diff options
Diffstat (limited to 'drivers/net/ethernet/netronome')
56 files changed, 6127 insertions, 966 deletions
diff --git a/drivers/net/ethernet/netronome/Kconfig b/drivers/net/ethernet/netronome/Kconfig index 549898d5d450..bac5be4d4f43 100644 --- a/drivers/net/ethernet/netronome/Kconfig +++ b/drivers/net/ethernet/netronome/Kconfig @@ -1,3 +1,4 @@ +# SPDX-License-Identifier: GPL-2.0-only # # Netronome device configuration # @@ -19,6 +20,8 @@ config NFP tristate "Netronome(R) NFP4000/NFP6000 NIC driver" depends on PCI && PCI_MSI depends on VXLAN || VXLAN=n + depends on TLS && TLS_DEVICE || TLS_DEVICE=n + select NET_DEVLINK ---help--- This driver supports the Netronome(R) NFP4000/NFP6000 based cards working as a advanced Ethernet NIC. It works with both diff --git a/drivers/net/ethernet/netronome/Makefile b/drivers/net/ethernet/netronome/Makefile index 7fb3b84b5556..d9a3948e8bde 100644 --- a/drivers/net/ethernet/netronome/Makefile +++ b/drivers/net/ethernet/netronome/Makefile @@ -1,3 +1,4 @@ +# SPDX-License-Identifier: GPL-2.0-only # # Makefile for the Netronome network device drivers # diff --git a/drivers/net/ethernet/netronome/nfp/Makefile b/drivers/net/ethernet/netronome/nfp/Makefile index 47c708f08ade..d31772ae511d 100644 --- a/drivers/net/ethernet/netronome/nfp/Makefile +++ b/drivers/net/ethernet/netronome/nfp/Makefile @@ -15,6 +15,9 @@ nfp-objs := \ nfpcore/nfp_resource.o \ nfpcore/nfp_rtsym.o \ nfpcore/nfp_target.o \ + ccm.o \ + ccm_mbox.o \ + devlink_param.o \ nfp_asm.o \ nfp_app.o \ nfp_app_nic.o \ @@ -33,6 +36,11 @@ nfp-objs := \ nfp_shared_buf.o \ nic/main.o +ifeq ($(CONFIG_TLS_DEVICE),y) +nfp-objs += \ + crypto/tls.o +endif + ifeq ($(CONFIG_NFP_APP_FLOWER),y) nfp-objs += \ flower/action.o \ @@ -42,7 +50,8 @@ nfp-objs += \ flower/match.o \ flower/metadata.o \ flower/offload.o \ - flower/tunnel_conf.o + flower/tunnel_conf.o \ + flower/qos_conf.o endif ifeq ($(CONFIG_BPF_SYSCALL),y) diff --git a/drivers/net/ethernet/netronome/nfp/abm/cls.c b/drivers/net/ethernet/netronome/nfp/abm/cls.c index 9852080cf454..9f8a1f69c0c4 100644 --- a/drivers/net/ethernet/netronome/nfp/abm/cls.c +++ b/drivers/net/ethernet/netronome/nfp/abm/cls.c @@ -39,7 +39,7 @@ nfp_abm_u32_check_knode(struct nfp_abm *abm, struct tc_cls_u32_knode *knode, } if (knode->sel->off || knode->sel->offshift || knode->sel->offmask || knode->sel->offoff || knode->fshift) { - NL_SET_ERR_MSG_MOD(extack, "variable offseting not supported"); + NL_SET_ERR_MSG_MOD(extack, "variable offsetting not supported"); return false; } if (knode->sel->hoff || knode->sel->hmask) { @@ -78,7 +78,7 @@ nfp_abm_u32_check_knode(struct nfp_abm *abm, struct tc_cls_u32_knode *knode, k = &knode->sel->keys[0]; if (k->offmask) { - NL_SET_ERR_MSG_MOD(extack, "offset mask - variable offseting not supported"); + NL_SET_ERR_MSG_MOD(extack, "offset mask - variable offsetting not supported"); return false; } if (k->off) { @@ -176,8 +176,10 @@ nfp_abm_u32_knode_replace(struct nfp_abm_link *alink, u8 mask, val; int err; - if (!nfp_abm_u32_check_knode(alink->abm, knode, proto, extack)) + if (!nfp_abm_u32_check_knode(alink->abm, knode, proto, extack)) { + err = -EOPNOTSUPP; goto err_delete; + } tos_off = proto == htons(ETH_P_IP) ? 16 : 20; @@ -198,14 +200,18 @@ nfp_abm_u32_knode_replace(struct nfp_abm_link *alink, if ((iter->val & cmask) == (val & cmask) && iter->band != knode->res->classid) { NL_SET_ERR_MSG_MOD(extack, "conflict with already offloaded filter"); + err = -EOPNOTSUPP; goto err_delete; } } if (!match) { match = kzalloc(sizeof(*match), GFP_KERNEL); - if (!match) - return -ENOMEM; + if (!match) { + err = -ENOMEM; + goto err_delete; + } + list_add(&match->list, &alink->dscp_map); } match->handle = knode->handle; @@ -221,7 +227,7 @@ nfp_abm_u32_knode_replace(struct nfp_abm_link *alink, err_delete: nfp_abm_u32_knode_delete(alink, knode); - return -EOPNOTSUPP; + return err; } static int nfp_abm_setup_tc_block_cb(enum tc_setup_type type, @@ -262,22 +268,12 @@ static int nfp_abm_setup_tc_block_cb(enum tc_setup_type type, } } +static LIST_HEAD(nfp_abm_block_cb_list); + int nfp_abm_setup_cls_block(struct net_device *netdev, struct nfp_repr *repr, - struct tc_block_offload *f) + struct flow_block_offload *f) { - if (f->binder_type != TCF_BLOCK_BINDER_TYPE_CLSACT_EGRESS) - return -EOPNOTSUPP; - - switch (f->command) { - case TC_BLOCK_BIND: - return tcf_block_cb_register(f->block, - nfp_abm_setup_tc_block_cb, - repr, repr, f->extack); - case TC_BLOCK_UNBIND: - tcf_block_cb_unregister(f->block, nfp_abm_setup_tc_block_cb, - repr); - return 0; - default: - return -EOPNOTSUPP; - } + return flow_block_cb_setup_simple(f, &nfp_abm_block_cb_list, + nfp_abm_setup_tc_block_cb, + repr, repr, true); } diff --git a/drivers/net/ethernet/netronome/nfp/abm/ctrl.c b/drivers/net/ethernet/netronome/nfp/abm/ctrl.c index 9584f03f3efa..69e84ff7f2e5 100644 --- a/drivers/net/ethernet/netronome/nfp/abm/ctrl.c +++ b/drivers/net/ethernet/netronome/nfp/abm/ctrl.c @@ -261,10 +261,15 @@ int nfp_abm_ctrl_qm_disable(struct nfp_abm *abm) int nfp_abm_ctrl_prio_map_update(struct nfp_abm_link *alink, u32 *packed) { + const u32 cmd = NFP_NET_CFG_MBOX_CMD_PCI_DSCP_PRIOMAP_SET; struct nfp_net *nn = alink->vnic; unsigned int i; int err; + err = nfp_net_mbox_lock(nn, alink->abm->prio_map_len); + if (err) + return err; + /* Write data_len and wipe reserved */ nn_writeq(nn, nn->tlv_caps.mbox_off + NFP_NET_ABM_MBOX_DATALEN, alink->abm->prio_map_len); @@ -273,8 +278,7 @@ int nfp_abm_ctrl_prio_map_update(struct nfp_abm_link *alink, u32 *packed) nn_writel(nn, nn->tlv_caps.mbox_off + NFP_NET_ABM_MBOX_DATA + i, packed[i / sizeof(u32)]); - err = nfp_net_reconfig_mbox(nn, - NFP_NET_CFG_MBOX_CMD_PCI_DSCP_PRIOMAP_SET); + err = nfp_net_mbox_reconfig_and_unlock(nn, cmd); if (err) nfp_err(alink->abm->app->cpp, "setting DSCP -> VQ map failed with error %d\n", err); diff --git a/drivers/net/ethernet/netronome/nfp/abm/main.c b/drivers/net/ethernet/netronome/nfp/abm/main.c index 4d4ff5844c47..9183b3e85d21 100644 --- a/drivers/net/ethernet/netronome/nfp/abm/main.c +++ b/drivers/net/ethernet/netronome/nfp/abm/main.c @@ -53,7 +53,8 @@ nfp_abm_setup_tc(struct nfp_app *app, struct net_device *netdev, } } -static struct net_device *nfp_abm_repr_get(struct nfp_app *app, u32 port_id) +static struct net_device * +nfp_abm_repr_get(struct nfp_app *app, u32 port_id, bool *redir_egress) { enum nfp_repr_type rtype; struct nfp_reprs *reprs; @@ -549,5 +550,5 @@ const struct nfp_app_type app_abm = { .eswitch_mode_get = nfp_abm_eswitch_mode_get, .eswitch_mode_set = nfp_abm_eswitch_mode_set, - .repr_get = nfp_abm_repr_get, + .dev_get = nfp_abm_repr_get, }; diff --git a/drivers/net/ethernet/netronome/nfp/abm/main.h b/drivers/net/ethernet/netronome/nfp/abm/main.h index 49749c60885e..48746c9c6224 100644 --- a/drivers/net/ethernet/netronome/nfp/abm/main.h +++ b/drivers/net/ethernet/netronome/nfp/abm/main.h @@ -247,7 +247,7 @@ int nfp_abm_setup_tc_mq(struct net_device *netdev, struct nfp_abm_link *alink, int nfp_abm_setup_tc_gred(struct net_device *netdev, struct nfp_abm_link *alink, struct tc_gred_qopt_offload *opt); int nfp_abm_setup_cls_block(struct net_device *netdev, struct nfp_repr *repr, - struct tc_block_offload *opt); + struct flow_block_offload *opt); int nfp_abm_ctrl_read_params(struct nfp_abm_link *alink); int nfp_abm_ctrl_find_addrs(struct nfp_abm *abm); diff --git a/drivers/net/ethernet/netronome/nfp/bpf/Makefile b/drivers/net/ethernet/netronome/nfp/bpf/Makefile deleted file mode 100644 index 805fa28f391a..000000000000 --- a/drivers/net/ethernet/netronome/nfp/bpf/Makefile +++ /dev/null @@ -1,2 +0,0 @@ -# SPDX-License-Identifier: GPL-2.0 -# kbuild requires Makefile in a directory to build individual objects diff --git a/drivers/net/ethernet/netronome/nfp/bpf/cmsg.c b/drivers/net/ethernet/netronome/nfp/bpf/cmsg.c index 9b6cfa697879..0e2db6ea79e9 100644 --- a/drivers/net/ethernet/netronome/nfp/bpf/cmsg.c +++ b/drivers/net/ethernet/netronome/nfp/bpf/cmsg.c @@ -6,48 +6,14 @@ #include <linux/bug.h> #include <linux/jiffies.h> #include <linux/skbuff.h> -#include <linux/wait.h> +#include <linux/timekeeping.h> +#include "../ccm.h" #include "../nfp_app.h" #include "../nfp_net.h" #include "fw.h" #include "main.h" -#define NFP_BPF_TAG_ALLOC_SPAN (U16_MAX / 4) - -static bool nfp_bpf_all_tags_busy(struct nfp_app_bpf *bpf) -{ - u16 used_tags; - - used_tags = bpf->tag_alloc_next - bpf->tag_alloc_last; - - return used_tags > NFP_BPF_TAG_ALLOC_SPAN; -} - -static int nfp_bpf_alloc_tag(struct nfp_app_bpf *bpf) -{ - /* All FW communication for BPF is request-reply. To make sure we - * don't reuse the message ID too early after timeout - limit the - * number of requests in flight. - */ - if (nfp_bpf_all_tags_busy(bpf)) { - cmsg_warn(bpf, "all FW request contexts busy!\n"); - return -EAGAIN; - } - - WARN_ON(__test_and_set_bit(bpf->tag_alloc_next, bpf->tag_allocator)); - return bpf->tag_alloc_next++; -} - -static void nfp_bpf_free_tag(struct nfp_app_bpf *bpf, u16 tag) -{ - WARN_ON(!__test_and_clear_bit(tag, bpf->tag_allocator)); - - while (!test_bit(bpf->tag_alloc_last, bpf->tag_allocator) && - bpf->tag_alloc_last != bpf->tag_alloc_next) - bpf->tag_alloc_last++; -} - static struct sk_buff * nfp_bpf_cmsg_alloc(struct nfp_app_bpf *bpf, unsigned int size) { @@ -87,149 +53,6 @@ nfp_bpf_cmsg_map_reply_size(struct nfp_app_bpf *bpf, unsigned int n) return size; } -static u8 nfp_bpf_cmsg_get_type(struct sk_buff *skb) -{ - struct cmsg_hdr *hdr; - - hdr = (struct cmsg_hdr *)skb->data; - - return hdr->type; -} - -static unsigned int nfp_bpf_cmsg_get_tag(struct sk_buff *skb) -{ - struct cmsg_hdr *hdr; - - hdr = (struct cmsg_hdr *)skb->data; - - return be16_to_cpu(hdr->tag); -} - -static struct sk_buff *__nfp_bpf_reply(struct nfp_app_bpf *bpf, u16 tag) -{ - unsigned int msg_tag; - struct sk_buff *skb; - - skb_queue_walk(&bpf->cmsg_replies, skb) { - msg_tag = nfp_bpf_cmsg_get_tag(skb); - if (msg_tag == tag) { - nfp_bpf_free_tag(bpf, tag); - __skb_unlink(skb, &bpf->cmsg_replies); - return skb; - } - } - - return NULL; -} - -static struct sk_buff *nfp_bpf_reply(struct nfp_app_bpf *bpf, u16 tag) -{ - struct sk_buff *skb; - - nfp_ctrl_lock(bpf->app->ctrl); - skb = __nfp_bpf_reply(bpf, tag); - nfp_ctrl_unlock(bpf->app->ctrl); - - return skb; -} - -static struct sk_buff *nfp_bpf_reply_drop_tag(struct nfp_app_bpf *bpf, u16 tag) -{ - struct sk_buff *skb; - - nfp_ctrl_lock(bpf->app->ctrl); - skb = __nfp_bpf_reply(bpf, tag); - if (!skb) - nfp_bpf_free_tag(bpf, tag); - nfp_ctrl_unlock(bpf->app->ctrl); - - return skb; -} - -static struct sk_buff * -nfp_bpf_cmsg_wait_reply(struct nfp_app_bpf *bpf, enum nfp_bpf_cmsg_type type, - int tag) -{ - struct sk_buff *skb; - int i, err; - - for (i = 0; i < 50; i++) { - udelay(4); - skb = nfp_bpf_reply(bpf, tag); - if (skb) - return skb; - } - - err = wait_event_interruptible_timeout(bpf->cmsg_wq, - skb = nfp_bpf_reply(bpf, tag), - msecs_to_jiffies(5000)); - /* We didn't get a response - try last time and atomically drop - * the tag even if no response is matched. - */ - if (!skb) - skb = nfp_bpf_reply_drop_tag(bpf, tag); - if (err < 0) { - cmsg_warn(bpf, "%s waiting for response to 0x%02x: %d\n", - err == ERESTARTSYS ? "interrupted" : "error", - type, err); - return ERR_PTR(err); - } - if (!skb) { - cmsg_warn(bpf, "timeout waiting for response to 0x%02x\n", - type); - return ERR_PTR(-ETIMEDOUT); - } - - return skb; -} - -static struct sk_buff * -nfp_bpf_cmsg_communicate(struct nfp_app_bpf *bpf, struct sk_buff *skb, - enum nfp_bpf_cmsg_type type, unsigned int reply_size) -{ - struct cmsg_hdr *hdr; - int tag; - - nfp_ctrl_lock(bpf->app->ctrl); - tag = nfp_bpf_alloc_tag(bpf); - if (tag < 0) { - nfp_ctrl_unlock(bpf->app->ctrl); - dev_kfree_skb_any(skb); - return ERR_PTR(tag); - } - - hdr = (void *)skb->data; - hdr->ver = CMSG_MAP_ABI_VERSION; - hdr->type = type; - hdr->tag = cpu_to_be16(tag); - - __nfp_app_ctrl_tx(bpf->app, skb); - - nfp_ctrl_unlock(bpf->app->ctrl); - - skb = nfp_bpf_cmsg_wait_reply(bpf, type, tag); - if (IS_ERR(skb)) - return skb; - - hdr = (struct cmsg_hdr *)skb->data; - if (hdr->type != __CMSG_REPLY(type)) { - cmsg_warn(bpf, "cmsg drop - wrong type 0x%02x != 0x%02lx!\n", - hdr->type, __CMSG_REPLY(type)); - goto err_free; - } - /* 0 reply_size means caller will do the validation */ - if (reply_size && skb->len != reply_size) { - cmsg_warn(bpf, "cmsg drop - type 0x%02x wrong size %d != %d!\n", - type, skb->len, reply_size); - goto err_free; - } - - return skb; -err_free: - dev_kfree_skb_any(skb); - return ERR_PTR(-EIO); -} - static int nfp_bpf_ctrl_rc_to_errno(struct nfp_app_bpf *bpf, struct cmsg_reply_map_simple *reply) @@ -275,8 +98,8 @@ nfp_bpf_ctrl_alloc_map(struct nfp_app_bpf *bpf, struct bpf_map *map) req->map_type = cpu_to_be32(map->map_type); req->map_flags = 0; - skb = nfp_bpf_cmsg_communicate(bpf, skb, CMSG_TYPE_MAP_ALLOC, - sizeof(*reply)); + skb = nfp_ccm_communicate(&bpf->ccm, skb, NFP_CCM_TYPE_BPF_MAP_ALLOC, + sizeof(*reply)); if (IS_ERR(skb)) return PTR_ERR(skb); @@ -310,8 +133,8 @@ void nfp_bpf_ctrl_free_map(struct nfp_app_bpf *bpf, struct nfp_bpf_map *nfp_map) req = (void *)skb->data; req->tid = cpu_to_be32(nfp_map->tid); - skb = nfp_bpf_cmsg_communicate(bpf, skb, CMSG_TYPE_MAP_FREE, - sizeof(*reply)); + skb = nfp_ccm_communicate(&bpf->ccm, skb, NFP_CCM_TYPE_BPF_MAP_FREE, + sizeof(*reply)); if (IS_ERR(skb)) { cmsg_warn(bpf, "leaking map - I/O error\n"); return; @@ -353,30 +176,151 @@ nfp_bpf_ctrl_reply_val(struct nfp_app_bpf *bpf, struct cmsg_reply_map_op *reply, return &reply->data[bpf->cmsg_key_sz * (n + 1) + bpf->cmsg_val_sz * n]; } +static bool nfp_bpf_ctrl_op_cache_invalidate(enum nfp_ccm_type op) +{ + return op == NFP_CCM_TYPE_BPF_MAP_UPDATE || + op == NFP_CCM_TYPE_BPF_MAP_DELETE; +} + +static bool nfp_bpf_ctrl_op_cache_capable(enum nfp_ccm_type op) +{ + return op == NFP_CCM_TYPE_BPF_MAP_LOOKUP || + op == NFP_CCM_TYPE_BPF_MAP_GETNEXT; +} + +static bool nfp_bpf_ctrl_op_cache_fill(enum nfp_ccm_type op) +{ + return op == NFP_CCM_TYPE_BPF_MAP_GETFIRST || + op == NFP_CCM_TYPE_BPF_MAP_GETNEXT; +} + +static unsigned int +nfp_bpf_ctrl_op_cache_get(struct nfp_bpf_map *nfp_map, enum nfp_ccm_type op, + const u8 *key, u8 *out_key, u8 *out_value, + u32 *cache_gen) +{ + struct bpf_map *map = &nfp_map->offmap->map; + struct nfp_app_bpf *bpf = nfp_map->bpf; + unsigned int i, count, n_entries; + struct cmsg_reply_map_op *reply; + + n_entries = nfp_bpf_ctrl_op_cache_fill(op) ? bpf->cmsg_cache_cnt : 1; + + spin_lock(&nfp_map->cache_lock); + *cache_gen = nfp_map->cache_gen; + if (nfp_map->cache_blockers) + n_entries = 1; + + if (nfp_bpf_ctrl_op_cache_invalidate(op)) + goto exit_block; + if (!nfp_bpf_ctrl_op_cache_capable(op)) + goto exit_unlock; + + if (!nfp_map->cache) + goto exit_unlock; + if (nfp_map->cache_to < ktime_get_ns()) + goto exit_invalidate; + + reply = (void *)nfp_map->cache->data; + count = be32_to_cpu(reply->count); + + for (i = 0; i < count; i++) { + void *cached_key; + + cached_key = nfp_bpf_ctrl_reply_key(bpf, reply, i); + if (memcmp(cached_key, key, map->key_size)) + continue; + + if (op == NFP_CCM_TYPE_BPF_MAP_LOOKUP) + memcpy(out_value, nfp_bpf_ctrl_reply_val(bpf, reply, i), + map->value_size); + if (op == NFP_CCM_TYPE_BPF_MAP_GETNEXT) { + if (i + 1 == count) + break; + + memcpy(out_key, + nfp_bpf_ctrl_reply_key(bpf, reply, i + 1), + map->key_size); + } + + n_entries = 0; + goto exit_unlock; + } + goto exit_unlock; + +exit_block: + nfp_map->cache_blockers++; +exit_invalidate: + dev_consume_skb_any(nfp_map->cache); + nfp_map->cache = NULL; +exit_unlock: + spin_unlock(&nfp_map->cache_lock); + return n_entries; +} + +static void +nfp_bpf_ctrl_op_cache_put(struct nfp_bpf_map *nfp_map, enum nfp_ccm_type op, + struct sk_buff *skb, u32 cache_gen) +{ + bool blocker, filler; + + blocker = nfp_bpf_ctrl_op_cache_invalidate(op); + filler = nfp_bpf_ctrl_op_cache_fill(op); + if (blocker || filler) { + u64 to = 0; + + if (filler) + to = ktime_get_ns() + NFP_BPF_MAP_CACHE_TIME_NS; + + spin_lock(&nfp_map->cache_lock); + if (blocker) { + nfp_map->cache_blockers--; + nfp_map->cache_gen++; + } + if (filler && !nfp_map->cache_blockers && + nfp_map->cache_gen == cache_gen) { + nfp_map->cache_to = to; + swap(nfp_map->cache, skb); + } + spin_unlock(&nfp_map->cache_lock); + } + + dev_consume_skb_any(skb); +} + static int -nfp_bpf_ctrl_entry_op(struct bpf_offloaded_map *offmap, - enum nfp_bpf_cmsg_type op, +nfp_bpf_ctrl_entry_op(struct bpf_offloaded_map *offmap, enum nfp_ccm_type op, u8 *key, u8 *value, u64 flags, u8 *out_key, u8 *out_value) { struct nfp_bpf_map *nfp_map = offmap->dev_priv; + unsigned int n_entries, reply_entries, count; struct nfp_app_bpf *bpf = nfp_map->bpf; struct bpf_map *map = &offmap->map; struct cmsg_reply_map_op *reply; struct cmsg_req_map_op *req; struct sk_buff *skb; + u32 cache_gen; int err; /* FW messages have no space for more than 32 bits of flags */ if (flags >> 32) return -EOPNOTSUPP; + /* Handle op cache */ + n_entries = nfp_bpf_ctrl_op_cache_get(nfp_map, op, key, out_key, + out_value, &cache_gen); + if (!n_entries) + return 0; + skb = nfp_bpf_cmsg_map_req_alloc(bpf, 1); - if (!skb) - return -ENOMEM; + if (!skb) { + err = -ENOMEM; + goto err_cache_put; + } req = (void *)skb->data; req->tid = cpu_to_be32(nfp_map->tid); - req->count = cpu_to_be32(1); + req->count = cpu_to_be32(n_entries); req->flags = cpu_to_be32(flags); /* Copy inputs */ @@ -386,16 +330,38 @@ nfp_bpf_ctrl_entry_op(struct bpf_offloaded_map *offmap, memcpy(nfp_bpf_ctrl_req_val(bpf, req, 0), value, map->value_size); - skb = nfp_bpf_cmsg_communicate(bpf, skb, op, - nfp_bpf_cmsg_map_reply_size(bpf, 1)); - if (IS_ERR(skb)) - return PTR_ERR(skb); + skb = nfp_ccm_communicate(&bpf->ccm, skb, op, 0); + if (IS_ERR(skb)) { + err = PTR_ERR(skb); + goto err_cache_put; + } + + if (skb->len < sizeof(*reply)) { + cmsg_warn(bpf, "cmsg drop - type 0x%02x too short %d!\n", + op, skb->len); + err = -EIO; + goto err_free; + } reply = (void *)skb->data; + count = be32_to_cpu(reply->count); err = nfp_bpf_ctrl_rc_to_errno(bpf, &reply->reply_hdr); + /* FW responds with message sized to hold the good entries, + * plus one extra entry if there was an error. + */ + reply_entries = count + !!err; + if (n_entries > 1 && count) + err = 0; if (err) goto err_free; + if (skb->len != nfp_bpf_cmsg_map_reply_size(bpf, reply_entries)) { + cmsg_warn(bpf, "cmsg drop - type 0x%02x too short %d for %d entries!\n", + op, skb->len, reply_entries); + err = -EIO; + goto err_free; + } + /* Copy outputs */ if (out_key) memcpy(out_key, nfp_bpf_ctrl_reply_key(bpf, reply, 0), @@ -404,106 +370,107 @@ nfp_bpf_ctrl_entry_op(struct bpf_offloaded_map *offmap, memcpy(out_value, nfp_bpf_ctrl_reply_val(bpf, reply, 0), map->value_size); - dev_consume_skb_any(skb); + nfp_bpf_ctrl_op_cache_put(nfp_map, op, skb, cache_gen); return 0; err_free: dev_kfree_skb_any(skb); +err_cache_put: + nfp_bpf_ctrl_op_cache_put(nfp_map, op, NULL, cache_gen); return err; } int nfp_bpf_ctrl_update_entry(struct bpf_offloaded_map *offmap, void *key, void *value, u64 flags) { - return nfp_bpf_ctrl_entry_op(offmap, CMSG_TYPE_MAP_UPDATE, + return nfp_bpf_ctrl_entry_op(offmap, NFP_CCM_TYPE_BPF_MAP_UPDATE, key, value, flags, NULL, NULL); } int nfp_bpf_ctrl_del_entry(struct bpf_offloaded_map *offmap, void *key) { - return nfp_bpf_ctrl_entry_op(offmap, CMSG_TYPE_MAP_DELETE, + return nfp_bpf_ctrl_entry_op(offmap, NFP_CCM_TYPE_BPF_MAP_DELETE, key, NULL, 0, NULL, NULL); } int nfp_bpf_ctrl_lookup_entry(struct bpf_offloaded_map *offmap, void *key, void *value) { - return nfp_bpf_ctrl_entry_op(offmap, CMSG_TYPE_MAP_LOOKUP, + return nfp_bpf_ctrl_entry_op(offmap, NFP_CCM_TYPE_BPF_MAP_LOOKUP, key, NULL, 0, NULL, value); } int nfp_bpf_ctrl_getfirst_entry(struct bpf_offloaded_map *offmap, void *next_key) { - return nfp_bpf_ctrl_entry_op(offmap, CMSG_TYPE_MAP_GETFIRST, + return nfp_bpf_ctrl_entry_op(offmap, NFP_CCM_TYPE_BPF_MAP_GETFIRST, NULL, NULL, 0, next_key, NULL); } int nfp_bpf_ctrl_getnext_entry(struct bpf_offloaded_map *offmap, void *key, void *next_key) { - return nfp_bpf_ctrl_entry_op(offmap, CMSG_TYPE_MAP_GETNEXT, + return nfp_bpf_ctrl_entry_op(offmap, NFP_CCM_TYPE_BPF_MAP_GETNEXT, key, NULL, 0, next_key, NULL); } +unsigned int nfp_bpf_ctrl_cmsg_min_mtu(struct nfp_app_bpf *bpf) +{ + return max(nfp_bpf_cmsg_map_req_size(bpf, 1), + nfp_bpf_cmsg_map_reply_size(bpf, 1)); +} + unsigned int nfp_bpf_ctrl_cmsg_mtu(struct nfp_app_bpf *bpf) { - return max3((unsigned int)NFP_NET_DEFAULT_MTU, - nfp_bpf_cmsg_map_req_size(bpf, 1), - nfp_bpf_cmsg_map_reply_size(bpf, 1)); + return max3(NFP_NET_DEFAULT_MTU, + nfp_bpf_cmsg_map_req_size(bpf, NFP_BPF_MAP_CACHE_CNT), + nfp_bpf_cmsg_map_reply_size(bpf, NFP_BPF_MAP_CACHE_CNT)); +} + +unsigned int nfp_bpf_ctrl_cmsg_cache_cnt(struct nfp_app_bpf *bpf) +{ + unsigned int mtu, req_max, reply_max, entry_sz; + + mtu = bpf->app->ctrl->dp.mtu; + entry_sz = bpf->cmsg_key_sz + bpf->cmsg_val_sz; + req_max = (mtu - sizeof(struct cmsg_req_map_op)) / entry_sz; + reply_max = (mtu - sizeof(struct cmsg_reply_map_op)) / entry_sz; + + return min3(req_max, reply_max, NFP_BPF_MAP_CACHE_CNT); } void nfp_bpf_ctrl_msg_rx(struct nfp_app *app, struct sk_buff *skb) { struct nfp_app_bpf *bpf = app->priv; - unsigned int tag; if (unlikely(skb->len < sizeof(struct cmsg_reply_map_simple))) { cmsg_warn(bpf, "cmsg drop - too short %d!\n", skb->len); - goto err_free; + dev_kfree_skb_any(skb); + return; } - if (nfp_bpf_cmsg_get_type(skb) == CMSG_TYPE_BPF_EVENT) { + if (nfp_ccm_get_type(skb) == NFP_CCM_TYPE_BPF_BPF_EVENT) { if (!nfp_bpf_event_output(bpf, skb->data, skb->len)) dev_consume_skb_any(skb); else dev_kfree_skb_any(skb); - return; - } - - nfp_ctrl_lock(bpf->app->ctrl); - - tag = nfp_bpf_cmsg_get_tag(skb); - if (unlikely(!test_bit(tag, bpf->tag_allocator))) { - cmsg_warn(bpf, "cmsg drop - no one is waiting for tag %u!\n", - tag); - goto err_unlock; } - __skb_queue_tail(&bpf->cmsg_replies, skb); - wake_up_interruptible_all(&bpf->cmsg_wq); - - nfp_ctrl_unlock(bpf->app->ctrl); - - return; -err_unlock: - nfp_ctrl_unlock(bpf->app->ctrl); -err_free: - dev_kfree_skb_any(skb); + nfp_ccm_rx(&bpf->ccm, skb); } void nfp_bpf_ctrl_msg_rx_raw(struct nfp_app *app, const void *data, unsigned int len) { + const struct nfp_ccm_hdr *hdr = data; struct nfp_app_bpf *bpf = app->priv; - const struct cmsg_hdr *hdr = data; if (unlikely(len < sizeof(struct cmsg_reply_map_simple))) { cmsg_warn(bpf, "cmsg drop - too short %d!\n", len); return; } - if (hdr->type == CMSG_TYPE_BPF_EVENT) + if (hdr->type == NFP_CCM_TYPE_BPF_BPF_EVENT) nfp_bpf_event_output(bpf, data, len); else cmsg_warn(bpf, "cmsg drop - msg type %d with raw buffer!\n", diff --git a/drivers/net/ethernet/netronome/nfp/bpf/fw.h b/drivers/net/ethernet/netronome/nfp/bpf/fw.h index 721921bcf120..a83a0ad5e27d 100644 --- a/drivers/net/ethernet/netronome/nfp/bpf/fw.h +++ b/drivers/net/ethernet/netronome/nfp/bpf/fw.h @@ -6,6 +6,7 @@ #include <linux/bitops.h> #include <linux/types.h> +#include "../ccm.h" /* Kernel's enum bpf_reg_type is not uABI so people may change it breaking * our FW ABI. In that case we will do translation in the driver. @@ -23,6 +24,7 @@ enum bpf_cap_tlv_type { NFP_BPF_CAP_TYPE_QUEUE_SELECT = 5, NFP_BPF_CAP_TYPE_ADJUST_TAIL = 6, NFP_BPF_CAP_TYPE_ABI_VERSION = 7, + NFP_BPF_CAP_TYPE_CMSG_MULTI_ENT = 8, }; struct nfp_bpf_cap_tlv_func { @@ -52,22 +54,6 @@ struct nfp_bpf_cap_tlv_maps { /* * Types defined for map related control messages */ -#define CMSG_MAP_ABI_VERSION 1 - -enum nfp_bpf_cmsg_type { - CMSG_TYPE_MAP_ALLOC = 1, - CMSG_TYPE_MAP_FREE = 2, - CMSG_TYPE_MAP_LOOKUP = 3, - CMSG_TYPE_MAP_UPDATE = 4, - CMSG_TYPE_MAP_DELETE = 5, - CMSG_TYPE_MAP_GETNEXT = 6, - CMSG_TYPE_MAP_GETFIRST = 7, - CMSG_TYPE_BPF_EVENT = 8, - __CMSG_TYPE_MAP_MAX, -}; - -#define CMSG_TYPE_MAP_REPLY_BIT 7 -#define __CMSG_REPLY(req) (BIT(CMSG_TYPE_MAP_REPLY_BIT) | (req)) /* BPF ABIv2 fixed-length control message fields */ #define CMSG_MAP_KEY_LW 16 @@ -84,19 +70,13 @@ enum nfp_bpf_cmsg_status { CMSG_RC_ERR_MAP_E2BIG = 7, }; -struct cmsg_hdr { - u8 type; - u8 ver; - __be16 tag; -}; - struct cmsg_reply_map_simple { - struct cmsg_hdr hdr; + struct nfp_ccm_hdr hdr; __be32 rc; }; struct cmsg_req_map_alloc_tbl { - struct cmsg_hdr hdr; + struct nfp_ccm_hdr hdr; __be32 key_size; /* in bytes */ __be32 value_size; /* in bytes */ __be32 max_entries; @@ -110,7 +90,7 @@ struct cmsg_reply_map_alloc_tbl { }; struct cmsg_req_map_free_tbl { - struct cmsg_hdr hdr; + struct nfp_ccm_hdr hdr; __be32 tid; }; @@ -120,7 +100,7 @@ struct cmsg_reply_map_free_tbl { }; struct cmsg_req_map_op { - struct cmsg_hdr hdr; + struct nfp_ccm_hdr hdr; __be32 tid; __be32 count; __be32 flags; @@ -135,7 +115,7 @@ struct cmsg_reply_map_op { }; struct cmsg_bpf_event { - struct cmsg_hdr hdr; + struct nfp_ccm_hdr hdr; __be32 cpu_id; __be64 map_ptr; __be32 data_size; diff --git a/drivers/net/ethernet/netronome/nfp/bpf/jit.c b/drivers/net/ethernet/netronome/nfp/bpf/jit.c index f272247d1708..c80bb83c8ac9 100644 --- a/drivers/net/ethernet/netronome/nfp/bpf/jit.c +++ b/drivers/net/ethernet/netronome/nfp/bpf/jit.c @@ -328,7 +328,18 @@ __emit_shf(struct nfp_prog *nfp_prog, u16 dst, enum alu_dst_ab dst_ab, return; } - if (sc == SHF_SC_L_SHF) + /* NFP shift instruction has something special. If shift direction is + * left then shift amount of 1 to 31 is specified as 32 minus the amount + * to shift. + * + * But no need to do this for indirect shift which has shift amount be + * 0. Even after we do this subtraction, shift amount 0 will be turned + * into 32 which will eventually be encoded the same as 0 because only + * low 5 bits are encoded, but shift amount be 32 will fail the + * FIELD_PREP check done later on shift mask (0x1f), due to 32 is out of + * mask range. + */ + if (sc == SHF_SC_L_SHF && shift) shift = 32 - shift; insn = OP_SHF_BASE | @@ -612,6 +623,13 @@ static void wrp_immed(struct nfp_prog *nfp_prog, swreg dst, u32 imm) } static void +wrp_zext(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta, u8 dst) +{ + if (meta->flags & FLAG_INSN_DO_ZEXT) + wrp_immed(nfp_prog, reg_both(dst + 1), 0); +} + +static void wrp_immed_relo(struct nfp_prog *nfp_prog, swreg dst, u32 imm, enum nfp_relo_type relo) { @@ -847,7 +865,8 @@ static int nfp_cpp_memcpy(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta) } static int -data_ld(struct nfp_prog *nfp_prog, swreg offset, u8 dst_gpr, int size) +data_ld(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta, swreg offset, + u8 dst_gpr, int size) { unsigned int i; u16 shift, sz; @@ -870,14 +889,15 @@ data_ld(struct nfp_prog *nfp_prog, swreg offset, u8 dst_gpr, int size) wrp_mov(nfp_prog, reg_both(dst_gpr + i), reg_xfer(i)); if (i < 2) - wrp_immed(nfp_prog, reg_both(dst_gpr + 1), 0); + wrp_zext(nfp_prog, meta, dst_gpr); return 0; } static int -data_ld_host_order(struct nfp_prog *nfp_prog, u8 dst_gpr, - swreg lreg, swreg rreg, int size, enum cmd_mode mode) +data_ld_host_order(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta, + u8 dst_gpr, swreg lreg, swreg rreg, int size, + enum cmd_mode mode) { unsigned int i; u8 mask, sz; @@ -900,33 +920,34 @@ data_ld_host_order(struct nfp_prog *nfp_prog, u8 dst_gpr, wrp_mov(nfp_prog, reg_both(dst_gpr + i), reg_xfer(i)); if (i < 2) - wrp_immed(nfp_prog, reg_both(dst_gpr + 1), 0); + wrp_zext(nfp_prog, meta, dst_gpr); return 0; } static int -data_ld_host_order_addr32(struct nfp_prog *nfp_prog, u8 src_gpr, swreg offset, - u8 dst_gpr, u8 size) +data_ld_host_order_addr32(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta, + u8 src_gpr, swreg offset, u8 dst_gpr, u8 size) { - return data_ld_host_order(nfp_prog, dst_gpr, reg_a(src_gpr), offset, - size, CMD_MODE_32b); + return data_ld_host_order(nfp_prog, meta, dst_gpr, reg_a(src_gpr), + offset, size, CMD_MODE_32b); } static int -data_ld_host_order_addr40(struct nfp_prog *nfp_prog, u8 src_gpr, swreg offset, - u8 dst_gpr, u8 size) +data_ld_host_order_addr40(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta, + u8 src_gpr, swreg offset, u8 dst_gpr, u8 size) { swreg rega, regb; addr40_offset(nfp_prog, src_gpr, offset, ®a, ®b); - return data_ld_host_order(nfp_prog, dst_gpr, rega, regb, + return data_ld_host_order(nfp_prog, meta, dst_gpr, rega, regb, size, CMD_MODE_40b_BA); } static int -construct_data_ind_ld(struct nfp_prog *nfp_prog, u16 offset, u16 src, u8 size) +construct_data_ind_ld(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta, + u16 offset, u16 src, u8 size) { swreg tmp_reg; @@ -942,10 +963,12 @@ construct_data_ind_ld(struct nfp_prog *nfp_prog, u16 offset, u16 src, u8 size) emit_br_relo(nfp_prog, BR_BLO, BR_OFF_RELO, 0, RELO_BR_GO_ABORT); /* Load data */ - return data_ld(nfp_prog, imm_b(nfp_prog), 0, size); + return data_ld(nfp_prog, meta, imm_b(nfp_prog), 0, size); } -static int construct_data_ld(struct nfp_prog *nfp_prog, u16 offset, u8 size) +static int +construct_data_ld(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta, + u16 offset, u8 size) { swreg tmp_reg; @@ -956,7 +979,7 @@ static int construct_data_ld(struct nfp_prog *nfp_prog, u16 offset, u8 size) /* Load data */ tmp_reg = re_load_imm_any(nfp_prog, offset, imm_b(nfp_prog)); - return data_ld(nfp_prog, tmp_reg, 0, size); + return data_ld(nfp_prog, meta, tmp_reg, 0, size); } static int @@ -1140,7 +1163,7 @@ mem_op_stack(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta, bool clr_gpr, lmem_step step) { s32 off = nfp_prog->stack_frame_depth + meta->insn.off + ptr_off; - bool first = true, last; + bool first = true, narrow_ld, last; bool needs_inc = false; swreg stack_off_reg; u8 prev_gpr = 255; @@ -1186,14 +1209,23 @@ mem_op_stack(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta, needs_inc = true; } + + narrow_ld = clr_gpr && size < 8; + if (lm3) { + unsigned int nop_cnt; + emit_csr_wr(nfp_prog, imm_b(nfp_prog), NFP_CSR_ACT_LM_ADDR3); - /* For size < 4 one slot will be filled by zeroing of upper. */ - wrp_nops(nfp_prog, clr_gpr && size < 8 ? 2 : 3); + /* For size < 4 one slot will be filled by zeroing of upper, + * but be careful, that zeroing could be eliminated by zext + * optimization. + */ + nop_cnt = narrow_ld && meta->flags & FLAG_INSN_DO_ZEXT ? 