOTSUPP; + } + + if ((attrs->flags & MLX5_ACCEL_ESP_FLAGS_ESN_TRIGGERED) && + (!MLX5_GET(ipsec_extended_cap, mdev->fpga->ipsec->caps, + v2_command))) { + mlx5_core_err(mdev, "Cannot offload xfrm states with AEAD key length other than 128/256 bit\n"); + return -EOPNOTSUPP; + } + + return 0; +} + +struct mlx5_accel_esp_xfrm * +mlx5_fpga_esp_create_xfrm(struct mlx5_core_dev *mdev, + const struct mlx5_accel_esp_xfrm_attrs *attrs, + u32 flags) +{ + struct mlx5_fpga_esp_xfrm *fpga_xfrm; + + if (!(flags & MLX5_ACCEL_XFRM_FLAG_REQUIRE_METADATA)) { + mlx5_core_warn(mdev, "Tried to create an esp action without metadata\n"); + return ERR_PTR(-EINVAL); + } + + if (mlx5_fpga_esp_validate_xfrm_attrs(mdev, attrs)) { + mlx5_core_warn(mdev, "Tried to create an esp with unsupported attrs\n"); + return ERR_PTR(-EOPNOTSUPP); + } + + fpga_xfrm = kzalloc(sizeof(*fpga_xfrm), GFP_KERNEL); + if (!fpga_xfrm) + return ERR_PTR(-ENOMEM); + + mutex_init(&fpga_xfrm->lock); + memcpy(&fpga_xfrm->accel_xfrm.attrs, attrs, + sizeof(fpga_xfrm->accel_xfrm.attrs)); + + return &fpga_xfrm->accel_xfrm; +} + +void mlx5_fpga_esp_destroy_xfrm(struct mlx5_accel_esp_xfrm *xfrm) +{ + struct mlx5_fpga_esp_xfrm *fpga_xfrm = + container_of(xfrm, struct mlx5_fpga_esp_xfrm, + accel_xfrm); + /* assuming no sa_ctx are connected to this xfrm_ctx */ + kfree(fpga_xfrm); +} + +int mlx5_fpga_esp_modify_xfrm(struct mlx5_accel_esp_xfrm *xfrm, + const struct mlx5_accel_esp_xfrm_attrs *attrs) +{ + struct mlx5_core_dev *mdev = xfrm->mdev; + struct mlx5_fpga_device *fdev = mdev->fpga; + struct mlx5_fpga_ipsec *fipsec = fdev->ipsec; + struct mlx5_fpga_esp_xfrm *fpga_xfrm; + struct mlx5_ifc_fpga_ipsec_sa org_hw_sa; + + int err = 0; + + if (!memcmp(&xfrm->attrs, attrs, sizeof(xfrm->attrs))) + return 0; + + if (!mlx5_fpga_esp_validate_xfrm_attrs(mdev, attrs)) { + mlx5_core_warn(mdev, "Tried to create an esp with unsupported attrs\n"); + return -EOPNOTSUPP; + } + + if (is_v2_sadb_supported(fipsec)) { + mlx5_core_warn(mdev, "Modify esp is not supported\n"); + return -EOPNOTSUPP; + } + + fpga_xfrm = container_of(xfrm, struct mlx5_fpga_esp_xfrm, accel_xfrm); + + mutex_lock(&fpga_xfrm->lock); + + if (!fpga_xfrm->sa_ctx) + /* Unbounded xfrm, chane only sw attrs */ + goto change_sw_xfrm_attrs; + + /* copy original hw sa */ + memcpy(&org_hw_sa, &fpga_xfrm->sa_ctx->hw_sa, sizeof(org_hw_sa)); + mutex_lock(&fipsec->sa_hash_lock); + /* remove original hw sa from hash */ + WARN_ON(rhashtable_remove_fast(&fipsec->sa_hash, + &fpga_xfrm->sa_ctx->hash, rhash_sa)); + /* update hw_sa with new xfrm attrs*/ + mlx5_fpga_ipsec_build_hw_xfrm(xfrm->mdev, attrs, + &fpga_xfrm->sa_ctx->hw_sa); + /* try to insert new hw_sa to hash */ + err = rhashtable_insert_fast(&fipsec->sa_hash, + &fpga_xfrm->sa_ctx->hash, rhash_sa); + if (err) + goto rollback_sa; + + /* modify device with new hw_sa */ + err = mlx5_fpga_ipsec_update_hw_sa(fdev, &fpga_xfrm->sa_ctx->hw_sa, + MLX5_FPGA_IPSEC_CMD_OP_MOD_SA_V2); + fpga_xfrm->sa_ctx->hw_sa.ipsec_sa_v1.cmd = 0; + if (err) + WARN_ON(rhashtable_remove_fast(&fipsec->sa_hash, + &fpga_xfrm->sa_ctx->hash, + rhash_sa)); +rollback_sa: + if (err) { + /* return original hw_sa to hash */ + memcpy(&fpga_xfrm->sa_ctx->hw_sa, &org_hw_sa, + sizeof(org_hw_sa)); + WARN_ON(rhashtable_insert_fast(&fipsec->sa_hash, + &fpga_xfrm->sa_ctx->hash, + rhash_sa)); + } + mutex_unlock(&fipsec->sa_hash_lock); + +change_sw_xfrm_attrs: + if (!err) + memcpy(&xfrm->attrs, attrs, sizeof(xfrm->attrs)); + mutex_unlock(&fpga_xfrm->lock); + return err; +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fpga/ipsec.h b/drivers/net/ethernet/mellanox/mlx5/core/fpga/ipsec.h index 26a3e4b56972..2b5e63b0d4d6 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fpga/ipsec.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/fpga/ipsec.h @@ -35,34 +35,39 @@ #define __MLX5_FPGA_IPSEC_H__ #include "accel/ipsec.h" +#include "fs_cmd.h" #ifdef CONFIG_MLX5_FPGA -void *mlx5_fpga_ipsec_sa_cmd_exec(struct mlx5_core_dev *mdev, - struct mlx5_accel_ipsec_sa *cmd); -int mlx5_fpga_ipsec_sa_cmd_wait(void *context); - u32 mlx5_fpga_ipsec_device_caps(struct mlx5_core_dev *mdev); unsigned int mlx5_fpga_ipsec_counters_count(struct mlx5_core_dev *mdev); int mlx5_fpga_ipsec_counters_read(struct mlx5_core_dev *mdev, u64 *counters, unsigned int counters_count); +void *mlx5_fpga_ipsec_create_sa_ctx(struct mlx5_core_dev *mdev, + struct mlx5_accel_esp_xfrm *accel_xfrm, + const __be32 saddr[4], + const __be32 daddr[4], + const __be32 spi, bool is_ipv6); +void mlx5_fpga_ipsec_delete_sa_ctx(void *context); + int mlx5_fpga_ipsec_init(struct mlx5_core_dev *mdev); void mlx5_fpga_ipsec_cleanup(struct mlx5_core_dev *mdev); +void mlx5_fpga_ipsec_build_fs_cmds(void); + +struct mlx5_accel_esp_xfrm * +mlx5_fpga_esp_create_xfrm(struct mlx5_core_dev *mdev, + const struct mlx5_accel_esp_xfrm_attrs *attrs, + u32 flags); +void mlx5_fpga_esp_destroy_xfrm(struct mlx5_accel_esp_xfrm *xfrm); +int mlx5_fpga_esp_modify_xfrm(struct mlx5_accel_esp_xfrm *xfrm, + const struct mlx5_accel_esp_xfrm_attrs *attrs); + +const struct mlx5_flow_cmds * +mlx5_fs_cmd_get_default_ipsec_fpga_cmds(enum fs_flow_table_type type); #else -static inline void *mlx5_fpga_ipsec_sa_cmd_exec(struct mlx5_core_dev *mdev, - struct mlx5_accel_ipsec_sa *cmd) -{ - return ERR_PTR(-EOPNOTSUPP); -} - -static inline int mlx5_fpga_ipsec_sa_cmd_wait(void *context) -{ - return -EOPNOTSUPP; -} - static inline u32 mlx5_fpga_ipsec_device_caps(struct mlx5_core_dev *mdev) { return 0; @@ -80,6 +85,20 @@ static inline int mlx5_fpga_ipsec_counters_read(struct mlx5_core_dev *mdev, return 0; } +static inline void * +mlx5_fpga_ipsec_create_sa_ctx(struct mlx5_core_dev *mdev, + struct mlx5_accel_esp_xfrm *accel_xfrm, + const __be32 saddr[4], + const __be32 daddr[4], + const __be32 spi, bool is_ipv6) +{ + return