2 : 3; + wrp_nops(nfp_prog, nop_cnt); } - if (clr_gpr && size < 8) - wrp_immed(nfp_prog, reg_both(gpr + 1), 0); + if (narrow_ld) + wrp_zext(nfp_prog, meta, gpr); while (size) { u32 slice_end; @@ -1294,9 +1326,10 @@ wrp_alu32_imm(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta, enum alu_op alu_op) { const struct bpf_insn *insn = &meta->insn; + u8 dst = insn->dst_reg * 2; - wrp_alu_imm(nfp_prog, insn->dst_reg * 2, alu_op, insn->imm); - wrp_immed(nfp_prog, reg_both(insn->dst_reg * 2 + 1), 0); + wrp_alu_imm(nfp_prog, dst, alu_op, insn->imm); + wrp_zext(nfp_prog, meta, dst); return 0; } @@ -1308,7 +1341,7 @@ wrp_alu32_reg(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta, u8 dst = meta->insn.dst_reg * 2, src = meta->insn.src_reg * 2; emit_alu(nfp_prog, reg_both(dst), reg_a(dst), alu_op, reg_b(src)); - wrp_immed(nfp_prog, reg_both(meta->insn.dst_reg * 2 + 1), 0); + wrp_zext(nfp_prog, meta, dst); return 0; } @@ -2385,12 +2418,14 @@ static int neg_reg(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta) u8 dst = meta->insn.dst_reg * 2; emit_alu(nfp_prog, reg_both(dst), reg_imm(0), ALU_OP_SUB, reg_b(dst)); - wrp_immed(nfp_prog, reg_both(meta->insn.dst_reg * 2 + 1), 0); + wrp_zext(nfp_prog, meta, dst); return 0; } -static int __ashr_imm(struct nfp_prog *nfp_prog, u8 dst, u8 shift_amt) +static int +__ashr_imm(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta, u8 dst, + u8 shift_amt) { if (shift_amt) { /* Set signedness bit (MSB of result). */ @@ -2399,7 +2434,7 @@ static int __ashr_imm(struct nfp_prog *nfp_prog, u8 dst, u8 shift_amt) emit_shf(nfp_prog, reg_both(dst), reg_none(), SHF_OP_ASHR, reg_b(dst), SHF_SC_R_SHF, shift_amt); } - wrp_immed(nfp_prog, reg_both(dst + 1), 0); + wrp_zext(nfp_prog, meta, dst); return 0; } @@ -2414,7 +2449,7 @@ static int ashr_reg(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta) umin = meta->umin_src; umax = meta->umax_src; if (umin == umax) - return __ashr_imm(nfp_prog, dst, umin); + return __ashr_imm(nfp_prog, meta, dst, umin); src = insn->src_reg * 2; /* NOTE: the first insn will set both indirect shift amount (source A) @@ -2423,7 +2458,7 @@ static int ashr_reg(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta) emit_alu(nfp_prog, reg_none(), reg_a(src), ALU_OP_OR, reg_b(dst)); emit_shf_indir(nfp_prog, reg_both(dst), reg_none(), SHF_OP_ASHR, reg_b(dst), SHF_SC_R_SHF); - wrp_immed(nfp_prog, reg_both(dst + 1), 0); + wrp_zext(nfp_prog, meta, dst); return 0; } @@ -2433,15 +2468,17 @@ static int ashr_imm(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta) const struct bpf_insn *insn = &meta->insn; u8 dst = insn->dst_reg * 2; - return __ashr_imm(nfp_prog, dst, insn->imm); + return __ashr_imm(nfp_prog, meta, dst, insn->imm); } -static int __shr_imm(struct nfp_prog *nfp_prog, u8 dst, u8 shift_amt) +static int +__shr_imm(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta, u8 dst, + u8 shift_amt) { if (shift_amt) emit_shf(nfp_prog, reg_both(dst), reg_none(), SHF_OP_NONE, reg_b(dst), SHF_SC_R_SHF, shift_amt); - wrp_immed(nfp_prog, reg_both(dst + 1), 0); + wrp_zext(nfp_prog, meta, dst); return 0; } @@ -2450,7 +2487,7 @@ static int shr_imm(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta) const struct bpf_insn *insn = &meta->insn; u8 dst = insn->dst_reg * 2; - return __shr_imm(nfp_prog, dst, insn->imm); + return __shr_imm(nfp_prog, meta, dst, insn->imm); } static int shr_reg(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta) @@ -2463,22 +2500,24 @@ static int shr_reg(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta) umin = meta->umin_src; umax = meta->umax_src; if (umin == umax) - return __shr_imm(nfp_prog, dst, umin); + return __shr_imm(nfp_prog, meta, dst, umin); src = insn->src_reg * 2; emit_alu(nfp_prog, reg_none(), reg_a(src), ALU_OP_OR, reg_imm(0)); emit_shf_indir(nfp_prog, reg_both(dst), reg_none(), SHF_OP_NONE, reg_b(dst), SHF_SC_R_SHF); - wrp_immed(nfp_prog, reg_both(dst + 1), 0); + wrp_zext(nfp_prog, meta, dst); return 0; } -static int __shl_imm(struct nfp_prog *nfp_prog, u8 dst, u8 shift_amt) +static int +__shl_imm(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta, u8 dst, + u8 shift_amt) { if (shift_amt) emit_shf(nfp_prog, reg_both(dst), reg_none(), SHF_OP_NONE, reg_b(dst), SHF_SC_L_SHF, shift_amt); - wrp_immed(nfp_prog, reg_both(dst + 1), 0); + wrp_zext(nfp_prog, meta, dst); return 0; } @@ -2487,7 +2526,7 @@ static int shl_imm(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta) const struct bpf_insn *insn = &meta->insn; u8 dst = insn->dst_reg * 2; - return __shl_imm(nfp_prog, dst, insn->imm); + return __shl_imm(nfp_prog, meta, dst, insn->imm); } static int shl_reg(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta) @@ -2500,11 +2539,11 @@ static int shl_reg(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta) umin = meta->umin_src; umax = meta->umax_src; if (umin == umax) - return __shl_imm(nfp_prog, dst, umin); + return __shl_imm(nfp_prog, meta, dst, umin); src = insn->src_reg * 2; shl_reg64_lt32_low(nfp_prog, dst, src); - wrp_immed(nfp_prog, reg_both(dst + 1), 0); + wrp_zext(nfp_prog, meta, dst); return 0; } @@ -2566,34 +2605,34 @@ static int imm_ld8(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta) static int data_ld1(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta) { - return construct_data_ld(nfp_prog, meta->insn.imm, 1); + return construct_data_ld(nfp_prog, meta, meta->insn.imm, 1); } static int data_ld2(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta) { - return construct_data_ld(nfp_prog, meta->insn.imm, 2); + return construct_data_ld(nfp_prog, meta, meta->insn.imm, 2); } static int data_ld4(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta) { - return construct_data_ld(nfp_prog, meta->insn.imm, 4); + return construct_data_ld(nfp_prog, meta, meta->insn.imm, 4); } static int data_ind_ld1(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta) { - return construct_data_ind_ld(nfp_prog, meta->insn.imm, + return construct_data_ind_ld(nfp_prog, meta, meta->insn.imm, meta->insn.src_reg * 2, 1); } static int data_ind_ld2(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta) { - return construct_data_ind_ld(nfp_prog, meta->insn.imm, + return construct_data_ind_ld(nfp_prog, meta, meta->insn.imm, meta->insn.src_reg * 2, 2); } static int data_ind_ld4(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta) { - return construct_data_ind_ld(nfp_prog, meta->insn.imm, + return construct_data_ind_ld(nfp_prog, meta, meta->insn.imm, meta->insn.src_reg * 2, 4); } @@ -2671,7 +2710,7 @@ mem_ldx_data(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta, tmp_reg = re_load_imm_any(nfp_prog, meta->insn.off, imm_b(nfp_prog)); - return data_ld_host_order_addr32(nfp_prog, meta->insn.src_reg * 2, + return data_ld_host_order_addr32(nfp_prog, meta, meta->insn.src_reg * 2, tmp_reg, meta->insn.dst_reg * 2, size); } @@ -2683,7 +2722,7 @@ mem_ldx_emem(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta, tmp_reg = re_load_imm_any(nfp_prog, meta->insn.off, imm_b(nfp_prog)); - return data_ld_host_order_addr40(nfp_prog, meta->insn.src_reg * 2, + return data_ld_host_order_addr40(nfp_prog, meta, meta->insn.src_reg * 2, tmp_reg, meta->insn.dst_reg * 2, size); } @@ -2744,7 +2783,7 @@ mem_ldx_data_from_pktcache_unaligned(struct nfp_prog *nfp_prog, wrp_reg_subpart(nfp_prog, dst_lo, src_lo, len_lo, off); if (!len_mid) { - wrp_immed(nfp_prog, dst_hi, 0); + wrp_zext(nfp_prog, meta, dst_gpr); return 0; } @@ -2752,7 +2791,7 @@ mem_ldx_data_from_pktcache_unaligned(struct nfp_prog *nfp_prog, if (size <= REG_WIDTH) { wrp_reg_or_subpart(nfp_prog, dst_lo, src_mid, len_mid, len_lo); - wrp_immed(nfp_prog, dst_hi, 0); + wrp_zext(nfp_prog, meta, dst_gpr); } else { swreg src_hi = reg_xfer(idx + 2); @@ -2783,10 +2822,10 @@ mem_ldx_data_from_pktcache_aligned(struct nfp_prog *nfp_prog, if (size < REG_WIDTH) { wrp_reg_subpart(nfp_prog, dst_lo, src_lo, size, 0); - wrp_immed(nfp_prog, dst_hi, 0); + wrp_zext(nfp_prog, meta, dst_gpr); } else if (size == REG_WIDTH) { wrp_mov(nfp_prog, dst_lo, src_lo); - wrp_immed(nfp_prog, dst_hi, 0); + wrp_zext(nfp_prog, meta, dst_gpr); } else { swreg src_hi = reg_xfer(idx + 1); @@ -3913,7 +3952,7 @@ static void nfp_bpf_opt_neg_add_sub(struct nfp_prog *nfp_prog) static void nfp_bpf_opt_ld_mask(struct nfp_prog *nfp_prog) { struct nfp_insn_meta *meta1, *meta2; - const s32 exp_mask[] = { + static const s32 exp_mask[] = { [BPF_B] = 0x000000ffU, [BPF_H] = 0x0000ffffU, [BPF_W] = 0xffffffffU, diff --git a/drivers/net/ethernet/netronome/nfp/bpf/main.c b/drivers/net/ethernet/netronome/nfp/bpf/main.c index 275de9f4c61c..8f732771d3fa 100644 --- a/drivers/net/ethernet/netronome/nfp/bpf/main.c +++ b/drivers/net/ethernet/netronome/nfp/bpf/main.c @@ -160,35 +160,19 @@ static int nfp_bpf_setup_tc_block_cb(enum tc_setup_type type, return 0; } -static int nfp_bpf_setup_tc_block(struct net_device *netdev, - struct tc_block_offload *f) -{ - struct nfp_net *nn = netdev_priv(netdev); - - if (f->binder_type != TCF_BLOCK_BINDER_TYPE_CLSACT_INGRESS) - return -EOPNOTSUPP; - - switch (f->command) { - case TC_BLOCK_BIND: - return tcf_block_cb_register(f->block, - nfp_bpf_setup_tc_block_cb, - nn, nn, f->extack); - case TC_BLOCK_UNBIND: - tcf_block_cb_unregister(f->block, - nfp_bpf_setup_tc_block_cb, - nn); - return 0; - default: - return -EOPNOTSUPP; - } -} +static LIST_HEAD(nfp_bpf_block_cb_list); static int nfp_bpf_setup_tc(struct nfp_app *app, struct net_device *netdev, enum tc_setup_type type, void *type_data) { + struct nfp_net *nn = netdev_priv(netdev); + switch (type) { case TC_SETUP_BLOCK: - return nfp_bpf_setup_tc_block(netdev, type_data); + return flow_block_cb_setup_simple(type_data, + &nfp_bpf_block_cb_list, + nfp_bpf_setup_tc_block_cb, + nn, nn, true); default: return -EOPNOTSUPP; } @@ -316,6 +300,14 @@ nfp_bpf_parse_cap_adjust_tail(struct nfp_app_bpf *bpf, void __iomem *value, } static int +nfp_bpf_parse_cap_cmsg_multi_ent(struct nfp_app_bpf *bpf, void __iomem *value, + u32 length) +{ + bpf->cmsg_multi_ent = true; + return 0; +} + +static int nfp_bpf_parse_cap_abi_version(struct nfp_app_bpf *bpf, void __iomem *value, u32 length) { @@ -391,6 +383,11 @@ static int nfp_bpf_parse_capabilities(struct nfp_app *app) length)) goto err_release_free; break; + case NFP_BPF_CAP_TYPE_CMSG_MULTI_ENT: + if (nfp_bpf_parse_cap_cmsg_multi_ent(app->priv, value, + length)) + goto err_release_free; + break; default: nfp_dbg(cpp, "unknown BPF capability: %d\n", type); break; @@ -431,6 +428,25 @@ static void nfp_bpf_ndo_uninit(struct nfp_app *app, struct net_device *netdev) bpf_offload_dev_netdev_unregister(bpf->bpf_dev, netdev); } +static int nfp_bpf_start(struct nfp_app *app) +{ + struct nfp_app_bpf *bpf = app->priv; + + if (app->ctrl->dp.mtu < nfp_bpf_ctrl_cmsg_min_mtu(bpf)) { + nfp_err(bpf->app->cpp, + "ctrl channel MTU below min required %u < %u\n", + app->ctrl->dp.mtu, nfp_bpf_ctrl_cmsg_min_mtu(bpf)); + return -EINVAL; + } + + if (bpf->cmsg_multi_ent) + bpf->cmsg_cache_cnt = nfp_bpf_ctrl_cmsg_cache_cnt(bpf); + else + bpf->cmsg_cache_cnt = 1; + + return 0; +} + static int nfp_bpf_init(struct nfp_app *app) { struct nfp_app_bpf *bpf; @@ -442,14 +458,16 @@ static int nfp_bpf_init(struct nfp_app *app) bpf->app = app; app->priv = bpf; - skb_queue_head_init(&bpf->cmsg_replies); - init_waitqueue_head(&bpf->cmsg_wq); INIT_LIST_HEAD(&bpf->map_list); - err = rhashtable_init(&bpf->maps_neutral, &nfp_bpf_maps_neutral_params); + err = nfp_ccm_init(&bpf->ccm, app); if (err) goto err_free_bpf; + err = rhashtable_init(&bpf->maps_neutral, &nfp_bpf_maps_neutral_params); + if (err) + goto err_clean_ccm; + nfp_bpf_init_capabilities(bpf); err = nfp_bpf_parse_capabilities(app); @@ -474,6 +492,8 @@ static int nfp_bpf_init(struct nfp_app *app) err_free_neutral_maps: rhashtable_destroy(&bpf->maps_neutral); +err_clean_ccm: + nfp_ccm_clean(&bpf->ccm); err_free_bpf: kfree(bpf); return err; @@ -484,7 +504,7 @@ static void nfp_bpf_clean(struct nfp_app *app) struct nfp_app_bpf *bpf = app->priv; bpf_offload_dev_destroy(bpf->bpf_dev); - WARN_ON(!skb_queue_empty(&bpf->cmsg_replies)); + nfp_ccm_clean(&bpf->ccm); WARN_ON(!list_empty(&bpf->map_list)); WARN_ON(bpf->maps_in_use || bpf->map_elems_in_use); rhashtable_free_and_destroy(&bpf->maps_neutral, @@ -500,6 +520,7 @@ const struct nfp_app_type app_bpf = { .init = nfp_bpf_init, .clean = nfp_bpf_clean, + .start = nfp_bpf_start, .check_mtu = nfp_bpf_check_mtu, diff --git a/drivers/net/ethernet/netronome/nfp/bpf/main.h b/drivers/net/ethernet/netronome/nfp/bpf/main.h index b25a48218bcf..fac9c6f9e197 100644 --- a/drivers/net/ethernet/netronome/nfp/bpf/main.h +++ b/drivers/net/ethernet/netronome/nfp/bpf/main.h @@ -14,6 +14,7 @@ #include <linux/types.h> #include <linux/wait.h> +#include "../ccm.h" #include "../nfp_asm.h" #include "fw.h" @@ -84,16 +85,10 @@ enum pkt_vec { /** * struct nfp_app_bpf - bpf app priv structure * @app: backpointer to the app + * @ccm: common control message handler data * * @bpf_dev: BPF offload device handle * - * @tag_allocator: bitmap of control message tags in use - * @tag_alloc_next: next tag bit to allocate - * @tag_alloc_last: next tag bit to be freed - * - * @cmsg_replies: received cmsg replies waiting to be consumed - * @cmsg_wq: work queue for waiting for cmsg replies - * * @cmsg_key_sz: size of key in cmsg element array * @cmsg_val_sz: size of value in cmsg element array * @@ -104,6 +99,7 @@ enum pkt_vec { * @maps_neutral: hash table of offload-neutral maps (on pointer) * * @abi_version: global BPF ABI version + * @cmsg_cache_cnt: number of entries to read for caching * * @adjust_head: adjust head capability * @adjust_head.flags: extra flags for adjust head @@ -129,22 +125,19 @@ enum pkt_vec { * @pseudo_random: FW initialized the pseudo-random machinery (CSRs) * @queue_select: BPF can set the RX queue ID in packet vector * @adjust_tail: BPF can simply trunc packet size for adjust tail + * @cmsg_multi_ent: FW can pack multiple map entries in a single cmsg */ struct nfp_app_bpf { struct nfp_app *app; + struct nfp_ccm ccm; struct bpf_offload_dev *bpf_dev; - DECLARE_BITMAP(tag_allocator, U16_MAX + 1); - u16 tag_alloc_next; - u16 tag_alloc_last; - - struct sk_buff_head cmsg_replies; - struct wait_queue_head cmsg_wq; - unsigned int cmsg_key_sz; unsigned int cmsg_val_sz; + unsigned int cmsg_cache_cnt; + struct list_head map_list; unsigned int maps_in_use; unsigned int map_elems_in_use; @@ -180,6 +173,7 @@ struct nfp_app_bpf { bool pseudo_random; bool queue_select; bool adjust_tail; + bool cmsg_multi_ent; }; enum nfp_bpf_map_use { @@ -194,11 +188,21 @@ struct nfp_bpf_map_word { unsigned char non_zero_update :1; }; +#define NFP_BPF_MAP_CACHE_CNT 4U +#define NFP_BPF_MAP_CACHE_TIME_NS (250 * 1000) + /** * struct nfp_bpf_map - private per-map data attached to BPF maps for offload * @offmap: pointer to the offloaded BPF map * @bpf: back pointer to bpf app private structure * @tid: table id identifying map on datapath + * + * @cache_lock: protects @cache_blockers, @cache_to, @cache + * @cache_blockers: number of ops in flight which block caching + * @cache_gen: counter incremented by every blocker on exit + * @cache_to: time when cache will no longer be valid (ns) + * @cache: skb with cached response + * * @l: link on the nfp_app_bpf->map_list list * @use_map: map of how the value is used (in 4B chunks) */ @@ -206,6 +210,13 @@ struct nfp_bpf_map { struct bpf_offloaded_map *offmap; struct nfp_app_bpf *bpf; u32 tid; + + spinlock_t cache_lock; + u32 cache_blockers; + u32 cache_gen; + u64 cache_to; + struct sk_buff *cache; + struct list_head l; struct nfp_bpf_map_word use_map[]; }; @@ -249,6 +260,8 @@ struct nfp_bpf_reg_state { #define FLAG_INSN_SKIP_PREC_DEPENDENT BIT(4) /* Instruction is optimized by the verifier */ #define FLAG_INSN_SKIP_VERIFIER_OPT BIT(5) +/* Instruction needs to zero extend to high 32-bit */ +#define FLAG_INSN_DO_ZEXT BIT(6) #define FLAG_INSN_SKIP_MASK (FLAG_INSN_SKIP_NOOP | \ FLAG_INSN_SKIP_PREC_DEPENDENT | \ @@ -573,7 +586,9 @@ nfp_bpf_goto_meta(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta, void *nfp_bpf_relo_for_vnic(struct nfp_prog *nfp_prog, struct nfp_bpf_vnic *bv); +unsigned int nfp_bpf_ctrl_cmsg_min_mtu(struct nfp_app_bpf *bpf); unsigned int nfp_bpf_ctrl_cmsg_mtu(struct nfp_app_bpf *bpf); +unsigned int nfp_bpf_ctrl_cmsg_cache_cnt(struct nfp_app_bpf *bpf); long long int nfp_bpf_ctrl_alloc_map(struct nfp_app_bpf *bpf, struct bpf_map *map); void diff --git a/drivers/net/ethernet/netronome/nfp/bpf/offload.c b/drivers/net/ethernet/netronome/nfp/bpf/offload.c index 15dce97650a5..95a0d3910e31 100644 --- a/drivers/net/ethernet/netronome/nfp/bpf/offload.c +++ b/drivers/net/ethernet/netronome/nfp/bpf/offload.c @@ -22,6 +22,7 @@ #include <net/tc_act/tc_mirred.h> #include "main.h" +#include "../ccm.h" #include "../nfp_app.h" #include "../nfp_net_ctrl.h" #include "../nfp_net.h" @@ -45,9 +46,7 @@ nfp_map_ptr_record(struct nfp_app_bpf *bpf, struct nfp_prog *nfp_prog, /* Grab a single ref to the map for our record. The prog destroy ndo * happens after free_used_maps(). */ - map = bpf_map_inc(map, false); - if (IS_ERR(map)) - return PTR_ERR(map); + bpf_map_inc(map); record = kmalloc(sizeof(*record), GFP_KERNEL); if (!record) { @@ -384,6 +383,7 @@ nfp_bpf_map_alloc(struct nfp_app_bpf *bpf, struct bpf_offloaded_map *offmap) offmap->dev_priv = nfp_map; nfp_map->offmap = offmap; nfp_map->bpf = bpf; + spin_lock_init(&nfp_map->cache_lock); res = nfp_bpf_ctrl_alloc_map(bpf, &offmap->map); if (res < 0) { @@ -406,6 +406,8 @@ nfp_bpf_map_free(struct nfp_app_bpf *bpf, struct bpf_offloaded_map *offmap) struct nfp_bpf_map *nfp_map = offmap->dev_priv; nfp_bpf_ctrl_free_map(bpf, nfp_map); + dev_consume_skb_any(nfp_map->cache); + WARN_ON_ONCE(nfp_map->cache_blockers); list_del_init(&nfp_map->l); bpf->map_elems_in_use -= offmap->map.max_entries; bpf->maps_in_use--; @@ -452,12 +454,12 @@ int nfp_bpf_event_output(struct nfp_app_bpf *bpf, const void *data, if (len < sizeof(struct cmsg_bpf_event) + pkt_size + data_size) return -EINVAL; - if (cbe->hdr.ver != CMSG_MAP_ABI_VERSION) + if (cbe->hdr.ver != NFP_CCM_ABI_VERSION) return -EINVAL; rcu_read_lock(); - record = rhashtable_lookup_fast(&bpf->maps_neutral, &map_id, - nfp_bpf_maps_neutral_params); + record = rhashtable_lookup(&bpf->maps_neutral, &map_id, + nfp_bpf_maps_neutral_params); if (!record || map_id_full > U32_MAX) { rcu_read_unlock(); cmsg_warn(bpf, "perf event: map id %lld (0x%llx) not recognized, dropping event\n", diff --git a/drivers/net/ethernet/netronome/nfp/bpf/verifier.c b/drivers/net/ethernet/netronome/nfp/bpf/verifier.c index 36f56eb4cbe2..e92ee510fd52 100644 --- a/drivers/net/ethernet/netronome/nfp/bpf/verifier.c +++ b/drivers/net/ethernet/netronome/nfp/bpf/verifier.c @@ -744,6 +744,17 @@ continue_subprog: goto continue_subprog; } +static void nfp_bpf_insn_flag_zext(struct nfp_prog *nfp_prog, + struct bpf_insn_aux_data *aux) +{ + struct nfp_insn_meta *meta; + + list_for_each_entry(meta, &nfp_prog->insns, l) { + if (aux[meta->n].zext_dst) + meta->flags |= FLAG_INSN_DO_ZEXT; + } +} + int nfp_bpf_finalize(struct bpf_verifier_env *env) { struct bpf_subprog_info *info; @@ -784,6 +795,7 @@ int nfp_bpf_finalize(struct bpf_verifier_env *env) return -EOPNOTSUPP; } + nfp_bpf_insn_flag_zext(nfp_prog, env->insn_aux_data); return 0; } diff --git a/drivers/net/ethernet/netronome/nfp/ccm.c b/drivers/net/ethernet/netronome/nfp/ccm.c new file mode 100644 index 000000000000..71afd111bae3 --- /dev/null +++ b/drivers/net/ethernet/netronome/nfp/ccm.c @@ -0,0 +1,217 @@ +// SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) +/* Copyright (C) 2016-2019 Netronome Systems, Inc. */ + +#include <linux/bitops.h> + +#include "ccm.h" +#include "nfp_app.h" +#include "nfp_net.h" + +#define ccm_warn(app, msg...) nn_dp_warn(&(app)->ctrl->dp, msg) + +#define NFP_CCM_TAG_ALLOC_SPAN (U16_MAX / 4) + +static bool nfp_ccm_all_tags_busy(struct nfp_ccm *ccm) +{ + u16 used_tags; + + used_tags = ccm->tag_alloc_next - ccm->tag_alloc_last; + + return used_tags > NFP_CCM_TAG_ALLOC_SPAN; +} + +static int nfp_ccm_alloc_tag(struct nfp_ccm *ccm) +{ + /* CCM is for FW communication which is request-reply. To make sure + * we don't reuse the message ID too early after timeout - limit the + * number of requests in flight. + */ + if (unlikely(nfp_ccm_all_tags_busy(ccm))) { + ccm_warn(ccm->app, "all FW request contexts busy!\n"); + return -EAGAIN; + } + + WARN_ON(__test_and_set_bit(ccm->tag_alloc_next, ccm->tag_allocator)); + return ccm->tag_alloc_next++; +} + +static void nfp_ccm_free_tag(struct nfp_ccm *ccm, u16 tag) +{ + WARN_ON(!__test_and_clear_bit(tag, ccm->tag_allocator)); + + while (!test_bit(ccm->tag_alloc_last, ccm->tag_allocator) && + ccm->tag_alloc_last != ccm->tag_alloc_next) + ccm->tag_alloc_last++; +} + +static struct sk_buff *__nfp_ccm_reply(struct nfp_ccm *ccm, u16 tag) +{ + unsigned int msg_tag; + struct sk_buff *skb; + + skb_queue_walk(&ccm->replies, skb) { + msg_tag = nfp_ccm_get_tag(skb); + if (msg_tag == tag) { + nfp_ccm_free_tag(ccm, tag); + __skb_unlink(skb, &ccm->replies); + return skb; + } + } + + return NULL; +} + +static struct sk_buff * +nfp_ccm_reply(struct nfp_ccm *ccm, struct nfp_app *app, u16 tag) +{ + struct sk_buff *skb; + + nfp_ctrl_lock(app->ctrl); + skb = __nfp_ccm_reply(ccm, tag); + nfp_ctrl_unlock(app->ctrl); + + return skb; +} + +static struct sk_buff * +nfp_ccm_reply_drop_tag(struct nfp_ccm *ccm, struct nfp_app *app, u16 tag) +{ + struct sk_buff *skb; + + nfp_ctrl_lock(app->ctrl); + skb = __nfp_ccm_reply(ccm, tag); + if (!skb) + nfp_ccm_free_tag(ccm, tag); + nfp_ctrl_unlock(app->ctrl); + + return skb; +} + +static struct sk_buff * +nfp_ccm_wait_reply(struct nfp_ccm *ccm, struct nfp_app *app, + enum nfp_ccm_type type, int tag) +{ + struct sk_buff *skb; + int i, err; + + for (i = 0; i < 50; i++) { + udelay(4); + skb = nfp_ccm_reply(ccm, app, tag); + if (skb) + return skb; + } + + err = wait_event_interruptible_timeout(ccm->wq, + skb = nfp_ccm_reply(ccm, app, + tag), + msecs_to_jiffies(5000)); + /* We didn't get a response - try last time and atomically drop + * the tag even if no response is matched. + */ + if (!skb) + skb = nfp_ccm_reply_drop_tag(ccm, app, tag); + if (err < 0) { + ccm_warn(app, "%s waiting for response to 0x%02x: %d\n", + err == ERESTARTSYS ? "interrupted" : "error", + type, err); + return ERR_PTR(err); + } + if (!skb) { + ccm_warn(app, "timeout waiting for response to 0x%02x\n", type); + return ERR_PTR(-ETIMEDOUT); + } + + return skb; +} + +struct sk_buff * +nfp_ccm_communicate(struct nfp_ccm *ccm, struct sk_buff *skb, + enum nfp_ccm_type type, unsigned int reply_size) +{ + struct nfp_app *app = ccm->app; + struct nfp_ccm_hdr *hdr; + int reply_type, tag; + + nfp_ctrl_lock(app->ctrl); + tag = nfp_ccm_alloc_tag(ccm); + if (tag < 0) { + nfp_ctrl_unlock(app->ctrl); + dev_kfree_skb_any(skb); + return ERR_PTR(tag); + } + + hdr = (void *)skb->data; + hdr->ver = NFP_CCM_ABI_VERSION; + hdr->type = type; + hdr->tag = cpu_to_be16(tag); + + __nfp_app_ctrl_tx(app, skb); + + nfp_ctrl_unlock(app->ctrl); + + skb = nfp_ccm_wait_reply(ccm, app, type, tag); + if (IS_ERR(skb)) + return skb; + + reply_type = nfp_ccm_get_type(skb); + if (reply_type != __NFP_CCM_REPLY(type)) { + ccm_warn(app, "cmsg drop - wrong type 0x%02x != 0x%02lx!\n", + reply_type, __NFP_CCM_REPLY(type)); + goto err_free; + } + /* 0 reply_size means caller will do the validation */ + if (reply_size && skb->len != reply_size) { + ccm_warn(app, "cmsg drop - type 0x%02x wrong size %d != %d!\n", + type, skb->len, reply_size); + goto err_free; + } + + return skb; +err_free: + dev_kfree_skb_any(skb); + return ERR_PTR(-EIO); +} + +void nfp_ccm_rx(struct nfp_ccm *ccm, struct sk_buff *skb) +{ + struct nfp_app *app = ccm->app; + unsigned int tag; + + if (unlikely(skb->len < sizeof(struct nfp_ccm_hdr))) { + ccm_warn(app, "cmsg drop - too short %d!\n", skb->len); + goto err_free; + } + + nfp_ctrl_lock(app->ctrl); + + tag = nfp_ccm_get_tag(skb); + if (unlikely(!test_bit(tag, ccm->tag_allocator))) { + ccm_warn(app, "cmsg drop - no one is waiting for tag %u!\n", + tag); + goto err_unlock; + } + + __skb_queue_tail(&ccm->replies, skb); + wake_up_interruptible_all(&ccm->wq); + + nfp_ctrl_unlock(app->ctrl); + return; + +err_unlock: + nfp_ctrl_unlock(app->ctrl); +err_free: + dev_kfree_skb_any(skb); +} + +int nfp_ccm_init(struct nfp_ccm *ccm, struct nfp_app *app) +{ + ccm->app = app; + skb_queue_head_init(&ccm->replies); + init_waitqueue_head(&ccm->wq); + return 0; +} + +void nfp_ccm_clean(struct nfp_ccm *ccm) +{ + WARN_ON(!skb_queue_empty(&ccm->replies)); +} diff --git a/drivers/net/ethernet/netronome/nfp/ccm.h b/drivers/net/ethernet/netronome/nfp/ccm.h new file mode 100644 index 000000000000..a460c75522be --- /dev/null +++ b/drivers/net/ethernet/netronome/nfp/ccm.h @@ -0,0 +1,131 @@ +/* SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) */ +/* Copyright (C) 2016-2019 Netronome Systems, Inc. */ + +#ifndef NFP_CCM_H +#define NFP_CCM_H 1 + +#include <linux/bitmap.h> +#include <linux/skbuff.h> +#include <linux/wait.h> + +struct nfp_app; +struct nfp_net; + +/* Firmware ABI */ + +enum nfp_ccm_type { + NFP_CCM_TYPE_BPF_MAP_ALLOC = 1, + NFP_CCM_TYPE_BPF_MAP_FREE = 2, + NFP_CCM_TYPE_BPF_MAP_LOOKUP = 3, + NFP_CCM_TYPE_BPF_MAP_UPDATE = 4, + NFP_CCM_TYPE_BPF_MAP_DELETE = 5, + NFP_CCM_TYPE_BPF_MAP_GETNEXT = 6, + NFP_CCM_TYPE_BPF_MAP_GETFIRST = 7, + NFP_CCM_TYPE_BPF_BPF_EVENT = 8, + NFP_CCM_TYPE_CRYPTO_RESET = 9, + NFP_CCM_TYPE_CRYPTO_ADD = 10, + NFP_CCM_TYPE_CRYPTO_DEL = 11, + NFP_CCM_TYPE_CRYPTO_UPDATE = 12, + __NFP_CCM_TYPE_MAX, +}; + +#define NFP_CCM_ABI_VERSION 1 + +#define NFP_CCM_TYPE_REPLY_BIT 7 +#define __NFP_CCM_REPLY(req) (BIT(NFP_CCM_TYPE_REPLY_BIT) | (req)) + +struct nfp_ccm_hdr { + union { + struct { + u8 type; + u8 ver; + __be16 tag; + }; + __be32 raw; + }; +}; + +static inline u8 nfp_ccm_get_type(struct sk_buff *skb) +{ + struct nfp_ccm_hdr *hdr; + + hdr = (struct nfp_ccm_hdr *)skb->data; + + return hdr->type; +} + +static inline __be16 __nfp_ccm_get_tag(struct sk_buff *skb) +{ + struct nfp_ccm_hdr *hdr; + + hdr = (struct nfp_ccm_hdr *)skb->data; + + return hdr->tag; +} + +static inline unsigned int nfp_ccm_get_tag(struct sk_buff *skb) +{ + return be16_to_cpu(__nfp_ccm_get_tag(skb)); +} + +#define NFP_NET_MBOX_TLV_TYPE GENMASK(31, 16) +#define NFP_NET_MBOX_TLV_LEN GENMASK(15, 0) + +enum nfp_ccm_mbox_tlv_type { + NFP_NET_MBOX_TLV_TYPE_UNKNOWN = 0, + NFP_NET_MBOX_TLV_TYPE_END = 1, + NFP_NET_MBOX_TLV_TYPE_MSG = 2, + NFP_NET_MBOX_TLV_TYPE_MSG_NOSUP = 3, + NFP_NET_MBOX_TLV_TYPE_RESV = 4, +}; + +/* Implementation */ + +/** + * struct nfp_ccm - common control message handling + * @app: APP handle + * + * @tag_allocator: bitmap of control message tags in use + * @tag_alloc_next: next tag bit to allocate + * @tag_alloc_last: next tag bit to be freed + * + * @replies: received cmsg replies waiting to be consumed + * @wq: work queue for waiting for cmsg replies + */ +struct nfp_ccm { + struct nfp_app *app; + + DECLARE_BITMAP(tag_allocator, U16_MAX + 1); + u16 tag_alloc_next; + u16 tag_alloc_last; + + struct sk_buff_head replies; + wait_queue_head_t wq; +}; + +int nfp_ccm_init(struct nfp_ccm *ccm, struct nfp_app *app); +void nfp_ccm_clean(struct nfp_ccm *ccm); +void nfp_ccm_rx(struct nfp_ccm *ccm, struct sk_buff *skb); +struct sk_buff * +nfp_ccm_communicate(struct nfp_ccm *ccm, struct sk_buff *skb, + enum nfp_ccm_type type, unsigned int reply_size); + +int nfp_ccm_mbox_alloc(struct nfp_net *nn); +void nfp_ccm_mbox_free(struct nfp_net *nn); +int nfp_ccm_mbox_init(struct nfp_net *nn); +void nfp_ccm_mbox_clean(struct nfp_net *nn); +bool nfp_ccm_mbox_fits(struct nfp_net *nn, unsigned int size); +struct sk_buff * +nfp_ccm_mbox_msg_alloc(struct nfp_net *nn, unsigned int req_size, + unsigned int reply_size, gfp_t flags); +int __nfp_ccm_mbox_communicate(struct nfp_net *nn, struct sk_buff *skb, + enum nfp_ccm_type type, + unsigned int reply_size, + unsigned int max_reply_size, bool critical); +int nfp_ccm_mbox_communicate(struct nfp_net *nn, struct sk_buff *skb, + enum nfp_ccm_type type, + unsigned int reply_size, + unsigned int max_reply_size); +int nfp_ccm_mbox_post(struct nfp_net *nn, struct sk_buff *skb, + enum nfp_ccm_type type, unsigned int max_reply_size); +#endif diff --git a/drivers/net/ethernet/netronome/nfp/ccm_mbox.c b/drivers/net/ethernet/netronome/nfp/ccm_mbox.c new file mode 100644 index 000000000000..f0783aa9e66e --- /dev/null +++ b/drivers/net/ethernet/netronome/nfp/ccm_mbox.c @@ -0,0 +1,743 @@ +// SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) +/* Copyright (C) 2019 Netronome Systems, Inc. */ + +#include <linux/bitfield.h> +#include <linux/io.h> +#include <linux/skbuff.h> + +#include "ccm.h" +#include "nfp_net.h" + +/* CCM messages via the mailbox. CMSGs get wrapped into simple TLVs + * and copied into the mailbox. Multiple messages can be copied to + * form a batch. Threads come in with CMSG formed in an skb, then + * enqueue that skb onto the request queue. If threads skb is first + * in queue this thread will handle the mailbox operation. It copies + * up to 64 messages into the mailbox (making sure that both requests + * and replies will fit. After FW is done processing the batch it + * copies the data out and wakes waiting threads. + * If a thread is waiting it either gets its the message completed + * (response is copied into the same skb as the request, overwriting + * it), or becomes the first in queue. + * Completions and next-to-run are signaled via the control buffer + * to limit potential cache line bounces. + */ + +#define NFP_CCM_MBOX_BATCH_LIMIT 64 +#define NFP_CCM_TIMEOUT (NFP_NET_POLL_TIMEOUT * 1000) +#define NFP_CCM_MAX_QLEN 1024 + +enum nfp_net_mbox_cmsg_state { + NFP_NET_MBOX_CMSG_STATE_QUEUED, + NFP_NET_MBOX_CMSG_STATE_NEXT, + NFP_NET_MBOX_CMSG_STATE_BUSY, + NFP_NET_MBOX_CMSG_STATE_REPLY_FOUND, + NFP_NET_MBOX_CMSG_STATE_DONE, +}; + +/** + * struct nfp_ccm_mbox_skb_cb - CCM mailbox specific info + * @state: processing state (/stage) of the message + * @err: error encountered during processing if any + * @max_len: max(request_len, reply_len) + * @exp_reply: expected reply length (0 means don't validate) + * @posted: the message was posted and nobody waits for the reply + */ +struct nfp_ccm_mbox_cmsg_cb { + enum nfp_net_mbox_cmsg_state state; + int err; + unsigned int max_len; + unsigned int exp_reply; + bool posted; +}; + +static u32 nfp_ccm_mbox_max_msg(struct nfp_net *nn) +{ + return round_down(nn->tlv_caps.mbox_len, 4) - + NFP_NET_CFG_MBOX_SIMPLE_VAL - /* common mbox command header */ + 4 * 2; /* Msg TLV plus End TLV headers */ +} + +static void +nfp_ccm_mbox_msg_init(struct sk_buff *skb, unsigned int exp_reply, int max_len) +{ + struct nfp_ccm_mbox_cmsg_cb *cb = (void *)skb->cb; + + cb->state = NFP_NET_MBOX_CMSG_STATE_QUEUED; + cb->err = 0; + cb->max_len = max_len; + cb->exp_reply = exp_reply; + cb->posted = false; +} + +static int nfp_ccm_mbox_maxlen(const struct sk_buff *skb) +{ + struct nfp_ccm_mbox_cmsg_cb *cb = (void *)skb->cb; + + return cb->max_len; +} + +static bool nfp_ccm_mbox_done(struct sk_buff *skb) +{ + struct nfp_ccm_mbox_cmsg_cb *cb = (void *)skb->cb; + + return cb->state == NFP_NET_MBOX_CMSG_STATE_DONE; +} + +static bool nfp_ccm_mbox_in_progress(struct sk_buff *skb) +{ + struct nfp_ccm_mbox_cmsg_cb *cb = (void *)skb->cb; + + return cb->state != NFP_NET_MBOX_CMSG_STATE_QUEUED && + cb->state != NFP_NET_MBOX_CMSG_STATE_NEXT; +} + +static void nfp_ccm_mbox_set_busy(struct sk_buff *skb) +{ + struct nfp_ccm_mbox_cmsg_cb *cb = (void *)skb->cb; + + cb->state = NFP_NET_MBOX_CMSG_STATE_BUSY; +} + +static bool nfp_ccm_mbox_is_posted(struct sk_buff *skb) +{ + struct nfp_ccm_mbox_cmsg_cb *cb = (void *)skb->cb; + + return cb->posted; +} + +static void nfp_ccm_mbox_mark_posted(struct sk_buff *skb) +{ + struct nfp_ccm_mbox_cmsg_cb *cb = (void *)skb->cb; + + cb->posted = true; +} + +static bool nfp_ccm_mbox_is_first(struct nfp_net *nn, struct sk_buff *skb) +{ + return skb_queue_is_first(&nn->mbox_cmsg.queue, skb); +} + +static bool nfp_ccm_mbox_should_run(struct nfp_net *nn, struct sk_buff *skb) +{ + struct nfp_ccm_mbox_cmsg_cb *cb = (void *)skb->cb; + + return cb->state == NFP_NET_MBOX_CMSG_STATE_NEXT; +} + +static void nfp_ccm_mbox_mark_next_runner(struct nfp_net *nn) +{ + struct nfp_ccm_mbox_cmsg_cb *cb; + struct sk_buff *skb; + + skb = skb_peek(&nn->mbox_cmsg.queue); + if (!skb) + return; + + cb = (void *)skb->cb; + cb->state = NFP_NET_MBOX_CMSG_STATE_NEXT; + if (cb->posted) + queue_work(nn->mbox_cmsg.workq, &nn->mbox_cmsg.runq_work); +} + +static void +nfp_ccm_mbox_write_tlv(struct nfp_net *nn, u32 off, u32 type, u32 len) +{ + nn_writel(nn, off, + FIELD_PREP(NFP_NET_MBOX_TLV_TYPE, type) | + FIELD_PREP(NFP_NET_MBOX_TLV_LEN, len)); +} + +static void nfp_ccm_mbox_copy_in(struct nfp_net *nn, struct sk_buff *last) +{ + struct sk_buff *skb; + int reserve, i, cnt; + __be32 *data; + u32 off, len; + + off = nn->tlv_caps.mbox_off + NFP_NET_CFG_MBOX_SIMPLE_VAL; + skb = __skb_peek(&nn->mbox_cmsg.queue); + while (true) { + nfp_ccm_mbox_write_tlv(nn, off, NFP_NET_MBOX_TLV_TYPE_MSG, + skb->len); + off += 4; + + /* Write data word by word, skb->data should be aligned */ + data = (__be32 *)skb->data; + cnt = skb->len / 4; + for (i = 0 ; i < cnt; i++) { + nn_writel(nn, off, be32_to_cpu(data[i])); + off += 4; + } + if (skb->len & 3) { + __be32 tmp = 0; + + memcpy(&tmp, &data[i], skb->len & 3); + nn_writel(nn, off, be32_to_cpu(tmp)); + off += 4; + } + + /* Reserve space if reply is bigger */ + len = round_up(skb->len, 4); + reserve = nfp_ccm_mbox_maxlen(skb) - len; + if (reserve > 0) { + nfp_ccm_mbox_write_tlv(nn, off, + NFP_NET_MBOX_TLV_TYPE_RESV, + reserve); + off += 4 + reserve; + } + + if (skb == last) + break; + skb = skb_queue_next(&nn->mbox_cmsg.queue, skb); + } + + nfp_ccm_mbox_write_tlv(nn, off, NFP_NET_MBOX_TLV_TYPE_END, 0); +} + +static struct sk_buff * +nfp_ccm_mbox_find_req(struct nfp_net *nn, __be16 tag, struct sk_buff *last) +{ + struct sk_buff *skb; + + skb = __skb_peek(&nn->mbox_cmsg.queue); + while (true) { + if (__nfp_ccm_get_tag(skb) == tag) + return skb; + + if (skb == last) + return NULL; + skb = skb_queue_next(&nn->mbox_cmsg.queue, skb); + } +} + +static void nfp_ccm_mbox_copy_out(struct nfp_net *nn, struct sk_buff *last) +{ + struct nfp_ccm_mbox_cmsg_cb *cb; + u8 __iomem *data, *end; + struct sk_buff *skb; + + data = nn->dp.ctrl_bar + nn->tlv_caps.mbox_off + + NFP_NET_CFG_MBOX_SIMPLE_VAL; + end = data + nn->tlv_caps.mbox_len; + + while (true) { + unsigned int length, offset, type; + struct nfp_ccm_hdr hdr; + u32 tlv_hdr; + + tlv_hdr = readl(data); + type = FIELD_GET(NFP_NET_MBOX_TLV_TYPE, tlv_hdr); + length = FIELD_GET(NFP_NET_MBOX_TLV_LEN, tlv_hdr); + offset = data - nn->dp.ctrl_bar; + + /* Advance past the header */ + data += 4; + + if (data + length > end) { + nn_dp_warn(&nn->dp, "mailbox oversized TLV type:%d offset:%u len:%u\n", + type, offset, length); + break; + } + + if (type == NFP_NET_MBOX_TLV_TYPE_END) + break; + if (type == NFP_NET_MBOX_TLV_TYPE_RESV) + goto next_tlv; + if (type != NFP_NET_MBOX_TLV_TYPE_MSG && + type != NFP_NET_MBOX_TLV_TYPE_MSG_NOSUP) { + nn_dp_warn(&nn->dp, "mailbox unknown TLV type:%d offset:%u len:%u\n", + type, offset, length); + break; + } + + if (length < 4) { + nn_dp_warn(&nn->dp, "mailbox msg too short to contain header TLV type:%d offset:%u len:%u\n", + type, offset, length); + break; + } + + hdr.raw = cpu_to_be32(readl(data)); + + skb = nfp_ccm_mbox_find_req(nn, hdr.tag, last); + if (!skb) { + nn_dp_warn(&nn->dp, "mailbox request not found:%u\n", + be16_to_cpu(hdr.tag)); + break; + } + cb = (void *)skb->cb; + + if (type == NFP_NET_MBOX_TLV_TYPE_MSG_NOSUP) { + nn_dp_warn(&nn->dp, + "mailbox msg not supported type:%d\n", + nfp_ccm_get_type(skb)); + cb->err = -EIO; + goto next_tlv; + } + + if (hdr.type != __NFP_CCM_REPLY(nfp_ccm_get_type(skb))) { + nn_dp_warn(&nn->dp, "mailbox msg reply wrong type:%u expected:%lu\n", + hdr.type, + __NFP_CCM_REPLY(nfp_ccm_get_type(skb))); + cb->err = -EIO; + goto next_tlv; + } + if (cb->exp_reply && length != cb->exp_reply) { + nn_dp_warn(&nn->dp, "mailbox msg reply wrong size type:%u expected:%u have:%u\n", + hdr.type, length, cb->exp_reply); + cb->err = -EIO; + goto next_tlv; + } + if (length > cb->max_len) { + nn_dp_warn(&nn->dp, "mailbox msg oversized reply type:%u max:%u have:%u\n", + hdr.type, cb->max_len, length); + cb->err = -EIO; + goto next_tlv; + } + + if (!cb->posted) { + __be32 *skb_data; + int i, cnt; + + if (length <= skb->len) + __skb_trim(skb, length); + else + skb_put(skb, length - skb->len); + + /* We overcopy here slightly, but that's okay, + * the skb is large enough, and the garbage will + * be ignored (beyond skb->len). + */ + skb_data = (__be32 *)skb->data; + memcpy(skb_data, &hdr, 4); + + cnt = DIV_ROUND_UP(length, 4); + for (i = 1 ; i < cnt; i++) + skb_data[i] = cpu_to_be32(readl(data + i * 4)); + } + + cb->state = NFP_NET_MBOX_CMSG_STATE_REPLY_FOUND; +next_tlv: + data += round_up(length, 4); + if (data + 4 > end) { + nn_dp_warn(&nn->dp, + "reached end of MBOX without END TLV\n"); + break; + } + } + + smp_wmb(); /* order the skb->data vs. cb->state */ + spin_lock_bh(&nn->mbox_cmsg.queue.lock); + do { + skb = __skb_dequeue(&nn->mbox_cmsg.queue); + cb = (void *)skb->cb; + + if (cb->state != NFP_NET_MBOX_CMSG_STATE_REPLY_FOUND) { + cb->err = -ENOENT; + smp_wmb(); /* order the cb->err vs. cb->state */ + } + cb->state = NFP_NET_MBOX_CMSG_STATE_DONE; + + if (cb->posted) { + if (cb->err) + nn_dp_warn(&nn->dp, + "mailbox posted msg failed type:%u err:%d\n", + nfp_ccm_get_type(skb), cb->err); + dev_consume_skb_any(skb); + } + } while (skb != last); + + nfp_ccm_mbox_mark_next_runner(nn); + spin_unlock_bh(&nn->mbox_cmsg.queue.lock); +} + +static void +nfp_ccm_mbox_mark_all_err(struct nfp_net *nn, struct sk_buff *last, int err) +{ + struct nfp_ccm_mbox_cmsg_cb *cb; + struct sk_buff *skb; + + spin_lock_bh(&nn->mbox_cmsg.queue.lock); + do { + skb = __skb_dequeue(&nn->mbox_cmsg.queue); + cb = (void *)skb->cb; + + cb->err = err; + smp_wmb(); /* order the cb->err vs. cb->state */ + cb->state = NFP_NET_MBOX_CMSG_STATE_DONE; + } while (skb != last); + + nfp_ccm_mbox_mark_next_runner(nn); + spin_unlock_bh(&nn->mbox_cmsg.queue.lock); +} + +static void nfp_ccm_mbox_run_queue_unlock(struct nfp_net *nn) + __releases(&nn->mbox_cmsg.queue.lock) +{ + int space = nn->tlv_caps.mbox_len - NFP_NET_CFG_MBOX_SIMPLE_VAL; + struct sk_buff *skb, *last; + int cnt, err; + + space -= 4; /* for End TLV */ + + /* First skb must fit, because it's ours and we checked it fits */ + cnt = 1; + last = skb = __skb_peek(&nn->mbox_cmsg.queue); + space -= 4 + nfp_ccm_mbox_maxlen(skb); + + while (!skb_queue_is_last(&nn->mbox_cmsg.queue, last)) { + skb = skb_queue_next(&nn->mbox_cmsg.queue, last); + space -= 4 + nfp_ccm_mbox_maxlen(skb); + if (space < 0) + break; + last = skb; + nfp_ccm_mbox_set_busy(skb); + cnt++; + if (cnt == NFP_CCM_MBOX_BATCH_LIMIT) + break; + } + spin_unlock_bh(&nn->mbox_cmsg.queue.lock); + + /* Now we own all skb's marked in progress, new requests may arrive + * at the end of the queue. + */ + + nn_ctrl_bar_lock(nn); + + nfp_ccm_mbox_copy_in(nn, last); + + err = nfp_net_mbox_reconfig(nn, NFP_NET_CFG_MBOX_CMD_TLV_CMSG); + if (!err) + nfp_ccm_mbox_copy_out(nn, last); + else + nfp_ccm_mbox_mark_all_err(nn, last, -EIO); + + nn_ctrl_bar_unlock(nn); + + wake_up_all(&nn->mbox_cmsg.wq); +} + +static int nfp_ccm_mbox_skb_return(struct sk_buff *skb) +{ + struct nfp_ccm_mbox_cmsg_cb *cb = (void *)skb->cb; + + if (cb->err) + dev_kfree_skb_any(skb); + return cb->err; +} + +/* If wait timed out but the command is already in progress we have + * to wait until it finishes. Runners has ownership of the skbs marked + * as busy. + */ +static int +nfp_ccm_mbox_unlink_unlock(struct nfp_net *nn, struct sk_buff *skb, + enum nfp_ccm_type type) + __releases(&nn->mbox_cmsg.queue.lock) +{ + bool was_first; + + if (nfp_ccm_mbox_in_progress(skb)) { + spin_unlock_bh(&nn->mbox_cmsg.queue.lock); + + wait_event(nn->mbox_cmsg.wq, nfp_ccm_mbox_done(skb)); + smp_rmb(); /* pairs with smp_wmb() after data is written */ + return nfp_ccm_mbox_skb_return(skb); + } + + was_first = nfp_ccm_mbox_should_run(nn, skb); + __skb_unlink(skb, &nn->mbox_cmsg.queue); + if (was_first) + nfp_ccm_mbox_mark_next_runner(nn); + + spin_unlock_bh(&nn->mbox_cmsg.queue.lock); + + if (was_first) + wake_up_all(&nn->mbox_cmsg.wq); + + nn_dp_warn(&nn->dp, "time out waiting for mbox response to 0x%02x\n", + type); + return -ETIMEDOUT; +} + +static int +nfp_ccm_mbox_msg_prepare(struct nfp_net *nn, struct sk_buff *skb, + enum nfp_ccm_type type, + unsigned int reply_size, unsigned int max_reply_size, + gfp_t flags) +{ + const unsigned int mbox_max = nfp_ccm_mbox_max_msg(nn); + unsigned int max_len; + ssize_t undersize; + int err; + + if (unlikely(!(nn->tlv_caps.mbox_cmsg_types & BIT(type)))) { + nn_dp_warn(&nn->dp, + "message type %d not supported by mailbox\n", type); + return -EINVAL; + } + + /* If the reply size is unknown assume it will take the entire + * mailbox, the callers should do their best for this to never + * happen. + */ + if (!max_reply_size) + max_reply_size = mbox_max; + max_reply_size = round_up(max_reply_size, 4); + + /* Make sure we can fit the entire reply into the skb, + * and that we don't have to slow down the mbox handler + * with allocations. + */ + undersize = max_reply_size - (skb_end_pointer(skb) - skb->data); + if (undersize > 0) { + err = pskb_expand_head(skb, 0, undersize, flags); + if (err) { + nn_dp_warn(&nn->dp, + "can't allocate reply buffer for mailbox\n"); + return err; + } + } + + /* Make sure that request and response both fit into the mailbox */ + max_len = max(max_reply_size, round_up(skb->len, 4)); + if (max_len > mbox_max) { + nn_dp_warn(&nn->dp, + "message too big for tha mailbox: %u/%u vs %u\n", + skb->len, max_reply_size, mbox_max); + return -EMSGSIZE; + } + + nfp_ccm_mbox_msg_init(skb, reply_size, max_len); + + return 0; +} + +static int +nfp_ccm_mbox_msg_enqueue(struct nfp_net *nn, struct sk_buff *skb, + enum nfp_ccm_type type, bool critical) +{ + struct nfp_ccm_hdr *hdr; + + assert_spin_locked(&nn->mbox_cmsg.queue.lock); + + if (!critical && nn->mbox_cmsg.queue.qlen >= NFP_CCM_MAX_QLEN) { + nn_dp_warn(&nn->dp, "mailbox request queue too long\n"); + return -EBUSY; + } + + hdr = (void *)skb->data; + hdr->ver = NFP_CCM_ABI_VERSION; + hdr->type = type; + hdr->tag = cpu_to_be16(nn->mbox_cmsg.tag++); + + __skb_queue_tail(&nn->mbox_cmsg.queue, skb); + + return 0; +} + +int __nfp_ccm_mbox_communicate(struct nfp_net *nn, struct sk_buff *skb, + enum nfp_ccm_type type, + unsigned int reply_size, + unsigned int max_reply_size, bool critical) +{ + int err; + + err = nfp_ccm_mbox_msg_prepare(nn, skb, type, reply_size, + max_reply_size, GFP_KERNEL); + if (err) + goto err_free_skb; + + spin_lock_bh(&nn->mbox_cmsg.queue.lock); + + err = nfp_ccm_mbox_msg_enqueue(nn, skb, type, critical); + if (err) + goto err_unlock; + + /* First in queue takes the mailbox lock and processes the batch */ + if (!nfp_ccm_mbox_is_first(nn, skb)) { + bool to; + + spin_unlock_bh(&nn->mbox_cmsg.queue.lock); + + to = !wait_event_timeout(nn->mbox_cmsg.wq, + nfp_ccm_mbox_done(skb) || + nfp_ccm_mbox_should_run(nn, skb), + msecs_to_jiffies(NFP_CCM_TIMEOUT)); + + /* fast path for those completed by another thread */ + if (nfp_ccm_mbox_done(skb)) { + smp_rmb(); /* pairs with wmb after data is written */ + return nfp_ccm_mbox_skb_return(skb); + } + + spin_lock_bh(&nn->mbox_cmsg.queue.lock); + + if (!nfp_ccm_mbox_is_first(nn, skb)) { + WARN_ON(!to); + + err = nfp_ccm_mbox_unlink_unlock(nn, skb, type); + if (err) + goto err_free_skb; + return 0; + } + } + + /* run queue expects the lock held */ + nfp_ccm_mbox_run_queue_unlock(nn); + return nfp_ccm_mbox_skb_return(skb); + +err_unlock: + spin_unlock_bh(&nn->mbox_cmsg.queue.lock); +err_free_skb: + dev_kfree_skb_any(skb); + return err; +} + +int nfp_ccm_mbox_communicate(struct nfp_net *nn, struct sk_buff *skb, + enum nfp_ccm_type type, + unsigned int reply_size, + unsigned int max_reply_size) +{ + return __nfp_ccm_mbox_communicate(nn, skb, type, reply_size, + max_reply_size, false); +} + +static void nfp_ccm_mbox_post_runq_work(struct work_struct *work) +{ + struct sk_buff *skb; + struct nfp_net *nn; + + nn = container_of(work, struct nfp_net, mbox_cmsg.runq_work); + + spin_lock_bh(&nn->mbox_cmsg.queue.lock); + + skb = __skb_peek(&nn->mbox_cmsg.queue); + if (WARN_ON(!skb || !nfp_ccm_mbox_is_posted(skb) || + !nfp_ccm_mbox_should_run(nn, skb))) { + spin_unlock_bh(&nn->mbox_cmsg.queue.lock); + return; + } + + nfp_ccm_mbox_run_queue_unlock(nn); +} + +static void nfp_ccm_mbox_post_wait_work(struct work_struct *work) +{ + struct sk_buff *skb; + struct nfp_net *nn; + int err; + + nn = container_of(work, struct nfp_net, mbox_cmsg.wait_work); + + skb = skb_peek(&nn->mbox_cmsg.queue); + if (WARN_ON(!skb || !nfp_ccm_mbox_is_posted(skb))) + /* Should never happen so it's unclear what to do here.. */ + goto exit_unlock_wake; + + err = nfp_net_mbox_reconfig_wait_posted(nn); + if (!err) + nfp_ccm_mbox_copy_out(nn, skb); + else + nfp_ccm_mbox_mark_all_err(nn, skb, -EIO); +exit_unlock_wake: + nn_ctrl_bar_unlock(nn); + wake_up_all(&nn->mbox_cmsg.wq); +} + +int nfp_ccm_mbox_post(struct nfp_net *nn, struct sk_buff *skb, + enum nfp_ccm_type type, unsigned int max_reply_size) +{ + int err; + + err = nfp_ccm_mbox_msg_prepare(nn, skb, type, 0, max_reply_size, + GFP_ATOMIC); + if (err) + goto err_free_skb; + + nfp_ccm_mbox_mark_posted(skb); + + spin_lock_bh(&nn->mbox_cmsg.queue.lock); + + err = nfp_ccm_mbox_msg_enqueue(nn, skb, type, false); + if (err) + goto err_unlock; + + if (nfp_ccm_mbox_is_first(nn, skb)) { + if (nn_ctrl_bar_trylock(nn)) { + nfp_ccm_mbox_copy_in(nn, skb); + nfp_net_mbox_reconfig_post(nn, + NFP_NET_CFG_MBOX_CMD_TLV_CMSG); + queue_work(nn->mbox_cmsg.workq, + &nn->mbox_cmsg.wait_work); + } else { + nfp_ccm_mbox_mark_next_runner(nn); + } + } + + spin_unlock_bh(&nn->mbox_cmsg.queue.lock); + + return 0; + +err_unlock: + spin_unlock_bh(&nn->mbox_cmsg.queue.lock); +err_free_skb: + dev_kfree_skb_any(skb); + return err; +} + +struct sk_buff * +nfp_ccm_mbox_msg_alloc(struct nfp_net *nn, unsigned int req_size, + unsigned int reply_size, gfp_t flags) +{ + unsigned int max_size; + struct sk_buff *skb; + + if (!reply_size) + max_size = nfp_ccm_mbox_max_msg(nn); + else + max_size = max(req_size, reply_size); + max_size = round_up(max_size, 4); + + skb = alloc_skb(max_size, flags); + if (!skb) + return NULL; + + skb_put(skb, req_size); + + return skb; +} + +bool nfp_ccm_mbox_fits(struct nfp_net *nn, unsigned int size) +{ + return nfp_ccm_mbox_max_msg(nn) >= size; +} + +int nfp_ccm_mbox_init(struct nfp_net *nn) +{ + return 0; +} + +void nfp_ccm_mbox_clean(struct nfp_net *nn) +{ + drain_workqueue(nn->mbox_cmsg.workq); +} + +int nfp_ccm_mbox_alloc(struct nfp_net *nn) +{ + skb_queue_head_init(&nn->mbox_cmsg.queue); + init_waitqueue_head(&nn->mbox_cmsg.wq); + INIT_WORK(&nn->mbox_cmsg.wait_work, nfp_ccm_mbox_post_wait_work); + INIT_WORK(&nn->mbox_cmsg.runq_work, nfp_ccm_mbox_post_runq_work); + + nn->mbox_cmsg.workq = alloc_workqueue("nfp-ccm-mbox", WQ_UNBOUND, 0); + if (!nn->mbox_cmsg.workq) + return -ENOMEM; + return 0; +} + +void nfp_ccm_mbox_free(struct nfp_net *nn) +{ + destroy_workqueue(nn->mbox_cmsg.workq); + WARN_ON(!skb_queue_empty(&nn->mbox_cmsg.queue)); +} diff --git a/drivers/net/ethernet/netronome/nfp/crypto/crypto.h b/drivers/net/ethernet/netronome/nfp/crypto/crypto.h new file mode 100644 index 000000000000..60372ddf69f0 --- /dev/null +++ b/drivers/net/ethernet/netronome/nfp/crypto/crypto.h @@ -0,0 +1,27 @@ +/* SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) */ +/* Copyright (C) 2019 Netronome Systems, Inc. */ + +#ifndef NFP_CRYPTO_H +#define NFP_CRYPTO_H 1 + +struct nfp_net_tls_offload_ctx { + __be32 fw_handle[2]; + + u8 rx_end[0]; + /* Tx only fields follow - Rx side does not have enough driver state + * to fit these + */ + + u32 next_seq; +}; + +#ifdef CONFIG_TLS_DEVICE +int nfp_net_tls_init(struct nfp_net *nn); +#else +static inline int nfp_net_tls_init(struct nfp_net *nn) +{ + return 0; +} +#endif + +#endif diff --git a/drivers/net/ethernet/netronome/nfp/crypto/fw.h b/drivers/net/ethernet/netronome/nfp/crypto/fw.h new file mode 100644 index 000000000000..67413d946c4a --- /dev/null +++ b/drivers/net/ethernet/netronome/nfp/crypto/fw.h @@ -0,0 +1,84 @@ +/* SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) */ +/* Copyright (C) 2019 Netronome Systems, Inc. */ + +#ifndef NFP_CRYPTO_FW_H +#define NFP_CRYPTO_FW_H 1 + +#include "../ccm.h" + +#define NFP_NET_CRYPTO_OP_TLS_1_2_AES_GCM_128_ENC 0 +#define NFP_NET_CRYPTO_OP_TLS_1_2_AES_GCM_128_DEC 1 + +struct nfp_crypto_reply_simple { + struct nfp_ccm_hdr hdr; + __be32 error; +}; + +struct nfp_crypto_req_reset { + struct nfp_ccm_hdr hdr; + __be32 ep_id; +}; + +#define NFP_NET_TLS_IPVER GENMASK(15, 12) +#define NFP_NET_TLS_VLAN GENMASK(11, 0) +#define NFP_NET_TLS_VLAN_UNUSED 4095 + +struct nfp_crypto_req_add_front { + struct nfp_ccm_hdr hdr; + __be32 ep_id; + u8 resv[3]; + u8 opcode; + u8 key_len; + __be16 ipver_vlan __packed; + u8 l4_proto; +#define NFP_NET_TLS_NON_ADDR_KEY_LEN 8 + u8 l3_addrs[0]; +}; + +struct nfp_crypto_req_add_back { + __be16 src_port; + __be16 dst_port; + __be32 key[8]; + __be32 salt; + __be32 iv[2]; + __be32 counter; + __be32 rec_no[2]; + __be32 tcp_seq; +}; + +struct nfp_crypto_req_add_v4 { + struct nfp_crypto_req_add_front front; + __be32 src_ip; + __be32 dst_ip; + struct nfp_crypto_req_add_back back; +}; + +struct nfp_crypto_req_add_v6 { + struct nfp_crypto_req_add_front front; + __be32 src_ip[4]; + __be32 dst_ip[4]; + struct nfp_crypto_req_add_back back; +}; + +struct nfp_crypto_reply_add { + struct nfp_ccm_hdr hdr; + __be32 error; + __be32 handle[2]; +}; + +struct nfp_crypto_req_del { + struct nfp_ccm_hdr hdr; + __be32 ep_id; + __be32 handle[2]; +}; + +struct nfp_crypto_req_update { + struct nfp_ccm_hdr hdr; + __be32 ep_id; + u8 resv[3]; + u8 opcode; + __be32 handle[2]; + __be32 rec_no[2]; + __be32 tcp_seq; +}; +#endif diff --git a/drivers/net/ethernet/netronome/nfp/crypto/tls.c b/drivers/net/ethernet/netronome/nfp/crypto/tls.c new file mode 100644 index 000000000000..96a96b35c0ca --- /dev/null +++ b/drivers/net/ethernet/netronome/nfp/crypto/tls.c @@ -0,0 +1,522 @@ +// SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) +/* Copyright (C) 2019 Netronome Systems, Inc. */ + +#include <linux/bitfield.h> +#include <linux/ipv6.h> +#include <linux/skbuff.h> +#include <linux/string.h> +#include <net/tls.h> + +#include "../ccm.h" +#include "../nfp_net.h" +#include "crypto.h" +#include "fw.h" + +#define NFP_NET_TLS_CCM_MBOX_OPS_MASK \ + (BIT(NFP_CCM_TYPE_CRYPTO_RESET) | \ + BIT(NFP_CCM_TYPE_CRYPTO_ADD) | \ + BIT(NFP_CCM_TYPE_CRYPTO_DEL) | \ + BIT(NFP_CCM_TYPE_CRYPTO_UPDATE)) + +#define NFP_NET_TLS_OPCODE_MASK_RX \ + BIT(NFP_NET_CRYPTO_OP_TLS_1_2_AES_GCM_128_DEC) + +#define NFP_NET_TLS_OPCODE_MASK_TX \ + BIT(NFP_NET_CRYPTO_OP_TLS_1_2_AES_GCM_128_ENC) + +#define NFP_NET_TLS_OPCODE_MASK \ + (NFP_NET_TLS_OPCODE_MASK_RX | NFP_NET_TLS_OPCODE_MASK_TX) + +static void nfp_net_crypto_set_op(struct nfp_net *nn, u8 opcode, bool on) +{ + u32 off, val; + + off = nn->tlv_caps.crypto_enable_off + round_down(opcode / 8, 4); + + val = nn_readl(nn, off); + if (on) + val |= BIT(opcode & 31); + else + val &= ~BIT(opcode & 31); + nn_writel(nn, off, val); +} + +static bool +__nfp_net_tls_conn_cnt_changed(struct nfp_net *nn, int add, + enum tls_offload_ctx_dir direction) +{ + u8 opcode; + int cnt; + + if (direction == TLS_OFFLOAD_CTX_DIR_TX) { + opcode = NFP_NET_CRYPTO_OP_TLS_1_2_AES_GCM_128_ENC; + nn->ktls_tx_conn_cnt += add; + cnt = nn->ktls_tx_conn_cnt; + nn->dp.ktls_tx = !!nn->ktls_tx_conn_cnt; + } else { + opcode = NFP_NET_CRYPTO_OP_TLS_1_2_AES_GCM_128_DEC; + nn->ktls_rx_conn_cnt += add; + cnt = nn->ktls_rx_conn_cnt; + } + + /* Care only about 0 -> 1 and 1 -> 0 transitions */ + if (cnt > 1) + return false; + + nfp_net_crypto_set_op(nn, opcode, cnt); + return true; +} + +static int +nfp_net_tls_conn_cnt_changed(struct nfp_net *nn, int add, + enum tls_offload_ctx_dir direction) +{ + int ret = 0; + + /* Use the BAR lock to protect the connection counts */ + nn_ctrl_bar_lock(nn); + if (__nfp_net_tls_conn_cnt_changed(nn, add, direction)) { + ret = __nfp_net_reconfig(nn, NFP_NET_CFG_UPDATE_CRYPTO); + /* Undo the cnt adjustment if failed */ + if (ret) + __nfp_net_tls_conn_cnt_changed(nn, -add, direction); + } + nn_ctrl_bar_unlock(nn); + + return ret; +} + +static int +nfp_net_tls_conn_add(struct nfp_net *nn, enum tls_offload_ctx_dir direction) +{ + return nfp_net_tls_conn_cnt_changed(nn, 1, direction); +} + +static int +nfp_net_tls_conn_remove(struct nfp_net *nn, enum tls_offload_ctx_dir direction) +{ + return nfp_net_tls_conn_cnt_changed(nn, -1, direction); +} + +static struct sk_buff * +nfp_net_tls_alloc_simple(struct nfp_net *nn, size_t req_sz, gfp_t flags) +{ + return nfp_ccm_mbox_msg_alloc(nn, req_sz, + sizeof(struct nfp_crypto_reply_simple), + flags); +} + +static int +nfp_net_tls_communicate_simple(struct nfp_net *nn, struct sk_buff *skb, + const char *name, enum nfp_ccm_type type) +{ + struct nfp_crypto_reply_simple *reply; + int err; + + err = __nfp_ccm_mbox_communicate(nn, skb, type, + sizeof(*reply), sizeof(*reply), + type == NFP_CCM_TYPE_CRYPTO_DEL); + if (err) { + nn_dp_warn(&nn->dp, "failed to %s TLS: %d\n", name, err); + return err; + } + + reply = (void *)skb->data; + err = -be32_to_cpu(reply->error); + if (err) + nn_dp_warn(&nn->dp, "failed to %s TLS, fw replied: %d\n", + name, err); + dev_consume_skb_any(skb); + + return err; +} + +static void nfp_net_tls_del_fw(struct nfp_net *nn, __be32 *fw_handle) +{ + struct nfp_crypto_req_del *req; + struct sk_buff *skb; + + skb = nfp_net_tls_alloc_simple(nn, sizeof(*req), GFP_KERNEL); + if (!skb) + return; + + req = (void *)skb->data; + req->ep_id = 0; + memcpy(req->handle, fw_handle, sizeof(req->handle)); + + nfp_net_tls_communicate_simple(nn, skb, "delete", + NFP_CCM_TYPE_CRYPTO_DEL); +} + +static void +nfp_net_tls_set_ipver_vlan(struct nfp_crypto_req_add_front *front, u8 ipver) +{ + front->ipver_vlan = cpu_to_be16(FIELD_PREP(NFP_NET_TLS_IPVER, ipver) | + FIELD_PREP(NFP_NET_TLS_VLAN, + NFP_NET_TLS_VLAN_UNUSED)); +} + +static void +nfp_net_tls_assign_conn_id(struct nfp_net *nn, + struct nfp_crypto_req_add_front *front) +{ + u32 len; + u64 id; + + id = atomic64_inc_return(&nn->ktls_conn_id_gen); + len = front->key_len - NFP_NET_TLS_NON_ADDR_KEY_LEN; + + memcpy(front->l3_addrs, &id, sizeof(id)); + memset(front->l3_addrs + sizeof(id), 0, len - sizeof(id)); +} + +static struct nfp_crypto_req_add_back * +nfp_net_tls_set_ipv4(struct nfp_net *nn, struct nfp_crypto_req_add_v4 *req, + struct sock *sk, int direction) +{ + struct inet_sock *inet = inet_sk(sk); + + req->front.key_len += sizeof(__be32) * 2; + + if (direction == TLS_OFFLOAD_CTX_DIR_TX) { + nfp_net_tls_assign_conn_id(nn, &req->front); + } else { + req->src_ip = inet->inet_daddr; + req->dst_ip = inet->inet_saddr; + } + + return &req->back; +} + +static struct nfp_crypto_req_add_back * +nfp_net_tls_set_ipv6(struct nfp_net *nn, struct nfp_crypto_req_add_v6 *req, + struct sock *sk, int direction) +{ +#if IS_ENABLED(CONFIG_IPV6) + struct ipv6_pinfo *np = inet6_sk(sk); + + req->front.key_len += sizeof(struct in6_addr) * 2; + + if (direction == TLS_OFFLOAD_CTX_DIR_TX) { + nfp_net_tls_assign_conn_id(nn, &req->front); + } else { + memcpy(req->src_ip, &sk->sk_v6_daddr, sizeof(req->src_ip)); + memcpy(req->dst_ip, &np->saddr, sizeof(req->dst_ip)); + } + +#endif + return &req->back; +} + +static void +nfp_net_tls_set_l4(struct nfp_crypto_req_add_front *front, + struct nfp_crypto_req_add_back *back, struct sock *sk, + int direction) +{ + struct inet_sock *inet = inet_sk(sk); + + front->l4_proto = IPPROTO_TCP; + + if (direction == TLS_OFFLOAD_CTX_DIR_TX) { + back->src_port = 0; + back->dst_port = 0; + } else { + back->src_port = inet->inet_dport; + back->dst_port = inet->inet_sport; + } +} + +static u8 nfp_tls_1_2_dir_to_opcode(enum tls_offload_ctx_dir direction) +{ + switch (direction) { + case TLS_OFFLOAD_CTX_DIR_TX: + return NFP_NET_CRYPTO_OP_TLS_1_2_AES_GCM_128_ENC; + case TLS_OFFLOAD_CTX_DIR_RX: + return NFP_NET_CRYPTO_OP_TLS_1_2_AES_GCM_128_DEC; + default: + WARN_ON_ONCE(1); + return 0; + } +} + +static bool +nfp_net_cipher_supported(struct nfp_net *nn, u16 cipher_type, + enum tls_offload_ctx_dir direction) +{ + u8 bit; + + switch (cipher_type) { + case TLS_CIPHER_AES_GCM_128: + if (direction == TLS_OFFLOAD_CTX_DIR_TX) + bit = NFP_NET_CRYPTO_OP_TLS_1_2_AES_GCM_128_ENC; + else + bit = NFP_NET_CRYPTO_OP_TLS_1_2_AES_GCM_128_DEC; + break; + default: + return false; + } + + return nn->tlv_caps.crypto_ops & BIT(bit); +} + +static int +nfp_net_tls_add(struct net_device *netdev, struct sock *sk, + enum tls_offload_ctx_dir direction, + struct tls_crypto_info *crypto_info, + u32 start_offload_tcp_sn) +{ + struct tls12_crypto_info_aes_gcm_128 *tls_ci; + struct nfp_net *nn = netdev_priv(netdev); + struct nfp_crypto_req_add_front *front; + struct nfp_net_tls_offload_ctx *ntls; + struct nfp_crypto_req_add_back *back; + struct nfp_crypto_reply_add *reply; + struct sk_buff *skb; + size_t req_sz; + void *req; + bool ipv6; + int err; + + BUILD_BUG_ON(sizeof(struct nfp_net_tls_offload_ctx) > + TLS_DRIVER_STATE_SIZE_TX); + BUILD_BUG_ON(offsetof(struct nfp_net_tls_offload_ctx, rx_end) > + TLS_DRIVER_STATE_SIZE_RX); + + if (!nfp_net_cipher_supported(nn, crypto_info->cipher_type, direction)) + return -EOPNOTSUPP; + + switch (sk->sk_family) { +#if IS_ENABLED(CONFIG_IPV6) + case AF_INET6: + if (sk->sk_ipv6only || + ipv6_addr_type(&sk->sk_v6_daddr) != IPV6_ADDR_MAPPED) { + req_sz = sizeof(struct nfp_crypto_req_add_v6); + ipv6 = true; + break; + } +#endif + /* fall through */ + case AF_INET: + req_sz = sizeof(struct nfp_crypto_req_add_v4); + ipv6 = false; + break; + default: + return -EOPNOTSUPP; + } + + err = nfp_net_tls_conn_add(nn, direction); + if (err) + return err; + + skb = nfp_ccm_mbox_msg_alloc(nn, req_sz, sizeof(*reply), GFP_KERNEL); + if (!skb) { + err = -ENOMEM; + goto err_conn_remove; + } + + front = (void *)skb->data; + front->ep_id = 0; + front->key_len = NFP_NET_TLS_NON_ADDR_KEY_LEN; + front->opcode = nfp_tls_1_2_dir_to_opcode(direction); + memset(front->resv, 0, sizeof(front->resv)); + + nfp_net_tls_set_ipver_vlan(front, ipv6 ? 6 : 4); + + req = (void *)skb->data; + if (ipv6) + back = nfp_net_tls_set_ipv6(nn, req, sk, direction); + else + back = nfp_net_tls_set_ipv4(nn, req, sk, direction); + + nfp_net_tls_set_l4(front, back, sk, direction); + + back->counter = 0; + back->tcp_seq = cpu_to_be32(start_offload_tcp_sn); + + tls_ci = (struct tls12_crypto_info_aes_gcm_128 *)crypto_info; + memcpy(back->key, tls_ci->key, TLS_CIPHER_AES_GCM_128_KEY_SIZE); + memset(&back->key[TLS_CIPHER_AES_GCM_128_KEY_SIZE / 4], 0, + sizeof(back->key) - TLS_CIPHER_AES_GCM_128_KEY_SIZE); + memcpy(back->iv, tls_ci->iv, TLS_CIPHER_AES_GCM_128_IV_SIZE); + memcpy(&back->salt, tls_ci->salt, TLS_CIPHER_AES_GCM_128_SALT_SIZE); + memcpy(back->rec_no, tls_ci->rec_seq, sizeof(tls_ci->rec_seq)); + + /* Get an extra ref on the skb so we can wipe the key after */ + skb_get(skb); + + err = nfp_ccm_mbox_communicate(nn, skb, NFP_CCM_TYPE_CRYPTO_ADD, + sizeof(*reply), sizeof(*reply)); + reply = (void *)skb->data; + + /* We depend on CCM MBOX code not reallocating skb we sent + * so we can clear the key material out of the memory. + */ + if (!WARN_ON_ONCE((u8 *)back < skb->head || + (u8 *)back > skb_end_pointer(skb)) && + !WARN_ON_ONCE((u8 *)&reply[1] > (u8 *)back)) + memzero_explicit(back, sizeof(*back)); + dev_consume_skb_any(skb); /* the extra ref from skb_get() above */ + + if (err) { + nn_dp_warn(&nn->dp, "failed to add TLS: %d (%d)\n", + err, direction == TLS_OFFLOAD_CTX_DIR_TX); + /* communicate frees skb on error */ + goto err_conn_remove; + } + + err = -be32_to_cpu(reply->error); + if (err) { + if (err == -ENOSPC) { + if (!atomic_fetch_inc(&nn->ktls_no_space)) + nn_info(nn, "HW TLS table full\n"); + } else { + nn_dp_warn(&nn->dp, + "failed to add TLS, FW replied: %d\n", err); + } + goto err_free_skb; + } + + if (!reply->handle[0] && !reply->handle[1]) { + nn_dp_warn(&nn->dp, "FW returned NULL handle\n"); + err = -EINVAL; + goto err_fw_remove; + } + + ntls = tls_driver_ctx(sk, direction); + memcpy(ntls->fw_handle, reply->handle, sizeof(ntls->fw_handle)); + if (direction == TLS_OFFLOAD_CTX_DIR_TX) + ntls->next_seq = start_offload_tcp_sn; + dev_consume_skb_any(skb); + + if (direction == TLS_OFFLOAD_CTX_DIR_TX) + return 0; + + tls_offload_rx_resync_set_type(sk, + TLS_OFFLOAD_SYNC_TYPE_CORE_NEXT_HINT); + return 0; + +err_fw_remove: + nfp_net_tls_del_fw(nn, reply->handle); +err_free_skb: + dev_consume_skb_any(skb); +err_conn_remove: + nfp_net_tls_conn_remove(nn, direction); + return err; +} + +static void +nfp_net_tls_del(struct net_device *netdev, struct tls_context *tls_ctx, + enum tls_offload_ctx_dir direction) +{ + struct nfp_net *nn = netdev_priv(netdev); + struct nfp_net_tls_offload_ctx *ntls; + + nfp_net_tls_conn_remove(nn, direction); + + ntls = __tls_driver_ctx(tls_ctx, direction); + nfp_net_tls_del_fw(nn, ntls->fw_handle); +} + +static int +nfp_net_tls_resync(struct net_device *netdev, struct sock *sk, u32 seq, + u8 *rcd_sn, enum tls_offload_ctx_dir direction) +{ + struct nfp_net *nn = netdev_priv(netdev); + struct nfp_net_tls_offload_ctx *ntls; + struct nfp_crypto_req_update *req; + struct sk_buff *skb; + gfp_t flags; + int err; + + flags = direction == TLS_OFFLOAD_CTX_DIR_TX ? GFP_KERNEL : GFP_ATOMIC; + skb = nfp_net_tls_alloc_simple(nn, sizeof(*req), flags); + if (!skb) + return -ENOMEM; + + ntls = tls_driver_ctx(sk, direction); + req = (void *)skb->data; + req->ep_id = 0; + req->opcode = nfp_tls_1_2_dir_to_opcode(direction); + memset(req->resv, 0, sizeof(req->resv)); + memcpy(req->handle, ntls->fw_handle, sizeof(ntls->fw_handle)); + req->tcp_seq = cpu_to_be32(seq); + memcpy(req->rec_no, rcd_sn, sizeof(req->rec_no)); + + if (direction == TLS_OFFLOAD_CTX_DIR_TX) { + err = nfp_net_tls_communicate_simple(nn, skb, "sync", + NFP_CCM_TYPE_CRYPTO_UPDATE); + if (err) + return err; + ntls->next_seq = seq; + } else { + nfp_ccm_mbox_post(nn, skb, NFP_CCM_TYPE_CRYPTO_UPDATE, + sizeof(struct nfp_crypto_reply_simple)); + } + + return 0; +} + +static const struct tlsdev_ops nfp_net_tls_ops = { + .tls_dev_add = nfp_net_tls_add, + .tls_dev_del = nfp_net_tls_del, + .tls_dev_resync = nfp_net_tls_resync, +}; + +static int nfp_net_tls_reset(struct nfp_net *nn) +{ + struct nfp_crypto_req_reset *req; + struct sk_buff *skb; + + skb = nfp_net_tls_alloc_simple(nn, sizeof(*req), GFP_KERNEL); + if (!skb) + return -ENOMEM; + + req = (void *)skb->data; + req->ep_id = 0; + + return nfp_net_tls_communicate_simple(nn, skb, "reset", + NFP_CCM_TYPE_CRYPTO_RESET); +} + +int nfp_net_tls_init(struct nfp_net *nn) +{ + struct net_device *netdev = nn->dp.netdev; + int err; + + if (!(nn->tlv_caps.crypto_ops & NFP_NET_TLS_OPCODE_MASK)) + return 0; + + if ((nn->tlv_caps.mbox_cmsg_types & NFP_NET_TLS_CCM_MBOX_OPS_MASK) != + NFP_NET_TLS_CCM_MBOX_OPS_MASK) + return 0; + + if (!nfp_ccm_mbox_fits(nn, sizeof(struct nfp_crypto_req_add_v6))) { + nn_warn(nn, "disabling TLS offload - mbox too small: %d\n", + nn->tlv_caps.mbox_len); + return 0; + } + + err = nfp_net_tls_reset(nn); + if (err) + return err; + + nn_ctrl_bar_lock(nn); + nn_writel(nn, nn->tlv_caps.crypto_enable_off, 0); + err = __nfp_net_reconfig(nn, NFP_NET_CFG_UPDATE_CRYPTO); + nn_ctrl_bar_unlock(nn); + if (err) + return err; + + if (nn->tlv_caps.crypto_ops & NFP_NET_TLS_OPCODE_MASK_RX) { + netdev->hw_features |= NETIF_F_HW_TLS_RX; + netdev->features |= NETIF_F_HW_TLS_RX; + } + if (nn->tlv_caps.crypto_ops & NFP_NET_TLS_OPCODE_MASK_TX) { + netdev->hw_features |= NETIF_F_HW_TLS_TX; + netdev->features |= NETIF_F_HW_TLS_TX; + } + + netdev->tlsdev_ops = &nfp_net_tls_ops; + + return 0; +} diff --git a/drivers/net/ethernet/netronome/nfp/devlink_param.c b/drivers/net/ethernet/netronome/nfp/devlink_param.c new file mode 100644 index 000000000000..36491835ac65 --- /dev/null +++ b/drivers/net/ethernet/netronome/nfp/devlink_param.c @@ -0,0 +1,255 @@ +// SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) +/* Copyright (C) 2019 Netronome Systems, Inc. */ + +#include <net/devlink.h> + +#include "nfpcore/nfp.h" +#include "nfpcore/nfp_nsp.h" +#include "nfp_main.h" + +/** + * struct nfp_devlink_param_u8_arg - Devlink u8 parameter get/set arguments + * @hwinfo_name: HWinfo key name + * @default_hi_val: Default HWinfo value if HWinfo doesn't exist + * @invalid_dl_val: Devlink value to use if HWinfo is unknown/invalid. + * -errno if there is no unknown/invalid value available + * @hi_to_dl: HWinfo to devlink value mapping + * @dl_to_hi: Devlink to hwinfo value mapping + * @max_dl_val: Maximum devlink value supported, for validation only + * @max_hi_val: Maximum HWinfo value supported, for validation only + */ +struct nfp_devlink_param_u8_arg { + const char *hwinfo_name; + const char *default_hi_val; + int invalid_dl_val; + u8 hi_to_dl[4]; + u8 dl_to_hi[4]; + u8 max_dl_val; + u8 max_hi_val; +}; + +static const struct nfp_devlink_param_u8_arg nfp_devlink_u8_args[] = { + [DEVLINK_PARAM_GENERIC_ID_FW_LOAD_POLICY] = { + .hwinfo_name = "app_fw_from_flash", + .default_hi_val = NFP_NSP_APP_FW_LOAD_DEFAULT, + .invalid_dl_val = + DEVLINK_PARAM_FW_LOAD_POLICY_VALUE_UNKNOWN, + .hi_to_dl = { + [NFP_NSP_APP_FW_LOAD_DISK] = + DEVLINK_PARAM_FW_LOAD_POLICY_VALUE_DISK, + [NFP_NSP_APP_FW_LOAD_FLASH] = + DEVLINK_PARAM_FW_LOAD_POLICY_VALUE_FLASH, + [NFP_NSP_APP_FW_LOAD_PREF] = + DEVLINK_PARAM_FW_LOAD_POLICY_VALUE_DRIVER, + }, + .dl_to_hi = { + [DEVLINK_PARAM_FW_LOAD_POLICY_VALUE_DRIVER] = + NFP_NSP_APP_FW_LOAD_PREF, + [DEVLINK_PARAM_FW_LOAD_POLICY_VALUE_FLASH] = + NFP_NSP_APP_FW_LOAD_FLASH, + [DEVLINK_PARAM_FW_LOAD_POLICY_VALUE_DISK] = + NFP_NSP_APP_FW_LOAD_DISK, + }, + .max_dl_val = DEVLINK_PARAM_FW_LOAD_POLICY_VALUE_DISK, + .max_hi_val = NFP_NSP_APP_FW_LOAD_PREF, + }, + [DEVLINK_PARAM_GENERIC_ID_RESET_DEV_ON_DRV_PROBE] = { + .hwinfo_name = "abi_drv_reset", + .default_hi_val = NFP_NSP_DRV_RESET_DEFAULT, + .invalid_dl_val = + DEVLINK_PARAM_RESET_DEV_ON_DRV_PROBE_VALUE_UNKNOWN, + .hi_to_dl = { + [NFP_NSP_DRV_RESET_ALWAYS] = + DEVLINK_PARAM_RESET_DEV_ON_DRV_PROBE_VALUE_ALWAYS, + [NFP_NSP_DRV_RESET_NEVER] = + DEVLINK_PARAM_RESET_DEV_ON_DRV_PROBE_VALUE_NEVER, + [NFP_NSP_DRV_RESET_DISK] = + DEVLINK_PARAM_RESET_DEV_ON_DRV_PROBE_VALUE_DISK, + }, + .dl_to_hi = { + [DEVLINK_PARAM_RESET_DEV_ON_DRV_PROBE_VALUE_ALWAYS] = + NFP_NSP_DRV_RESET_ALWAYS, + [DEVLINK_PARAM_RESET_DEV_ON_DRV_PROBE_VALUE_NEVER] = + NFP_NSP_DRV_RESET_NEVER, + [DEVLINK_PARAM_RESET_DEV_ON_DRV_PROBE_VALUE_DISK] = + NFP_NSP_DRV_RESET_DISK, + }, + .max_dl_val = DEVLINK_PARAM_RESET_DEV_ON_DRV_PROBE_VALUE_DISK, + .max_hi_val = NFP_NSP_DRV_RESET_NEVER, + } +}; + +static int +nfp_devlink_param_u8_get(struct devlink *devlink, u32 id, + struct devlink_param_gset_ctx *ctx) +{ + const struct nfp_devlink_param_u8_arg *arg; + struct nfp_pf *pf = devlink_priv(devlink); + struct nfp_nsp *nsp; + char hwinfo[32]; + long value; + int err; + + if (id >= ARRAY_SIZE(nfp_devlink_u8_args)) + return -EOPNOTSUPP; + + arg = &nfp_devlink_u8_args[id]; + + nsp = nfp_nsp_open(pf->cpp); + if (IS_ERR(nsp)) { + err = PTR_ERR(nsp); + nfp_warn(pf->cpp, "can't access NSP: %d\n", err); + return err; + } + + snprintf(hwinfo, sizeof(hwinfo), arg->hwinfo_name); + err = nfp_nsp_hwinfo_lookup_optional(nsp, hwinfo, sizeof(hwinfo), + arg->default_hi_val); + if (err) { + nfp_warn(pf->cpp, "HWinfo lookup failed: %d\n", err); + goto exit_close_nsp; + } + + err = kstrtol(hwinfo, 0, &value); + if (err || value < 0 || value > arg->max_hi_val) { + nfp_warn(pf->cpp, "HWinfo '%s' value %li invalid\n", + arg->hwinfo_name, value); + + if (arg->invalid_dl_val >= 0) + ctx->val.vu8 = arg->invalid_dl_val; + else + err = arg->invalid_dl_val; + + goto exit_close_nsp; + } + + ctx->val.vu8 = arg->hi_to_dl[value]; + +exit_close_nsp: + nfp_nsp_close(nsp); + return err; +} + +static int +nfp_devlink_param_u8_set(struct devlink *devlink, u32 id, + struct devlink_param_gset_ctx *ctx) +{ + const struct nfp_devlink_param_u8_arg *arg; + struct nfp_pf *pf = devlink_priv(devlink); + struct nfp_nsp *nsp; + char hwinfo[32]; + int err; + + if (id >= ARRAY_SIZE(nfp_devlink_u8_args)) + return -EOPNOTSUPP; + + arg = &nfp_devlink_u8_args[id]; + + nsp = nfp_nsp_open(pf->cpp); + if (IS_ERR(nsp)) { + err = PTR_ERR(nsp); + nfp_warn(pf->cpp, "can't access NSP: %d\n", err); + return err; + } + + /* Note the value has already been validated. */ + snprintf(hwinfo, sizeof(hwinfo), "%s=%u", + arg->hwinfo_name, arg->dl_to_hi[ctx->val.vu8]); + err = nfp_nsp_hwinfo_set(nsp, hwinfo, sizeof(hwinfo)); + if (err) { + nfp_warn(pf->cpp, "HWinfo set failed: %d\n", err); + goto exit_close_nsp; + } + +exit_close_nsp: + nfp_nsp_close(nsp); + return err; +} + +static int +nfp_devlink_param_u8_validate(struct devlink *devlink, u32 id, + union devlink_param_value val, + struct netlink_ext_ack *extack) +{ + const struct nfp_devlink_param_u8_arg *arg; + + if (id >= ARRAY_SIZE(nfp_devlink_u8_args)) + return -EOPNOTSUPP; + + arg = &nfp_devlink_u8_args[id]; + + if (val.vu8 > arg->max_dl_val) { + NL_SET_ERR_MSG_MOD(extack, "parameter out of range"); + return -EINVAL; + } + + if (val.vu8 == arg->invalid_dl_val) { + NL_SET_ERR_MSG_MOD(extack, "unknown/invalid value specified"); + return -EINVAL; + } + + return 0; +} + +static const struct devlink_param nfp_devlink_params[] = { + DEVLINK_PARAM_GENERIC(FW_LOAD_POLICY, + BIT(DEVLINK_PARAM_CMODE_PERMANENT), + nfp_devlink_param_u8_get, + nfp_devlink_param_u8_set, + nfp_devlink_param_u8_validate), + DEVLINK_PARAM_GENERIC(RESET_DEV_ON_DRV_PROBE, + BIT(DEVLINK_PARAM_CMODE_PERMANENT), + nfp_devlink_param_u8_get, + nfp_devlink_param_u8_set, + nfp_devlink_param_u8_validate), +}; + +static int nfp_devlink_supports_params(struct nfp_pf *pf) +{ + struct nfp_nsp *nsp; + bool supported; + int err; + + nsp = nfp_nsp_open(pf->cpp); + if (IS_ERR(nsp)) { + err = PTR_ERR(nsp); + dev_err(&pf->pdev->dev, "Failed to access the NSP: %d\n", err); + return err; + } + + supported = nfp_nsp_has_hwinfo_lookup(nsp) && + nfp_nsp_has_hwinfo_set(nsp); + + nfp_nsp_close(nsp); + return supported; +} + +int nfp_devlink_params_register(struct nfp_pf *pf) +{ + struct devlink *devlink = priv_to_devlink(pf); + int err; + + err = nfp_devlink_supports_params(pf); + if (err <= 0) + return err; + + err = devlink_params_register(devlink, nfp_devlink_params, + ARRAY_SIZE(nfp_devlink_params)); + if (err) + return err; + + devlink_params_publish(devlink); + return 0; +} + +void nfp_devlink_params_unregister(struct nfp_pf *pf) +{ + int err; + + err = nfp_devlink_supports_params(pf); + if (err <= 0) + return; + + devlink_params_unregister(priv_to_devlink(pf), nfp_devlink_params, + ARRAY_SIZE(nfp_devlink_params)); +} diff --git a/drivers/net/ethernet/netronome/nfp/flower/Makefile b/drivers/net/ethernet/netronome/nfp/flower/Makefile deleted file mode 100644 index 805fa28f391a..000000000000 --- a/drivers/net/ethernet/netronome/nfp/flower/Makefile +++ /dev/null @@ -1,2 +0,0 @@ -# SPDX-License-Identifier: GPL-2.0 -# kbuild requires Makefile in a directory to build individual objects diff --git a/drivers/net/ethernet/netronome/nfp/flower/action.c b/drivers/net/ethernet/netronome/nfp/flower/action.c index e336f6ee94f5..1b019fdfcd97 100644 --- a/drivers/net/ethernet/netronome/nfp/flower/action.c +++ b/drivers/net/ethernet/netronome/nfp/flower/action.c @@ -2,10 +2,12 @@ /* Copyright (C) 2017-2018 Netronome Systems, Inc. */ #include <linux/bitfield.h> +#include <linux/mpls.h> #include <net/pkt_cls.h> #include <net/tc_act/tc_csum.h> #include <net/tc_act/tc_gact.h> #include <net/tc_act/tc_mirred.h> +#include <net/tc_act/tc_mpls.h> #include <net/tc_act/tc_pedit.h> #include <net/tc_act/tc_vlan.h> #include <net/tc_act/tc_tunnel_key.h> @@ -25,6 +27,80 @@ NFP_FL_TUNNEL_KEY | \ NFP_FL_TUNNEL_GENEVE_OPT) +static int +nfp_fl_push_mpls(struct nfp_fl_push_mpls *push_mpls, + const struct flow_action_entry *act, + struct netlink_ext_ack *extack) +{ + size_t act_size = sizeof(struct nfp_fl_push_mpls); + u32 mpls_lse = 0; + + push_mpls->head.jump_id = NFP_FL_ACTION_OPCODE_PUSH_MPLS; + push_mpls->head.len_lw = act_size >> NFP_FL_LW_SIZ; + + /* BOS is optional in the TC action but required for offload. */ + if (act->mpls_push.bos != ACT_MPLS_BOS_NOT_SET) { + mpls_lse |= act->mpls_push.bos << MPLS_LS_S_SHIFT; + } else { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: BOS field must explicitly be set for MPLS push"); + return -EOPNOTSUPP; + } + + /* Leave MPLS TC as a default value of 0 if not explicitly set. */ + if (act->mpls_push.tc != ACT_MPLS_TC_NOT_SET) + mpls_lse |= act->mpls_push.tc << MPLS_LS_TC_SHIFT; + + /* Proto, label and TTL are enforced and verified for MPLS push. */ + mpls_lse |= act->mpls_push.label << MPLS_LS_LABEL_SHIFT; + mpls_lse |= act->mpls_push.ttl << MPLS_LS_TTL_SHIFT; + push_mpls->ethtype = act->mpls_push.proto; + push_mpls->lse = cpu_to_be32(mpls_lse); + + return 0; +} + +static void +nfp_fl_pop_mpls(struct nfp_fl_pop_mpls *pop_mpls, + const struct flow_action_entry *act) +{ + size_t act_size = sizeof(struct nfp_fl_pop_mpls); + + pop_mpls->head.jump_id = NFP_FL_ACTION_OPCODE_POP_MPLS; + pop_mpls->head.len_lw = act_size >> NFP_FL_LW_SIZ; + pop_mpls->ethtype = act->mpls_pop.proto; +} + +static void +nfp_fl_set_mpls(struct nfp_fl_set_mpls *set_mpls, + const struct flow_action_entry *act) +{ + size_t act_size = sizeof(struct nfp_fl_set_mpls); + u32 mpls_lse = 0, mpls_mask = 0; + + set_mpls->head.jump_id = NFP_FL_ACTION_OPCODE_SET_MPLS; + set_mpls->head.len_lw = act_size >> NFP_FL_LW_SIZ; + + if (act->mpls_mangle.label != ACT_MPLS_LABEL_NOT_SET) { + mpls_lse |= act->mpls_mangle.label << MPLS_LS_LABEL_SHIFT; + mpls_mask |= MPLS_LS_LABEL_MASK; + } + if (act->mpls_mangle.tc != ACT_MPLS_TC_NOT_SET) { + mpls_lse |= act->mpls_mangle.tc << MPLS_LS_TC_SHIFT; + mpls_mask |= MPLS_LS_TC_MASK; + } + if (act->mpls_mangle.bos != ACT_MPLS_BOS_NOT_SET) { + mpls_lse |= act->mpls_mangle.bos << MPLS_LS_S_SHIFT; + mpls_mask |= MPLS_LS_S_MASK; + } + if (act->mpls_mangle.ttl) { + mpls_lse |= act->mpls_mangle.ttl << MPLS_LS_TTL_SHIFT; + mpls_mask |= MPLS_LS_TTL_MASK; + } + + set_mpls->lse = cpu_to_be32(mpls_lse); + set_mpls->lse_mask = cpu_to_be32(mpls_mask); +} + static void nfp_fl_pop_vlan(struct nfp_fl_pop_vlan *pop_vlan) { size_t act_size = sizeof(struct nfp_fl_pop_vlan); @@ -54,7 +130,8 @@ nfp_fl_push_vlan(struct nfp_fl_push_vlan *push_vlan, static int nfp_fl_pre_lag(struct nfp_app *app, const struct flow_action_entry *act, - struct nfp_fl_payload *nfp_flow, int act_len) + struct nfp_fl_payload *nfp_flow, int act_len, + struct netlink_ext_ack *extack) { size_t act_size = sizeof(struct nfp_fl_pre_lag); struct nfp_fl_pre_lag *pre_lag; @@ -65,8 +142,10 @@ nfp_fl_pre_lag(struct nfp_app *app, const struct flow_action_entry *act, if (!out_dev || !netif_is_lag_master(out_dev)) return 0; - if (act_len + act_size > NFP_FL_MAX_A_SIZ) + if (act_len + act_size > NFP_FL_MAX_A_SIZ) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: maximum allowed action list size exceeded at LAG action"); return -EOPNOTSUPP; + } /* Pre_lag action must be first on action list. * If other actions already exist they need pushed forward. @@ -76,7 +155,7 @@ nfp_fl_pre_lag(struct nfp_app *app, const struct flow_action_entry *act, nfp_flow->action_data, act_len); pre_lag = (struct nfp_fl_pre_lag *)nfp_flow->action_data; - err = nfp_flower_lag_populate_pre_action(app, out_dev, pre_lag); + err = nfp_flower_lag_populate_pre_action(app, out_dev, pre_lag, extack); if (err) return err; @@ -93,7 +172,8 @@ nfp_fl_output(struct nfp_app *app, struct nfp_fl_output *output, const struct flow_action_entry *act, struct nfp_fl_payload *nfp_flow, bool last, struct net_device *in_dev, - enum nfp_flower_tun_type tun_type, int *tun_out_cnt) + enum nfp_flower_tun_type tun_type, int *tun_out_cnt, + bool pkt_host, struct netlink_ext_ack *extack) { size_t act_size = sizeof(struct nfp_fl_output); struct nfp_flower_priv *priv = app->priv; @@ -104,18 +184,24 @@ nfp_fl_output(struct nfp_app *app, struct nfp_fl_output *output, output->head.len_lw = act_size >> NFP_FL_LW_SIZ; out_dev = act->dev; - if (!out_dev) + if (!out_dev) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: invalid egress interface for mirred action"); return -EOPNOTSUPP; + } tmp_flags = last ? NFP_FL_OUT_FLAGS_LAST : 0; if (tun_type) { /* Verify the egress netdev matches the tunnel type. */ - if (!nfp_fl_netdev_is_tunnel_type(out_dev, tun_type)) + if (!nfp_fl_netdev_is_tunnel_type(out_dev, tun_type)) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: egress interface does not match the required tunnel type"); return -EOPNOTSUPP; + } - if (*tun_out_cnt) + if (*tun_out_cnt) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: cannot offload more than one tunnel mirred output per filter"); return -EOPNOTSUPP; + } (*tun_out_cnt)++; output->flags = cpu_to_be16(tmp_flags | @@ -127,42 +213,87 @@ nfp_fl_output(struct nfp_app *app, struct nfp_fl_output *output, output->flags = cpu_to_be16(tmp_flags); gid = nfp_flower_lag_get_output_id(app, out_dev); - if (gid < 0) + if (gid < 0) { + NL_SET_ERR_MSG_MOD(extack, "invalid entry: cannot find group id for LAG action"); return gid; + } output->port = cpu_to_be32(NFP_FL_LAG_OUT | gid); + } else if (nfp_flower_internal_port_can_offload(app, out_dev)) { + if (!