NULL; +} + +static inline void mlx5_fpga_ipsec_delete_sa_ctx(void *context) +{ +} + static inline int mlx5_fpga_ipsec_init(struct mlx5_core_dev *mdev) { return 0; @@ -89,6 +108,35 @@ static inline void mlx5_fpga_ipsec_cleanup(struct mlx5_core_dev *mdev) { } +static inline void mlx5_fpga_ipsec_build_fs_cmds(void) +{ +} + +static inline struct mlx5_accel_esp_xfrm * +mlx5_fpga_esp_create_xfrm(struct mlx5_core_dev *mdev, + const struct mlx5_accel_esp_xfrm_attrs *attrs, + u32 flags) +{ + return ERR_PTR(-EOPNOTSUPP); +} + +static inline void mlx5_fpga_esp_destroy_xfrm(struct mlx5_accel_esp_xfrm *xfrm) +{ +} + +static inline int +mlx5_fpga_esp_modify_xfrm(struct mlx5_accel_esp_xfrm *xfrm, + const struct mlx5_accel_esp_xfrm_attrs *attrs) +{ + return -EOPNOTSUPP; +} + +static inline const struct mlx5_flow_cmds * +mlx5_fs_cmd_get_default_ipsec_fpga_cmds(enum fs_flow_table_type type) +{ + return mlx5_fs_cmd_get_default(type); +} + #endif /* CONFIG_MLX5_FPGA */ #endif /* __MLX5_FPGA_SADB_H__ */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c index f9b82fad2446..10e16381f20a 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c @@ -38,6 +38,7 @@ #include "fs_cmd.h" #include "diag/fs_tracepoint.h" #include "accel/ipsec.h" +#include "fpga/ipsec.h" #define INIT_TREE_NODE_ARRAY_SIZE(...) (sizeof((struct init_tree_node[]){__VA_ARGS__}) /\ sizeof(struct init_tree_node)) @@ -2258,6 +2259,10 @@ static struct mlx5_flow_root_namespace struct mlx5_flow_root_namespace *root_ns; struct mlx5_flow_namespace *ns; + if (mlx5_accel_ipsec_device_caps(steering->dev) & MLX5_ACCEL_IPSEC_CAP_DEVICE && + (table_type == FS_FT_NIC_RX || table_type == FS_FT_NIC_TX)) + cmds = mlx5_fs_cmd_get_default_ipsec_fpga_cmds(table_type); + /* Create the root namespace */ root_ns = kvzalloc(sizeof(*root_ns), GFP_KERNEL); if (!root_ns) @@ -2649,8 +2654,7 @@ int mlx5_init_fs(struct mlx5_core_dev *dev) goto err; } - if (mlx5_accel_ipsec_device_caps(steering->dev) & - MLX5_ACCEL_IPSEC_DEVICE) { + if (MLX5_IPSEC_DEV(dev)) { err = init_egress_root_ns(steering); if (err) goto err; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index a968badc09b1..13b6f66310c9 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -58,6 +58,7 @@ #include "eswitch.h" #include "lib/mlx5.h" #include "fpga/core.h" +#include "fpga/ipsec.h" #include "accel/ipsec.h" #include "lib/clock.h" @@ -1658,6 +1659,7 @@ static int __init init(void) get_random_bytes(&sw_owner_id, sizeof(sw_owner_id)); mlx5_core_verify_params(); + mlx5_fpga_ipsec_build_fs_cmds(); mlx5_register_debugfs(); err = pci_register_driver(&mlx5_core_driver); diff --git a/include/linux/mlx5/accel.h b/include/linux/mlx5/accel.h new file mode 100644 index 000000000000..70e7e5673ce9 --- /dev/null +++ b/include/linux/mlx5/accel.h @@ -0,0 +1,144 @@ +/* + * Copyright (c) 2018 Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + * + */ + +#ifndef __MLX5_ACCEL_H__ +#define __MLX5_ACCEL_H__ + +#include + +enum mlx5_accel_esp_aes_gcm_keymat_iv_algo { + MLX5_ACCEL_ESP_AES_GCM_IV_ALGO_SEQ, +}; + +enum mlx5_accel_esp_flags { + MLX5_ACCEL_ESP_FLAGS_TUNNEL = 0, /* Default */ + MLX5_ACCEL_ESP_FLAGS_TRANSPORT = 1UL << 0, + MLX5_ACCEL_ESP_FLAGS_ESN_TRIGGERED = 1UL << 1, + MLX5_ACCEL_ESP_FLAGS_ESN_STATE_OVERLAP = 1UL << 2, +}; + +enum mlx5_accel_esp_action { + MLX5_ACCEL_ESP_ACTION_DECRYPT, + MLX5_ACCEL_ESP_ACTION_ENCRYPT, +}; + +enum mlx5_accel_esp_keymats { + MLX5_ACCEL_ESP_KEYMAT_AES_NONE, + MLX5_ACCEL_ESP_KEYMAT_AES_GCM, +}; + +enum mlx5_accel_esp_replay { + MLX5_ACCEL_ESP_REPLAY_NONE, + MLX5_ACCEL_ESP_REPLAY_BMP, +}; + +struct aes_gcm_keymat { + u64 seq_iv; + enum mlx5_accel_esp_aes_gcm_keymat_iv_algo iv_algo; + + u32 salt; + u32 icv_len; + + u32 key_len; + u32 aes_key[256 / 32]; +}; + +struct mlx5_accel_esp_xfrm_attrs { + enum mlx5_accel_esp_action action; + u32 esn; + u32 spi; + u32 seq; + u32 tfc_pad; + u32 flags; + u32 sa_handle; + enum mlx5_accel_esp_replay replay_type; + union { + struct { + u32 size; + + } bmp; + } replay; + enum mlx5_accel_esp_keymats keymat_type; + union { + struct aes_gcm_keymat aes_gcm; + } keymat; +}; + +struct mlx5_accel_esp_xfrm { + struct mlx5_core_dev *mdev; + struct mlx5_accel_esp_xfrm_attrs attrs; +}; + +enum { + MLX5_ACCEL_XFRM_FLAG_REQUIRE_METADATA = 1UL << 0, +}; + +enum mlx5_accel_ipsec_cap { + MLX5_ACCEL_IPSEC_CAP_DEVICE = 1 << 0, + MLX5_ACCEL_IPSEC_CAP_REQUIRED_METADATA = 1 << 1, + MLX5_ACCEL_IPSEC_CAP_ESP = 1 << 2, + MLX5_ACCEL_IPSEC_CAP_IPV6 = 1 << 3, + MLX5_ACCEL_IPSEC_CAP_LSO = 1 << 4, + MLX5_ACCEL_IPSEC_CAP_RX_NO_TRAILER = 1 << 5, + MLX5_ACCEL_IPSEC_CAP_ESN = 1 << 6, + MLX5_ACCEL_IPSEC_CAP_TX_IV_IS_ESN = 1 << 7, +}; + +#ifdef CONFIG_MLX5_ACCEL + +u32 mlx5_accel_ipsec_device_caps(struct mlx5_core_dev *mdev); + +struct mlx5_accel_esp_xfrm * +mlx5_accel_esp_create_xfrm(struct mlx5_core_dev *mdev, + const struct mlx5_accel_esp_xfrm_attrs *attrs, + u32 flags); +void mlx5_accel_esp_destroy_xfrm(struct mlx5_accel_esp_xfrm *xfrm); +int mlx5_accel_esp_modify_xfrm(struct mlx5_accel_esp_xfrm *xfrm, + const struct mlx5_accel_esp_xfrm_attrs *attrs); + +#else + +static inline u32 mlx5_accel_ipsec_device_caps(struct mlx5_core_dev *mdev) { return 0; } + +static inline struct mlx5_accel_esp_xfrm * +mlx5_accel_esp_create_xfrm(struct mlx5_core_dev *mdev, + const struct mlx5_accel_esp_xfrm_attrs *attrs, + u32 flags) { return ERR_PTR(-EOPNOTSUPP); } +static inline void +mlx5_accel_esp_destroy_xfrm(struct mlx5_accel_esp_xfrm *xfrm) {} +static inline int +mlx5_accel_esp_modify_xfrm(struct mlx5_accel_esp_xfrm *xfrm, + const struct mlx5_accel_esp_xfrm_attrs *attrs) { return -EOPNOTSUPP; } + +#endif +#endif diff --git a/include/linux/mlx5/fs.h b/include/linux/mlx5/fs.h index 744ea228acea..b957e52434f8 100644 --- a/include/linux/mlx5/fs.h +++ b/include/linux/mlx5/fs.h @@ -40,6 +40,8 @@ enum { MLX5_FLOW_CONTEXT_ACTION_FWD_NEXT_PRIO = 1 << 16, + MLX5_FLOW_CONTEXT_ACTION_ENCRYPT = 1 << 17, + MLX5_FLOW_CONTEXT_ACTION_DECRYPT = 1 << 18, }; enum { @@ -146,6 +148,7 @@ struct mlx5_flow_act { u32 flow_tag; u32 encap_id; u32 modify_id; + uintptr_t esp_id; }; #define MLX5_DECLARE_FLOW_ACT(name) \ diff --git a/include/linux/mlx5/mlx5_ifc_fpga.h b/include/linux/mlx5/mlx5_ifc_fpga.h index 255a88d08078..ec052491ba3d 100644 --- a/include/linux/mlx5/mlx5_ifc_fpga.h +++ b/include/linux/mlx5/mlx5_ifc_fpga.h @@ -373,7 +373,10 @@ struct mlx5_ifc_fpga_destroy_qp_out_bits { struct mlx5_ifc_ipsec_extended_cap_bits { u8 encapsulation[0x20]; - u8 reserved_0[0x15]; + u8 reserved_0[0x12]; + u8 v2_command[0x1]; + u8 udp_encap[0x1]; + u8 rx_no_trailer[0x1]; u8 ipv4_fragment[0x1]; u8 ipv6[0x1]; u8 esn[0x1]; @@ -429,4 +432,91 @@ struct mlx5_ifc_ipsec_counters_bits { u8 dropped_cmd[0x40]; }; +enum mlx5_ifc_fpga_ipsec_response_syndrome { + MLX5_FPGA_IPSEC_RESPONSE_SUCCESS = 0, + MLX5_FPGA_IPSEC_RESPONSE_ILLEGAL_REQUEST = 1, + MLX5_FPGA_IPSEC_RESPONSE_SADB_ISSUE = 2, + MLX5_FPGA_IPSEC_RESPONSE_WRITE_RESPONSE_ISSUE = 3, +}; + +struct mlx5_ifc_fpga_ipsec_cmd_resp { + __be32 syndrome; + union { + __be32 sw_sa_handle; + __be32 flags; + }; + u8 reserved[24]; +} __packed; + +enum mlx5_ifc_fpga_ipsec_cmd_opcode { + MLX5_FPGA_IPSEC_CMD_OP_ADD_SA = 0, + MLX5_FPGA_IPSEC_CMD_OP_DEL_SA = 1, + MLX5_FPGA_IPSEC_CMD_OP_ADD_SA_V2 = 2, + MLX5_FPGA_IPSEC_CMD_OP_DEL_SA_V2 = 3, + MLX5_FPGA_IPSEC_CMD_OP_MOD_SA_V2 = 4, + MLX5_FPGA_IPSEC_CMD_OP_SET_CAP = 5, +}; + +enum mlx5_ifc_fpga_ipsec_cap { + MLX5_FPGA_IPSEC_CAP_NO_TRAILER = BIT(0), +}; + +struct mlx5_ifc_fpga_ipsec_cmd_cap { + __be32 cmd; + __be32 flags; + u8 reserved[24]; +} __packed; + +enum mlx5_ifc_fpga_ipsec_sa_flags { + MLX5_FPGA_IPSEC_SA_ESN_EN = BIT(0), + MLX5_FPGA_IPSEC_SA_ESN_OVERLAP = BIT(1), + MLX5_FPGA_IPSEC_SA_IPV6 = BIT(2), + MLX5_FPGA_IPSEC_SA_DIR_SX = BIT(3), + MLX5_FPGA_IPSEC_SA_SPI_EN = BIT(4), + MLX5_FPGA_IPSEC_SA_SA_VALID = BIT(5), + MLX5_FPGA_IPSEC_SA_IP_ESP = BIT(6), + MLX5_FPGA_IPSEC_SA_IP_AH = BIT(7), +}; + +enum mlx5_ifc_fpga_ipsec_sa_enc_mode { + MLX5_FPGA_IPSEC_SA_ENC_MODE_NONE = 0, + MLX5_FPGA_IPSEC_SA_ENC_MODE_AES_GCM_128_AUTH_128 = 1, + MLX5_FPGA_IPSEC_SA_ENC_MODE_AES_GCM_256_AUTH_128 = 3, +}; + +struct mlx5_ifc_fpga_ipsec_sa_v1 { + __be32 cmd; + u8 key_enc[32]; + u8 key_auth[32]; + __be32 sip[4]; + __be32 dip[4]; + union { + struct { + __be32 reserved; + u8 salt_iv[8]; + __be32 salt; + } __packed gcm; + struct { + u8 salt[16]; + } __packed cbc; + }; + __be32 spi; + __be32 sw_sa_handle; + __be16 tfclen; + u8 enc_mode; + u8 reserved1[2]; + u8 flags; + u8 reserved2[2]; +}; + +struct mlx5_ifc_fpga_ipsec_sa { + struct mlx5_ifc_fpga_ipsec_sa_v1 ipsec_sa_v1; + __be16 udp_sp; + __be16 udp_dp; + u8 reserved1[4]; + __be32 esn; + __be16 vid; /* only 12 bits, rest is reserved */ + __be16 reserved2; +} __packed; + #endif /* MLX5_IFC_FPGA_H */