(priv->flower_ext_feats & NFP_FL_FEATS_PRE_TUN_RULES)) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: pre-tunnel rules not supported in loaded firmware"); + return -EOPNOTSUPP; + } + + if (nfp_flow->pre_tun_rule.dev || !pkt_host) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: pre-tunnel rules require single egress dev and ptype HOST action"); + return -EOPNOTSUPP; + } + + nfp_flow->pre_tun_rule.dev = out_dev; + + return 0; } else { /* Set action output parameters. */ output->flags = cpu_to_be16(tmp_flags); if (nfp_netdev_is_nfp_repr(in_dev)) { /* Confirm ingress and egress are on same device. */ - if (!netdev_port_same_parent_id(in_dev, out_dev)) + if (!netdev_port_same_parent_id(in_dev, out_dev)) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: ingress and egress interfaces are on different devices"); return -EOPNOTSUPP; + } } - if (!nfp_netdev_is_nfp_repr(out_dev)) + if (!nfp_netdev_is_nfp_repr(out_dev)) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: egress interface is not an nfp port"); return -EOPNOTSUPP; + } output->port = cpu_to_be32(nfp_repr_get_port_id(out_dev)); - if (!output->port) + if (!output->port) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: invalid port id for egress interface"); return -EOPNOTSUPP; + } } nfp_flow->meta.shortcut = output->port; return 0; } +static bool +nfp_flower_tun_is_gre(struct flow_cls_offload *flow, int start_idx) +{ + struct flow_action_entry *act = flow->rule->action.entries; + int num_act = flow->rule->action.num_entries; + int act_idx; + + /* Preparse action list for next mirred or redirect action */ + for (act_idx = start_idx + 1; act_idx < num_act; act_idx++) + if (act[act_idx].id == FLOW_ACTION_REDIRECT || + act[act_idx].id == FLOW_ACTION_MIRRED) + return netif_is_gretap(act[act_idx].dev); + + return false; +} + static enum nfp_flower_tun_type -nfp_fl_get_tun_from_act_l4_port(struct nfp_app *app, - const struct flow_action_entry *act) +nfp_fl_get_tun_from_act(struct nfp_app *app, + struct flow_cls_offload *flow, + const struct flow_action_entry *act, int act_idx) { const struct ip_tunnel_info *tun = act->tunnel; struct nfp_flower_priv *priv = app->priv; + /* Determine the tunnel type based on the egress netdev + * in the mirred action for tunnels without l4. + */ + if (nfp_flower_tun_is_gre(flow, act_idx)) + return NFP_FL_TUNNEL_GRE; + switch (tun->key.tp_dst) { - case htons(NFP_FL_VXLAN_PORT): + case htons(IANA_VXLAN_UDP_PORT): return NFP_FL_TUNNEL_VXLAN; - case htons(NFP_FL_GENEVE_PORT): + case htons(GENEVE_UDP_PORT): if (priv->flower_ext_feats & NFP_FL_FEATS_GENEVE) return NFP_FL_TUNNEL_GENEVE; /* FALLTHROUGH */ @@ -194,7 +325,8 @@ static struct nfp_fl_pre_tunnel *nfp_fl_pre_tunnel(char *act_data, int act_len) static int nfp_fl_push_geneve_options(struct nfp_fl_payload *nfp_fl, int *list_len, - const struct flow_action_entry *act) + const struct flow_action_entry *act, + struct netlink_ext_ack *extack) { struct ip_tunnel_info *ip_tun = (struct ip_tunnel_info *)act->tunnel; int opt_len, opt_cnt, act_start, tot_push_len; @@ -212,20 +344,26 @@ nfp_fl_push_geneve_options(struct nfp_fl_payload *nfp_fl, int *list_len, struct geneve_opt *opt = (struct geneve_opt *)src; opt_cnt++; - if (opt_cnt > NFP_FL_MAX_GENEVE_OPT_CNT) + if (opt_cnt > NFP_FL_MAX_GENEVE_OPT_CNT) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: maximum allowed number of geneve options exceeded"); return -EOPNOTSUPP; + } tot_push_len += sizeof(struct nfp_fl_push_geneve) + opt->length * 4; - if (tot_push_len > NFP_FL_MAX_GENEVE_OPT_ACT) + if (tot_push_len > NFP_FL_MAX_GENEVE_OPT_ACT) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: maximum allowed action list size exceeded at push geneve options"); return -EOPNOTSUPP; + } opt_len -= sizeof(struct geneve_opt) + opt->length * 4; src += sizeof(struct geneve_opt) + opt->length * 4; } - if (*list_len + tot_push_len > NFP_FL_MAX_A_SIZ) + if (*list_len + tot_push_len > NFP_FL_MAX_A_SIZ) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: maximum allowed action list size exceeded at push geneve options"); return -EOPNOTSUPP; + } act_start = *list_len; *list_len += tot_push_len; @@ -256,14 +394,13 @@ nfp_fl_push_geneve_options(struct nfp_fl_payload *nfp_fl, int *list_len, } static int -nfp_fl_set_ipv4_udp_tun(struct nfp_app *app, - struct nfp_fl_set_ipv4_udp_tun *set_tun, - const struct flow_action_entry *act, - struct nfp_fl_pre_tunnel *pre_tun, - enum nfp_flower_tun_type tun_type, - struct net_device *netdev) +nfp_fl_set_ipv4_tun(struct nfp_app *app, struct nfp_fl_set_ipv4_tun *set_tun, + const struct flow_action_entry *act, + struct nfp_fl_pre_tunnel *pre_tun, + enum nfp_flower_tun_type tun_type, + struct net_device *netdev, struct netlink_ext_ack *extack) { - size_t act_size = sizeof(struct nfp_fl_set_ipv4_udp_tun); + size_t act_size = sizeof(struct nfp_fl_set_ipv4_tun); const struct ip_tunnel_info *ip_tun = act->tunnel; struct nfp_flower_priv *priv = app->priv; u32 tmp_set_ip_tun_type_index = 0; @@ -275,8 +412,10 @@ nfp_fl_set_ipv4_udp_tun(struct nfp_app *app, NFP_FL_TUNNEL_GENEVE_OPT != TUNNEL_GENEVE_OPT); if (ip_tun->options_len && (tun_type != NFP_FL_TUNNEL_GENEVE || - !(priv->flower_ext_feats & NFP_FL_FEATS_GENEVE_OPT))) + !(priv->flower_ext_feats & NFP_FL_FEATS_GENEVE_OPT))) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: loaded firmware does not support geneve options offload"); return -EOPNOTSUPP; + } set_tun->head.jump_id = NFP_FL_ACTION_OPCODE_SET_IPV4_TUNNEL; set_tun->head.len_lw = act_size >> NFP_FL_LW_SIZ; @@ -316,8 +455,10 @@ nfp_fl_set_ipv4_udp_tun(struct nfp_app *app, set_tun->tos = ip_tun->key.tos; if (!(ip_tun->key.tun_flags & NFP_FL_TUNNEL_KEY) || - ip_tun->key.tun_flags & ~NFP_FL_SUPPORTED_IPV4_UDP_TUN_FLAGS) + ip_tun->key.tun_flags & ~NFP_FL_SUPPORTED_IPV4_UDP_TUN_FLAGS) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: loaded firmware does not support tunnel flag offload"); return -EOPNOTSUPP; + } set_tun->tun_flags = ip_tun->key.tun_flags; if (tun_type == NFP_FL_TUNNEL_GENEVE) { @@ -345,18 +486,22 @@ static void nfp_fl_set_helper32(u32 value, u32 mask, u8 *p_exact, u8 *p_mask) static int nfp_fl_set_eth(const struct flow_action_entry *act, u32 off, - struct nfp_fl_set_eth *set_eth) + struct nfp_fl_set_eth *set_eth, struct netlink_ext_ack *extack) { u32 exact, mask; - if (off + 4 > ETH_ALEN * 2) + if (off + 4 > ETH_ALEN * 2) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: invalid pedit ethernet action"); return -EOPNOTSUPP; + } mask = ~act->mangle.mask; exact = act->mangle.val; - if (exact & ~mask) + if (exact & ~mask) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: invalid pedit ethernet action"); return -EOPNOTSUPP; + } nfp_fl_set_helper32(exact, mask, &set_eth->eth_addr_val[off], &set_eth->eth_addr_mask[off]); @@ -377,7 +522,8 @@ struct ipv4_ttl_word { static int nfp_fl_set_ip4(const struct flow_action_entry *act, u32 off, struct nfp_fl_set_ip4_addrs *set_ip_addr, - struct nfp_fl_set_ip4_ttl_tos *set_ip_ttl_tos) + struct nfp_fl_set_ip4_ttl_tos *set_ip_ttl_tos, + struct netlink_ext_ack *extack) { struct ipv4_ttl_word *ttl_word_mask; struct ipv4_ttl_word *ttl_word; @@ -389,8 +535,10 @@ nfp_fl_set_ip4(const struct flow_action_entry *act, u32 off, mask = (__force __be32)~act->mangle.mask; exact = (__force __be32)act->mangle.val; - if (exact & ~mask) + if (exact & ~mask) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: invalid pedit IPv4 action"); return -EOPNOTSUPP; + } switch (off) { case offsetof(struct iphdr, daddr): @@ -413,8 +561,10 @@ nfp_fl_set_ip4(const struct flow_action_entry *act, u32 off, ttl_word_mask = (struct ipv4_ttl_word *)&mask; ttl_word = (struct ipv4_ttl_word *)&exact; - if (ttl_word_mask->protocol || ttl_word_mask->check) + if (ttl_word_mask->protocol || ttl_word_mask->check) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: invalid pedit IPv4 ttl action"); return -EOPNOTSUPP; + } set_ip_ttl_tos->ipv4_ttl_mask |= ttl_word_mask->ttl; set_ip_ttl_tos->ipv4_ttl &= ~ttl_word_mask->ttl; @@ -429,8 +579,10 @@ nfp_fl_set_ip4(const struct flow_action_entry *act, u32 off, tos_word = (struct iphdr *)&exact; if (tos_word_mask->version || tos_word_mask->ihl || - tos_word_mask->tot_len) + tos_word_mask->tot_len) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: invalid pedit IPv4 tos action"); return -EOPNOTSUPP; + } set_ip_ttl_tos->ipv4_tos_mask |= tos_word_mask->tos; set_ip_ttl_tos->ipv4_tos &= ~tos_word_mask->tos; @@ -441,6 +593,7 @@ nfp_fl_set_ip4(const struct flow_action_entry *act, u32 off, NFP_FL_LW_SIZ; break; default: + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: pedit on unsupported section of IPv4 header"); return -EOPNOTSUPP; } @@ -468,7 +621,8 @@ struct ipv6_hop_limit_word { static int nfp_fl_set_ip6_hop_limit_flow_label(u32 off, __be32 exact, __be32 mask, - struct nfp_fl_set_ipv6_tc_hl_fl *ip_hl_fl) + struct nfp_fl_set_ipv6_tc_hl_fl *ip_hl_fl, + struct netlink_ext_ack *extack) { struct ipv6_hop_limit_word *fl_hl_mask; struct ipv6_hop_limit_word *fl_hl; @@ -478,8 +632,10 @@ nfp_fl_set_ip6_hop_limit_flow_label(u32 off, __be32 exact, __be32 mask, fl_hl_mask = (struct ipv6_hop_limit_word *)&mask; fl_hl = (struct ipv6_hop_limit_word *)&exact; - if (fl_hl_mask->nexthdr || fl_hl_mask->payload_len) + if (fl_hl_mask->nexthdr || fl_hl_mask->payload_len) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: invalid pedit IPv6 hop limit action"); return -EOPNOTSUPP; + } ip_hl_fl->ipv6_hop_limit_mask |= fl_hl_mask->hop_limit; ip_hl_fl->ipv6_hop_limit &= ~fl_hl_mask->hop_limit; @@ -488,8 +644,10 @@ nfp_fl_set_ip6_hop_limit_flow_label(u32 off, __be32 exact, __be32 mask, break; case round_down(offsetof(struct ipv6hdr, flow_lbl), 4): if (mask & ~IPV6_FLOW_LABEL_MASK || - exact & ~IPV6_FLOW_LABEL_MASK) + exact & ~IPV6_FLOW_LABEL_MASK) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: invalid pedit IPv6 flow label action"); return -EOPNOTSUPP; + } ip_hl_fl->ipv6_label_mask |= mask; ip_hl_fl->ipv6_label &= ~mask; @@ -507,7 +665,8 @@ static int nfp_fl_set_ip6(const struct flow_action_entry *act, u32 off, struct nfp_fl_set_ipv6_addr *ip_dst, struct nfp_fl_set_ipv6_addr *ip_src, - struct nfp_fl_set_ipv6_tc_hl_fl *ip_hl_fl) + struct nfp_fl_set_ipv6_tc_hl_fl *ip_hl_fl, + struct netlink_ext_ack *extack) { __be32 exact, mask; int err = 0; @@ -517,12 +676,14 @@ nfp_fl_set_ip6(const struct flow_action_entry *act, u32 off, mask = (__force __be32)~act->mangle.mask; exact = (__force __be32)act->mangle.val; - if (exact & ~mask) + if (exact & ~mask) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: invalid pedit IPv6 action"); return -EOPNOTSUPP; + } if (off < offsetof(struct ipv6hdr, saddr)) { err = nfp_fl_set_ip6_hop_limit_flow_label(off, exact, mask, - ip_hl_fl); + ip_hl_fl, extack); } else if (off < offsetof(struct ipv6hdr, daddr)) { word = (off - offsetof(struct ipv6hdr, saddr)) / sizeof(exact); nfp_fl_set_ip6_helper(NFP_FL_ACTION_OPCODE_SET_IPV6_SRC, word, @@ -533,6 +694,7 @@ nfp_fl_set_ip6(const struct flow_action_entry *act, u32 off, nfp_fl_set_ip6_helper(NFP_FL_ACTION_OPCODE_SET_IPV6_DST, word, exact, mask, ip_dst); } else { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: pedit on unsupported section of IPv6 header"); return -EOPNOTSUPP; } @@ -541,18 +703,23 @@ nfp_fl_set_ip6(const struct flow_action_entry *act, u32 off, static int nfp_fl_set_tport(const struct flow_action_entry *act, u32 off, - struct nfp_fl_set_tport *set_tport, int opcode) + struct nfp_fl_set_tport *set_tport, int opcode, + struct netlink_ext_ack *extack) { u32 exact, mask; - if (off) + if (off) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: pedit on unsupported section of L4 header"); return -EOPNOTSUPP; + } mask = ~act->mangle.mask; exact = act->mangle.val; - if (exact & ~mask) + if (exact & ~mask) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: invalid pedit L4 action"); return -EOPNOTSUPP; + } nfp_fl_set_helper32(exact, mask, set_tport->tp_port_val, set_tport->tp_port_mask); @@ -582,60 +749,23 @@ static u32 nfp_fl_csum_l4_to_flag(u8 ip_proto) } } -static int -nfp_fl_pedit(const struct flow_action_entry *act, - struct tc_cls_flower_offload *flow, - char *nfp_action, int *a_len, u32 *csum_updated) -{ - struct flow_rule *rule = tc_cls_flower_offload_flow_rule(flow); +struct nfp_flower_pedit_acts { struct nfp_fl_set_ipv6_addr set_ip6_dst, set_ip6_src; struct nfp_fl_set_ipv6_tc_hl_fl set_ip6_tc_hl_fl; struct nfp_fl_set_ip4_ttl_tos set_ip_ttl_tos; struct nfp_fl_set_ip4_addrs set_ip_addr; - enum flow_action_mangle_base htype; struct nfp_fl_set_tport set_tport; struct nfp_fl_set_eth set_eth; +}; + +static int +nfp_fl_commit_mangle(struct flow_cls_offload *flow, char *nfp_action, + int *a_len, struct nfp_flower_pedit_acts *set_act, + u32 *csum_updated) +{ + struct flow_rule *rule = flow_cls_offload_flow_rule(flow); size_t act_size = 0; u8 ip_proto = 0; - u32 offset; - int err; - - memset(&set_ip6_tc_hl_fl, 0, sizeof(set_ip6_tc_hl_fl)); - memset(&set_ip_ttl_tos, 0, sizeof(set_ip_ttl_tos)); - memset(&set_ip6_dst, 0, sizeof(set_ip6_dst)); - memset(&set_ip6_src, 0, sizeof(set_ip6_src)); - memset(&set_ip_addr, 0, sizeof(set_ip_addr)); - memset(&set_tport, 0, sizeof(set_tport)); - memset(&set_eth, 0, sizeof(set_eth)); - - htype = act->mangle.htype; - offset = act->mangle.offset; - - switch (htype) { - case TCA_PEDIT_KEY_EX_HDR_TYPE_ETH: - err = nfp_fl_set_eth(act, offset, &set_eth); - break; - case TCA_PEDIT_KEY_EX_HDR_TYPE_IP4: - err = nfp_fl_set_ip4(act, offset, &set_ip_addr, - &set_ip_ttl_tos); - break; - case TCA_PEDIT_KEY_EX_HDR_TYPE_IP6: - err = nfp_fl_set_ip6(act, offset, &set_ip6_dst, - &set_ip6_src, &set_ip6_tc_hl_fl); - break; - case TCA_PEDIT_KEY_EX_HDR_TYPE_TCP: - err = nfp_fl_set_tport(act, offset, &set_tport, - NFP_FL_ACTION_OPCODE_SET_TCP); - break; - case TCA_PEDIT_KEY_EX_HDR_TYPE_UDP: - err = nfp_fl_set_tport(act, offset, &set_tport, - NFP_FL_ACTION_OPCODE_SET_UDP); - break; - default: - return -EOPNOTSUPP; - } - if (err) - return err; if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_BASIC)) { struct flow_match_basic match; @@ -644,77 +774,82 @@ nfp_fl_pedit(const struct flow_action_entry *act, ip_proto = match.key->ip_proto; } - if (set_eth.head.len_lw) { - act_size = sizeof(set_eth); - memcpy(nfp_action, &set_eth, act_size); + if (set_act->set_eth.head.len_lw) { + act_size = sizeof(set_act->set_eth); + memcpy(nfp_action, &set_act->set_eth, act_size); *a_len += act_size; } - if (set_ip_ttl_tos.head.len_lw) { + + if (set_act->set_ip_ttl_tos.head.len_lw) { nfp_action += act_size; - act_size = sizeof(set_ip_ttl_tos); - memcpy(nfp_action, &set_ip_ttl_tos, act_size); + act_size = sizeof(set_act->set_ip_ttl_tos); + memcpy(nfp_action, &set_act->set_ip_ttl_tos, act_size); *a_len += act_size; /* Hardware will automatically fix IPv4 and TCP/UDP checksum. */ *csum_updated |= TCA_CSUM_UPDATE_FLAG_IPV4HDR | nfp_fl_csum_l4_to_flag(ip_proto); } - if (set_ip_addr.head.len_lw) { + + if (set_act->set_ip_addr.head.len_lw) { nfp_action += act_size; - act_size = sizeof(set_ip_addr); - memcpy(nfp_action, &set_ip_addr, act_size); + act_size = sizeof(set_act->set_ip_addr); + memcpy(nfp_action, &set_act->set_ip_addr, act_size); *a_len += act_size; /* Hardware will automatically fix IPv4 and TCP/UDP checksum. */ *csum_updated |= TCA_CSUM_UPDATE_FLAG_IPV4HDR | nfp_fl_csum_l4_to_flag(ip_proto); } - if (set_ip6_tc_hl_fl.head.len_lw) { + + if (set_act->set_ip6_tc_hl_fl.head.len_lw) { nfp_action += act_size; - act_size = sizeof(set_ip6_tc_hl_fl); - memcpy(nfp_action, &set_ip6_tc_hl_fl, act_size); + act_size = sizeof(set_act->set_ip6_tc_hl_fl); + memcpy(nfp_action, &set_act->set_ip6_tc_hl_fl, act_size); *a_len += act_size; /* Hardware will automatically fix TCP/UDP checksum. */ *csum_updated |= nfp_fl_csum_l4_to_flag(ip_proto); } - if (set_ip6_dst.head.len_lw && set_ip6_src.head.len_lw) { + + if (set_act->set_ip6_dst.head.len_lw && + set_act->set_ip6_src.head.len_lw) { /* TC compiles set src and dst IPv6 address as a single action, * the hardware requires this to be 2 separate actions. */ nfp_action += act_size; - act_size = sizeof(set_ip6_src); - memcpy(nfp_action, &set_ip6_src, act_size); + act_size = sizeof(set_act->set_ip6_src); + memcpy(nfp_action, &set_act->set_ip6_src, act_size); *a_len += act_size; - act_size = sizeof(set_ip6_dst); - memcpy(&nfp_action[sizeof(set_ip6_src)], &set_ip6_dst, - act_size); + act_size = sizeof(set_act->set_ip6_dst); + memcpy(&nfp_action[sizeof(set_act->set_ip6_src)], + &set_act->set_ip6_dst, act_size); *a_len += act_size; /* Hardware will automatically fix TCP/UDP checksum. */ *csum_updated |= nfp_fl_csum_l4_to_flag(ip_proto); - } else if (set_ip6_dst.head.len_lw) { + } else if (set_act->set_ip6_dst.head.len_lw) { nfp_action += act_size; - act_size = sizeof(set_ip6_dst); - memcpy(nfp_action, &set_ip6_dst, act_size); + act_size = sizeof(set_act->set_ip6_dst); + memcpy(nfp_action, &set_act->set_ip6_dst, act_size); *a_len += act_size; /* Hardware will automatically fix TCP/UDP checksum. */ *csum_updated |= nfp_fl_csum_l4_to_flag(ip_proto); - } else if (set_ip6_src.head.len_lw) { + } else if (set_act->set_ip6_src.head.len_lw) { nfp_action += act_size; - act_size = sizeof(set_ip6_src); - memcpy(nfp_action, &set_ip6_src, act_size); + act_size = sizeof(set_act->set_ip6_src); + memcpy(nfp_action, &set_act->set_ip6_src, act_size); *a_len += act_size; /* Hardware will automatically fix TCP/UDP checksum. */ *csum_updated |= nfp_fl_csum_l4_to_flag(ip_proto); } - if (set_tport.head.len_lw) { + if (set_act->set_tport.head.len_lw) { nfp_action += act_size; - act_size = sizeof(set_tport); - memcpy(nfp_action, &set_tport, act_size); + act_size = sizeof(set_act->set_tport); + memcpy(nfp_action, &set_act->set_tport, act_size); *a_len += act_size; /* Hardware will automatically fix TCP/UDP checksum. */ @@ -725,11 +860,47 @@ nfp_fl_pedit(const struct flow_action_entry *act, } static int -nfp_flower_output_action(struct nfp_app *app, const struct flow_action_entry *act, +nfp_fl_pedit(const struct flow_action_entry *act, + struct flow_cls_offload *flow, char *nfp_action, int *a_len, + u32 *csum_updated, struct nfp_flower_pedit_acts *set_act, + struct netlink_ext_ack *extack) +{ + enum flow_action_mangle_base htype; + u32 offset; + + htype = act->mangle.htype; + offset = act->mangle.offset; + + switch (htype) { + case TCA_PEDIT_KEY_EX_HDR_TYPE_ETH: + return nfp_fl_set_eth(act, offset, &set_act->set_eth, extack); + case TCA_PEDIT_KEY_EX_HDR_TYPE_IP4: + return nfp_fl_set_ip4(act, offset, &set_act->set_ip_addr, + &set_act->set_ip_ttl_tos, extack); + case TCA_PEDIT_KEY_EX_HDR_TYPE_IP6: + return nfp_fl_set_ip6(act, offset, &set_act->set_ip6_dst, + &set_act->set_ip6_src, + &set_act->set_ip6_tc_hl_fl, extack); + case TCA_PEDIT_KEY_EX_HDR_TYPE_TCP: + return nfp_fl_set_tport(act, offset, &set_act->set_tport, + NFP_FL_ACTION_OPCODE_SET_TCP, extack); + case TCA_PEDIT_KEY_EX_HDR_TYPE_UDP: + return nfp_fl_set_tport(act, offset, &set_act->set_tport, + NFP_FL_ACTION_OPCODE_SET_UDP, extack); + default: + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: pedit on unsupported header"); + return -EOPNOTSUPP; + } +} + +static int +nfp_flower_output_action(struct nfp_app *app, + const struct flow_action_entry *act, struct nfp_fl_payload *nfp_fl, int *a_len, struct net_device *netdev, bool last, enum nfp_flower_tun_type *tun_type, int *tun_out_cnt, - int *out_cnt, u32 *csum_updated) + int *out_cnt, u32 *csum_updated, bool pkt_host, + struct netlink_ext_ack *extack) { struct nfp_flower_priv *priv = app->priv; struct nfp_fl_output *output; @@ -738,15 +909,19 @@ nfp_flower_output_action(struct nfp_app *app, const struct flow_action_entry *ac /* If csum_updated has not been reset by now, it means HW will * incorrectly update csums when they are not requested. */ - if (*csum_updated) + if (*csum_updated) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: set actions without updating checksums are not supported"); return -EOPNOTSUPP; + } - if (*a_len + sizeof(struct nfp_fl_output) > NFP_FL_MAX_A_SIZ) + if (*a_len + sizeof(struct nfp_fl_output) > NFP_FL_MAX_A_SIZ) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: mirred output increases action list size beyond the allowed maximum"); return -EOPNOTSUPP; + } output = (struct nfp_fl_output *)&nfp_fl->action_data[*a_len]; err = nfp_fl_output(app, output, act, nfp_fl, last, netdev, *tun_type, - tun_out_cnt); + tun_out_cnt, pkt_host, extack); if (err) return err; @@ -756,11 +931,13 @@ nfp_flower_output_action(struct nfp_app *app, const struct flow_action_entry *ac /* nfp_fl_pre_lag returns -err or size of prelag action added. * This will be 0 if it is not egressing to a lag dev. */ - prelag_size = nfp_fl_pre_lag(app, act, nfp_fl, *a_len); - if (prelag_size < 0) + prelag_size = nfp_fl_pre_lag(app, act, nfp_fl, *a_len, extack); + if (prelag_size < 0) { return prelag_size; - else if (prelag_size > 0 && (!last || *out_cnt)) + } else if (prelag_size > 0 && (!last || *out_cnt)) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: LAG action has to be last action in action list"); return -EOPNOTSUPP; + } *a_len += prelag_size; } @@ -771,39 +948,51 @@ nfp_flower_output_action(struct nfp_app *app, const struct flow_action_entry *ac static int nfp_flower_loop_action(struct nfp_app *app, const struct flow_action_entry *act, - struct tc_cls_flower_offload *flow, + struct flow_cls_offload *flow, struct nfp_fl_payload *nfp_fl, int *a_len, struct net_device *netdev, enum nfp_flower_tun_type *tun_type, int *tun_out_cnt, - int *out_cnt, u32 *csum_updated) + int *out_cnt, u32 *csum_updated, + struct nfp_flower_pedit_acts *set_act, bool *pkt_host, + struct netlink_ext_ack *extack, int act_idx) { - struct nfp_fl_set_ipv4_udp_tun *set_tun; + struct nfp_fl_set_ipv4_tun *set_tun; struct nfp_fl_pre_tunnel *pre_tun; struct nfp_fl_push_vlan *psh_v; + struct nfp_fl_push_mpls *psh_m; struct nfp_fl_pop_vlan *pop_v; + struct nfp_fl_pop_mpls *pop_m; + struct nfp_fl_set_mpls *set_m; int err; switch (act->id) { case FLOW_ACTION_DROP: nfp_fl->meta.shortcut = cpu_to_be32(NFP_FL_SC_ACT_DROP); break; + case FLOW_ACTION_REDIRECT_INGRESS: case FLOW_ACTION_REDIRECT: err = nfp_flower_output_action(app, act, nfp_fl, a_len, netdev, true, tun_type, tun_out_cnt, - out_cnt, csum_updated); + out_cnt, csum_updated, *pkt_host, + extack); if (err) return err; break; + case FLOW_ACTION_MIRRED_INGRESS: case FLOW_ACTION_MIRRED: err = nfp_flower_output_action(app, act, nfp_fl, a_len, netdev, false, tun_type, tun_out_cnt, - out_cnt, csum_updated); + out_cnt, csum_updated, *pkt_host, + extack); if (err) return err; break; case FLOW_ACTION_VLAN_POP: - if (*a_len + sizeof(struct nfp_fl_pop_vlan) > NFP_FL_MAX_A_SIZ) + if (*a_len + + sizeof(struct nfp_fl_pop_vlan) > NFP_FL_MAX_A_SIZ) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: maximum allowed action list size exceeded at pop vlan"); return -EOPNOTSUPP; + } pop_v = (struct nfp_fl_pop_vlan *)&nfp_fl->action_data[*a_len]; nfp_fl->meta.shortcut = cpu_to_be32(NFP_FL_SC_ACT_POPV); @@ -812,8 +1001,11 @@ nfp_flower_loop_action(struct nfp_app *app, const struct flow_action_entry *act, *a_len += sizeof(struct nfp_fl_pop_vlan); break; case FLOW_ACTION_VLAN_PUSH: - if (*a_len + sizeof(struct nfp_fl_push_vlan) > NFP_FL_MAX_A_SIZ) + if (*a_len + + sizeof(struct nfp_fl_push_vlan) > NFP_FL_MAX_A_SIZ) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: maximum allowed action list size exceeded at push vlan"); return -EOPNOTSUPP; + } psh_v = (struct nfp_fl_push_vlan *)&nfp_fl->action_data[*a_len]; nfp_fl->meta.shortcut = cpu_to_be32(NFP_FL_SC_ACT_NULL); @@ -824,35 +1016,41 @@ nfp_flower_loop_action(struct nfp_app *app, const struct flow_action_entry *act, case FLOW_ACTION_TUNNEL_ENCAP: { const struct ip_tunnel_info *ip_tun = act->tunnel; - *tun_type = nfp_fl_get_tun_from_act_l4_port(app, act); - if (*tun_type == NFP_FL_TUNNEL_NONE) + *tun_type = nfp_fl_get_tun_from_act(app, flow, act, act_idx); + if (*tun_type == NFP_FL_TUNNEL_NONE) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: unsupported tunnel type in action list"); return -EOPNOTSUPP; + } - if (ip_tun->mode & ~NFP_FL_SUPPORTED_TUNNEL_INFO_FLAGS) + if (ip_tun->mode & ~NFP_FL_SUPPORTED_TUNNEL_INFO_FLAGS) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: unsupported tunnel flags in action list"); return -EOPNOTSUPP; + } /* Pre-tunnel action is required for tunnel encap. * This checks for next hop entries on NFP. * If none, the packet falls back before applying other actions. */ if (*a_len + sizeof(struct nfp_fl_pre_tunnel) + - sizeof(struct nfp_fl_set_ipv4_udp_tun) > NFP_FL_MAX_A_SIZ) + sizeof(struct nfp_fl_set_ipv4_tun) > NFP_FL_MAX_A_SIZ) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: maximum allowed action list size exceeded at tunnel encap"); return -EOPNOTSUPP; + } pre_tun = nfp_fl_pre_tunnel(nfp_fl->action_data, *a_len); nfp_fl->meta.shortcut = cpu_to_be32(NFP_FL_SC_ACT_NULL); *a_len += sizeof(struct nfp_fl_pre_tunnel); - err = nfp_fl_push_geneve_options(nfp_fl, a_len, act); + err = nfp_fl_push_geneve_options(nfp_fl, a_len, act, extack); if (err) return err; set_tun = (void *)&nfp_fl->action_data[*a_len]; - err = nfp_fl_set_ipv4_udp_tun(app, set_tun, act, pre_tun, - *tun_type, netdev); + err = nfp_fl_set_ipv4_tun(app, set_tun, act, pre_tun, + *tun_type, netdev, extack); if (err) return err; - *a_len += sizeof(struct nfp_fl_set_ipv4_udp_tun); + *a_len += sizeof(struct nfp_fl_set_ipv4_tun); } break; case FLOW_ACTION_TUNNEL_DECAP: @@ -860,34 +1058,124 @@ nfp_flower_loop_action(struct nfp_app *app, const struct flow_action_entry *act, return 0; case FLOW_ACTION_MANGLE: if (nfp_fl_pedit(act, flow, &nfp_fl->action_data[*a_len], - a_len, csum_updated)) + a_len, csum_updated, set_act, extack)) return -EOPNOTSUPP; break; case FLOW_ACTION_CSUM: /* csum action requests recalc of something we have not fixed */ - if (act->csum_flags & ~*csum_updated) + if (act->csum_flags & ~*csum_updated) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: unsupported csum update action in action list"); return -EOPNOTSUPP; + } /* If we will correctly fix the csum we can remove it from the * csum update list. Which will later be used to check support. */ *csum_updated &= ~act->csum_flags; break; + case FLOW_ACTION_MPLS_PUSH: + if (*a_len + + sizeof(struct nfp_fl_push_mpls) > NFP_FL_MAX_A_SIZ) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: maximum allowed action list size exceeded at push MPLS"); + return -EOPNOTSUPP; + } + + psh_m = (struct nfp_fl_push_mpls *)&nfp_fl->action_data[*a_len]; + nfp_fl->meta.shortcut = cpu_to_be32(NFP_FL_SC_ACT_NULL); + + err = nfp_fl_push_mpls(psh_m, act, extack); + if (err) + return err; + *a_len += sizeof(struct nfp_fl_push_mpls); + break; + case FLOW_ACTION_MPLS_POP: + if (*a_len + + sizeof(struct nfp_fl_pop_mpls) > NFP_FL_MAX_A_SIZ) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: maximum allowed action list size exceeded at pop MPLS"); + return -EOPNOTSUPP; + } + + pop_m = (struct nfp_fl_pop_mpls *)&nfp_fl->action_data[*a_len]; + nfp_fl->meta.shortcut = cpu_to_be32(NFP_FL_SC_ACT_NULL); + + nfp_fl_pop_mpls(pop_m, act); + *a_len += sizeof(struct nfp_fl_pop_mpls); + break; + case FLOW_ACTION_MPLS_MANGLE: + if (*a_len + + sizeof(struct nfp_fl_set_mpls) > NFP_FL_MAX_A_SIZ) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: maximum allowed action list size exceeded at set MPLS"); + return -EOPNOTSUPP; + } + + set_m = (struct nfp_fl_set_mpls *)&nfp_fl->action_data[*a_len]; + nfp_fl->meta.shortcut = cpu_to_be32(NFP_FL_SC_ACT_NULL); + + nfp_fl_set_mpls(set_m, act); + *a_len += sizeof(struct nfp_fl_set_mpls); + break; + case FLOW_ACTION_PTYPE: + /* TC ptype skbedit sets PACKET_HOST for ingress redirect. */ + if (act->ptype != PACKET_HOST) + return -EOPNOTSUPP; + + *pkt_host = true; + break; default: /* Currently we do not handle any other actions. */ + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: unsupported action in action list"); return -EOPNOTSUPP; } return 0; } +static bool nfp_fl_check_mangle_start(struct flow_action *flow_act, + int current_act_idx) +{ + struct flow_action_entry current_act; + struct flow_action_entry prev_act; + + current_act = flow_act->entries[current_act_idx]; + if (current_act.id != FLOW_ACTION_MANGLE) + return false; + + if (current_act_idx == 0) + return true; + + prev_act = flow_act->entries[current_act_idx - 1]; + + return prev_act.id != FLOW_ACTION_MANGLE; +} + +static bool nfp_fl_check_mangle_end(struct flow_action *flow_act, + int current_act_idx) +{ + struct flow_action_entry current_act; + struct flow_action_entry next_act; + + current_act = flow_act->entries[current_act_idx]; + if (current_act.id != FLOW_ACTION_MANGLE) + return false; + + if (current_act_idx == flow_act->num_entries) + return true; + + next_act = flow_act->entries[current_act_idx + 1]; + + return next_act.id != FLOW_ACTION_MANGLE; +} + int nfp_flower_compile_action(struct nfp_app *app, - struct tc_cls_flower_offload *flow, + struct flow_cls_offload *flow, struct net_device *netdev, - struct nfp_fl_payload *nfp_flow) + struct nfp_fl_payload *nfp_flow, + struct netlink_ext_ack *extack) { int act_len, act_cnt, err, tun_out_cnt, out_cnt, i; + struct nfp_flower_pedit_acts set_act; enum nfp_flower_tun_type tun_type; struct flow_action_entry *act; + bool pkt_host = false; u32 csum_updated = 0; memset(nfp_flow->action_data, 0, NFP_FL_MAX_A_SIZ); @@ -899,12 +1187,19 @@ int nfp_flower_compile_action(struct nfp_app *app, out_cnt = 0; flow_action_for_each(i, act, &flow->rule->action) { + if (nfp_fl_check_mangle_start(&flow->rule->action, i)) + memset(&set_act, 0, sizeof(set_act)); err = nfp_flower_loop_action(app, act, flow, nfp_flow, &act_len, netdev, &tun_type, &tun_out_cnt, - &out_cnt, &csum_updated); + &out_cnt, &csum_updated, + &set_act, &pkt_host, extack, i); if (err) return err; act_cnt++; + if (nfp_fl_check_mangle_end(&flow->rule->action, i)) + nfp_fl_commit_mangle(flow, + &nfp_flow->action_data[act_len], + &act_len, &set_act, &csum_updated); } /* We optimise when the action list is small, this can unfortunately diff --git a/drivers/net/ethernet/netronome/nfp/flower/cmsg.c b/drivers/net/ethernet/netronome/nfp/flower/cmsg.c index cf9e1118ee8f..05981b54eaab 100644 --- a/drivers/net/ethernet/netronome/nfp/flower/cmsg.c +++ b/drivers/net/ethernet/netronome/nfp/flower/cmsg.c @@ -159,7 +159,7 @@ nfp_flower_cmsg_portmod_rx(struct nfp_app *app, struct sk_buff *skb) rtnl_lock(); rcu_read_lock(); - netdev = nfp_app_repr_get(app, be32_to_cpu(msg->portnum)); + netdev = nfp_app_dev_get(app, be32_to_cpu(msg->portnum), NULL); rcu_read_unlock(); if (!netdev) { nfp_flower_cmsg_warn(app, "ctrl msg for unknown port 0x%08x\n", @@ -192,7 +192,7 @@ nfp_flower_cmsg_portreify_rx(struct nfp_app *app, struct sk_buff *skb) msg = nfp_flower_cmsg_get_data(skb); rcu_read_lock(); - exists = !!nfp_app_repr_get(app, be32_to_cpu(msg->portnum)); + exists = !!nfp_app_dev_get(app, be32_to_cpu(msg->portnum), NULL); rcu_read_unlock(); if (!exists) { nfp_flower_cmsg_warn(app, "ctrl msg for unknown port 0x%08x\n", @@ -205,6 +205,50 @@ nfp_flower_cmsg_portreify_rx(struct nfp_app *app, struct sk_buff *skb) } static void +nfp_flower_cmsg_merge_hint_rx(struct nfp_app *app, struct sk_buff *skb) +{ + unsigned int msg_len = nfp_flower_cmsg_get_data_len(skb); + struct nfp_flower_cmsg_merge_hint *msg; + struct nfp_fl_payload *sub_flows[2]; + int err, i, flow_cnt; + + msg = nfp_flower_cmsg_get_data(skb); + /* msg->count starts at 0 and always assumes at least 1 entry. */ + flow_cnt = msg->count + 1; + + if (msg_len < struct_size(msg, flow, flow_cnt)) { + nfp_flower_cmsg_warn(app, "Merge hint ctrl msg too short - %d bytes but expect %zd\n", + msg_len, struct_size(msg, flow, flow_cnt)); + return; + } + + if (flow_cnt != 2) { + nfp_flower_cmsg_warn(app, "Merge hint contains %d flows - two are expected\n", + flow_cnt); + return; + } + + rtnl_lock(); + for (i = 0; i < flow_cnt; i++) { + u32 ctx = be32_to_cpu(msg->flow[i].host_ctx); + + sub_flows[i] = nfp_flower_get_fl_payload_from_ctx(app, ctx); + if (!sub_flows[i]) { + nfp_flower_cmsg_warn(app, "Invalid flow in merge hint\n"); + goto err_rtnl_unlock; + } + } + + err = nfp_flower_merge_offloaded_flows(app, sub_flows[0], sub_flows[1]); + /* Only warn on memory fail. Hint veto will not break functionality. */ + if (err == -ENOMEM) + nfp_flower_cmsg_warn(app, "Flow merge memory fail.\n"); + +err_rtnl_unlock: + rtnl_unlock(); +} + +static void nfp_flower_cmsg_process_one_rx(struct nfp_app *app, struct sk_buff *skb) { struct nfp_flower_priv *app_priv = app->priv; @@ -216,18 +260,24 @@ nfp_flower_cmsg_process_one_rx(struct nfp_app *app, struct sk_buff *skb) type = cmsg_hdr->type; switch (type) { - case NFP_FLOWER_CMSG_TYPE_PORT_REIFY: - nfp_flower_cmsg_portreify_rx(app, skb); - break; case NFP_FLOWER_CMSG_TYPE_PORT_MOD: nfp_flower_cmsg_portmod_rx(app, skb); break; + case NFP_FLOWER_CMSG_TYPE_MERGE_HINT: + if (app_priv->flower_ext_feats & NFP_FL_FEATS_FLOW_MERGE) { + nfp_flower_cmsg_merge_hint_rx(app, skb); + break; + } + goto err_default; case NFP_FLOWER_CMSG_TYPE_NO_NEIGH: nfp_tunnel_request_route(app, skb); break; case NFP_FLOWER_CMSG_TYPE_ACTIVE_TUNS: nfp_tunnel_keep_alive(app, skb); break; + case NFP_FLOWER_CMSG_TYPE_QOS_STATS: + nfp_flower_stats_rlim_reply(app, skb); + break; case NFP_FLOWER_CMSG_TYPE_LAG_CONFIG: if (app_priv->flower_ext_feats & NFP_FL_FEATS_LAG) { skb_stored = nfp_flower_lag_unprocessed_msg(app, skb); @@ -235,6 +285,7 @@ nfp_flower_cmsg_process_one_rx(struct nfp_app *app, struct sk_buff *skb) } /* fall through */ default: +err_default: nfp_flower_cmsg_warn(app, "Cannot handle invalid repr control type %u\n", type); goto out; @@ -274,8 +325,7 @@ nfp_flower_queue_ctl_msg(struct nfp_app *app, struct sk_buff *skb, int type) struct nfp_flower_priv *priv = app->priv; struct sk_buff_head *skb_head; - if (type == NFP_FLOWER_CMSG_TYPE_PORT_REIFY || - type == NFP_FLOWER_CMSG_TYPE_PORT_MOD) + if (type == NFP_FLOWER_CMSG_TYPE_PORT_MOD) skb_head = &priv->cmsg_skbs_high; else skb_head = &priv->cmsg_skbs_low; @@ -314,6 +364,10 @@ void nfp_flower_cmsg_rx(struct nfp_app *app, struct sk_buff *skb) } else if (cmsg_hdr->type == NFP_FLOWER_CMSG_TYPE_TUN_NEIGH) { /* Acks from the NFP that the route is added - ignore. */ dev_consume_skb_any(skb); + } else if (cmsg_hdr->type == NFP_FLOWER_CMSG_TYPE_PORT_REIFY) { + /* Handle REIFY acks outside wq to prevent RTNL conflict. */ + nfp_flower_cmsg_portreify_rx(app, skb); + dev_consume_skb_any(skb); } else { nfp_flower_queue_ctl_msg(app, skb, cmsg_hdr->type); } diff --git a/drivers/net/ethernet/netronome/nfp/flower/cmsg.h b/drivers/net/ethernet/netronome/nfp/flower/cmsg.h index 0ed51e79db00..7eb2ec8969c3 100644 --- a/drivers/net/ethernet/netronome/nfp/flower/cmsg.h +++ b/drivers/net/ethernet/netronome/nfp/flower/cmsg.h @@ -8,6 +8,7 @@ #include <linux/skbuff.h> #include <linux/types.h> #include <net/geneve.h> +#include <net/gre.h> #include <net/vxlan.h> #include "../nfp_app.h" @@ -22,6 +23,7 @@ #define NFP_FLOWER_LAYER_CT BIT(6) #define NFP_FLOWER_LAYER_VXLAN BIT(7) +#define NFP_FLOWER_LAYER2_GRE BIT(0) #define NFP_FLOWER_LAYER2_GENEVE BIT(5) #define NFP_FLOWER_LAYER2_GENEVE_OP BIT(6) @@ -37,6 +39,9 @@ #define NFP_FL_IP_FRAG_FIRST BIT(7) #define NFP_FL_IP_FRAGMENTED BIT(6) +/* GRE Tunnel flags */ +#define NFP_FL_GRE_FLAG_KEY BIT(2) + /* Compressed HW representation of TCP Flags */ #define NFP_FL_TCP_FLAG_URG BIT(4) #define NFP_FL_TCP_FLAG_PSH BIT(3) @@ -63,8 +68,11 @@ #define NFP_FL_ACTION_OPCODE_OUTPUT 0 #define NFP_FL_ACTION_OPCODE_PUSH_VLAN 1 #define NFP_FL_ACTION_OPCODE_POP_VLAN 2 +#define NFP_FL_ACTION_OPCODE_PUSH_MPLS 3 +#define NFP_FL_ACTION_OPCODE_POP_MPLS 4 #define NFP_FL_ACTION_OPCODE_SET_IPV4_TUNNEL 6 #define NFP_FL_ACTION_OPCODE_SET_ETHERNET 7 +#define NFP_FL_ACTION_OPCODE_SET_MPLS 8 #define NFP_FL_ACTION_OPCODE_SET_IPV4_ADDRS 9 #define NFP_FL_ACTION_OPCODE_SET_IPV4_TTL_TOS 10 #define NFP_FL_ACTION_OPCODE_SET_IPV6_SRC 11 @@ -107,6 +115,7 @@ enum nfp_flower_tun_type { NFP_FL_TUNNEL_NONE = 0, + NFP_FL_TUNNEL_GRE = 1, NFP_FL_TUNNEL_VXLAN = 2, NFP_FL_TUNNEL_GENEVE = 4, }; @@ -203,7 +212,7 @@ struct nfp_fl_pre_tunnel { __be32 extra[3]; }; -struct nfp_fl_set_ipv4_udp_tun { +struct nfp_fl_set_ipv4_tun { struct nfp_fl_act_head head; __be16 reserved; __be64 tun_id __packed; @@ -211,7 +220,8 @@ struct nfp_fl_set_ipv4_udp_tun { __be16 tun_flags; u8 ttl; u8 tos; - __be32 extra; + __be16 outer_vlan_tpid; + __be16 outer_vlan_tci; u8 tun_len; u8 res2; __be16 tun_proto; @@ -226,6 +236,24 @@ struct nfp_fl_push_geneve { u8 opt_data[]; }; +struct nfp_fl_push_mpls { + struct nfp_fl_act_head head; + __be16 ethtype; + __be32 lse; +}; + +struct nfp_fl_pop_mpls { + struct nfp_fl_act_head head; + __be16 ethtype; +}; + +struct nfp_fl_set_mpls { + struct nfp_fl_act_head head; + __be16 reserved; + __be32 lse_mask; + __be32 lse; +}; + /* Metadata with L2 (1W/4B) * ---------------------------------------------------------------- * 3 2 1 @@ -354,6 +382,16 @@ struct nfp_flower_ipv6 { struct in6_addr ipv6_dst; }; +struct nfp_flower_tun_ipv4 { + __be32 src; + __be32 dst; +}; + +struct nfp_flower_tun_ip_ext { + u8 tos; + u8 ttl; +}; + /* Flow Frame IPv4 UDP TUNNEL --> Tunnel details (4W/16B) * ----------------------------------------------------------------- * 3 2 1 @@ -371,15 +409,42 @@ struct nfp_flower_ipv6 { * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ */ struct nfp_flower_ipv4_udp_tun { - __be32 ip_src; - __be32 ip_dst; + struct nfp_flower_tun_ipv4 ipv4; __be16 reserved1; - u8 tos; - u8 ttl; + struct nfp_flower_tun_ip_ext ip_ext; __be32 reserved2; __be32 tun_id; }; +/* Flow Frame GRE TUNNEL --> Tunnel details (6W/24B) + * ----------------------------------------------------------------- + * 3 2 1 + * 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + * | ipv4_addr_src | + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + * | ipv4_addr_dst | + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + * | tun_flags | tos | ttl | + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + * | Reserved | Ethertype | + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + * | Key | + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + * | Reserved | + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + */ + +struct nfp_flower_ipv4_gre_tun { + struct nfp_flower_tun_ipv4 ipv4; + __be16 tun_flags; + struct nfp_flower_tun_ip_ext ip_ext; + __be16 reserved1; + __be16 ethertype; + __be32 tun_key; + __be32 reserved2; +}; + struct nfp_flower_geneve_options { u8 data[NFP_FL_MAX_GENEVE_OPT_KEY]; }; @@ -402,11 +467,13 @@ struct nfp_flower_cmsg_hdr { /* Types defined for port related control messages */ enum nfp_flower_cmsg_type_port { NFP_FLOWER_CMSG_TYPE_FLOW_ADD = 0, + NFP_FLOWER_CMSG_TYPE_FLOW_MOD = 1, NFP_FLOWER_CMSG_TYPE_FLOW_DEL = 2, NFP_FLOWER_CMSG_TYPE_LAG_CONFIG = 4, NFP_FLOWER_CMSG_TYPE_PORT_REIFY = 6, NFP_FLOWER_CMSG_TYPE_MAC_REPR = 7, NFP_FLOWER_CMSG_TYPE_PORT_MOD = 8, + NFP_FLOWER_CMSG_TYPE_MERGE_HINT = 9, NFP_FLOWER_CMSG_TYPE_NO_NEIGH = 10, NFP_FLOWER_CMSG_TYPE_TUN_MAC = 11, NFP_FLOWER_CMSG_TYPE_ACTIVE_TUNS = 12, @@ -414,6 +481,10 @@ enum nfp_flower_cmsg_type_port { NFP_FLOWER_CMSG_TYPE_TUN_IPS = 14, NFP_FLOWER_CMSG_TYPE_FLOW_STATS = 15, NFP_FLOWER_CMSG_TYPE_PORT_ECHO = 16, + NFP_FLOWER_CMSG_TYPE_QOS_MOD = 18, + NFP_FLOWER_CMSG_TYPE_QOS_DEL = 19, + NFP_FLOWER_CMSG_TYPE_QOS_STATS = 20, + NFP_FLOWER_CMSG_TYPE_PRE_TUN_RULE = 21, NFP_FLOWER_CMSG_TYPE_MAX = 32, }; @@ -451,6 +522,16 @@ struct nfp_flower_cmsg_portreify { #define NFP_FLOWER_CMSG_PORTREIFY_INFO_EXIST BIT(0) +/* NFP_FLOWER_CMSG_TYPE_FLOW_MERGE_HINT */ +struct nfp_flower_cmsg_merge_hint { + u8 reserved[3]; + u8 count; + struct { + __be32 host_ctx; + __be64 host_cookie; + } __packed flow[0]; +}; + enum nfp_flower_cmsg_port_type { NFP_FLOWER_CMSG_PORT_TYPE_UNSPEC = 0x0, NFP_FLOWER_CMSG_PORT_TYPE_PHYS_PORT = 0x1, @@ -473,6 +554,13 @@ enum nfp_flower_cmsg_port_vnic_type { #define NFP_FLOWER_CMSG_PORT_PCIE_Q GENMASK(5, 0) #define NFP_FLOWER_CMSG_PORT_PHYS_PORT_NUM GENMASK(7, 0) +static inline u32 nfp_flower_internal_port_get_port_id(u8 internal_port) +{ + return FIELD_PREP(NFP_FLOWER_CMSG_PORT_PHYS_PORT_NUM, internal_port) | + FIELD_PREP(NFP_FLOWER_CMSG_PORT_TYPE, + NFP_FLOWER_CMSG_PORT_TYPE_OTHER_PORT); +} + static inline u32 nfp_flower_cmsg_phys_port(u8 phys_port) { return FIELD_PREP(NFP_FLOWER_CMSG_PORT_PHYS_PORT_NUM, phys_port) | @@ -508,6 +596,8 @@ nfp_fl_netdev_is_tunnel_type(struct net_device *netdev, { if (netif_is_vxlan(netdev)) return tun_type == NFP_FL_TUNNEL_VXLAN; + if (netif_is_gretap(netdev)) + return tun_type == NFP_FL_TUNNEL_GRE; if (netif_is_geneve(netdev)) return tun_type == NFP_FL_TUNNEL_GENEVE; @@ -524,6 +614,8 @@ static inline bool nfp_fl_is_netdev_to_offload(struct net_device *netdev) return true; if (netif_is_geneve(netdev)) return true; + if (netif_is_gretap(netdev)) + return true; return false; } diff --git a/drivers/net/ethernet/netronome/nfp/flower/lag_conf.c b/drivers/net/ethernet/netronome/nfp/flower/lag_conf.c index 5db838f45694..63907aeb3884 100644 --- a/drivers/net/ethernet/netronome/nfp/flower/lag_conf.c +++ b/drivers/net/ethernet/netronome/nfp/flower/lag_conf.c @@ -156,7 +156,8 @@ nfp_fl_lag_find_group_for_master_with_lag(struct nfp_fl_lag *lag, int nfp_flower_lag_populate_pre_action(struct nfp_app *app, struct net_device *master, - struct nfp_fl_pre_lag *pre_act) + struct nfp_fl_pre_lag *pre_act, + struct netlink_ext_ack *extack) { struct nfp_flower_priv *priv = app->priv; struct nfp_fl_lag_group *group = NULL; @@ -167,6 +168,7 @@ int nfp_flower_lag_populate_pre_action(struct nfp_app *app, master); if (!group) { mutex_unlock(&priv->nfp_lag.lock); + NL_SET_ERR_MSG_MOD(extack, "invalid entry: group does not exist for LAG action"); return -ENOENT; } diff --git a/drivers/net/ethernet/netronome/nfp/flower/main.c b/drivers/net/ethernet/netronome/nfp/flower/main.c index 408089133599..d8ad9346a26a 100644 --- a/drivers/net/ethernet/netronome/nfp/flower/main.c +++ b/drivers/net/ethernet/netronome/nfp/flower/main.c @@ -22,6 +22,9 @@ #define NFP_FLOWER_ALLOWED_VER 0x0001000000010000UL +#define NFP_MIN_INT_PORT_ID 1 +#define NFP_MAX_INT_PORT_ID 256 + static const char *nfp_flower_extra_cap(struct nfp_app *app, struct nfp_net *nn) { return "FLOWER"; @@ -32,6 +35,113 @@ static enum devlink_eswitch_mode eswitch_mode_get(struct nfp_app *app) return DEVLINK_ESWITCH_MODE_SWITCHDEV; } +static int +nfp_flower_lookup_internal_port_id(struct nfp_flower_priv *priv, + struct net_device *netdev) +{ + struct net_device *entry; + int i, id = 0; + + rcu_read_lock(); + idr_for_each_entry(&priv->internal_ports.port_ids, entry, i) + if (entry == netdev) { + id = i; + break; + } + rcu_read_unlock(); + + return id; +} + +static int +nfp_flower_get_internal_port_id(struct nfp_app *app, struct net_device *netdev) +{ + struct nfp_flower_priv *priv = app->priv; + int id; + + id = nfp_flower_lookup_internal_port_id(priv, netdev); + if (id > 0) + return id; + + idr_preload(GFP_ATOMIC); + spin_lock_bh(&priv->internal_ports.lock); + id = idr_alloc(&priv->internal_ports.port_ids, netdev, + NFP_MIN_INT_PORT_ID, NFP_MAX_INT_PORT_ID, GFP_ATOMIC); + spin_unlock_bh(&priv->internal_ports.lock); + idr_preload_end(); + + return id; +} + +u32 nfp_flower_get_port_id_from_netdev(struct nfp_app *app, + struct net_device *netdev) +{ + int ext_port; + + if (nfp_netdev_is_nfp_repr(netdev)) { + return nfp_repr_get_port_id(netdev); + } else if (nfp_flower_internal_port_can_offload(app, netdev)) { + ext_port = nfp_flower_get_internal_port_id(app, netdev); + if (ext_port < 0) + return 0; + + return nfp_flower_internal_port_get_port_id(ext_port); + } + + return 0; +} + +static struct net_device * +nfp_flower_get_netdev_from_internal_port_id(struct nfp_app *app, int port_id) +{ + struct nfp_flower_priv *priv = app->priv; + struct net_device *netdev; + + rcu_read_lock(); + netdev = idr_find(&priv->internal_ports.port_ids, port_id); + rcu_read_unlock(); + + return netdev; +} + +static void +nfp_flower_free_internal_port_id(struct nfp_app *app, struct net_device *netdev) +{ + struct nfp_flower_priv *priv = app->priv; + int id; + + id = nfp_flower_lookup_internal_port_id(priv, netdev); + if (!id) + return; + + spin_lock_bh(&priv->internal_ports.lock); + idr_remove(&priv->internal_ports.port_ids, id); + spin_unlock_bh(&priv->internal_ports.lock); +} + +static int +nfp_flower_internal_port_event_handler(struct nfp_app *app, + struct net_device *netdev, + unsigned long event) +{ + if (event == NETDEV_UNREGISTER && + nfp_flower_internal_port_can_offload(app, netdev)) + nfp_flower_free_internal_port_id(app, netdev); + + return NOTIFY_OK; +} + +static void nfp_flower_internal_port_init(struct nfp_flower_priv *priv) +{ + spin_lock_init(&priv->internal_ports.lock); + idr_init(&priv->internal_ports.port_ids); +} + +static void nfp_flower_internal_port_cleanup(struct nfp_flower_priv *priv) +{ + idr_destroy(&priv->internal_ports.port_ids); +} + static struct nfp_flower_non_repr_priv * nfp_flower_non_repr_priv_lookup(struct nfp_app *app, struct net_device *netdev) { @@ -119,12 +229,21 @@ nfp_flower_repr_get_type_and_port(struct nfp_app *app, u32 port_id, u8 *port) } static struct net_device * -nfp_flower_repr_get(struct nfp_app *app, u32 port_id) +nfp_flower_dev_get(struct nfp_app *app, u32 port_id, bool *redir_egress) { enum nfp_repr_type repr_type; struct nfp_reprs *reprs; u8 port = 0; + /* Check if the port is internal. */ + if (FIELD_GET(NFP_FLOWER_CMSG_PORT_TYPE, port_id) == + NFP_FLOWER_CMSG_PORT_TYPE_OTHER_PORT) { + if (redir_egress) + *redir_egress = true; + port = FIELD_GET(NFP_FLOWER_CMSG_PORT_PHYS_PORT_NUM, port_id); + return nfp_flower_get_netdev_from_internal_port_id(app, port); + } + repr_type = nfp_flower_repr_get_type_and_port(app, port_id, &port); if (repr_type > NFP_REPR_TYPE_MAX) return NULL; @@ -281,6 +400,7 @@ nfp_flower_spawn_vnic_reprs(struct nfp_app *app, repr_priv = kzalloc(sizeof(*repr_priv), GFP_KERNEL); if (!repr_priv) { err = -ENOMEM; + nfp_repr_free(repr); goto err_reprs_clean; } @@ -294,6 +414,7 @@ nfp_flower_spawn_vnic_reprs(struct nfp_app *app, port = nfp_port_alloc(app, port_type, repr); if (IS_ERR(port)) { err = PTR_ERR(port); + kfree(repr_priv); nfp_repr_free(repr); goto err_reprs_clean; } @@ -314,6 +435,7 @@ nfp_flower_spawn_vnic_reprs(struct nfp_app *app, err = nfp_repr_init(app, repr, port_id, port, priv->nn->dp.netdev); if (err) { + kfree(repr_priv); nfp_port_free(port); nfp_repr_free(repr); goto err_reprs_clean; @@ -396,6 +518,7 @@ nfp_flower_spawn_phy_reprs(struct nfp_app *app, struct nfp_flower_priv *priv) repr_priv = kzalloc(sizeof(*repr_priv), GFP_KERNEL); if (!repr_priv) { err = -ENOMEM; + nfp_repr_free(repr); goto err_reprs_clean; } @@ -406,11 +529,13 @@ nfp_flower_spawn_phy_reprs(struct nfp_app *app, struct nfp_flower_priv *priv) port = nfp_port_alloc(app, NFP_PORT_PHYS_PORT, repr); if (IS_ERR(port)) { err = PTR_ERR(port); + kfree(repr_priv); nfp_repr_free(repr); goto err_reprs_clean; } err = nfp_port_init_phy_port(app->pf, app, port, i); if (err) { + kfree(repr_priv); nfp_port_free(port); nfp_repr_free(repr); goto err_reprs_clean; @@ -423,6 +548,7 @@ nfp_flower_spawn_phy_reprs(struct nfp_app *app, struct nfp_flower_priv *priv) err = nfp_repr_init(app, repr, cmsg_port_id, port, priv->nn->dp.netdev); if (err) { + kfree(repr_priv); nfp_port_free(port); nfp_repr_free(repr); goto err_reprs_clean; @@ -641,11 +767,34 @@ static int nfp_flower_init(struct nfp_app *app) goto err_cleanup_metadata; } + if (app_priv->flower_ext_feats & NFP_FL_FEATS_FLOW_MOD) { + /* Tell the firmware that the driver supports flow merging. */ + err = nfp_rtsym_write_le(app->pf->rtbl, + "_abi_flower_merge_hint_enable", 1); + if (!err) { + app_priv->flower_ext_feats |= NFP_FL_FEATS_FLOW_MERGE; + nfp_flower_internal_port_init(app_priv); + } else if (err == -ENOENT) { + nfp_warn(app->cpp, "Flow merge not supported by FW.\n"); + } else { + goto err_lag_clean; + } + } else { + nfp_warn(app->cpp, "Flow mod/merge not supported by FW.\n"); + } + + if (app_priv->flower_ext_feats & NFP_FL_FEATS_VF_RLIM) + nfp_flower_qos_init(app); + INIT_LIST_HEAD(&app_priv->indr_block_cb_priv); INIT_LIST_HEAD(&app_priv->non_repr_priv); + app_priv->pre_tun_rule_cnt = 0; return 0; +err_lag_clean: + if (app_priv->flower_ext_feats & NFP_FL_FEATS_LAG) + nfp_flower_lag_cleanup(&app_priv->nfp_lag); err_cleanup_metadata: nfp_flower_metadata_cleanup(app); err_free_app_priv: @@ -661,9 +810,15 @@ static void nfp_flower_clean(struct nfp_app *app) skb_queue_purge(&app_priv->cmsg_skbs_low); flush_work(&app_priv->cmsg_work); + if (app_priv->flower_ext_feats & NFP_FL_FEATS_VF_RLIM) + nfp_flower_qos_cleanup(app); + if (app_priv->flower_ext_feats & NFP_FL_FEATS_LAG) nfp_flower_lag_cleanup(&app_priv->nfp_lag); + if (app_priv->flower_ext_feats & NFP_FL_FEATS_FLOW_MERGE) + nfp_flower_internal_port_cleanup(app_priv); + nfp_flower_metadata_cleanup(app); vfree(app->priv); app->priv = NULL; @@ -762,6 +917,10 @@ nfp_flower_netdev_event(struct nfp_app *app, struct net_device *netdev, if (ret & NOTIFY_STOP_MASK) return ret; + ret = nfp_flower_internal_port_event_handler(app, netdev, event); + if (ret & NOTIFY_STOP_MASK) + return ret; + return nfp_tunnel_mac_event_handler(app, netdev, event, ptr); } @@ -800,7 +959,7 @@ const struct nfp_app_type app_flower = { .sriov_disable = nfp_flower_sriov_disable, .eswitch_mode_get = eswitch_mode_get, - .repr_get = nfp_flower_repr_get, + .dev_get = nfp_flower_dev_get, .setup_tc = nfp_flower_setup_tc, }; diff --git a/drivers/net/ethernet/netronome/nfp/flower/main.h b/drivers/net/ethernet/netronome/nfp/flower/main.h index c0945a5fd1a4..31d94592a7c0 100644 --- a/drivers/net/ethernet/netronome/nfp/flower/main.h +++ b/drivers/net/ethernet/netronome/nfp/flower/main.h @@ -5,6 +5,7 @@ #define __NFP_FLOWER_H__ 1 #include "cmsg.h" +#include "../nfp_net.h" #include <linux/circ_buf.h> #include <linux/hashtable.h> @@ -34,14 +35,15 @@ struct nfp_app; #define NFP_FL_MASK_REUSE_TIME_NS 40000 #define NFP_FL_MASK_ID_LOCATION 1 -#define NFP_FL_VXLAN_PORT 4789 -#define NFP_FL_GENEVE_PORT 6081 - /* Extra features bitmap. */ #define NFP_FL_FEATS_GENEVE BIT(0) #define NFP_FL_NBI_MTU_SETTING BIT(1) #define NFP_FL_FEATS_GENEVE_OPT BIT(2) #define NFP_FL_FEATS_VLAN_PCP BIT(3) +#define NFP_FL_FEATS_VF_RLIM BIT(4) +#define NFP_FL_FEATS_FLOW_MOD BIT(5) +#define NFP_FL_FEATS_PRE_TUN_RULES BIT(6) +#define NFP_FL_FEATS_FLOW_MERGE BIT(30) #define NFP_FL_FEATS_LAG BIT(31) struct nfp_fl_mask_id { @@ -118,6 +120,16 @@ struct nfp_fl_lag { }; /** + * struct nfp_fl_internal_ports - Flower APP priv data for additional ports + * @port_ids: Assignment of ids to any additional ports + * @lock: Lock for extra ports list + */ +struct nfp_fl_internal_ports { + struct idr port_ids; + spinlock_t lock; +}; + +/** * struct nfp_flower_priv - Flower APP per-vNIC priv data * @app: Back pointer to app * @nn: Pointer to vNIC @@ -131,6 +143,7 @@ struct nfp_fl_lag { * @flow_table: Hash table used to store flower rules * @stats: Stored stats updates for flower rules * @stats_lock: Lock for flower rule stats updates + * @stats_ctx_table: Hash table to map stats contexts to its flow rule * @cmsg_work: Workqueue for control messages processing * @cmsg_skbs_high: List of higher priority skbs for control message * processing @@ -146,6 +159,11 @@ struct nfp_fl_lag { * @non_repr_priv: List of offloaded non-repr ports and their priv data * @active_mem_unit: Current active memory unit for flower rules * @total_mem_units: Total number of available memory units for flower rules + * @internal_ports: Internal port ids used in offloaded rules + * @qos_stats_work: Workqueue for qos stats processing + * @qos_rate_limiters: Current active qos rate limiters + * @qos_stats_lock: Lock on qos stats updates + * @pre_tun_rule_cnt: Number of pre-tunnel rules offloaded */ struct nfp_flower_priv { struct nfp_app *app; @@ -160,6 +178,7 @@ struct nfp_flower_priv { struct rhashtable flow_table; struct nfp_fl_stats *stats; spinlock_t stats_lock; /* lock stats */ + struct rhashtable stats_ctx_table; struct work_struct cmsg_work; struct sk_buff_head cmsg_skbs_high; struct sk_buff_head cmsg_skbs_low; @@ -172,6 +191,25 @@ struct nfp_flower_priv { struct list_head non_repr_priv; unsigned int active_mem_unit; unsigned int total_mem_units; + struct nfp_fl_internal_ports internal_ports; + struct delayed_work qos_stats_work; + unsigned int qos_rate_limiters; + spinlock_t qos_stats_lock; /* Protect the qos stats */ + int pre_tun_rule_cnt; +}; + +/** + * struct nfp_fl_qos - Flower APP priv data for quality of service + * @netdev_port_id: NFP port number of repr with qos info + * @curr_stats: Currently stored stats updates for qos info + * @prev_stats: Previously stored updates for qos info + * @last_update: Stored time when last stats were updated + */ +struct nfp_fl_qos { + u32 netdev_port_id; + struct nfp_stat_pair curr_stats; + struct nfp_stat_pair prev_stats; + u64 last_update; }; /** @@ -180,14 +218,20 @@ struct nfp_flower_priv { * @lag_port_flags: Extended port flags to record lag state of repr * @mac_offloaded: Flag indicating a MAC address is offloaded for repr * @offloaded_mac_addr: MAC address that has been offloaded for repr + * @block_shared: Flag indicating if offload applies to shared blocks * @mac_list: List entry of reprs that share the same offloaded MAC + * @qos_table: Stored info on filters implementing qos + * @on_bridge: Indicates if the repr is attached to a bridge */ struct nfp_flower_repr_priv { struct nfp_repr *nfp_repr; unsigned long lag_port_flags; bool mac_offloaded; u8 offloaded_mac_addr[ETH_ALEN]; + bool block_shared; struct list_head mac_list; + struct nfp_fl_qos qos_table; + bool on_bridge; }; /** @@ -239,6 +283,30 @@ struct nfp_fl_payload { char *unmasked_data; char *mask_data; char *action_data; + struct list_head linked_flows; + bool in_hw; + struct { + struct net_device *dev; + __be16 vlan_tci; + __be16 port_idx; + } pre_tun_rule; +}; + +struct nfp_fl_payload_link { + /* A link contains a pointer to a merge flow and an associated sub_flow. + * Each merge flow will feature in 2 links to its underlying sub_flows. + * A sub_flow will have at least 1 link to a merge flow or more if it + * has been used to create multiple merge flows. + * + * For a merge flow, 'linked_flows' in its nfp_fl_payload struct lists + * all links to sub_flows (sub_flow.flow) via merge.list. + * For a sub_flow, 'linked_flows' gives all links to merge flows it has + * formed (merge_flow.flow) via sub_flow.list. + */ + struct { + struct list_head list; + struct nfp_fl_payload *flow; + } merge_flow, sub_flow; }; extern const struct rhashtable_params nfp_flower_table_params; @@ -250,26 +318,64 @@ struct nfp_fl_stats_frame { __be64 stats_cookie; }; +static inline bool +nfp_flower_internal_port_can_offload(struct nfp_app *app, + struct net_device *netdev) +{ + struct nfp_flower_priv *app_priv = app->priv; + + if (!(app_priv->flower_ext_feats & NFP_FL_FEATS_FLOW_MERGE)) + return false; + if (!netdev->rtnl_link_ops) + return false; + if (!strcmp(netdev->rtnl_link_ops->kind, "openvswitch")) + return true; + + return false; +} + +/* The address of the merged flow acts as its cookie. + * Cookies supplied to us by TC flower are also addresses to allocated + * memory and thus this scheme should not generate any collisions. + */ +static inline bool nfp_flower_is_merge_flow(struct nfp_fl_payload *flow_pay) +{ + return flow_pay->tc_flower_cookie == (unsigned long)flow_pay; +} + +static inline bool nfp_flower_is_supported_bridge(struct net_device *netdev) +{ + return netif_is_ovs_master(netdev); +} + int nfp_flower_metadata_init(struct nfp_app *app, u64 host_ctx_count, unsigned int host_ctx_split); void nfp_flower_metadata_cleanup(struct nfp_app *app); int nfp_flower_setup_tc(struct nfp_app *app, struct net_device *netdev, enum tc_setup_type type, void *type_data); +int nfp_flower_merge_offloaded_flows(struct nfp_app *app, + struct nfp_fl_payload *sub_flow1, + struct nfp_fl_payload *sub_flow2); int nfp_flower_compile_flow_match(struct nfp_app *app, - struct tc_cls_flower_offload *flow, + struct flow_cls_offload *flow, struct nfp_fl_key_ls *key_ls, struct net_device *netdev, struct nfp_fl_payload *nfp_flow, - enum nfp_flower_tun_type tun_type); + enum nfp_flower_tun_type tun_type, + struct netlink_ext_ack *extack); int nfp_flower_compile_action(struct nfp_app *app, - struct tc_cls_flower_offload *flow, + struct flow_cls_offload *flow, struct net_device *netdev, - struct nfp_fl_payload *nfp_flow); + struct nfp_fl_payload *nfp_flow, + struct netlink_ext_ack *extack); int nfp_compile_flow_metadata(struct nfp_app *app, - struct tc_cls_flower_offload *flow, + struct flow_cls_offload *flow, struct nfp_fl_payload *nfp_flow, - struct net_device *netdev); + struct net_device *netdev, + struct netlink_ext_ack *extack); +void __nfp_modify_flow_metadata(struct nfp_flower_priv *priv, + struct nfp_fl_payload *nfp_flow); int nfp_modify_flow_metadata(struct nfp_app *app, struct nfp_fl_payload *nfp_flow); @@ -277,6 +383,8 @@ struct nfp_fl_payload * nfp_flower_search_fl_table(struct nfp_app *app, unsigned long tc_flower_cookie, struct net_device *netdev); struct nfp_fl_payload * +nfp_flower_get_fl_payload_from_ctx(struct nfp_app *app, u32 ctx_id); +struct nfp_fl_payload * nfp_flower_remove_fl_table(struct nfp_app *app, unsigned long tc_flower_cookie); void nfp_flower_rx_flow_stats(struct nfp_app *app, struct sk_buff *skb); @@ -299,9 +407,15 @@ int nfp_flower_lag_netdev_event(struct nfp_flower_priv *priv, bool nfp_flower_lag_unprocessed_msg(struct nfp_app *app, struct sk_buff *skb); int nfp_flower_lag_populate_pre_action(struct nfp_app *app, struct net_device *master, - struct nfp_fl_pre_lag *pre_act); + struct nfp_fl_pre_lag *pre_act, + struct netlink_ext_ack *extack); int nfp_flower_lag_get_output_id(struct nfp_app *app, struct net_device *master); +void nfp_flower_qos_init(struct nfp_app *app); +void nfp_flower_qos_cleanup(struct nfp_app *app); +int nfp_flower_setup_qos_offload(struct nfp_app *app, struct net_device *netdev, + struct tc_cls_matchall_offload *flow); +void nfp_flower_stats_rlim_reply(struct nfp_app *app, struct sk_buff *skb); int nfp_flower_reg_indir_block_handler(struct nfp_app *app, struct net_device *netdev, unsigned long event); @@ -314,4 +428,10 @@ void __nfp_flower_non_repr_priv_put(struct nfp_flower_non_repr_priv *non_repr_priv); void nfp_flower_non_repr_priv_put(struct nfp_app *app, struct net_device *netdev); +u32 nfp_flower_get_port_id_from_netdev(struct nfp_app *app, + struct net_device *netdev); +int nfp_flower_xmit_pre_tun_flow(struct nfp_app *app, + struct nfp_fl_payload *flow); +int nfp_flower_xmit_pre_tun_del_flow(struct nfp_app *app, + struct nfp_fl_payload *flow); #endif diff --git a/drivers/net/ethernet/netronome/nfp/flower/match.c b/drivers/net/ethernet/netronome/nfp/flower/match.c index 9b8b843d0340..9cc3ba17ff69 100644 --- a/drivers/net/ethernet/netronome/nfp/flower/match.c +++ b/drivers/net/ethernet/netronome/nfp/flower/match.c @@ -10,9 +10,9 @@ static void nfp_flower_compile_meta_tci(struct nfp_flower_meta_tci *ext, struct nfp_flower_meta_tci *msk, - struct tc_cls_flower_offload *flow, u8 key_type) + struct flow_cls_offload *flow, u8 key_type) { - struct flow_rule *rule = tc_cls_flower_offload_flow_rule(flow); + struct flow_rule *rule = flow_cls_offload_flow_rule(flow); u16 tmp_tci; memset(ext, 0, sizeof(struct nfp_flower_meta_tci)); @@ -54,7 +54,8 @@ nfp_flower_compile_ext_meta(struct nfp_flower_ext_meta *frame, u32 key_ext) static int nfp_flower_compile_port(struct nfp_flower_in_port *frame, u32 cmsg_port, - bool mask_version, enum nfp_flower_tun_type tun_type) + bool mask_version, enum nfp_flower_tun_type tun_type, + struct netlink_ext_ack *extack) { if (mask_version) { frame->in_port = cpu_to_be32(~0); @@ -64,8 +65,10 @@ nfp_flower_compile_port(struct nfp_flower_in_port *frame, u32 cmsg_port, if (tun_type) { frame->in_port = cpu_to_be32(NFP_FL_PORT_TYPE_TUN | tun_type); } else { - if (!cmsg_port) + if (!cmsg_port) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: invalid ingress interface for match offload"); return -EOPNOTSUPP; + } frame->in_port = cpu_to_be32(cmsg_port); } @@ -75,9 +78,9 @@ nfp_flower_compile_port(struct nfp_flower_in_port *frame, u32 cmsg_port, static void nfp_flower_compile_mac(struct nfp_flower_mac_mpls *ext, struct nfp_flower_mac_mpls *msk, - struct tc_cls_flower_offload *flow) + struct flow_cls_offload *flow) { - struct flow_rule *rule = tc_cls_flower_offload_flow_rule(flow); + struct flow_rule *rule = flow_cls_offload_flow_rule(flow); memset(ext, 0, sizeof(struct nfp_flower_mac_mpls)); memset(msk, 0, sizeof(struct nfp_flower_mac_mpls)); @@ -127,9 +130,9 @@ nfp_flower_compile_mac(struct nfp_flower_mac_mpls *ext, static void nfp_flower_compile_tport(struct nfp_flower_tp_ports *ext, struct nfp_flower_tp_ports *msk, - struct tc_cls_flower_offload *flow) + struct flow_cls_offload *flow) { - struct flow_rule *rule = tc_cls_flower_offload_flow_rule(flow); + struct flow_rule *rule = flow_cls_offload_flow_rule(flow); memset(ext, 0, sizeof(struct nfp_flower_tp_ports)); memset(msk, 0, sizeof(struct nfp_flower_tp_ports)); @@ -148,9 +151,9 @@ nfp_flower_compile_tport(struct nfp_flower_tp_ports *ext, static void nfp_flower_compile_ip_ext(struct nfp_flower_ip_ext *ext, struct nfp_flower_ip_ext *msk, - struct tc_cls_flower_offload *flow) + struct flow_cls_offload *flow) { - struct flow_rule *rule = tc_cls_flower_offload_flow_rule(flow); + struct flow_rule *rule = flow_cls_offload_flow_rule(flow); if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_BASIC)) { struct flow_match_basic match; @@ -222,9 +225,9 @@ nfp_flower_compile_ip_ext(struct nfp_flower_ip_ext *ext, static void nfp_flower_compile_ipv4(struct nfp_flower_ipv4 *ext, struct nfp_flower_ipv4 *msk, - struct tc_cls_flower_offload *flow) + struct flow_cls_offload *flow) { - struct flow_rule *rule = tc_cls_flower_offload_flow_rule(flow); + struct flow_rule *rule = flow_cls_offload_flow_rule(flow); struct flow_match_ipv4_addrs match; memset(ext, 0, sizeof(struct nfp_flower_ipv4)); @@ -244,9 +247,9 @@ nfp_flower_compile_ipv4(struct nfp_flower_ipv4 *ext, static void nfp_flower_compile_ipv6(struct nfp_flower_ipv6 *ext, struct nfp_flower_ipv6 *msk, - struct tc_cls_flower_offload *flow) + struct flow_cls_offload *flow) { - struct flow_rule *rule = tc_cls_flower_offload_flow_rule(flow); + struct flow_rule *rule = flow_cls_offload_flow_rule(flow); memset(ext, 0, sizeof(struct nfp_flower_ipv6)); memset(msk, 0, sizeof(struct nfp_flower_ipv6)); @@ -266,7 +269,7 @@ nfp_flower_compile_ipv6(struct nfp_flower_ipv6 *ext, static int nfp_flower_compile_geneve_opt(void *ext, void *msk, - struct tc_cls_flower_offload *flow) + struct flow_cls_offload *flow) { struct flow_match_enc_opts match; @@ -278,11 +281,76 @@ nfp_flower_compile_geneve_opt(void *ext, void *msk, } static void +nfp_flower_compile_tun_ipv4_addrs(struct nfp_flower_tun_ipv4 *ext, + struct nfp_flower_tun_ipv4 *msk, + struct flow_cls_offload *flow) +{ + struct flow_rule *rule = flow_cls_offload_flow_rule(flow); + + if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS)) { + struct flow_match_ipv4_addrs match; + + flow_rule_match_enc_ipv4_addrs(rule, &match); + ext->src = match.key->src; + ext->dst = match.key->dst; + msk->src = match.mask->src; + msk->dst = match.mask->dst; + } +} + +static void +nfp_flower_compile_tun_ip_ext(struct nfp_flower_tun_ip_ext *ext, + struct nfp_flower_tun_ip_ext *msk, + struct flow_cls_offload *flow) +{ + struct flow_rule *rule = flow_cls_offload_flow_rule(flow); + + if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_IP)) { + struct flow_match_ip match; + + flow_rule_match_enc_ip(rule, &match); + ext->tos = match.key->tos; + ext->ttl = match.key->ttl; + msk->tos = match.mask->tos; + msk->ttl = match.mask->ttl; + } +} + +static void +nfp_flower_compile_ipv4_gre_tun(struct nfp_flower_ipv4_gre_tun *ext, + struct nfp_flower_ipv4_gre_tun *msk, + struct flow_cls_offload *flow) +{ + struct flow_rule *rule = flow_cls_offload_flow_rule(flow); + + memset(ext, 0, sizeof(struct nfp_flower_ipv4_gre_tun)); + memset(msk, 0, sizeof(struct nfp_flower_ipv4_gre_tun)); + + /* NVGRE is the only supported GRE tunnel type */ + ext->ethertype = cpu_to_be16(ETH_P_TEB); + msk->ethertype = cpu_to_be16(~0); + + if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_KEYID)) { + struct flow_match_enc_keyid match; + + flow_rule_match_enc_keyid(rule, &match); + ext->tun_key = match.key->keyid; + msk->tun_key = match.mask->keyid; + + ext->tun_flags = cpu_to_be16(NFP_FL_GRE_FLAG_KEY); + msk->tun_flags = cpu_to_be16(NFP_FL_GRE_FLAG_KEY); + } + + nfp_flower_compile_tun_ipv4_addrs(&ext->ipv4, &msk->ipv4, flow); + nfp_flower_compile_tun_ip_ext(&ext->ip_ext, &msk->ip_ext, flow); +} + +static void nfp_flower_compile_ipv4_udp_tun(struct nfp_flower_ipv4_udp_tun *ext, struct nfp_flower_ipv4_udp_tun *msk, - struct tc_cls_flower_offload *flow) + struct flow_cls_offload *flow) { - struct flow_rule *rule = tc_cls_flower_offload_flow_rule(flow); + struct flow_rule *rule = flow_cls_offload_flow_rule(flow); memset(ext, 0, sizeof(struct nfp_flower_ipv4_udp_tun)); memset(msk, 0, sizeof(struct nfp_flower_ipv4_udp_tun)); @@ -298,41 +366,24 @@ nfp_flower_compile_ipv4_udp_tun(struct nfp_flower_ipv4_udp_tun *ext, msk->tun_id = cpu_to_be32(temp_vni); } - if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS)) { - struct flow_match_ipv4_addrs match; - - flow_rule_match_enc_ipv4_addrs(rule, &match); - ext->ip_src = match.key->src; - ext->ip_dst = match.key->dst; - msk->ip_src = match.mask->src; - msk->ip_dst = match.mask->dst; - } - - if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_IP)) { - struct flow_match_ip match; - - flow_rule_match_enc_ip(rule, &match); - ext->tos = match.key->tos; - ext->ttl = match.key->ttl; - msk->tos = match.mask->tos; - msk->ttl = match.mask->ttl; - } + nfp_flower_compile_tun_ipv4_addrs(&ext->ipv4, &msk->ipv4, flow); + nfp_flower_compile_tun_ip_ext(&ext->ip_ext, &msk->ip_ext, flow); } int nfp_flower_compile_flow_match(struct nfp_app *app, - struct tc_cls_flower_offload *flow, + struct flow_cls_offload *flow, struct nfp_fl_key_ls *key_ls, struct net_device *netdev, struct nfp_fl_payload *nfp_flow, - enum nfp_flower_tun_type tun_type) + enum nfp_flower_tun_type tun_type, + struct netlink_ext_ack *extack) { - u32 cmsg_port = 0; + u32 port_id; int err; u8 *ext; u8 *msk; - if (nfp_netdev_is_nfp_repr(netdev)) - cmsg_port = nfp_repr_get_port_id(netdev); + port_id = nfp_flower_get_port_id_from_netdev(app, netdev); memset(nfp_flow->unmasked_data, 0, key_ls->key_size); memset(nfp_flow->mask_data, 0, key_ls->key_size); @@ -358,13 +409,13 @@ int nfp_flower_compile_flow_match(struct nfp_app *app, /* Populate Exact Port data. */ err = nfp_flower_compile_port((struct nfp_flower_in_port *)ext, - cmsg_port, false, tun_type); + port_id, false, tun_type, extack); if (err) return err; /* Populate Mask Port Data. */ err = nfp_flower_compile_port((struct nfp_flower_in_port *)msk, - cmsg_port, true, tun_type); + port_id, true, tun_type, extack); if (err) return err; @@ -403,12 +454,27 @@ int nfp_flower_compile_flow_match(struct nfp_app *app, msk += sizeof(struct nfp_flower_ipv6); } + if (key_ls->key_layer_two & NFP_FLOWER_LAYER2_GRE) { + __be32 tun_dst; + + nfp_flower_compile_ipv4_gre_tun((void *)ext, (void *)msk, flow); + tun_dst = ((struct nfp_flower_ipv4_gre_tun *)ext)->ipv4.dst; + ext += sizeof(struct nfp_flower_ipv4_gre_tun); + msk += sizeof(struct nfp_flower_ipv4_gre_tun); + + /* Store the tunnel destination in the rule data. + * This must be present and be an exact match. + */ + nfp_flow->nfp_tun_ipv4_addr = tun_dst; + nfp_tunnel_add_ipv4_off(app, tun_dst); + } + if (key_ls->key_layer & NFP_FLOWER_LAYER_VXLAN || key_ls->key_layer_two & NFP_FLOWER_LAYER2_GENEVE) { __be32 tun_dst; nfp_flower_compile_ipv4_udp_tun((void *)ext, (void *)msk, flow); - tun_dst = ((struct nfp_flower_ipv4_udp_tun *)ext)->ip_dst; + tun_dst = ((struct nfp_flower_ipv4_udp_tun *)ext)->ipv4.dst; ext += sizeof(struct nfp_flower_ipv4_udp_tun); msk += sizeof(struct nfp_flower_ipv4_udp_tun); diff --git a/drivers/net/ethernet/netronome/nfp/flower/metadata.c b/drivers/net/ethernet/netronome/nfp/flower/metadata.c index 492837b852b6..7c4a15e967df 100644 --- a/drivers/net/ethernet/netronome/nfp/flower/metadata.c +++ b/drivers/net/ethernet/netronome/nfp/flower/metadata.c @@ -24,6 +24,18 @@ struct nfp_fl_flow_table_cmp_arg { unsigned long cookie; }; +struct nfp_fl_stats_ctx_to_flow { + struct rhash_head ht_node; + u32 stats_cxt; + struct nfp_fl_payload *flow; +}; + +static const struct rhashtable_params stats_ctx_table_params = { + .key_offset = offsetof(struct nfp_fl_stats_ctx_to_flow, stats_cxt), + .head_offset = offsetof(struct nfp_fl_stats_ctx_to_flow, ht_node), + .key_len = sizeof(u32), +}; + static int nfp_release_stats_entry(struct nfp_app *app, u32 stats_context_id) { struct nfp_flower_priv *priv = app->priv; @@ -264,9 +276,6 @@ nfp_check_mask_remove(struct nfp_app *app, char *mask_data, u32 mask_len, if (!mask_entry) return false; - if (meta_flags) - *meta_flags &= ~NFP_FL_META_FLAG_MANAGE_MASK; - *mask_id = mask_entry->mask_id; mask_entry->ref_cnt--; if (!mask_entry->ref_cnt) { @@ -281,29 +290,55 @@ nfp_check_mask_remove(struct nfp_app *app, char *mask_data, u32 mask_len, } int nfp_compile_flow_metadata(struct nfp_app *app, - struct tc_cls_flower_offload *flow, + struct flow_cls_offload *flow, struct nfp_fl_payload *nfp_flow, - struct net_device *netdev) + struct net_device *netdev, + struct netlink_ext_ack *extack) { + struct nfp_fl_stats_ctx_to_flow *ctx_entry; struct nfp_flower_priv *priv = app->priv; struct nfp_fl_payload *check_entry; u8 new_mask_id; u32 stats_cxt; + int err; - if (nfp_get_stats_entry(app, &stats_cxt)) - return -ENOENT; + err = nfp_get_stats_entry(app, &stats_cxt); + if (err) { + NL_SET_ERR_MSG_MOD(extack, "invalid entry: cannot allocate new stats context"); + return err; + } nfp_flow->meta.host_ctx_id = cpu_to_be32(stats_cxt); nfp_flow->meta.host_cookie = cpu_to_be64(flow->cookie); nfp_flow->ingress_dev = netdev; + ctx_entry = kzalloc(sizeof(*ctx_entry), GFP_KERNEL); + if (!ctx_entry) { + err = -ENOMEM; + goto err_release_stats; + } + + ctx_entry->stats_cxt = stats_cxt; + ctx_entry->flow = nfp_flow; + + if (rhashtable_insert_fast(&priv->stats_ctx_table, &ctx_entry->ht_node, + stats_ctx_table_params)) { + err = -ENOMEM; + goto err_free_ctx_entry; + } + new_mask_id = 0; if (!nfp_check_mask_add(app, nfp_flow->mask_data, nfp_flow->meta.mask_len, &nfp_flow->meta.flags, &new_mask_id)) { - if (nfp_release_stats_entry(app, stats_cxt)) - return -EINVAL; - return -ENOENT; + NL_SET_ERR_MSG_MOD(extack, "invalid entry: cannot allocate a new mask id"); + if (nfp_release_stats_entry(app, stats_cxt)) { + NL_SET_ERR_MSG_MOD(extack, "invalid entry: cannot release stats context"); + err = -EINVAL; + goto err_remove_rhash; + } + err = -ENOENT; + goto err_remove_rhash; } nfp_flow->meta.flow_version = cpu_to_be64(priv->flower_version); @@ -317,43 +352,97 @@ int nfp_compile_flow_metadata(struct nfp_app *app, check_entry = nfp_flower_search_fl_table(app, flow->cookie, netdev); if (check_entry) { - if (nfp_release_stats_entry(app, stats_cxt)) - return -EINVAL; + NL_SET_ERR_MSG_MOD(extack, "invalid entry: cannot offload duplicate flow entry"); + if (nfp_release_stats_entry(app, stats_cxt)) { + NL_SET_ERR_MSG_MOD(extack, "invalid entry: cannot release stats context"); + err = -EINVAL; + goto err_remove_mask; + } if (!nfp_check_mask_remove(app, nfp_flow->mask_data, nfp_flow->meta.mask_len, - NULL, &new_mask_id)) - return -EINVAL; + NULL, &new_mask_id)) { + NL_SET_ERR_MSG_MOD(extack, "invalid entry: cannot release mask id"); + err = -EINVAL; + goto err_remove_mask; + } - return -EEXIST; + err = -EEXIST; + goto err_remove_mask; } return 0; + +err_remove_mask: + nfp_check_mask_remove(app, nfp_flow->mask_data, nfp_flow->meta.mask_len, + NULL, &new_mask_id); +err_remove_rhash: + WARN_ON_ONCE(rhashtable_remove_fast(&priv->stats_ctx_table, + &ctx_entry->ht_node, + stats_ctx_table_params)); +err_free_ctx_entry: + kfree(ctx_entry); +err_release_stats: + nfp_release_stats_entry(app, stats_cxt); + + return err; +} + +void __nfp_modify_flow_metadata(struct nfp_flower_priv *priv, + struct nfp_fl_payload *nfp_flow) +{ + nfp_flow->meta.flags &= ~NFP_FL_META_FLAG_MANAGE_MASK; + nfp_flow->meta.flow_version = cpu_to_be64(priv->flower_version); + priv->flower_version++; } int nfp_modify_flow_metadata(struct nfp_app *app, struct nfp_fl_payload *nfp_flow) { + struct nfp_fl_stats_ctx_to_flow *ctx_entry; struct nfp_flower_priv *priv = app->priv; u8 new_mask_id = 0; u32 temp_ctx_id; + __nfp_modify_flow_metadata(priv, nfp_flow); + nfp_check_mask_remove(app, nfp_flow->mask_data, nfp_flow->meta.mask_len, &nfp_flow->meta.flags, &new_mask_id); - nfp_flow->meta.flow_version = cpu_to_be64(priv->flower_version); - priv->flower_version++; - /* Update flow payload with mask ids. */ nfp_flow->unmasked_data[NFP_FL_MASK_ID_LOCATION] = new_mask_id; - /* Release the stats ctx id. */ + /* Release the stats ctx id and ctx to flow table entry. */ temp_ctx_id = be32_to_cpu(nfp_flow->meta.host_ctx_id); + ctx_entry = rhashtable_lookup_fast(&priv->stats_ctx_table, &temp_ctx_id, + stats_ctx_table_params); + if (!ctx_entry) + return -ENOENT; + + WARN_ON_ONCE(rhashtable_remove_fast(&priv->stats_ctx_table, + &ctx_entry->ht_node, + stats_ctx_table_params)); + kfree(ctx_entry); + return nfp_release_stats_entry(app, temp_ctx_id); } +struct nfp_fl_payload * +nfp_flower_get_fl_payload_from_ctx(struct nfp_app *app, u32 ctx_id) +{ + struct nfp_fl_stats_ctx_to_flow *ctx_entry; + struct nfp_flower_priv *priv = app->priv; + + ctx_entry = rhashtable_lookup_fast(&priv->stats_ctx_table, &ctx_id, + stats_ctx_table_params); + if (!ctx_entry) + return NULL; + + return ctx_entry->flow; +} + static int nfp_fl_obj_cmpfn(struct rhashtable_compare_arg *arg, const void *obj) { @@ -403,6 +492,10 @@ int nfp_flower_metadata_init(struct nfp_app *app, u64 host_ctx_count, if (err) return err; + err = rhashtable_init(&priv->stats_ctx_table, &stats_ctx_table_params); + if (err) + goto err_free_flow_table; + get_random_bytes(&priv->mask_id_seed, sizeof(priv->mask_id_seed)); /* Init ring buffer and unallocated mask_ids. */ @@ -410,7 +503,7 @@ int nfp_flower_metadata_init(struct nfp_app *app, u64 host_ctx_count, kmalloc_array(NFP_FLOWER_MASK_ENTRY_RS, NFP_FLOWER_MASK_ELEMENT_RS, GFP_KERNEL); if (!priv->mask_ids.mask_id_free_list.buf) - goto err_free_flow_table; + goto err_free_stats_ctx_table; priv->mask_ids.init_unallocated = NFP_FLOWER_MASK_ENTRY_RS - 1; @@ -447,6 +540,8 @@ err_free_last_used: kfree(priv->mask_ids.last_used); err_free_mask_id: kfree(priv->mask_ids.mask_id_free_list.buf); +err_free_stats_ctx_table: + rhashtable_destroy(&priv->stats_ctx_table); err_free_flow_table: rhashtable_destroy(&priv->flow_table); return -ENOMEM; @@ -461,6 +556,8 @@ void nfp_flower_metadata_cleanup(struct nfp_app *app) rhashtable_free_and_destroy(&priv->flow_table, nfp_check_rhashtable_empty, NULL); + rhashtable_free_and_destroy(&priv->stats_ctx_table, + nfp_check_rhashtable_empty, NULL); kvfree(priv->stats); kfree(priv->mask_ids.mask_id_free_list.buf); kfree(priv->mask_ids.last_used); diff --git a/drivers/net/ethernet/netronome/nfp/flower/offload.c b/drivers/net/ethernet/netronome/nfp/flower/offload.c index 450d7296fd57..987ae221f6be 100644 --- a/drivers/net/ethernet/netronome/nfp/flower/offload.c +++ b/drivers/net/ethernet/netronome/nfp/flower/offload.c @@ -52,8 +52,34 @@ #define NFP_FLOWER_WHITELIST_TUN_DISSECTOR_R \ (BIT(FLOW_DISSECTOR_KEY_ENC_CONTROL) | \ - BIT(FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS) | \ - BIT(FLOW_DISSECTOR_KEY_ENC_PORTS)) + BIT(FLOW_DISSECTOR_KEY_ENC_IPV4_ADDRS)) + +#define NFP_FLOWER_MERGE_FIELDS \ + (NFP_FLOWER_LAYER_PORT | \ + NFP_FLOWER_LAYER_MAC | \ + NFP_FLOWER_LAYER_TP | \ + NFP_FLOWER_LAYER_IPV4 | \ + NFP_FLOWER_LAYER_IPV6) + +#define NFP_FLOWER_PRE_TUN_RULE_FIELDS \ + (NFP_FLOWER_LAYER_PORT | \ + NFP_FLOWER_LAYER_MAC | \ + NFP_FLOWER_LAYER_IPV4) + +struct nfp_flower_merge_check { + union { + struct { + __be16 tci; + struct nfp_flower_mac_mpls l2; + struct nfp_flower_tp_ports l4; + union { + struct nfp_flower_ipv4 ipv4; + struct nfp_flower_ipv6 ipv6; + }; + }; + unsigned long vals[8]; + }; +}; static int nfp_flower_xmit_flow(struct nfp_app *app, struct nfp_fl_payload *nfp_flow, @@ -100,9 +126,9 @@ nfp_flower_xmit_flow(struct nfp_app *app, struct nfp_fl_payload *nfp_flow, return 0; } -static bool nfp_flower_check_higher_than_mac(struct tc_cls_flower_offload *f) +static bool nfp_flower_check_higher_than_mac(struct flow_cls_offload *f) { - struct flow_rule *rule = tc_cls_flower_offload_flow_rule(f); + struct flow_rule *rule = flow_cls_offload_flow_rule(f); return flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_IPV4_ADDRS) || flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_IPV6_ADDRS) || @@ -110,14 +136,25 @@ static bool nfp_flower_check_higher_than_mac(struct tc_cls_flower_offload *f) flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ICMP); } +static bool nfp_flower_check_higher_than_l3(struct flow_cls_offload *f) +{ + struct flow_rule *rule = flow_cls_offload_flow_rule(f); + + return flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_PORTS) || + flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ICMP); +} + static int -nfp_flower_calc_opt_layer(struct flow_match_enc_opts *enc_opts, - u32 *key_layer_two, int *key_size) +nfp_flower_calc_opt_layer(struct flow_dissector_key_enc_opts *enc_opts, + u32 *key_layer_two, int *key_size, + struct netlink_ext_ack *extack) { - if (enc_opts->key->len > NFP_FL_MAX_GENEVE_OPT_KEY) + if (enc_opts->len > NFP_FL_MAX_GENEVE_OPT_KEY) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: geneve options exceed maximum length"); return -EOPNOTSUPP; + } - if (enc_opts->key->len > 0) { + if (enc_opts->len > 0) { *key_layer_two |= NFP_FLOWER_LAYER2_GENEVE_OP; *key_size += sizeof(struct nfp_flower_geneve_options); } @@ -126,13 +163,65 @@ nfp_flower_calc_opt_layer(struct flow_match_enc_opts *enc_opts, } static int +nfp_flower_calc_udp_tun_layer(struct flow_dissector_key_ports *enc_ports, + struct flow_dissector_key_enc_opts *enc_op, + u32 *key_layer_two, u8 *key_layer, int *key_size, + struct nfp_flower_priv *priv, + enum nfp_flower_tun_type *tun_type, + struct netlink_ext_ack *extack) +{ + int err; + + switch (enc_ports->dst) { + case htons(IANA_VXLAN_UDP_PORT): + *tun_type = NFP_FL_TUNNEL_VXLAN; + *key_layer |= NFP_FLOWER_LAYER_VXLAN; + *key_size += sizeof(struct nfp_flower_ipv4_udp_tun); + + if (enc_op) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: encap options not supported on vxlan tunnels"); + return -EOPNOTSUPP; + } + break; + case htons(GENEVE_UDP_PORT): + if (!(priv->flower_ext_feats & NFP_FL_FEATS_GENEVE)) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: loaded firmware does not support geneve offload"); + return -EOPNOTSUPP; + } + *tun_type = NFP_FL_TUNNEL_GENEVE; + *key_layer |= NFP_FLOWER_LAYER_EXT_META; + *key_size += sizeof(struct nfp_flower_ext_meta); + *key_layer_two |= NFP_FLOWER_LAYER2_GENEVE; + *key_size += sizeof(struct nfp_flower_ipv4_udp_tun); + + if (!enc_op) + break; + if (!(priv->flower_ext_feats & NFP_FL_FEATS_GENEVE_OPT)) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: loaded firmware does not support geneve option offload"); + return -EOPNOTSUPP; + } + err = nfp_flower_calc_opt_layer(enc_op, key_layer_two, + key_size, extack); + if (err) + return err; + break; + default: + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: tunnel type unknown"); + return -EOPNOTSUPP; + } + + return 0; +} + +static int nfp_flower_calculate_key_layers(struct nfp_app *app, struct net_device *netdev, struct nfp_fl_key_ls *ret_key_ls, - struct tc_cls_flower_offload *flow, - enum nfp_flower_tun_type *tun_type) + struct flow_cls_offload *flow, + enum nfp_flower_tun_type *tun_type, + struct netlink_ext_ack *extack) { - struct flow_rule *rule = tc_cls_flower_offload_flow_rule(flow); + struct flow_rule *rule = flow_cls_offload_flow_rule(flow); struct flow_dissector *dissector = rule->match.dissector; struct flow_match_basic basic = { NULL, NULL}; struct nfp_flower_priv *priv = app->priv; @@ -141,14 +230,18 @@ nfp_flower_calculate_key_layers(struct nfp_app *app, int key_size; int err; - if (dissector->used_keys & ~NFP_FLOWER_WHITELIST_DISSECTOR) + if (dissector->used_keys & ~NFP_FLOWER_WHITELIST_DISSECTOR) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: match not supported"); return -EOPNOTSUPP; + } /* If any tun dissector is used then the required set must be used. */ if (dissector->used_keys & NFP_FLOWER_WHITELIST_TUN_DISSECTOR && (dissector->used_keys & NFP_FLOWER_WHITELIST_TUN_DISSECTOR_R) - != NFP_FLOWER_WHITELIST_TUN_DISSECTOR_R) + != NFP_FLOWER_WHITELIST_TUN_DISSECTOR_R) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: tunnel match not supported"); return -EOPNOTSUPP; + } key_layer_two = 0; key_layer = NFP_FLOWER_LAYER_PORT; @@ -166,8 +259,10 @@ nfp_flower_calculate_key_layers(struct nfp_app *app, flow_rule_match_vlan(rule, &vlan); if (!(priv->flower_ext_feats & NFP_FL_FEATS_VLAN_PCP) && - vlan.key->vlan_priority) + vlan.key->vlan_priority) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: loaded firmware does not support VLAN PCP offload"); return -EOPNOTSUPP; + } } if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_CONTROL)) { @@ -178,56 +273,68 @@ nfp_flower_calculate_key_layers(struct nfp_app *app, flow_rule_match_enc_control(rule, &enc_ctl); - if (enc_ctl.mask->addr_type != 0xffff || - enc_ctl.key->addr_type != FLOW_DISSECTOR_KEY_IPV4_ADDRS) + if (enc_ctl.mask->addr_type != 0xffff) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: wildcarded protocols on tunnels are not supported"); return -EOPNOTSUPP; + } + if (enc_ctl.key->addr_type != FLOW_DISSECTOR_KEY_IPV4_ADDRS) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: only IPv4 tunnels are supported"); + return -EOPNOTSUPP; + } /* These fields are already verified as used. */ flow_rule_match_enc_ipv4_addrs(rule, &ipv4_addrs); - if (ipv4_addrs.mask->dst != cpu_to_be32(~0)) - return -EOPNOTSUPP; - - flow_rule_match_enc_ports(rule, &enc_ports); - if (enc_ports.mask->dst != cpu_to_be16(~0)) + if (ipv4_addrs.mask->dst != cpu_to_be32(~0)) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: only an exact match IPv4 destination address is supported"); return -EOPNOTSUPP; + } if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_OPTS)) flow_rule_match_enc_opts(rule, &enc_op); - switch (enc_ports.key->dst) { - case htons(NFP_FL_VXLAN_PORT): - *tun_type = NFP_FL_TUNNEL_VXLAN; - key_layer |= NFP_FLOWER_LAYER_VXLAN; - key_size += sizeof(struct nfp_flower_ipv4_udp_tun); - if (enc_op.key) + if (!flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_ENC_PORTS)) { + /* check if GRE, which has no enc_ports */ + if (netif_is_gretap(netdev)) { + *tun_type = NFP_FL_TUNNEL_GRE; + key_layer |= NFP_FLOWER_LAYER_EXT_META; + key_size += sizeof(struct nfp_flower_ext_meta); + key_layer_two |= NFP_FLOWER_LAYER2_GRE; + key_size += + sizeof(struct nfp_flower_ipv4_gre_tun); + + if (enc_op.key) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: encap options not supported on GRE tunnels"); + return -EOPNOTSUPP; + } + } else { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: an exact match on L4 destination port is required for non-GRE tunnels"); return -EOPNOTSUPP; - break; - case htons(NFP_FL_GENEVE_PORT): - if (!(priv->flower_ext_feats & NFP_FL_FEATS_GENEVE)) + } + } else { + flow_rule_match_enc_ports(rule, &enc_ports); + if (enc_ports.mask->dst != cpu_to_be16(~0)) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: only an exact match L4 destination port is supported"); return -EOPNOTSUPP; - *tun_type = NFP_FL_TUNNEL_GENEVE; - key_layer |= NFP_FLOWER_LAYER_EXT_META; - key_size += sizeof(struct nfp_flower_ext_meta); - key_layer_two |= NFP_FLOWER_LAYER2_GENEVE; - key_size += sizeof(struct nfp_flower_ipv4_udp_tun); + } - if (!enc_op.key) - break; - if (!(priv->flower_ext_feats & NFP_FL_FEATS_GENEVE_OPT)) - return -EOPNOTSUPP; - err = nfp_flower_calc_opt_layer(&enc_op, &key_layer_two, - &key_size); + err = nfp_flower_calc_udp_tun_layer(enc_ports.key, + enc_op.key, + &key_layer_two, + &key_layer, + &key_size, priv, + tun_type, extack); if (err) return err; - break; - default: - return -EOPNOTSUPP; - } - /* Ensure the ingress netdev matches the expected tun type. */ - if (!nfp_fl_netdev_is_tunnel_type(netdev, *tun_type)) - return -EOPNOTSUPP; + /* Ensure the ingress netdev matches the expected + * tun type. + */ + if (!nfp_fl_netdev_is_tunnel_type(netdev, *tun_type)) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: ingress netdev does not match the expected tunnel type"); + return -EOPNOTSUPP; + } + } } if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_BASIC)) @@ -250,6 +357,7 @@ nfp_flower_calculate_key_layers(struct nfp_app *app, * because we rely on it to get to the host. */ case cpu_to_be16(ETH_P_ARP): + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: ARP not supported"); return -EOPNOTSUPP; case cpu_to_be16(ETH_P_MPLS_UC): @@ -265,17 +373,15 @@ nfp_flower_calculate_key_layers(struct nfp_app *app, break; default: - /* Other ethtype - we need check the masks for the - * remainder of the key to ensure we can offload. - */ - if (nfp_flower_check_higher_than_mac(flow)) - return -EOPNOTSUPP; - break; + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: match on given EtherType is not supported"); + return -EOPNOTSUPP; } + } else if (nfp_flower_check_higher_than_mac(flow)) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: cannot match above L2 without specified EtherType"); + return -EOPNOTSUPP; } if (basic.mask && basic.mask->ip_proto) { - /* Ethernet type is present in the key. */ switch (basic.key->ip_proto) { case IPPROTO_TCP: case IPPROTO_UDP: @@ -285,14 +391,15 @@ nfp_flower_calculate_key_layers(struct nfp_app *app, key_layer |= NFP_FLOWER_LAYER_TP; key_size += sizeof(struct nfp_flower_tp_ports); break; - default: - /* Other ip proto - we need check the masks for the - * remainder of the key to ensure we can offload. - */ - return -EOPNOTSUPP; } } + if (!(key_layer & NFP_FLOWER_LAYER_TP) && + nfp_flower_check_higher_than_l3(flow)) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: cannot match on L4 information without specified IP protocol type"); + return -EOPNOTSUPP; + } + if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_TCP)) { struct flow_match_tcp tcp; u32 tcp_flags; @@ -300,22 +407,28 @@ nfp_flower_calculate_key_layers(struct nfp_app *app, flow_rule_match_tcp(rule, &tcp); tcp_flags = be16_to_cpu(tcp.key->flags); - if (tcp_flags & ~NFP_FLOWER_SUPPORTED_TCPFLAGS) + if (tcp_flags & ~NFP_FLOWER_SUPPORTED_TCPFLAGS) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: no match support for selected TCP flags"); return -EOPNOTSUPP; + } /* We only support PSH and URG flags when either * FIN, SYN or RST is present as well. */ if ((tcp_flags & (TCPHDR_PSH | TCPHDR_URG)) && - !(tcp_flags & (TCPHDR_FIN | TCPHDR_SYN | TCPHDR_RST))) + !(tcp_flags & (TCPHDR_FIN | TCPHDR_SYN | TCPHDR_RST))) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: PSH and URG is only supported when used with FIN, SYN or RST"); return -EOPNOTSUPP; + } /* We need to store TCP flags in the either the IPv4 or IPv6 key * space, thus we need to ensure we include a IPv4/IPv6 key * layer if we have not done so already. */ - if (!basic.key) + if (!basic.key) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: match on TCP flags requires a match on L3 protocol"); return -EOPNOTSUPP; + } if (!(key_layer & NFP_FLOWER_LAYER_IPV4) && !(key_layer & NFP_FLOWER_LAYER_IPV6)) { @@ -326,11 +439,12 @@ nfp_flower_calculate_key_layers(struct nfp_app *app, break; case cpu_to_be16(ETH_P_IPV6): - key_layer |= NFP_FLOWER_LAYER_IPV6; + key_layer |= NFP_FLOWER_LAYER_IPV6; key_size += sizeof(struct nfp_flower_ipv6); break; default: + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: match on TCP flags requires a match on IPv4/IPv6"); return -EOPNOTSUPP; } } @@ -340,8 +454,10 @@ nfp_flower_calculate_key_layers(struct nfp_app *app, struct flow_match_control ctl; flow_rule_match_control(rule, &ctl); - if (ctl.key->flags & ~NFP_FLOWER_SUPPORTED_CTLFLAGS) + if (ctl.key->flags & ~NFP_FLOWER_SUPPORTED_CTLFLAGS) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: match on unknown control flag"); return -EOPNOTSUPP; + } } ret_key_ls->key_layer = key_layer; @@ -376,6 +492,9 @@ nfp_flower_allocate_new(struct nfp_fl_key_ls *key_layer) flow_pay->nfp_tun_ipv4_addr = 0; flow_pay->meta.flags = 0; + INIT_LIST_HEAD(&flow_pay->linked_flows); + flow_pay->in_hw = false; + flow_pay->pre_tun_rule.dev = NULL; return flow_pay; @@ -388,6 +507,601 @@ err_free_flow: return NULL; } +static int +nfp_flower_update_merge_with_actions(struct nfp_fl_payload *flow, + struct nfp_flower_merge_check *merge, + u8 *last_act_id, int *act_out) +{ + struct nfp_fl_set_ipv6_tc_hl_fl *ipv6_tc_hl_fl; + struct nfp_fl_set_ip4_ttl_tos *ipv4_ttl_tos; + struct nfp_fl_set_ip4_addrs *ipv4_add; + struct nfp_fl_set_ipv6_addr *ipv6_add; + struct nfp_fl_push_vlan *push_vlan; + struct nfp_fl_set_tport *tport; + struct nfp_fl_set_eth *eth; + struct nfp_fl_act_head *a; + unsigned int act_off = 0; + u8 act_id = 0; + u8 *ports; + int i; + + while (act_off < flow->meta.act_len) { + a = (struct nfp_fl_act_head *)&flow->action_data[act_off]; + act_id = a->jump_id; + + switch (act_id) { + case NFP_FL_ACTION_OPCODE_OUTPUT: + if (act_out) + (*act_out)++; + break; + case NFP_FL_ACTION_OPCODE_PUSH_VLAN: + push_vlan = (struct nfp_fl_push_vlan *)a; + if (push_vlan->vlan_tci) + merge->tci = cpu_to_be16(0xffff); + break; + case NFP_FL_ACTION_OPCODE_POP_VLAN: + merge->tci = cpu_to_be16(0); + break; + case NFP_FL_ACTION_OPCODE_SET_IPV4_TUNNEL: + /* New tunnel header means l2 to l4 can be matched. */ + eth_broadcast_addr(&merge->l2.mac_dst[0]); + eth_broadcast_addr(&merge->l2.mac_src[0]); + memset(&merge->l4, 0xff, + sizeof(struct nfp_flower_tp_ports)); + memset(&merge->ipv4, 0xff, + sizeof(struct nfp_flower_ipv4)); + break; + case NFP_FL_ACTION_OPCODE_SET_ETHERNET: + eth = (struct nfp_fl_set_eth *)a; + for (i = 0; i < ETH_ALEN; i++) + merge->l2.mac_dst[i] |= eth->eth_addr_mask[i]; + for (i = 0; i < ETH_ALEN; i++) + merge->l2.mac_src[i] |= + eth->eth_addr_mask[ETH_ALEN + i]; + break; + case NFP_FL_ACTION_OPCODE_SET_IPV4_ADDRS: + ipv4_add = (struct nfp_fl_set_ip4_addrs *)a; + merge->ipv4.ipv4_src |= ipv4_add->ipv4_src_mask; + merge->ipv4.ipv4_dst |= ipv4_add->ipv4_dst_mask; + break; + case NFP_FL_ACTION_OPCODE_SET_IPV4_TTL_TOS: + ipv4_ttl_tos = (struct nfp_fl_set_ip4_ttl_tos *)a; + merge->ipv4.ip_ext.ttl |= ipv4_ttl_tos->ipv4_ttl_mask; + merge->ipv4.ip_ext.tos |= ipv4_ttl_tos->ipv4_tos_mask; + break; + case NFP_FL_ACTION_OPCODE_SET_IPV6_SRC: + ipv6_add = (struct nfp_fl_set_ipv6_addr *)a; + for (i = 0; i < 4; i++) + merge->ipv6.ipv6_src.in6_u.u6_addr32[i] |= + ipv6_add->ipv6[i].mask; + break; + case NFP_FL_ACTION_OPCODE_SET_IPV6_DST: + ipv6_add = (struct nfp_fl_set_ipv6_addr *)a; + for (i = 0; i < 4; i++) + merge->ipv6.ipv6_dst.in6_u.u6_addr32[i] |= + ipv6_add->ipv6[i].mask; + break; + case NFP_FL_ACTION_OPCODE_SET_IPV6_TC_HL_FL: + ipv6_tc_hl_fl = (struct nfp_fl_set_ipv6_tc_hl_fl *)a; + merge->ipv6.ip_ext.ttl |= + ipv6_tc_hl_fl->ipv6_hop_limit_mask; + merge->ipv6.ip_ext.tos |= ipv6_tc_hl_fl->ipv6_tc_mask; + merge->ipv6.ipv6_flow_label_exthdr |= + ipv6_tc_hl_fl->ipv6_label_mask; + break; + case NFP_FL_ACTION_OPCODE_SET_UDP: + case NFP_FL_ACTION_OPCODE_SET_TCP: + tport = (struct nfp_fl_set_tport *)a; + ports = (u8 *)&merge->l4.port_src; + for (i = 0; i < 4; i++) + ports[i] |= tport->tp_port_mask[i]; + break; + case NFP_FL_ACTION_OPCODE_PRE_TUNNEL: + case NFP_FL_ACTION_OPCODE_PRE_LAG: + case NFP_FL_ACTION_OPCODE_PUSH_GENEVE: + break; + default: + return -EOPNOTSUPP; + } + + act_off += a->len_lw << NFP_FL_LW_SIZ; + } + + if (last_act_id) + *last_act_id = act_id; + + return 0; +} + +static int +nfp_flower_populate_merge_match(struct nfp_fl_payload *flow, + struct nfp_flower_merge_check *merge, + bool extra_fields) +{ + struct nfp_flower_meta_tci *meta_tci; + u8 *mask = flow->mask_data; + u8 key_layer, match_size; + + memset(merge, 0, sizeof(struct nfp_flower_merge_check)); + + meta_tci = (struct nfp_flower_meta_tci *)mask; + key_layer = meta_tci->nfp_flow_key_layer; + + if (key_layer & ~NFP_FLOWER_MERGE_FIELDS && !extra_fields) + return -EOPNOTSUPP; + + merge->tci = meta_tci->tci; + mask += sizeof(struct nfp_flower_meta_tci); + + if (key_layer & NFP_FLOWER_LAYER_EXT_META) + mask += sizeof(struct nfp_flower_ext_meta); + + mask += sizeof(struct nfp_flower_in_port); + + if (key_layer & NFP_FLOWER_LAYER_MAC) { + match_size = sizeof(struct nfp_flower_mac_mpls); + memcpy(&merge->l2, mask, match_size); + mask += match_size; + } + + if (key_layer & NFP_FLOWER_LAYER_TP) { + match_size = sizeof(struct nfp_flower_tp_ports); + memcpy(&merge->l4, mask, match_size); + mask += match_size; + } + + if (key_layer & NFP_FLOWER_LAYER_IPV4) { + match_size = sizeof(struct nfp_flower_ipv4); + memcpy(&merge->ipv4, mask, match_size); + } + + if (key_layer & NFP_FLOWER_LAYER_IPV6) { + match_size = sizeof(struct nfp_flower_ipv6); + memcpy(&merge->ipv6, mask, match_size); + } + + return 0; +} + +static int +nfp_flower_can_merge(struct nfp_fl_payload *sub_flow1, + struct nfp_fl_payload *sub_flow2) +{ + /* Two flows can be merged if sub_flow2 only matches on bits that are + * either matched by sub_flow1 or set by a sub_flow1 action. This + * ensures that every packet that hits sub_flow1 and recirculates is + * guaranteed to hit sub_flow2. + */ + struct nfp_flower_merge_check sub_flow1_merge, sub_flow2_merge; + int err, act_out = 0; + u8 last_act_id = 0; + + err = nfp_flower_populate_merge_match(sub_flow1, &sub_flow1_merge, + true); + if (err) + return err; + + err = nfp_flower_populate_merge_match(sub_flow2, &sub_flow2_merge, + false); + if (err) + return err; + + err = nfp_flower_update_merge_with_actions(sub_flow1, &sub_flow1_merge, + &last_act_id, &act_out); + if (err) + return err; + + /* Must only be 1 output action and it must be the last in sequence. */ + if (act_out != 1 || last_act_id != NFP_FL_ACTION_OPCODE_OUTPUT) + return -EOPNOTSUPP; + + /* Reject merge if sub_flow2 matches on something that is not matched + * on or set in an action by sub_flow1. + */ + err = bitmap_andnot(sub_flow2_merge.vals, sub_flow2_merge.vals, + sub_flow1_merge.vals, + sizeof(struct nfp_flower_merge_check) * 8); + if (err) + return -EINVAL; + + return 0; +} + +static unsigned int +nfp_flower_copy_pre_actions(char *act_dst, char *act_src, int len, + bool *tunnel_act) +{ + unsigned int act_off = 0, act_len; + struct nfp_fl_act_head *a; + u8 act_id = 0; + + while (act_off < len) { + a = (struct nfp_fl_act_head *)&act_src[act_off]; + act_len = a->len_lw << NFP_FL_LW_SIZ; + act_id = a->jump_id; + + switch (act_id) { + case NFP_FL_ACTION_OPCODE_PRE_TUNNEL: + if (tunnel_act) + *tunnel_act = true; + /* fall through */ + case NFP_FL_ACTION_OPCODE_PRE_LAG: + memcpy(act_dst + act_off, act_src + act_off, act_len); + break; + default: + return act_off; + } + + act_off += act_len; + } + + return act_off; +} + +static int +nfp_fl_verify_post_tun_acts(char *acts, int len, struct nfp_fl_push_vlan **vlan) +{ + struct nfp_fl_act_head *a; + unsigned int act_off = 0; + + while (act_off < len) { + a = (struct nfp_fl_act_head *)&acts[act_off]; + + if (a->jump_id == NFP_FL_ACTION_OPCODE_PUSH_VLAN && !act_off) + *vlan = (struct nfp_fl_push_vlan *)a; + else if (a->jump_id != NFP_FL_ACTION_OPCODE_OUTPUT) + return -EOPNOTSUPP; + + act_off += a->len_lw << NFP_FL_LW_SIZ; + } + + /* Ensure any VLAN push also has an egress action. */ + if (*vlan && act_off <= sizeof(struct nfp_fl_push_vlan)) + return -EOPNOTSUPP; + + return 0; +} + +static int +nfp_fl_push_vlan_after_tun(char *acts, int len, struct nfp_fl_push_vlan *vlan) +{ + struct nfp_fl_set_ipv4_tun *tun; + struct nfp_fl_act_head *a; + unsigned int act_off = 0; + + while (act_off < len) { + a = (struct nfp_fl_act_head *)&acts[act_off]; + + if (a->jump_id == NFP_FL_ACTION_OPCODE_SET_IPV4_TUNNEL) { + tun = (struct nfp_fl_set_ipv4_tun *)a; + tun->outer_vlan_tpid = vlan->vlan_tpid; + tun->outer_vlan_tci = vlan->vlan_tci; + + return 0; + } + + act_off += a->len_lw << NFP_FL_LW_SIZ; + } + + /* Return error if no tunnel action is found. */ + return -EOPNOTSUPP; +} + +static int +nfp_flower_merge_action(struct nfp_fl_payload *sub_flow1, + struct nfp_fl_payload *sub_flow2, + struct nfp_fl_payload *merge_flow) +{ + unsigned int sub1_act_len, sub2_act_len, pre_off1, pre_off2; + struct nfp_fl_push_vlan *post_tun_push_vlan = NULL; + bool tunnel_act = false; + char *merge_act; + int err; + + /* The last action of sub_flow1 must be output - do not merge this. */ + sub1_act_len = sub_flow1->meta.act_len - sizeof(struct nfp_fl_output); + sub2_act_len = sub_flow2->meta.act_len; + + if (!sub2_act_len) + return -EINVAL; + + if (sub1_act_len + sub2_act_len > NFP_FL_MAX_A_SIZ) + return -EINVAL; + + /* A shortcut can only be applied if there is a single action. */ + if (sub1_act_len) + merge_flow->meta.shortcut = cpu_to_be32(NFP_FL_SC_ACT_NULL); + else + merge_flow->meta.shortcut = sub_flow2->meta.shortcut; + + merge_flow->meta.act_len = sub1_act_len + sub2_act_len; + merge_act = merge_flow->action_data; + + /* Copy any pre-actions to the start of merge flow action list. */ + pre_off1 = nfp_flower_copy_pre_actions(merge_act, + sub_flow1->action_data, + sub1_act_len, &tunnel_act); + merge_act += pre_off1; + sub1_act_len -= pre_off1; + pre_off2 = nfp_flower_copy_pre_actions(merge_act, + sub_flow2->action_data, + sub2_act_len, NULL); + merge_act += pre_off2; + sub2_act_len -= pre_off2; + + /* FW does a tunnel push when egressing, therefore, if sub_flow 1 pushes + * a tunnel, there are restrictions on what sub_flow 2 actions lead to a + * valid merge. + */ + if (tunnel_act) { + char *post_tun_acts = &sub_flow2->action_data[pre_off2]; + + err = nfp_fl_verify_post_tun_acts(post_tun_acts, sub2_act_len, + &post_tun_push_vlan); + if (err) + return err; + + if (post_tun_push_vlan) { + pre_off2 += sizeof(*post_tun_push_vlan); + sub2_act_len -= sizeof(*post_tun_push_vlan); + } + } + + /* Copy remaining actions from sub_flows 1 and 2. */ + memcpy(merge_act, sub_flow1->action_data + pre_off1, sub1_act_len); + + if (post_tun_push_vlan) { + /* Update tunnel action in merge to include VLAN push. */ + err = nfp_fl_push_vlan_after_tun(merge_act, sub1_act_len, + post_tun_push_vlan); + if (err) + return err; + + merge_flow->meta.act_len -= sizeof(*post_tun_push_vlan); + } + + merge_act += sub1_act_len; + memcpy(merge_act, sub_flow2->action_data + pre_off2, sub2_act_len); + + return 0; +} + +/* Flow link code should only be accessed under RTNL. */ +static void nfp_flower_unlink_flow(struct nfp_fl_payload_link *link) +{ + list_del(&link->merge_flow.list); + list_del(&link->sub_flow.list); + kfree(link); +} + +static void nfp_flower_unlink_flows(struct nfp_fl_payload *merge_flow, + struct nfp_fl_payload *sub_flow) +{ + struct nfp_fl_payload_link *link; + + list_for_each_entry(link, &merge_flow->linked_flows, merge_flow.list) + if (link->sub_flow.flow == sub_flow) { + nfp_flower_unlink_flow(link); + return; + } +} + +static int nfp_flower_link_flows(struct nfp_fl_payload *merge_flow, + struct nfp_fl_payload *sub_flow) +{ + struct nfp_fl_payload_link *link; + + link = kmalloc(sizeof(*link), GFP_KERNEL); + if (!link) + return -ENOMEM; + + link->merge_flow.flow = merge_flow; + list_add_tail(&link->merge_flow.list, &merge_flow->linked_flows); + link->sub_flow.flow = sub_flow; + list_add_tail(&link->sub_flow.list, &sub_flow->linked_flows); + + return 0; +} + +/** + * nfp_flower_merge_offloaded_flows() - Merge 2 existing flows to single flow. + * @app: Pointer to the APP handle + * @sub_flow1: Initial flow matched to produce merge hint + * @sub_flow2: Post recirculation flow matched in merge hint + * + * Combines 2 flows (if valid) to a single flow, removing the initial from hw + * and offloading the new, merged flow. + * + * Return: negative value on error, 0 in success. + */ +int nfp_flower_merge_offloaded_flows(struct nfp_app *app, + struct nfp_fl_payload *sub_flow1, + struct nfp_fl_payload *sub_flow2) +{ + struct flow_cls_offload merge_tc_off; + struct nfp_flower_priv *priv = app->priv; + struct netlink_ext_ack *extack = NULL; + struct nfp_fl_payload *merge_flow; + struct nfp_fl_key_ls merge_key_ls; + int err; + + ASSERT_RTNL(); + + extack = merge_tc_off.common.extack; + if (sub_flow1 == sub_flow2 || + nfp_flower_is_merge_flow(sub_flow1) || + nfp_flower_is_merge_flow(sub_flow2)) + return -EINVAL; + + err = nfp_flower_can_merge(sub_flow1, sub_flow2); + if (err) + return err; + + merge_key_ls.key_size = sub_flow1->meta.key_len; + + merge_flow = nfp_flower_allocate_new(&merge_key_ls); + if (!merge_flow) + return -ENOMEM; + + merge_flow->tc_flower_cookie = (unsigned long)merge_flow; + merge_flow->ingress_dev = sub_flow1->ingress_dev; + + memcpy(merge_flow->unmasked_data, sub_flow1->unmasked_data, + sub_flow1->meta.key_len); + memcpy(merge_flow->mask_data, sub_flow1->mask_data, + sub_flow1->meta.mask_len); + + err = nfp_flower_merge_action(sub_flow1, sub_flow2, merge_flow); + if (err) + goto err_destroy_merge_flow; + + err = nfp_flower_link_flows(merge_flow, sub_flow1); + if (err) + goto err_destroy_merge_flow; + + err = nfp_flower_link_flows(merge_flow, sub_flow2); + if (err) + goto err_unlink_sub_flow1; + + merge_tc_off.cookie = merge_flow->tc_flower_cookie; + err = nfp_compile_flow_metadata(app, &merge_tc_off, merge_flow, + merge_flow->ingress_dev, extack); + if (err) + goto err_unlink_sub_flow2; + + err = rhashtable_insert_fast(&priv->flow_table, &merge_flow->fl_node, + nfp_flower_table_params); + if (err) + goto err_release_metadata; + + err = nfp_flower_xmit_flow(app, merge_flow, + NFP_FLOWER_CMSG_TYPE_FLOW_MOD); + if (err) + goto err_remove_rhash; + + merge_flow->in_hw = true; + sub_flow1->in_hw = false; + + return 0; + +err_remove_rhash: + WARN_ON_ONCE(rhashtable_remove_fast(&priv->flow_table, + &merge_flow->fl_node, + nfp_flower_table_params)); +err_release_metadata: + nfp_modify_flow_metadata(app, merge_flow); +err_unlink_sub_flow2: + nfp_flower_unlink_flows(merge_flow, sub_flow2); +err_unlink_sub_flow1: + nfp_flower_unlink_flows(merge_flow, sub_flow1); +err_destroy_merge_flow: + kfree(merge_flow->action_data); + kfree(merge_flow->mask_data); + kfree(merge_flow->unmasked_data); + kfree(merge_flow); + return err; +} + +/** + * nfp_flower_validate_pre_tun_rule() + * @app: Pointer to the APP handle + * @flow: Pointer to NFP flow representation of rule + * @extack: Netlink extended ACK report + * + * Verifies the flow as a pre-tunnel rule. + * + * Return: negative value on error, 0 if verified. + */ +static int +nfp_flower_validate_pre_tun_rule(struct nfp_app *app, + struct nfp_fl_payload *flow, + struct netlink_ext_ack *extack) +{ + struct nfp_flower_meta_tci *meta_tci; + struct nfp_flower_mac_mpls *mac; + struct nfp_fl_act_head *act; + u8 *mask = flow->mask_data; + bool vlan = false; + int act_offset; + u8 key_layer; + + meta_tci = (struct nfp_flower_meta_tci *)flow->unmasked_data; + if (meta_tci->tci & cpu_to_be16(NFP_FLOWER_MASK_VLAN_PRESENT)) { + u16 vlan_tci = be16_to_cpu(meta_tci->tci); + + vlan_tci &= ~NFP_FLOWER_MASK_VLAN_PRESENT; + flow->pre_tun_rule.vlan_tci = cpu_to_be16(vlan_tci); + vlan = true; + } else { + flow->pre_tun_rule.vlan_tci = cpu_to_be16(0xffff); + } + + key_layer = meta_tci->nfp_flow_key_layer; + if (key_layer & ~NFP_FLOWER_PRE_TUN_RULE_FIELDS) { + NL_SET_ERR_MSG_MOD(extack, "unsupported pre-tunnel rule: too many match fields"); + return -EOPNOTSUPP; + } + + if (!(key_layer & NFP_FLOWER_LAYER_MAC)) { + NL_SET_ERR_MSG_MOD(extack, "unsupported pre-tunnel rule: MAC fields match required"); + return -EOPNOTSUPP; + } + + /* Skip fields known to exist. */ + mask += sizeof(struct nfp_flower_meta_tci); + mask += sizeof(struct nfp_flower_in_port); + + /* Ensure destination MAC address is fully matched. */ + mac = (struct nfp_flower_mac_mpls *)mask; + if (!is_broadcast_ether_addr(&mac->mac_dst[0])) { + NL_SET_ERR_MSG_MOD(extack, "unsupported pre-tunnel rule: dest MAC field must not be masked"); + return -EOPNOTSUPP; + } + + if (key_layer & NFP_FLOWER_LAYER_IPV4) { + int ip_flags = offsetof(struct nfp_flower_ipv4, ip_ext.flags); + int ip_proto = offsetof(struct nfp_flower_ipv4, ip_ext.proto); + int i; + + mask += sizeof(struct nfp_flower_mac_mpls); + + /* Ensure proto and flags are the only IP layer fields. */ + for (i = 0; i < sizeof(struct nfp_flower_ipv4); i++) + if (mask[i] && i != ip_flags && i != ip_proto) { + NL_SET_ERR_MSG_MOD(extack, "unsupported pre-tunnel rule: only flags and proto can be matched in ip header"); + return -EOPNOTSUPP; + } + } + + /* Action must be a single egress or pop_vlan and egress. */ + act_offset = 0; + act = (struct nfp_fl_act_head *)&flow->action_data[act_offset]; + if (vlan) { + if (act->jump_id != NFP_FL_ACTION_OPCODE_POP_VLAN) { + NL_SET_ERR_MSG_MOD(extack, "unsupported pre-tunnel rule: match on VLAN must have VLAN pop as first action"); + return -EOPNOTSUPP; + } + + act_offset += act->len_lw << NFP_FL_LW_SIZ; + act = (struct nfp_fl_act_head *)&flow->action_data[act_offset]; + } + + if (act->jump_id != NFP_FL_ACTION_OPCODE_OUTPUT) { + NL_SET_ERR_MSG_MOD(extack, "unsupported pre-tunnel rule: non egress action detected where egress was expected"); + return -EOPNOTSUPP; + } + + act_offset += act->len_lw << NFP_FL_LW_SIZ; + + /* Ensure there are no more actions after egress. */ + if (act_offset != flow->meta.act_len) { + NL_SET_ERR_MSG_MOD(extack, "unsupported pre-tunnel rule: egress is not the last action"); + return -EOPNOTSUPP; + } + + return 0; +} + /** * nfp_flower_add_offload() - Adds a new flow to hardware. * @app: Pointer to the APP handle @@ -400,15 +1114,17 @@ err_free_flow: */ static int nfp_flower_add_offload(struct nfp_app *app, struct net_device *netdev, - struct tc_cls_flower_offload *flow) + struct flow_cls_offload *flow) { enum nfp_flower_tun_type tun_type = NFP_FL_TUNNEL_NONE; struct nfp_flower_priv *priv = app->priv; + struct netlink_ext_ack *extack = NULL; struct nfp_fl_payload *flow_pay; struct nfp_fl_key_ls *key_layer; struct nfp_port *port = NULL; int err; + extack = flow->common.extack; if (nfp_netdev_is_nfp_repr(netdev)) port = nfp_port_from_netdev(netdev); @@ -417,7 +1133,7 @@ nfp_flower_add_offload(struct nfp_app *app, struct net_device *netdev, return -ENOMEM; err = nfp_flower_calculate_key_layers(app, netdev, key_layer, flow, - &tun_type); + &tun_type, extack); if (err) goto err_free_key_ls; @@ -428,32 +1144,45 @@ nfp_flower_add_offload(struct nfp_app *app, struct net_device *netdev, } err = nfp_flower_compile_flow_match(app, flow, key_layer, netdev, - flow_pay, tun_type); + flow_pay, tun_type, extack); if (err) goto err_destroy_flow; - err = nfp_flower_compile_action(app, flow, netdev, flow_pay); + err = nfp_flower_compile_action(app, flow, netdev, flow_pay, extack); if (err) goto err_destroy_flow; - err = nfp_compile_flow_metadata(app, flow, flow_pay, netdev); + if (flow_pay->pre_tun_rule.dev) { + err = nfp_flower_validate_pre_tun_rule(app, flow_pay, extack); + if (err) + goto err_destroy_flow; + } + + err = nfp_compile_flow_metadata(app, flow, flow_pay, netdev, extack); if (err) goto err_destroy_flow; flow_pay->tc_flower_cookie = flow->cookie; err = rhashtable_insert_fast(&priv->flow_table, &flow_pay->fl_node, nfp_flower_table_params); - if (err) + if (err) { + NL_SET_ERR_MSG_MOD(extack, "invalid entry: cannot insert flow into tables for offloads"); goto err_release_metadata; + } - err = nfp_flower_xmit_flow(app, flow_pay, - NFP_FLOWER_CMSG_TYPE_FLOW_ADD); + if (flow_pay->pre_tun_rule.dev) + err = nfp_flower_xmit_pre_tun_flow(app, flow_pay); + else + err = nfp_flower_xmit_flow(app, flow_pay, + NFP_FLOWER_CMSG_TYPE_FLOW_ADD); if (err) goto err_remove_rhash; if (port) port->tc_offload_cnt++; + flow_pay->in_hw = true; + /* Deallocate flow payload when flower rule has been destroyed. */ kfree(key_layer); @@ -475,6 +1204,75 @@ err_free_key_ls: return err; } +static void +nfp_flower_remove_merge_flow(struct nfp_app *app, + struct nfp_fl_payload *del_sub_flow, + struct nfp_fl_payload *merge_flow) +{ + struct nfp_flower_priv *priv = app->priv; + struct nfp_fl_payload_link *link, *temp; + struct nfp_fl_payload *origin; + bool mod = false; + int err; + + link = list_first_entry(&merge_flow->linked_flows, + struct nfp_fl_payload_link, merge_flow.list); + origin = link->sub_flow.flow; + + /* Re-add rule the merge had overwritten if it has not been deleted. */ + if (origin != del_sub_flow) + mod = true; + + err = nfp_modify_flow_metadata(app, merge_flow); + if (err) { + nfp_flower_cmsg_warn(app, "Metadata fail for merge flow delete.\n"); + goto err_free_links; + } + + if (!mod) { + err = nfp_flower_xmit_flow(app, merge_flow, + NFP_FLOWER_CMSG_TYPE_FLOW_DEL); + if (err) { + nfp_flower_cmsg_warn(app, "Failed to delete merged flow.\n"); + goto err_free_links; + } + } else { + __nfp_modify_flow_metadata(priv, origin); + err = nfp_flower_xmit_flow(app, origin, + NFP_FLOWER_CMSG_TYPE_FLOW_MOD); + if (err) + nfp_flower_cmsg_warn(app, "Failed to revert merge flow.\n"); + origin->in_hw = true; + } + +err_free_links: + /* Clean any links connected with the merged flow. */ + list_for_each_entry_safe(link, temp, &merge_flow->linked_flows, + merge_flow.list) + nfp_flower_unlink_flow(link); + + kfree(merge_flow->action_data); + kfree(merge_flow->mask_data); + kfree(merge_flow->unmasked_data); + WARN_ON_ONCE(rhashtable_remove_fast(&priv->flow_table, + &merge_flow->fl_node, + nfp_flower_table_params)); + kfree_rcu(merge_flow, rcu); +} + +static void +nfp_flower_del_linked_merge_flows(struct nfp_app *app, + struct nfp_fl_payload *sub_flow) +{ + struct nfp_fl_payload_link *link, *temp; + + /* Remove any merge flow formed from the deleted sub_flow. */ + list_for_each_entry_safe(link, temp, &sub_flow->linked_flows, + sub_flow.list) + nfp_flower_remove_merge_flow(app, sub_flow, + link->merge_flow.flow); +} + /** * nfp_flower_del_offload() - Removes a flow from hardware. * @app: Pointer to the APP handle @@ -482,39 +1280,51 @@ err_free_key_ls: * @flow: TC flower classifier offload structure * * Removes a flow from the repeated hash structure and clears the - * action payload. + * action payload. Any flows merged from this are also deleted. * * Return: negative value on error, 0 if removed successfully. */ static int nfp_flower_del_offload(struct nfp_app *app, struct net_device *netdev, - struct tc_cls_flower_offload *flow) + struct flow_cls_offload *flow) { struct nfp_flower_priv *priv = app->priv; + struct netlink_ext_ack *extack = NULL; struct nfp_fl_payload *nfp_flow; struct nfp_port *port = NULL; int err; + extack = flow->common.extack; if (nfp_netdev_is_nfp_repr(netdev)) port = nfp_port_from_netdev(netdev); nfp_flow = nfp_flower_search_fl_table(app, flow->cookie, netdev); - if (!nfp_flow) + if (!nfp_flow) { + NL_SET_ERR_MSG_MOD(extack, "invalid entry: cannot remove flow that does not exist"); return -ENOENT; + } err = nfp_modify_flow_metadata(app, nfp_flow); if (err) - goto err_free_flow; + goto err_free_merge_flow; if (nfp_flow->nfp_tun_ipv4_addr) nfp_tunnel_del_ipv4_off(app, nfp_flow->nfp_tun_ipv4_addr); - err = nfp_flower_xmit_flow(app, nfp_flow, - NFP_FLOWER_CMSG_TYPE_FLOW_DEL); - if (err) - goto err_free_flow; + if (!nfp_flow->in_hw) { + err = 0; + goto err_free_merge_flow; + } -err_free_flow: + if (nfp_flow->pre_tun_rule.dev) + err = nfp_flower_xmit_pre_tun_del_flow(app, nfp_flow); + else + err = nfp_flower_xmit_flow(app, nfp_flow, + NFP_FLOWER_CMSG_TYPE_FLOW_DEL); + /* Fall through on error. */ + +err_free_merge_flow: + nfp_flower_del_linked_merge_flows(app, nfp_flow); if (port) port->tc_offload_cnt--; kfree(nfp_flow->action_data); @@ -527,6 +1337,52 @@ err_free_flow: return err; } +static void +__nfp_flower_update_merge_stats(struct nfp_app *app, + struct nfp_fl_payload *merge_flow) +{ + struct nfp_flower_priv *priv = app->priv; + struct nfp_fl_payload_link *link; + struct nfp_fl_payload *sub_flow; + u64 pkts, bytes, used; + u32 ctx_id; + + ctx_id = be32_to_cpu(merge_flow->meta.host_ctx_id); + pkts = priv->stats[ctx_id].pkts; + /* Do not cycle subflows if no stats to distribute. */ + if (!pkts) + return; + bytes = priv->stats[ctx_id].bytes; + used = priv->stats[ctx_id].used; + + /* Reset stats for the merge flow. */ + priv->stats[ctx_id].pkts = 0; + priv->stats[ctx_id].bytes = 0; + + /* The merge flow has received stats updates from firmware. + * Distribute these stats to all subflows that form the merge. + * The stats will collected from TC via the subflows. + */ + list_for_each_entry(link, &merge_flow->linked_flows, merge_flow.list) { + sub_flow = link->sub_flow.flow; + ctx_id = be32_to_cpu(sub_flow->meta.host_ctx_id); + priv->stats[ctx_id].pkts += pkts; + priv->stats[ctx_id].bytes += bytes; + max_t(u64, priv->stats[ctx_id].used, used); + } +} + +static void +nfp_flower_update_merge_stats(struct nfp_app *app, + struct nfp_fl_payload *sub_flow) +{ + struct nfp_fl_payload_link *link; + + /* Get merge flows that the subflow forms to distribute their stats. */ + list_for_each_entry(link, &sub_flow->linked_flows, sub_flow.list) + __nfp_flower_update_merge_stats(app, link->merge_flow.flow); +} + /** * nfp_flower_get_stats() - Populates flow stats obtained from hardware. * @app: Pointer to the APP handle @@ -540,19 +1396,27 @@ err_free_flow: */ static int nfp_flower_get_stats(struct nfp_app *app, struct net_device *netdev, - struct tc_cls_flower_offload *flow) + struct flow_cls_offload *flow) { struct nfp_flower_priv *priv = app->priv; + struct netlink_ext_ack *extack = NULL; struct nfp_fl_payload *nfp_flow; u32 ctx_id; + extack = flow->common.extack; nfp_flow = nfp_flower_search_fl_table(app, flow->cookie, netdev); - if (!nfp_flow) + if (!nfp_flow) { + NL_SET_ERR_MSG_MOD(extack, "invalid entry: cannot dump stats for flow that does not exist"); return -EINVAL; + } ctx_id = be32_to_cpu(nfp_flow->meta.host_ctx_id); spin_lock_bh(&priv->stats_lock); + /* If request is for a sub_flow, update stats from merged flows. */ + if (!list_empty(&nfp_flow->linked_flows)) + nfp_flower_update_merge_stats(app, nfp_flow); + flow_stats_update(&flow->stats, priv->stats[ctx_id].bytes, priv->stats[ctx_id].pkts, priv->stats[ctx_id].used); @@ -565,17 +1429,17 @@ nfp_flower_get_stats(struct nfp_app *app, struct net_device *netdev, static int nfp_flower_repr_offload(struct nfp_app *app, struct net_device *netdev, - struct tc_cls_flower_offload *flower) + struct flow_cls_offload *flower) { if (!eth_proto_is_802_3(flower->common.protocol)) return -EOPNOTSUPP; switch (flower->command) { - case TC_CLSFLOWER_REPLACE: + case FLOW_CLS_REPLACE: return nfp_flower_add_offload(app, netdev, flower); - case TC_CLSFLOWER_DESTROY: + case FLOW_CLS_DESTROY: return nfp_flower_del_offload(app, netdev, flower); - case TC_CLSFLOWER_STATS: + case FLOW_CLS_STATS: return nfp_flower_get_stats(app, netdev, flower); default: return -EOPNOTSUPP; @@ -594,28 +1458,53 @@ static int nfp_flower_setup_tc_block_cb(enum tc_setup_type type, case TC_SETUP_CLSFLOWER: return nfp_flower_repr_offload(repr->app, repr->netdev, type_data); + case TC_SETUP_CLSMATCHALL: + return nfp_flower_setup_qos_offload(repr->app, repr->netdev, + type_data); default: return -EOPNOTSUPP; } } +static LIST_HEAD(nfp_block_cb_list); + static int nfp_flower_setup_tc_block(struct net_device *netdev, - struct tc_block_offload *f) + struct flow_block_offload *f) { struct nfp_repr *repr = netdev_priv(netdev); + struct nfp_flower_repr_priv *repr_priv; + struct flow_block_cb *block_cb; - if (f->binder_type != TCF_BLOCK_BINDER_TYPE_CLSACT_INGRESS) + if (f->binder_type != FLOW_BLOCK_BINDER_TYPE_CLSACT_INGRESS) return -EOPNOTSUPP; + repr_priv = repr->app_priv; + repr_priv->block_shared = f->block_shared; + f->driver_block_list = &nfp_block_cb_list; + switch (f->command) { - case TC_BLOCK_BIND: - return tcf_block_cb_register(f->block, - nfp_flower_setup_tc_block_cb, - repr, repr, f->extack); - case TC_BLOCK_UNBIND: - tcf_block_cb_unregister(f->block, - nfp_flower_setup_tc_block_cb, - repr); + case FLOW_BLOCK_BIND: + if (flow_block_cb_is_busy(nfp_flower_setup_tc_block_cb, repr, + &nfp_block_cb_list)) + return -EBUSY; + + block_cb = flow_block_cb_alloc(nfp_flower_setup_tc_block_cb, + repr, repr, NULL); + if (IS_ERR(block_cb)) + return PTR_ERR(block_cb); + + flow_block_cb_add(block_cb, f); + list_add_tail(&block_cb->driver_list, &nfp_block_cb_list); + return 0; + case FLOW_BLOCK_UNBIND: + block_cb = flow_block_cb_lookup(f->block, + nfp_flower_setup_tc_block_cb, + repr); + if (!block_cb) + return -ENOENT; + + flow_block_cb_remove(block_cb, f); + list_del(&block_cb->driver_list); return 0; default: return -EOPNOTSUPP; @@ -660,7 +1549,7 @@ static int nfp_flower_setup_indr_block_cb(enum tc_setup_type type, void *type_data, void *cb_priv) { struct nfp_flower_indr_block_cb_priv *priv = cb_priv; - struct tc_cls_flower_offload *flower = type_data; + struct flow_cls_offload *flower = type_data; if (flower->common.chain_index) return -EOPNOTSUPP; @@ -674,19 +1563,37 @@ static int nfp_flower_setup_indr_block_cb(enum tc_setup_type type, } } +static void nfp_flower_setup_indr_tc_release(void *cb_priv) +{ + struct nfp_flower_indr_block_cb_priv *priv = cb_priv; + + list_del(&priv->list); + kfree(priv); +} + static int nfp_flower_setup_indr_tc_block(struct net_device *netdev, struct nfp_app *app, - struct tc_block_offload *f) + struct flow_block_offload *f) { struct nfp_flower_indr_block_cb_priv *cb_priv; struct nfp_flower_priv *priv = app->priv; - int err; + struct flow_block_cb *block_cb; - if (f->binder_type != TCF_BLOCK_BINDER_TYPE_CLSACT_INGRESS) + if ((f->binder_type != FLOW_BLOCK_BINDER_TYPE_CLSACT_INGRESS && + !nfp_flower_internal_port_can_offload(app, netdev)) || + (f->binder_type != FLOW_BLOCK_BINDER_TYPE_CLSACT_EGRESS && + nfp_flower_internal_port_can_offload(app, netdev))) return -EOPNOTSUPP; switch (f->command) { - case TC_BLOCK_BIND: + case FLOW_BLOCK_BIND: + cb_priv = nfp_flower_indr_block_cb_priv_lookup(app, netdev); + if (cb_priv && + flow_block_cb_is_busy(nfp_flower_setup_indr_block_cb, + cb_priv, + &nfp_block_cb_list)) + return -EBUSY; + cb_priv = kmalloc(sizeof(*cb_priv), GFP_KERNEL); if (!cb_priv) return -ENOMEM; @@ -695,26 +1602,31 @@ nfp_flower_setup_indr_tc_block(struct net_device *netdev, struct nfp_app *app, cb_priv->app = app; list_add(&cb_priv->list, &priv->indr_block_cb_priv); - err = tcf_block_cb_register(f->block, - nfp_flower_setup_indr_block_cb, - cb_priv, cb_priv, f->extack); - if (err) { + block_cb = flow_block_cb_alloc(nfp_flower_setup_indr_block_cb, + cb_priv, cb_priv, + nfp_flower_setup_indr_tc_release); + if (IS_ERR(block_cb)) { list_del(&cb_priv->list); kfree(cb_priv); + return PTR_ERR(block_cb); } - return err; - case TC_BLOCK_UNBIND: + flow_block_cb_add(block_cb, f); + list_add_tail(&block_cb->driver_list, &nfp_block_cb_list); + return 0; + case FLOW_BLOCK_UNBIND: cb_priv = nfp_flower_indr_block_cb_priv_lookup(app, netdev); if (!cb_priv) return -ENOENT; - tcf_block_cb_unregister(f->block, - nfp_flower_setup_indr_block_cb, - cb_priv); - list_del(&cb_priv->list); - kfree(cb_priv); + block_cb = flow_block_cb_lookup(f->block, + nfp_flower_setup_indr_block_cb, + cb_priv); + if (!block_cb) + return -ENOENT; + flow_block_cb_remove(block_cb, f); + list_del(&block_cb->driver_list); return 0; default: return -EOPNOTSUPP; @@ -745,16 +1657,17 @@ int nfp_flower_reg_indir_block_handler(struct nfp_app *app, return NOTIFY_OK; if (event == NETDEV_REGISTER) { - err = __tc_indr_block_cb_register(netdev, app, - nfp_flower_indr_setup_tc_cb, - app); + err = __flow_indr_block_cb_register(netdev, app, + nfp_flower_indr_setup_tc_cb, + app); if (err) nfp_flower_cmsg_warn(app, "Indirect block reg failed - %s\n", netdev->name); } else if (event == NETDEV_UNREGISTER) { - __tc_indr_block_cb_unregister(netdev, - nfp_flower_indr_setup_tc_cb, app); + __flow_indr_block_cb_unregister(netdev, + nfp_flower_indr_setup_tc_cb, + app); } return NOTIFY_OK; diff --git a/drivers/net/ethernet/netronome/nfp/flower/qos_conf.c b/drivers/net/ethernet/netronome/nfp/flower/qos_conf.c new file mode 100644 index 000000000000..124a43dc136a --- /dev/null +++ b/drivers/net/ethernet/netronome/nfp/flower/qos_conf.c @@ -0,0 +1,366 @@ +// SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) +/* Copyright (C) 2019 Netronome Systems, Inc. */ + +#include <linux/math64.h> +#include <net/pkt_cls.h> +#include <net/pkt_sched.h> + +#include "cmsg.h" +#include "main.h" +#include "../nfp_port.h" + +#define NFP_FL_QOS_UPDATE msecs_to_jiffies(1000) + +struct nfp_police_cfg_head { + __be32 flags_opts; + __be32 port; +}; + +/* Police cmsg for configuring a trTCM traffic conditioner (8W/32B) + * See RFC 2698 for more details. + * ---------------------------------------------------------------- + * 3 2 1 + * 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + * | Flag options | + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + * | Port Ingress | + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + * | Token Bucket Peak | + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + * | Token Bucket Committed | + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + * | Peak Burst Size | + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + * | Committed Burst Size | + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + * | Peak Information Rate | + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + * | Committed Information Rate | + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + */ +struct nfp_police_config { + struct nfp_police_cfg_head head; + __be32 bkt_tkn_p; + __be32 bkt_tkn_c; + __be32 pbs; + __be32 cbs; + __be32 pir; + __be32 cir; +}; + +struct nfp_police_stats_reply { + struct nfp_police_cfg_head head; + __be64 pass_bytes; + __be64 pass_pkts; + __be64 drop_bytes; + __be64 drop_pkts; +}; + +static int +nfp_flower_install_rate_limiter(struct nfp_app *app, struct net_device *netdev, + struct tc_cls_matchall_offload *flow, + struct netlink_ext_ack *extack) +{ + struct flow_action_entry *action = &flow->rule->action.entries[0]; + struct nfp_flower_priv *fl_priv = app->priv; + struct nfp_flower_repr_priv *repr_priv; + struct nfp_police_config *config; + struct nfp_repr *repr; + struct sk_buff *skb; + u32 netdev_port_id; + u64 burst, rate; + + if (!nfp_netdev_is_nfp_repr(netdev)) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: qos rate limit offload not supported on higher level port"); + return -EOPNOTSUPP; + } + repr = netdev_priv(netdev); + repr_priv = repr->app_priv; + + if (repr_priv->block_shared) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: qos rate limit offload not supported on shared blocks"); + return -EOPNOTSUPP; + } + + if (repr->port->type != NFP_PORT_VF_PORT) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: qos rate limit offload not supported on non-VF ports"); + return -EOPNOTSUPP; + } + + if (!flow_offload_has_one_action(&flow->rule->action)) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: qos rate limit offload requires a single action"); + return -EOPNOTSUPP; + } + + if (flow->common.prio != 1) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: qos rate limit offload requires highest priority"); + return -EOPNOTSUPP; + } + + if (action->id != FLOW_ACTION_POLICE) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: qos rate limit offload requires police action"); + return -EOPNOTSUPP; + } + + rate = action->police.rate_bytes_ps; + burst = div_u64(rate * PSCHED_NS2TICKS(action->police.burst), + PSCHED_TICKS_PER_SEC); + netdev_port_id = nfp_repr_get_port_id(netdev); + + skb = nfp_flower_cmsg_alloc(repr->app, sizeof(struct nfp_police_config), + NFP_FLOWER_CMSG_TYPE_QOS_MOD, GFP_KERNEL); + if (!skb) + return -ENOMEM; + + config = nfp_flower_cmsg_get_data(skb); + memset(config, 0, sizeof(struct nfp_police_config)); + config->head.port = cpu_to_be32(netdev_port_id); + config->bkt_tkn_p = cpu_to_be32(burst); + config->bkt_tkn_c = cpu_to_be32(burst); + config->pbs = cpu_to_be32(burst); + config->cbs = cpu_to_be32(burst); + config->pir = cpu_to_be32(rate); + config->cir = cpu_to_be32(rate); + nfp_ctrl_tx(repr->app->ctrl, skb); + + repr_priv->qos_table.netdev_port_id = netdev_port_id; + fl_priv->qos_rate_limiters++; + if (fl_priv->qos_rate_limiters == 1) + schedule_delayed_work(&fl_priv->qos_stats_work, + NFP_FL_QOS_UPDATE); + + return 0; +} + +static int +nfp_flower_remove_rate_limiter(struct nfp_app *app, struct net_device *netdev, + struct tc_cls_matchall_offload *flow, + struct netlink_ext_ack *extack) +{ + struct nfp_flower_priv *fl_priv = app->priv; + struct nfp_flower_repr_priv *repr_priv; + struct nfp_police_config *config; + struct nfp_repr *repr; + struct sk_buff *skb; + u32 netdev_port_id; + + if (!nfp_netdev_is_nfp_repr(netdev)) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: qos rate limit offload not supported on higher level port"); + return -EOPNOTSUPP; + } + repr = netdev_priv(netdev); + + netdev_port_id = nfp_repr_get_port_id(netdev); + repr_priv = repr->app_priv; + + if (!repr_priv->qos_table.netdev_port_id) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: cannot remove qos entry that does not exist"); + return -EOPNOTSUPP; + } + + skb = nfp_flower_cmsg_alloc(repr->app, sizeof(struct nfp_police_config), + NFP_FLOWER_CMSG_TYPE_QOS_DEL, GFP_KERNEL); + if (!skb) + return -ENOMEM; + + /* Clear all qos associate data for this interface */ + memset(&repr_priv->qos_table, 0, sizeof(struct nfp_fl_qos)); + fl_priv->qos_rate_limiters--; + if (!fl_priv->qos_rate_limiters) + cancel_delayed_work_sync(&fl_priv->qos_stats_work); + + config = nfp_flower_cmsg_get_data(skb); + memset(config, 0, sizeof(struct nfp_police_config)); + config->head.port = cpu_to_be32(netdev_port_id); + nfp_ctrl_tx(repr->app->ctrl, skb); + + return 0; +} + +void nfp_flower_stats_rlim_reply(struct nfp_app *app, struct sk_buff *skb) +{ + struct nfp_flower_priv *fl_priv = app->priv; + struct nfp_flower_repr_priv *repr_priv; + struct nfp_police_stats_reply *msg; + struct nfp_stat_pair *curr_stats; + struct nfp_stat_pair *prev_stats; + struct net_device *netdev; + struct nfp_repr *repr; + u32 netdev_port_id; + + msg = nfp_flower_cmsg_get_data(skb); + netdev_port_id = be32_to_cpu(msg->head.port); + rcu_read_lock(); + netdev = nfp_app_dev_get(app, netdev_port_id, NULL); + if (!netdev) + goto exit_unlock_rcu; + + repr = netdev_priv(netdev); + repr_priv = repr->app_priv; + curr_stats = &repr_priv->qos_table.curr_stats; + prev_stats = &repr_priv->qos_table.prev_stats; + + spin_lock_bh(&fl_priv->qos_stats_lock); + curr_stats->pkts = be64_to_cpu(msg->pass_pkts) + + be64_to_cpu(msg->drop_pkts); + curr_stats->bytes = be64_to_cpu(msg->pass_bytes) + + be64_to_cpu(msg->drop_bytes); + + if (!repr_priv->qos_table.last_update) { + prev_stats->pkts = curr_stats->pkts; + prev_stats->bytes = curr_stats->bytes; + } + + repr_priv->qos_table.last_update = jiffies; + spin_unlock_bh(&fl_priv->qos_stats_lock); + +exit_unlock_rcu: + rcu_read_unlock(); +} + +static void +nfp_flower_stats_rlim_request(struct nfp_flower_priv *fl_priv, + u32 netdev_port_id) +{ + struct nfp_police_cfg_head *head; + struct sk_buff *skb; + + skb = nfp_flower_cmsg_alloc(fl_priv->app, + sizeof(struct nfp_police_cfg_head), + NFP_FLOWER_CMSG_TYPE_QOS_STATS, + GFP_ATOMIC); + if (!skb) + return; + + head = nfp_flower_cmsg_get_data(skb); + memset(head, 0, sizeof(struct nfp_police_cfg_head)); + head->port = cpu_to_be32(netdev_port_id); + + nfp_ctrl_tx(fl_priv->app->ctrl, skb); +} + +static void +nfp_flower_stats_rlim_request_all(struct nfp_flower_priv *fl_priv) +{ + struct nfp_reprs *repr_set; + int i; + + rcu_read_lock(); + repr_set = rcu_dereference(fl_priv->app->reprs[NFP_REPR_TYPE_VF]); + if (!repr_set) + goto exit_unlock_rcu; + + for (i = 0; i < repr_set->num_reprs; i++) { + struct net_device *netdev; + + netdev = rcu_dereference(repr_set->reprs[i]); + if (netdev) { + struct nfp_repr *priv = netdev_priv(netdev); + struct nfp_flower_repr_priv *repr_priv; + u32 netdev_port_id; + + repr_priv = priv->app_priv; + netdev_port_id = repr_priv->qos_table.netdev_port_id; + if (!netdev_port_id) + continue; + + nfp_flower_stats_rlim_request(fl_priv, netdev_port_id); + } + } + +exit_unlock_rcu: + rcu_read_unlock(); +} + +static void update_stats_cache(struct work_struct *work) +{ + struct delayed_work *delayed_work; + struct nfp_flower_priv *fl_priv; + + delayed_work = to_delayed_work(work); + fl_priv = container_of(delayed_work, struct nfp_flower_priv, + qos_stats_work); + + nfp_flower_stats_rlim_request_all(fl_priv); + schedule_delayed_work(&fl_priv->qos_stats_work, NFP_FL_QOS_UPDATE); +} + +static int +nfp_flower_stats_rate_limiter(struct nfp_app *app, struct net_device *netdev, + struct tc_cls_matchall_offload *flow, + struct netlink_ext_ack *extack) +{ + struct nfp_flower_priv *fl_priv = app->priv; + struct nfp_flower_repr_priv *repr_priv; + struct nfp_stat_pair *curr_stats; + struct nfp_stat_pair *prev_stats; + u64 diff_bytes, diff_pkts; + struct nfp_repr *repr; + + if (!nfp_netdev_is_nfp_repr(netdev)) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: qos rate limit offload not supported on higher level port"); + return -EOPNOTSUPP; + } + repr = netdev_priv(netdev); + + repr_priv = repr->app_priv; + if (!repr_priv->qos_table.netdev_port_id) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: cannot find qos entry for stats update"); + return -EOPNOTSUPP; + } + + spin_lock_bh(&fl_priv->qos_stats_lock); + curr_stats = &repr_priv->qos_table.curr_stats; + prev_stats = &repr_priv->qos_table.prev_stats; + diff_pkts = curr_stats->pkts - prev_stats->pkts; + diff_bytes = curr_stats->bytes - prev_stats->bytes; + prev_stats->pkts = curr_stats->pkts; + prev_stats->bytes = curr_stats->bytes; + spin_unlock_bh(&fl_priv->qos_stats_lock); + + flow_stats_update(&flow->stats, diff_bytes, diff_pkts, + repr_priv->qos_table.last_update); + return 0; +} + +void nfp_flower_qos_init(struct nfp_app *app) +{ + struct nfp_flower_priv *fl_priv = app->priv; + + spin_lock_init(&fl_priv->qos_stats_lock); + INIT_DELAYED_WORK(&fl_priv->qos_stats_work, &update_stats_cache); +} + +void nfp_flower_qos_cleanup(struct nfp_app *app) +{ + struct nfp_flower_priv *fl_priv = app->priv; + + cancel_delayed_work_sync(&fl_priv->qos_stats_work); +} + +int nfp_flower_setup_qos_offload(struct nfp_app *app, struct net_device *netdev, + struct tc_cls_matchall_offload *flow) +{ + struct netlink_ext_ack *extack = flow->common.extack; + struct nfp_flower_priv *fl_priv = app->priv; + + if (!(fl_priv->flower_ext_feats & NFP_FL_FEATS_VF_RLIM)) { + NL_SET_ERR_MSG_MOD(extack, "unsupported offload: loaded firmware does not support qos rate limit offload"); + return -EOPNOTSUPP; + } + + switch (flow->command) { + case TC_CLSMATCHALL_REPLACE: + return nfp_flower_install_rate_limiter(app, netdev, flow, + extack); + case TC_CLSMATCHALL_DESTROY: + return nfp_flower_remove_rate_limiter(app, netdev, flow, + extack); + case TC_CLSMATCHALL_STATS: + return nfp_flower_stats_rate_limiter(app, netdev, flow, + extack); + default: + return -EOPNOTSUPP; + } +} diff --git a/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c b/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c index 4d78be4ec4e9..2600ce476d6b 100644 --- a/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c +++ b/drivers/net/ethernet/netronome/nfp/flower/tunnel_conf.c @@ -15,6 +15,24 @@ #define NFP_FL_MAX_ROUTES 32 +#define NFP_TUN_PRE_TUN_RULE_LIMIT 32 +#define NFP_TUN_PRE_TUN_RULE_DEL 0x1 +#define NFP_TUN_PRE_TUN_IDX_BIT 0x8 + +/** + * struct nfp_tun_pre_run_rule - rule matched before decap + * @flags: options for the rule offset + * @port_idx: index of destination MAC address for the rule + * @vlan_tci: VLAN info associated with MAC + * @host_ctx_id: stats context of rule to update + */ +struct nfp_tun_pre_tun_rule { + __be32 flags; + __be16 port_idx; + __be16 vlan_tci; + __be32 host_ctx_id; +}; + /** * struct nfp_tun_active_tuns - periodic message of active tunnels * @seq: sequence number of the message @@ -124,11 +142,12 @@ enum nfp_flower_mac_offload_cmd { /** * struct nfp_tun_offloaded_mac - hashtable entry for an offloaded MAC - * @ht_node: Hashtable entry - * @addr: Offloaded MAC address - * @index: Offloaded index for given MAC address - * @ref_count: Number of devs using this MAC address - * @repr_list: List of reprs sharing this MAC address + * @ht_node: Hashtable entry + * @addr: Offloaded MAC address + * @index: Offloaded index for given MAC address + * @ref_count: Number of devs using this MAC address + * @repr_list: List of reprs sharing this MAC address + * @bridge_count: Number of bridge/internal devs with MAC */ struct nfp_tun_offloaded_mac { struct rhash_head ht_node; @@ -136,6 +155,7 @@ struct nfp_tun_offloaded_mac { u16 index; int ref_count; struct list_head repr_list; + int bridge_count; }; static const struct rhashtable_params offloaded_macs_params = { @@ -162,16 +182,16 @@ void nfp_tunnel_keep_alive(struct nfp_app *app, struct sk_buff *skb) } pay_len = nfp_flower_cmsg_get_data_len(skb); - if (pay_len != sizeof(struct nfp_tun_active_tuns) + - sizeof(struct route_ip_info) * count) { + if (pay_len != struct_size(payload, tun_info, count)) { nfp_flower_cmsg_warn(app, "Corruption in tunnel keep-alive message.\n"); return; } + rcu_read_lock(); for (i = 0; i < count; i++) { ipv4_addr = payload->tun_info[i].ipv4; port = be32_to_cpu(payload->tun_info[i].egress_port); - netdev = nfp_app_repr_get(app, port); + netdev = nfp_app_dev_get(app, port, NULL); if (!netdev) continue; @@ -183,6 +203,7 @@ void nfp_tunnel_keep_alive(struct nfp_app *app, struct sk_buff *skb) neigh_event_send(n, NULL); neigh_release(n); } + rcu_read_unlock(); } static int @@ -270,9 +291,10 @@ nfp_tun_write_neigh(struct net_device *netdev, struct nfp_app *app, struct flowi4 *flow, struct neighbour *neigh, gfp_t flag) { struct nfp_tun_neigh payload; + u32 port_id; - /* Only offload representor IPv4s for now. */ - if (!nfp_netdev_is_nfp_repr(netdev)) + port_id = nfp_flower_get_port_id_from_netdev(app, netdev); + if (!port_id) return; memset(&payload, 0, sizeof(struct nfp_tun_neigh)); @@ -290,7 +312,7 @@ nfp_tun_write_neigh(struct net_device *netdev, struct nfp_app *app, payload.src_ipv4 = flow->saddr; ether_addr_copy(payload.src_addr, netdev->dev_addr); neigh_ha_snapshot(payload.dst_addr, neigh, netdev); - payload.port_id = cpu_to_be32(nfp_repr_get_port_id(netdev)); + payload.port_id = cpu_to_be32(port_id); /* Add destination of new route to NFP cache. */ nfp_tun_add_route_to_cache(app, payload.dst_ipv4); @@ -326,13 +348,13 @@ nfp_tun_neigh_event_handler(struct notifier_block *nb, unsigned long event, flow.daddr = *(__be32 *)n->primary_key; - /* Only concerned with route changes for representors. */ - if (!nfp_netdev_is_nfp_repr(n->dev)) - return NOTIFY_DONE; - app_priv = container_of(nb, struct nfp_flower_priv, tun.neigh_nb); app = app_priv->app; + if (!nfp_netdev_is_nfp_repr(n->dev) && + !nfp_flower_internal_port_can_offload(app, n->dev)) + return NOTIFY_DONE; + /* Only concerned with changes to routes already added to NFP. */ if (!nfp_tun_has_route(app, flow.daddr)) return NOTIFY_DONE; @@ -366,9 +388,10 @@ void nfp_tunnel_request_route(struct nfp_app *app, struct sk_buff *skb) payload = nfp_flower_cmsg_get_data(skb); - netdev = nfp_app_repr_get(app, be32_to_cpu(payload->ingress_port)); + rcu_read_lock(); + netdev = nfp_app_dev_get(app, be32_to_cpu(payload->ingress_port), NULL); if (!netdev) - goto route_fail_warning; + goto fail_rcu_unlock; flow.daddr = payload->ipv4_addr; flow.flowi4_proto = IPPROTO_UDP; @@ -378,21 +401,23 @@ void nfp_tunnel_request_route(struct nfp_app *app, struct sk_buff *skb) rt = ip_route_output_key(dev_net(netdev), &flow); err = PTR_ERR_OR_ZERO(rt); if (err) - goto route_fail_warning; + goto fail_rcu_unlock; #else - goto route_fail_warning; + goto fail_rcu_unlock; #endif /* Get the neighbour entry for the lookup */ n = dst_neigh_lookup(&rt->dst, &flow.daddr); ip_rt_put(rt); if (!n) - goto route_fail_warning; - nfp_tun_write_neigh(n->dev, app, &flow, n, GFP_KERNEL); + goto fail_rcu_unlock; + nfp_tun_write_neigh(n->dev, app, &flow, n, GFP_ATOMIC); neigh_release(n); + rcu_read_unlock(); return; -route_fail_warning: +fail_rcu_unlock: + rcu_read_unlock(); nfp_flower_cmsg_warn(app, "Requested route not found.\n"); } @@ -551,6 +576,8 @@ nfp_tunnel_offloaded_macs_inc_ref_and_link(struct nfp_tun_offloaded_mac *entry, list_del(&repr_priv->mac_list); list_add_tail(&repr_priv->mac_list, &entry->repr_list); + } else if (nfp_flower_is_supported_bridge(netdev)) { + entry->bridge_count++; } entry->ref_count++; @@ -567,20 +594,35 @@ nfp_tunnel_add_shared_mac(struct nfp_app *app, struct net_device *netdev, entry = nfp_tunnel_lookup_offloaded_macs(app, netdev->dev_addr); if (entry && nfp_tunnel_is_mac_idx_global(entry->index)) { - nfp_tunnel_offloaded_macs_inc_ref_and_link(entry, netdev, mod); - return 0; + if (entry->bridge_count || + !nfp_flower_is_supported_bridge(netdev)) { + nfp_tunnel_offloaded_macs_inc_ref_and_link(entry, + netdev, mod); + return 0; + } + + /* MAC is global but matches need to go to pre_tun table. */ + nfp_mac_idx = entry->index | NFP_TUN_PRE_TUN_IDX_BIT; } - /* Assign a global index if non-repr or MAC address is now shared. */ - if (entry || !port) { - ida_idx = ida_simple_get(&priv->tun.mac_off_ids, 0, - NFP_MAX_MAC_INDEX, GFP_KERNEL); - if (ida_idx < 0) - return ida_idx; + if (!nfp_mac_idx) { + /* Assign a global index if non-repr or MAC is now shared. */ + if (entry || !port) { + ida_idx = ida_simple_get(&priv->tun.mac_off_ids, 0, + NFP_MAX_MAC_INDEX, GFP_KERNEL); + if (ida_idx < 0) + return ida_idx; - nfp_mac_idx = nfp_tunnel_get_global_mac_idx_from_ida(ida_idx); - } else { - nfp_mac_idx = nfp_tunnel_get_mac_idx_from_phy_port_id(port); + nfp_mac_idx = + nfp_tunnel_get_global_mac_idx_from_ida(ida_idx); + + if (nfp_flower_is_supported_bridge(netdev)) + nfp_mac_idx |= NFP_TUN_PRE_TUN_IDX_BIT; + + } else { + nfp_mac_idx = + nfp_tunnel_get_mac_idx_from_phy_port_id(port); + } } if (!entry) { @@ -649,6 +691,25 @@ nfp_tunnel_del_shared_mac(struct nfp_app *app, struct net_device *netdev, list_del(&repr_priv->mac_list); } + if (nfp_flower_is_supported_bridge(netdev)) { + entry->bridge_count--; + + if (!entry->bridge_count && entry->ref_count) { + u16 nfp_mac_idx; + + nfp_mac_idx = entry->index & ~NFP_TUN_PRE_TUN_IDX_BIT; + if (__nfp_tunnel_offload_mac(app, mac, nfp_mac_idx, + false)) { + nfp_flower_cmsg_warn(app, "MAC offload index revert failed on %s.\n", + netdev_name(netdev)); + return 0; + } + + entry->index = nfp_mac_idx; + return 0; + } + } + /* If MAC is now used by 1 repr set the offloaded MAC index to port. */ if (entry->ref_count == 1 && list_is_singular(&entry->repr_list)) { u16 nfp_mac_idx; @@ -708,6 +769,9 @@ nfp_tunnel_offload_mac(struct nfp_app *app, struct net_device *netdev, return 0; repr_priv = repr->app_priv; + if (repr_priv->on_bridge) + return 0; + mac_offloaded = &repr_priv->mac_offloaded; off_mac = &repr_priv->offloaded_mac_addr[0]; port = nfp_repr_get_port_id(netdev); @@ -823,10 +887,119 @@ int nfp_tunnel_mac_event_handler(struct nfp_app *app, if (err) nfp_flower_cmsg_warn(app, "Failed to offload MAC change on %s.\n", netdev_name(netdev)); + } else if (event == NETDEV_CHANGEUPPER) { + /* If a repr is attached to a bridge then tunnel packets + * entering the physical port are directed through the bridge + * datapath and cannot be directly detunneled. Therefore, + * associated offloaded MACs and indexes should not be used + * by fw for detunneling. + */ + struct netdev_notifier_changeupper_info *info = ptr; + struct net_device *upper = info->upper_dev; + struct nfp_flower_repr_priv *repr_priv; + struct nfp_repr *repr; + + if (!nfp_netdev_is_nfp_repr(netdev) || + !nfp_flower_is_supported_bridge(upper)) + return NOTIFY_OK; + + repr = netdev_priv(netdev); + if (repr->app != app) + return NOTIFY_OK; + + repr_priv = repr->app_priv; + + if (info->linking) { + if (nfp_tunnel_offload_mac(app, netdev, + NFP_TUNNEL_MAC_OFFLOAD_DEL)) + nfp_flower_cmsg_warn(app, "Failed to delete offloaded MAC on %s.\n", + netdev_name(netdev)); + repr_priv->on_bridge = true; + } else { + repr_priv->on_bridge = false; + + if (!(netdev->flags & IFF_UP)) + return NOTIFY_OK; + + if (nfp_tunnel_offload_mac(app, netdev, + NFP_TUNNEL_MAC_OFFLOAD_ADD)) + nfp_flower_cmsg_warn(app, "Failed to offload MAC on %s.\n", + netdev_name(netdev)); + } } return NOTIFY_OK; } +int nfp_flower_xmit_pre_tun_flow(struct nfp_app *app, + struct nfp_fl_payload *flow) +{ + struct nfp_flower_priv *app_priv = app->priv; + struct nfp_tun_offloaded_mac *mac_entry; + struct nfp_tun_pre_tun_rule payload; + struct net_device *internal_dev; + int err; + + if (app_priv->pre_tun_rule_cnt == NFP_TUN_PRE_TUN_RULE_LIMIT) + return -ENOSPC; + + memset(&payload, 0, sizeof(struct nfp_tun_pre_tun_rule)); + + internal_dev = flow->pre_tun_rule.dev; + payload.vlan_tci = flow->pre_tun_rule.vlan_tci; + payload.host_ctx_id = flow->meta.host_ctx_id; + + /* Lookup MAC index for the pre-tunnel rule egress device. + * Note that because the device is always an internal port, it will + * have a constant global index so does not need to be tracked. + */ + mac_entry = nfp_tunnel_lookup_offloaded_macs(app, + internal_dev->dev_addr); + if (!mac_entry) + return -ENOENT; + + payload.port_idx = cpu_to_be16(mac_entry->index); + + /* Copy mac id and vlan to flow - dev may not exist at delete time. */ + flow->pre_tun_rule.vlan_tci = payload.vlan_tci; + flow->pre_tun_rule.port_idx = payload.port_idx; + + err = nfp_flower_xmit_tun_conf(app, NFP_FLOWER_CMSG_TYPE_PRE_TUN_RULE, + sizeof(struct nfp_tun_pre_tun_rule), + (unsigned char *)&payload, GFP_KERNEL); + if (err) + return err; + + app_priv->pre_tun_rule_cnt++; + + return 0; +} + +int nfp_flower_xmit_pre_tun_del_flow(struct nfp_app *app, + struct nfp_fl_payload *flow) +{ + struct nfp_flower_priv *app_priv = app->priv; + struct nfp_tun_pre_tun_rule payload; + u32 tmp_flags = 0; + int err; + + memset(&payload, 0, sizeof(struct nfp_tun_pre_tun_rule)); + + tmp_flags |= NFP_TUN_PRE_TUN_RULE_DEL; + payload.flags = cpu_to_be32(tmp_flags); + payload.vlan_tci = flow->pre_tun_rule.vlan_tci; + payload.port_idx = flow->pre_tun_rule.port_idx; + + err = nfp_flower_xmit_tun_conf(app, NFP_FLOWER_CMSG_TYPE_PRE_TUN_RULE, + sizeof(struct nfp_tun_pre_tun_rule), + (unsigned char *)&payload, GFP_KERNEL); + if (err) + return err; + + app_priv->pre_tun_rule_cnt--; + + return 0; +} + int nfp_tunnel_config_start(struct nfp_app *app) { struct nfp_flower_priv *priv = app->priv; diff --git a/drivers/net/ethernet/netronome/nfp/nfp_app.h b/drivers/net/ethernet/netronome/nfp/nfp_app.h index f8d422713705..76d13af46a7a 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_app.h +++ b/drivers/net/ethernet/netronome/nfp/nfp_app.h @@ -79,7 +79,7 @@ extern const struct nfp_app_type app_abm; * @eswitch_mode_set: set SR-IOV eswitch mode (under pf->lock) * @sriov_enable: app-specific sriov initialisation * @sriov_disable: app-specific sriov clean-up - * @repr_get: get representor netdev + * @dev_get: get representor or internal port representing netdev */ struct nfp_app_type { enum nfp_app_id id; @@ -143,7 +143,8 @@ struct nfp_app_type { enum devlink_eswitch_mode (*eswitch_mode_get)(struct nfp_app *app); int (*eswitch_mode_set)(struct nfp_app *app, u16 mode); - struct net_device *(*repr_get)(struct nfp_app *app, u32 id); + struct net_device *(*dev_get)(struct nfp_app *app, u32 id, + bool *redir_egress); }; /** @@ -397,12 +398,14 @@ static inline void nfp_app_sriov_disable(struct nfp_app *app) app->type->sriov_disable(app); } -static inline struct net_device *nfp_app_repr_get(struct nfp_app *app, u32 id) +static inline +struct net_device *nfp_app_dev_get(struct nfp_app *app, u32 id, + bool *redir_egress) { - if (unlikely(!app || !app->type->repr_get)) + if (unlikely(!app || !app->type->dev_get)) return NULL; - return app->type->repr_get(app, id); + return app->type->dev_get(app, id, redir_egress); } struct nfp_app *nfp_app_from_netdev(struct net_device *netdev); @@ -433,6 +436,6 @@ int nfp_app_nic_vnic_alloc(struct nfp_app *app, struct nfp_net *nn, int nfp_app_nic_vnic_init_phy_port(struct nfp_pf *pf, struct nfp_app *app, struct nfp_net *nn, unsigned int id); -struct devlink *nfp_devlink_get_devlink(struct net_device *netdev); +struct devlink_port *nfp_devlink_get_devlink_port(struct net_device *netdev); #endif diff --git a/drivers/net/ethernet/netronome/nfp/nfp_devlink.c b/drivers/net/ethernet/netronome/nfp/nfp_devlink.c index e9eca99cf493..c50fce42f473 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_devlink.c +++ b/drivers/net/ethernet/netronome/nfp/nfp_devlink.c @@ -144,7 +144,8 @@ nfp_devlink_sb_pool_get(struct devlink *devlink, unsigned int sb_index, static int nfp_devlink_sb_pool_set(struct devlink *devlink, unsigned int sb_index, u16 pool_index, - u32 size, enum devlink_sb_threshold_type threshold_type) + u32 size, enum devlink_sb_threshold_type threshold_type, + struct netlink_ext_ack *extack) { struct nfp_pf *pf = devlink_priv(devlink); @@ -354,6 +355,8 @@ int nfp_devlink_port_register(struct nfp_app *app, struct nfp_port *port) { struct nfp_eth_table_port eth_port; struct devlink *devlink; + const u8 *serial; + int serial_len; int ret; rtnl_lock(); @@ -362,10 +365,10 @@ int nfp_devlink_port_register(struct nfp_app *app, struct nfp_port *port) if (ret) return ret; - devlink_port_type_eth_set(&port->dl_port, port->netdev); + serial_len = nfp_cpp_serial(port->app->cpp, &serial); devlink_port_attrs_set(&port->dl_port, DEVLINK_PORT_FLAVOUR_PHYSICAL, eth_port.label_port, eth_port.is_split, - eth_port.label_subport); + eth_port.label_subport, serial, serial_len); devlink = priv_to_devlink(app->pf); @@ -377,13 +380,23 @@ void nfp_devlink_port_unregister(struct nfp_port *port) devlink_port_unregister(&port->dl_port); } -struct devlink *nfp_devlink_get_devlink(struct net_device *netdev) +void nfp_devlink_port_type_eth_set(struct nfp_port *port) +{ + devlink_port_type_eth_set(&port->dl_port, port->netdev); +} + +void nfp_devlink_port_type_clear(struct nfp_port *port) { - struct nfp_app *app; + devlink_port_type_clear(&port->dl_port); +} + +struct devlink_port *nfp_devlink_get_devlink_port(struct net_device *netdev) +{ + struct nfp_port *port; - app = nfp_app_from_netdev(netdev); - if (!app) + port = nfp_port_from_netdev(netdev); + if (!port) return NULL; - return priv_to_devlink(app->pf); + return &port->dl_port; } diff --git a/drivers/net/ethernet/netronome/nfp/nfp_main.c b/drivers/net/ethernet/netronome/nfp/nfp_main.c index f4c8776e42b6..4d282fc56009 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_main.c +++ b/drivers/net/ethernet/netronome/nfp/nfp_main.c @@ -294,6 +294,9 @@ static int nfp_pcie_sriov_disable(struct pci_dev *pdev) static int nfp_pcie_sriov_configure(struct pci_dev *pdev, int num_vfs) { + if (!pci_get_drvdata(pdev)) + return -ENOENT; + if (num_vfs == 0) return nfp_pcie_sriov_disable(pdev); else @@ -349,7 +352,7 @@ nfp_net_fw_request(struct pci_dev *pdev, struct nfp_pf *pf, const char *name) err = request_firmware_direct(&fw, name, &pdev->dev); nfp_info(pf->cpp, " %s: %s\n", - name, err ? "not found" : "found, loading..."); + name, err ? "not found" : "found"); if (err) return NULL; @@ -427,8 +430,35 @@ nfp_net_fw_find(struct pci_dev *pdev, struct nfp_pf *pf) return nfp_net_fw_request(pdev, pf, fw_name); } +static int +nfp_get_fw_policy_value(struct pci_dev *pdev, struct nfp_nsp *nsp, + const char *key, const char *default_val, int max_val, + int *value) +{ + char hwinfo[64]; + long hi_val; + int err; + + snprintf(hwinfo, sizeof(hwinfo), key); + err = nfp_nsp_hwinfo_lookup_optional(nsp, hwinfo, sizeof(hwinfo), + default_val); + if (err) + return err; + + err = kstrtol(hwinfo, 0, &hi_val); + if (err || hi_val < 0 || hi_val > max_val) { + dev_warn(&pdev->dev, + "Invalid value '%s' from '%s', ignoring\n", + hwinfo, key); + err = kstrtol(default_val, 0, &hi_val); + } + + *value = hi_val; + return err; +} + /** - * nfp_net_fw_load() - Load the firmware image + * nfp_fw_load() - Load the firmware image * @pdev: PCI Device structure * @pf: NFP PF Device structure * @nsp: NFP SP handle @@ -438,44 +468,107 @@ nfp_net_fw_find(struct pci_dev *pdev, struct nfp_pf *pf) static int nfp_fw_load(struct pci_dev *pdev, struct nfp_pf *pf, struct nfp_nsp *nsp) { - const struct firmware *fw; + bool do_reset, fw_loaded = false; + const struct firmware *fw = NULL; + int err, reset, policy, ifcs = 0; + char *token, *ptr; + char hwinfo[64]; u16 interface; - int err; + + snprintf(hwinfo, sizeof(hwinfo), "abi_drv_load_ifc"); + err = nfp_nsp_hwinfo_lookup_optional(nsp, hwinfo, sizeof(hwinfo), + NFP_NSP_DRV_LOAD_IFC_DEFAULT); + if (err) + return err; interface = nfp_cpp_interface(pf->cpp); - if (NFP_CPP_INTERFACE_UNIT_of(interface) != 0) { - /* Only Unit 0 should reset or load firmware */ + ptr = hwinfo; + while ((token = strsep(&ptr, ","))) { + unsigned long interface_hi; + + err = kstrtoul(token, 0, &interface_hi); + if (err) { + dev_err(&pdev->dev, + "Failed to parse interface '%s': %d\n", + token, err); + return err; + } + + ifcs++; + if (interface == interface_hi) + break; + } + + if (!token) { dev_info(&pdev->dev, "Firmware will be loaded by partner\n"); return 0; } + err = nfp_get_fw_policy_value(pdev, nsp, "abi_drv_reset", + NFP_NSP_DRV_RESET_DEFAULT, + NFP_NSP_DRV_RESET_NEVER, &reset); + if (err) + return err; + + err = nfp_get_fw_policy_value(pdev, nsp, "app_fw_from_flash", + NFP_NSP_APP_FW_LOAD_DEFAULT, + NFP_NSP_APP_FW_LOAD_PREF, &policy); + if (err) + return err; + fw = nfp_net_fw_find(pdev, pf); - if (!fw) { - if (nfp_nsp_has_stored_fw_load(nsp)) - nfp_nsp_load_stored_fw(nsp); - return 0; + do_reset = reset == NFP_NSP_DRV_RESET_ALWAYS || + (fw && reset == NFP_NSP_DRV_RESET_DISK); + + if (do_reset) { + dev_info(&pdev->dev, "Soft-resetting the NFP\n"); + err = nfp_nsp_device_soft_reset(nsp); + if (err < 0) { + dev_err(&pdev->dev, + "Failed to soft reset the NFP: %d\n", err); + goto exit_release_fw; + } } - dev_info(&pdev->dev, "Soft-reset, loading FW image\n"); - err = nfp_nsp_device_soft_reset(nsp); - if (err < 0) { - dev_err(&pdev->dev, "Failed to soft reset the NFP: %d\n", - err); - goto exit_release_fw; - } + if (fw && policy != NFP_NSP_APP_FW_LOAD_FLASH) { + if (nfp_nsp_has_fw_loaded(nsp) && nfp_nsp_fw_loaded(nsp)) + goto exit_release_fw; - err = nfp_nsp_load_fw(nsp, fw); - if (err < 0) { - dev_err(&pdev->dev, "FW loading failed: %d\n", err); - goto exit_release_fw; + err = nfp_nsp_load_fw(nsp, fw); + if (err < 0) { + dev_err(&pdev->dev, "FW loading failed: %d\n", + err); + goto exit_release_fw; + } + dev_info(&pdev->dev, "Finished loading FW image\n"); + fw_loaded = true; + } else if (policy != NFP_NSP_APP_FW_LOAD_DISK && + nfp_nsp_has_stored_fw_load(nsp)) { + + /* Don't propagate this error to stick with legacy driver + * behavior, failure will be detected later during init. + */ + if (!nfp_nsp_load_stored_fw(nsp)) + dev_info(&pdev->dev, "Finished loading stored FW image\n"); + + /* Don't flag the fw_loaded in this case since other devices + * may reuse the firmware when configured this way + */ + } else { + dev_warn(&pdev->dev, "Didn't load firmware, please update flash or reconfigure card\n"); } - dev_info(&pdev->dev, "Finished loading FW image\n"); - exit_release_fw: release_firmware(fw); - return err < 0 ? err : 1; + /* We don't want to unload firmware when other devices may still be + * dependent on it, which could be the case if there are multiple + * devices that could load firmware. + */ + if (fw_loaded && ifcs == 1) + pf->unload_fw_on_remove = true; + + return err < 0 ? err : fw_loaded; } static void @@ -593,6 +686,10 @@ static int nfp_pci_probe(struct pci_dev *pdev, struct nfp_pf *pf; int err; + if (pdev->vendor == PCI_VENDOR_ID_NETRONOME && + pdev->device == PCI_DEVICE_ID_NETRONOME_NFP6000_VF) + dev_warn(&pdev->dev, "Binding NFP VF device to the NFP PF driver, the VF driver is called 'nfp_netvf'\n"); + err = pci_enable_device(pdev); if (err < 0) return err; @@ -697,7 +794,7 @@ err_net_remove: err_fw_unload: kfree(pf->rtbl); nfp_mip_close(pf->mip); - if (pf->fw_loaded) + if (pf->unload_fw_on_remove) nfp_fw_unload(pf); kfree(pf->eth_tbl); kfree(pf->nspi); @@ -720,9 +817,13 @@ err_pci_disable: return err; } -static void nfp_pci_remove(struct pci_dev *pdev) +static void __nfp_pci_shutdown(struct pci_dev *pdev, bool unload_fw) { - struct nfp_pf *pf = pci_get_drvdata(pdev); + struct nfp_pf *pf; + + pf = pci_get_drvdata(pdev); + if (!pf) + return; nfp_hwmon_unregister(pf); @@ -733,7 +834,7 @@ static void nfp_pci_remove(struct pci_dev *pdev) vfree(pf->dumpspec); kfree(pf->rtbl); nfp_mip_close(pf->mip); - if (pf->fw_loaded) + if (unload_fw && pf->unload_fw_on_remove) nfp_fw_unload(pf); destroy_workqueue(pf->wq); @@ -749,11 +850,22 @@ static void nfp_pci_remove(struct pci_dev *pdev) pci_disable_device(pdev); } +static void nfp_pci_remove(struct pci_dev *pdev) +{ + __nfp_pci_shutdown(pdev, true); +} + +static void nfp_pci_shutdown(struct pci_dev *pdev) +{ + __nfp_pci_shutdown(pdev, false); +} + static struct pci_driver nfp_pci_driver = { .name = nfp_driver_name, .id_table = nfp_pci_device_ids, .probe = nfp_pci_probe, .remove = nfp_pci_remove, + .shutdown = nfp_pci_shutdown, .sriov_configure = nfp_pcie_sriov_configure, }; @@ -793,6 +905,8 @@ static void __exit nfp_main_exit(void) module_init(nfp_main_init); module_exit(nfp_main_exit); +MODULE_FIRMWARE("netronome/nic_AMDA0058-0011_2x40.nffw"); +MODULE_FIRMWARE("netronome/nic_AMDA0058-0012_2x40.nffw"); MODULE_FIRMWARE("netronome/nic_AMDA0081-0001_1x40.nffw"); MODULE_FIRMWARE("netronome/nic_AMDA0081-0001_4x10.nffw"); MODULE_FIRMWARE("netronome/nic_AMDA0096-0001_2x10.nffw"); diff --git a/drivers/net/ethernet/netronome/nfp/nfp_main.h b/drivers/net/ethernet/netronome/nfp/nfp_main.h index b7211f200d22..5d5812fd9317 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_main.h +++ b/drivers/net/ethernet/netronome/nfp/nfp_main.h @@ -64,6 +64,7 @@ struct nfp_dumpspec { * @limit_vfs: Number of VFs supported by firmware (~0 for PCI limit) * @num_vfs: Number of SR-IOV VFs enabled * @fw_loaded: Is the firmware loaded? + * @unload_fw_on_remove:Do we need to unload firmware on driver removal? * @ctrl_vnic: Pointer to the control vNIC if available * @mip: MIP handle * @rtbl: RTsym table @@ -110,6 +111,7 @@ struct nfp_pf { unsigned int num_vfs; bool fw_loaded; + bool unload_fw_on_remove; struct nfp_net *ctrl_vnic; @@ -185,4 +187,7 @@ int nfp_shared_buf_pool_get(struct nfp_pf *pf, unsigned int sb, u16 pool_index, int nfp_shared_buf_pool_set(struct nfp_pf *pf, unsigned int sb, u16 pool_index, u32 size, enum devlink_sb_threshold_type threshold_type); + +int nfp_devlink_params_register(struct nfp_pf *pf); +void nfp_devlink_params_unregister(struct nfp_pf *pf); #endif /* NFP_MAIN_H */ diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net.h b/drivers/net/ethernet/netronome/nfp/nfp_net.h index be37c2d6151c..250f510b1d21 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_net.h +++ b/drivers/net/ethernet/netronome/nfp/nfp_net.h @@ -12,11 +12,14 @@ #ifndef _NFP_NET_H_ #define _NFP_NET_H_ +#include <linux/atomic.h> #include <linux/interrupt.h> #include <linux/list.h> #include <linux/netdevice.h> #include <linux/pci.h> #include <linux/io-64-nonatomic-hi-lo.h> +#include <linux/semaphore.h> +#include <linux/workqueue.h> #include <net/xdp.h> #include "nfp_net_ctrl.h" @@ -63,7 +66,7 @@ #define NFP_NET_MAX_DMA_BITS 40 /* Default size for MTU and freelist buffer sizes */ -#define NFP_NET_DEFAULT_MTU 1500 +#define NFP_NET_DEFAULT_MTU 1500U /* Maximum number of bytes prepended to a packet */ #define NFP_NET_MAX_PREPEND 64 @@ -238,7 +241,7 @@ struct nfp_net_tx_ring { #define PCIE_DESC_RX_I_TCP_CSUM_OK cpu_to_le16(BIT(11)) #define PCIE_DESC_RX_I_UDP_CSUM cpu_to_le16(BIT(10)) #define PCIE_DESC_RX_I_UDP_CSUM_OK cpu_to_le16(BIT(9)) -#define PCIE_DESC_RX_BPF cpu_to_le16(BIT(8)) +#define PCIE_DESC_RX_DECRYPTED cpu_to_le16(BIT(8)) #define PCIE_DESC_RX_EOP cpu_to_le16(BIT(7)) #define PCIE_DESC_RX_IP4_CSUM cpu_to_le16(BIT(6)) #define PCIE_DESC_RX_IP4_CSUM_OK cpu_to_le16(BIT(5)) @@ -365,6 +368,7 @@ struct nfp_net_rx_ring { * @hw_csum_rx_inner_ok: Counter of packets where the inner HW checksum was OK * @hw_csum_rx_complete: Counter of packets with CHECKSUM_COMPLETE reported * @hw_csum_rx_error: Counter of packets with bad checksums + * @hw_tls_rx: Number of packets with TLS decrypted by hardware * @tx_sync: Seqlock for atomic updates of TX stats * @tx_pkts: Number of Transmitted packets * @tx_bytes: Number of Transmitted bytes @@ -372,6 +376,11 @@ struct nfp_net_rx_ring { * @hw_csum_tx_inner: Counter of inner TX checksum offload requests * @tx_gather: Counter of packets with Gather DMA * @tx_lso: Counter of LSO packets sent + * @hw_tls_tx: Counter of TLS packets sent with crypto offloaded to HW + * @tls_tx_fallback: Counter of TLS packets sent which had to be encrypted + * by the fallback path because packets came out of order + * @tls_tx_no_fallback: Counter of TLS packets not sent because the fallback + * path could not encrypt them * @tx_errors: How many TX errors were encountered * @tx_busy: How often was TX busy (no space)? * @rx_replace_buf_alloc_fail: Counter of RX buffer allocation failures @@ -392,7 +401,7 @@ struct nfp_net_r_vector { struct { struct tasklet_struct tasklet; struct sk_buff_head queue; - struct spinlock lock; + spinlock_t lock; }; }; @@ -408,22 +417,30 @@ struct nfp_net_r_vector { u64 hw_csum_rx_ok; u64 hw_csum_rx_inner_ok; u64 hw_csum_rx_complete; + u64 hw_tls_rx; + + u64 hw_csum_rx_error; + u64 rx_replace_buf_alloc_fail; struct nfp_net_tx_ring *xdp_ring; struct u64_stats_sync tx_sync; u64 tx_pkts; u64 tx_bytes; - u64 hw_csum_tx; + + u64 ____cacheline_aligned_in_smp hw_csum_tx; u64 hw_csum_tx_inner; u64 tx_gather; u64 tx_lso; + u64 hw_tls_tx; - u64 hw_csum_rx_error; - u64 rx_replace_buf_alloc_fail; + u64 tls_tx_fallback; + u64 tls_tx_no_fallback; u64 tx_errors; u64 tx_busy; + /* Cold data follows */ + u32 irq_vector; irq_handler_t handler; char name[IFNAMSIZ + 8]; @@ -458,6 +475,7 @@ struct nfp_stat_pair { * @netdev: Backpointer to net_device structure * @is_vf: Is the driver attached to a VF? * @chained_metadata_format: Firemware will use new metadata format + * @ktls_tx: Is kTLS TX enabled? * @rx_dma_dir: Mapping direction for RX buffers * @rx_dma_off: Offset at which DMA packets (for XDP headroom) * @rx_offset: Offset in the RX buffers where packet data starts @@ -482,6 +500,7 @@ struct nfp_net_dp { u8 is_vf:1; u8 chained_metadata_format:1; + u8 ktls_tx:1; u8 rx_dma_dir; u8 rx_offset; @@ -539,12 +558,17 @@ struct nfp_net_dp { * @shared_handler: Handler for shared interrupts * @shared_name: Name for shared interrupt * @me_freq_mhz: ME clock_freq (MHz) - * @reconfig_lock: Protects HW reconfiguration request regs/machinery + * @reconfig_lock: Protects @reconfig_posted, @reconfig_timer_active, + * @reconfig_sync_present and HW reconfiguration request + * regs/machinery from async requests (sync must take + * @bar_lock) * @reconfig_posted: Pending reconfig bits coming from async sources * @reconfig_timer_active: Timer for reading reconfiguration results is pending * @reconfig_sync_present: Some thread is performing synchronous reconfig * @reconfig_timer: Timer for async reading of reconfig results * @reconfig_in_progress_update: Update FW is processing now (debug only) + * @bar_lock: vNIC config BAR access lock, protects: update, + * mailbox area, crypto TLV * @link_up: Is the link up? * @link_status_lock: Protects @link_* and ensures atomicity with BAR reading * @rx_coalesce_usecs: RX interrupt moderation usecs delay parameter @@ -557,6 +581,18 @@ struct nfp_net_dp { * @tx_bar: Pointer to mapped TX queues * @rx_bar: Pointer to mapped FL/RX queues * @tlv_caps: Parsed TLV capabilities + * @ktls_tx_conn_cnt: Number of offloaded kTLS TX connections + * @ktls_rx_conn_cnt: Number of offloaded kTLS RX connections + * @ktls_conn_id_gen: Trivial generator for kTLS connection ids (for TX) + * @ktls_no_space: Counter of firmware rejecting kTLS connection due to + * lack of space + * @mbox_cmsg: Common Control Message via vNIC mailbox state + * @mbox_cmsg.queue: CCM mbox queue of pending messages + * @mbox_cmsg.wq: CCM mbox wait queue of waiting processes + * @mbox_cmsg.workq: CCM mbox work queue for @wait_work and @runq_work + * @mbox_cmsg.wait_work: CCM mbox posted msg reconfig wait work + * @mbox_cmsg.runq_work: CCM mbox posted msg queue runner work + * @mbox_cmsg.tag: CCM mbox message tag allocator * @debugfs_dir: Device directory in debugfs * @vnic_list: Entry on device vNIC list * @pdev: Backpointer to PCI device @@ -615,6 +651,8 @@ struct nfp_net { struct timer_list reconfig_timer; u32 reconfig_in_progress_update; + struct semaphore bar_lock; + u32 rx_coalesce_usecs; u32 rx_coalesce_max_frames; u32 tx_coalesce_usecs; @@ -630,6 +668,22 @@ struct nfp_net { struct nfp_net_tlv_caps tlv_caps; + unsigned int ktls_tx_conn_cnt; + unsigned int ktls_rx_conn_cnt; + + atomic64_t ktls_conn_id_gen; + + atomic_t ktls_no_space; + + struct { + struct sk_buff_head queue; + wait_queue_head_t wq; + struct workqueue_struct *workq; + struct work_struct wait_work; + struct work_struct runq_work; + u16 tag; + } mbox_cmsg; + struct dentry *debugfs_dir; struct list_head vnic_list; @@ -839,6 +893,21 @@ static inline void nfp_ctrl_unlock(struct nfp_net *nn) spin_unlock_bh(&nn->r_vecs[0].lock); } +static inline void nn_ctrl_bar_lock(struct nfp_net *nn) +{ + down(&nn->bar_lock); +} + +static inline bool nn_ctrl_bar_trylock(struct nfp_net *nn) +{ + return !down_trylock(&nn->bar_lock); +} + +static inline void nn_ctrl_bar_unlock(struct nfp_net *nn) +{ + up(&nn->bar_lock); +} + /* Globals */ extern const char nfp_driver_version[]; @@ -866,12 +935,17 @@ void nfp_ctrl_close(struct nfp_net *nn); void nfp_net_set_ethtool_ops(struct net_device *netdev); void nfp_net_info(struct nfp_net *nn); +int __nfp_net_reconfig(struct nfp_net *nn, u32 update); int nfp_net_reconfig(struct nfp_net *nn, u32 update); unsigned int nfp_net_rss_key_sz(struct nfp_net *nn); void nfp_net_rss_write_itbl(struct nfp_net *nn); void nfp_net_rss_write_key(struct nfp_net *nn); void nfp_net_coalesce_write_cfg(struct nfp_net *nn); -int nfp_net_reconfig_mbox(struct nfp_net *nn, u32 mbox_cmd); +int nfp_net_mbox_lock(struct nfp_net *nn, unsigned int data_size); +int nfp_net_mbox_reconfig(struct nfp_net *nn, u32 mbox_cmd); +int nfp_net_mbox_reconfig_and_unlock(struct nfp_net *nn, u32 mbox_cmd); +void nfp_net_mbox_reconfig_post(struct nfp_net *nn, u32 update); +int nfp_net_mbox_reconfig_wait_posted(struct nfp_net *nn); unsigned int nfp_net_irqs_alloc(struct pci_dev *pdev, struct msix_entry *irq_entries, diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c index 6d1b8816552e..bcdcd6de7dea 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c +++ b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c @@ -36,14 +36,17 @@ #include <linux/vmalloc.h> #include <linux/ktime.h> +#include <net/tls.h> #include <net/vxlan.h> #include "nfpcore/nfp_nsp.h" +#include "ccm.h" #include "nfp_app.h" #include "nfp_net_ctrl.h" #include "nfp_net.h" #include "nfp_net_sriov.h" #include "nfp_port.h" +#include "crypto/crypto.h" /** * nfp_net_get_fw_version() - Read and parse the FW version @@ -137,20 +140,37 @@ static bool nfp_net_reconfig_check_done(struct nfp_net *nn, bool last_check) return false; } -static int nfp_net_reconfig_wait(struct nfp_net *nn, unsigned long deadline) +static bool __nfp_net_reconfig_wait(struct nfp_net *nn, unsigned long deadline) { bool timed_out = false; + int i; + + /* Poll update field, waiting for NFP to ack the config. + * Do an opportunistic wait-busy loop, afterward sleep. + */ + for (i = 0; i < 50; i++) { + if (nfp_net_reconfig_check_done(nn, false)) + return false; + udelay(4); + } - /* Poll update field, waiting for NFP to ack the config */ while (!nfp_net_reconfig_check_done(nn, timed_out)) { - msleep(1); + usleep_range(250, 500); timed_out = time_is_before_eq_jiffies(deadline); } + return timed_out; +} + +static int nfp_net_reconfig_wait(struct nfp_net *nn, unsigned long deadline) +{ + if (__nfp_net_reconfig_wait(nn, deadline)) + return -EIO; + if (nn_readl(nn, NFP_NET_CFG_UPDATE) & NFP_NET_CFG_UPDATE_ERR) return -EIO; - return timed_out ? -EIO : 0; + return 0; } static void nfp_net_reconfig_timer(struct timer_list *t) @@ -210,6 +230,7 @@ static void nfp_net_reconfig_sync_enter(struct nfp_net *nn) spin_lock_bh(&nn->reconfig_lock); + WARN_ON(nn->reconfig_sync_present); nn->reconfig_sync_present = true; if (nn->reconfig_timer_active) { @@ -243,7 +264,7 @@ static void nfp_net_reconfig_wait_posted(struct nfp_net *nn) } /** - * nfp_net_reconfig() - Reconfigure the firmware + * __nfp_net_reconfig() - Reconfigure the firmware * @nn: NFP Net device to reconfigure * @update: The value for the update field in the BAR config * @@ -253,7 +274,7 @@ static void nfp_net_reconfig_wait_posted(struct nfp_net *nn) * * Return: Negative errno on error, 0 on success */ -int nfp_net_reconfig(struct nfp_net *nn, u32 update) +int __nfp_net_reconfig(struct nfp_net *nn, u32 update) { int ret; @@ -274,8 +295,31 @@ int nfp_net_reconfig(struct nfp_net *nn, u32 update) return ret; } +int nfp_net_reconfig(struct nfp_net *nn, u32 update) +{ + int ret; + + nn_ctrl_bar_lock(nn); + ret = __nfp_net_reconfig(nn, update); + nn_ctrl_bar_unlock(nn); + + return ret; +} + +int nfp_net_mbox_lock(struct nfp_net *nn, unsigned int data_size) +{ + if (nn->tlv_caps.mbox_len < NFP_NET_CFG_MBOX_SIMPLE_VAL + data_size) { + nn_err(nn, "mailbox too small for %u of data (%u)\n", + data_size, nn->tlv_caps.mbox_len); + return -EIO; + } + + nn_ctrl_bar_lock(nn); + return 0; +} + /** - * nfp_net_reconfig_mbox() - Reconfigure the firmware via the mailbox + * nfp_net_mbox_reconfig() - Reconfigure the firmware via the mailbox * @nn: NFP Net device to reconfigure * @mbox_cmd: The value for the mailbox command * @@ -283,19 +327,14 @@ int nfp_net_reconfig(struct nfp_net *nn, u32 update) * * Return: Negative errno on error, 0 on success */ -int nfp_net_reconfig_mbox(struct nfp_net *nn, u32 mbox_cmd) +int nfp_net_mbox_reconfig(struct nfp_net *nn, u32 mbox_cmd) { u32 mbox = nn->tlv_caps.mbox_off; int ret; - if (!nfp_net_has_mbox(&nn->tlv_caps)) { - nn_err(nn, "no mailbox present, command: %u\n", mbox_cmd); - return -EIO; - } - nn_writeq(nn, mbox + NFP_NET_CFG_MBOX_SIMPLE_CMD, mbox_cmd); - ret = nfp_net_reconfig(nn, NFP_NET_CFG_UPDATE_MBOX); + ret = __nfp_net_reconfig(nn, NFP_NET_CFG_UPDATE_MBOX); if (ret) { nn_err(nn, "Mailbox update error\n"); return ret; @@ -304,6 +343,33 @@ int nfp_net_reconfig_mbox(struct nfp_net *nn, u32 mbox_cmd) return -nn_readl(nn, mbox + NFP_NET_CFG_MBOX_SIMPLE_RET); } +void nfp_net_mbox_reconfig_post(struct nfp_net *nn, u32 mbox_cmd) +{ + u32 mbox = nn->tlv_caps.mbox_off; + + nn_writeq(nn, mbox + NFP_NET_CFG_MBOX_SIMPLE_CMD, mbox_cmd); + + nfp_net_reconfig_post(nn, NFP_NET_CFG_UPDATE_MBOX); +} + +int nfp_net_mbox_reconfig_wait_posted(struct nfp_net *nn) +{ + u32 mbox = nn->tlv_caps.mbox_off; + + nfp_net_reconfig_wait_posted(nn); + + return -nn_readl(nn, mbox + NFP_NET_CFG_MBOX_SIMPLE_RET); +} + +int nfp_net_mbox_reconfig_and_unlock(struct nfp_net *nn, u32 mbox_cmd) +{ + int ret; + + ret = nfp_net_mbox_reconfig(nn, mbox_cmd); + nn_ctrl_bar_unlock(nn); + return ret; +} + /* Interrupt configuration and handling */ @@ -756,6 +822,100 @@ static void nfp_net_tx_csum(struct nfp_net_dp *dp, u64_stats_update_end(&r_vec->tx_sync); } +static struct sk_buff * +nfp_net_tls_tx(struct nfp_net_dp *dp, struct nfp_net_r_vector *r_vec, + struct sk_buff *skb, u64 *tls_handle, int *nr_frags) +{ +#ifdef CONFIG_TLS_DEVICE + struct nfp_net_tls_offload_ctx *ntls; + struct sk_buff *nskb; + bool resync_pending; + u32 datalen, seq; + + if (likely(!dp->ktls_tx)) + return skb; + if (!skb->sk || !tls_is_sk_tx_device_offloaded(skb->sk)) + return skb; + + datalen = skb->len - (skb_transport_offset(skb) + tcp_hdrlen(skb)); + seq = ntohl(tcp_hdr(skb)->seq); + ntls = tls_driver_ctx(skb->sk, TLS_OFFLOAD_CTX_DIR_TX); + resync_pending = tls_offload_tx_resync_pending(skb->sk); + if (unlikely(resync_pending || ntls->next_seq != seq)) { + /* Pure ACK out of order already */ + if (!datalen) + return skb; + + u64_stats_update_begin(&r_vec->tx_sync); + r_vec->tls_tx_fallback++; + u64_stats_update_end(&r_vec->tx_sync); + + nskb = tls_encrypt_skb(skb); + if (!nskb) { + u64_stats_update_begin(&r_vec->tx_sync); + r_vec->tls_tx_no_fallback++; + u64_stats_update_end(&r_vec->tx_sync); + return NULL; + } + /* encryption wasn't necessary */ + if (nskb == skb) + return skb; + /* we don't re-check ring space */ + if (unlikely(skb_is_nonlinear(nskb))) { + nn_dp_warn(dp, "tls_encrypt_skb() produced fragmented frame\n"); + u64_stats_update_begin(&r_vec->tx_sync); + r_vec->tx_errors++; + u64_stats_update_end(&r_vec->tx_sync); + dev_kfree_skb_any(nskb); + return NULL; + } + + /* jump forward, a TX may have gotten lost, need to sync TX */ + if (!resync_pending && seq - ntls->next_seq < U32_MAX / 4) + tls_offload_tx_resync_request(nskb->sk, seq, + ntls->next_seq); + + *nr_frags = 0; + return nskb; + } + + if (datalen) { + u64_stats_update_begin(&r_vec->tx_sync); + if (!skb_is_gso(skb)) + r_vec->hw_tls_tx++; + else + r_vec->hw_tls_tx += skb_shinfo(skb)->gso_segs; + u64_stats_update_end(&r_vec->tx_sync); + } + + memcpy(tls_handle, ntls->fw_handle, sizeof(ntls->fw_handle)); + ntls->next_seq += datalen; +#endif + return skb; +} + +static void nfp_net_tls_tx_undo(struct sk_buff *skb, u64 tls_handle) +{ +#ifdef CONFIG_TLS_DEVICE + struct nfp_net_tls_offload_ctx *ntls; + u32 datalen, seq; + + if (!tls_handle) + return; + if (WARN_ON_ONCE(!skb->sk || !tls_is_sk_tx_device_offloaded(skb->sk))) + return; + + datalen = skb->len - (skb_transport_offset(skb) + tcp_hdrlen(skb)); + seq = ntohl(tcp_hdr(skb)->seq); + + ntls = tls_driver_ctx(skb->sk, TLS_OFFLOAD_CTX_DIR_TX); + if (ntls->next_seq == seq + datalen) + ntls->next_seq = seq; + else + WARN_ON_ONCE(1); +#endif +} + static void nfp_net_tx_xmit_more_flush(struct nfp_net_tx_ring *tx_ring) { wmb(); @@ -763,24 +923,47 @@ static void nfp_net_tx_xmit_more_flush(struct nfp_net_tx_ring *tx_ring) tx_ring->wr_ptr_add = 0; } -static int nfp_net_prep_port_id(struct sk_buff *skb) +static int nfp_net_prep_tx_meta(struct sk_buff *skb, u64 tls_handle) { struct metadata_dst *md_dst = skb_metadata_dst(skb); unsigned char *data; + u32 meta_id = 0; + int md_bytes; - if (likely(!md_dst)) - return 0; - if (unlikely(md_dst->type != METADATA_HW_PORT_MUX)) + if (likely(!md_dst && !tls_handle)) return 0; + if (unlikely(md_dst && md_dst->type != METADATA_HW_PORT_MUX)) { + if (!tls_handle) + return 0; + md_dst = NULL; + } + + md_bytes = 4 + !!md_dst * 4 + !!tls_handle * 8; - if (unlikely(skb_cow_head(skb, 8))) + if (unlikely(skb_cow_head(skb, md_bytes))) return -ENOMEM; - data = skb_push(skb, 8); - put_unaligned_be32(NFP_NET_META_PORTID, data); - put_unaligned_be32(md_dst->u.port_info.port_id, data + 4); + meta_id = 0; + data = skb_push(skb, md_bytes) + md_bytes; + if (md_dst) { + data -= 4; + put_unaligned_be32(md_dst->u.port_info.port_id, data); + meta_id = NFP_NET_META_PORTID; + } + if (tls_handle) { + /* conn handle is opaque, we just use u64 to be able to quickly + * compare it to zero + */ + data -= 8; + memcpy(data, &tls_handle, sizeof(tls_handle)); + meta_id <<= NFP_NET_META_FIELD_SIZE; + meta_id |= NFP_NET_META_CONN_HANDLE; + } + + data -= 4; + put_unaligned_be32(meta_id, data); - return 8; + return md_bytes; } /** @@ -793,7 +976,7 @@ static int nfp_net_prep_port_id(struct sk_buff *skb) static int nfp_net_tx(struct sk_buff *skb, struct net_device *netdev) { struct nfp_net *nn = netdev_priv(netdev); - const struct skb_frag_struct *frag; + const skb_frag_t *frag; int f, nr_frags, wr_idx, md_bytes; struct nfp_net_tx_ring *tx_ring; struct nfp_net_r_vector *r_vec; @@ -803,6 +986,7 @@ static int nfp_net_tx(struct sk_buff *skb, struct net_device *netdev) struct nfp_net_dp *dp; dma_addr_t dma_addr; unsigned int fsize; + u64 tls_handle = 0; u16 qidx; dp = &nn->dp; @@ -824,18 +1008,21 @@ static int nfp_net_tx(struct sk_buff *skb, struct net_device *netdev) return NETDEV_TX_BUSY; } - md_bytes = nfp_net_prep_port_id(skb); - if (unlikely(md_bytes < 0)) { + skb = nfp_net_tls_tx(dp, r_vec, skb, &tls_handle, &nr_frags); + if (unlikely(!skb)) { nfp_net_tx_xmit_more_flush(tx_ring); - dev_kfree_skb_any(skb); return NETDEV_TX_OK; } + md_bytes = nfp_net_prep_tx_meta(skb, tls_handle); + if (unlikely(md_bytes < 0)) + goto err_flush; + /* Start with the head skbuf */ dma_addr = dma_map_single(dp->dev, skb->data, skb_headlen(skb), DMA_TO_DEVICE); if (dma_mapping_error(dp->dev, dma_addr)) - goto err_free; + goto err_dma_err; wr_idx = D_IDX(tx_ring, tx_ring->wr_p); @@ -909,7 +1096,7 @@ static int nfp_net_tx(struct sk_buff *skb, struct net_device *netdev) nfp_net_tx_ring_stop(nd_q, tx_ring); tx_ring->wr_ptr_add += nr_frags + 1; - if (__netdev_tx_sent_queue(nd_q, txbuf->real_len, skb->xmit_more)) + if (__netdev_tx_sent_queue(nd_q, txbuf->real_len, netdev_xmit_more())) nfp_net_tx_xmit_more_flush(tx_ring); return NETDEV_TX_OK; @@ -931,12 +1118,14 @@ err_unmap: tx_ring->txbufs[wr_idx].skb = NULL; tx_ring->txbufs[wr_idx].dma_addr = 0; tx_ring->txbufs[wr_idx].fidx = -2; -err_free: +err_dma_err: nn_dp_warn(dp, "Failed to map DMA TX buffer\n"); +err_flush: nfp_net_tx_xmit_more_flush(tx_ring); u64_stats_update_begin(&r_vec->tx_sync); r_vec->tx_errors++; u64_stats_update_end(&r_vec->tx_sync); + nfp_net_tls_tx_undo(skb, tls_handle); dev_kfree_skb_any(skb); return NETDEV_TX_OK; } @@ -967,7 +1156,7 @@ static void nfp_net_tx_complete(struct nfp_net_tx_ring *tx_ring, int budget) todo = D_IDX(tx_ring, qcp_rd_p - tx_ring->qcp_rd_p); while (todo--) { - const struct skb_frag_struct *frag; + const skb_frag_t *frag; struct nfp_net_tx_buf *tx_buf; struct sk_buff *skb; int fidx, nr_frags; @@ -1082,7 +1271,7 @@ static bool nfp_net_xdp_complete(struct nfp_net_tx_ring *tx_ring) static void nfp_net_tx_ring_reset(struct nfp_net_dp *dp, struct nfp_net_tx_ring *tx_ring) { - const struct skb_frag_struct *frag; + const skb_frag_t *frag; struct netdev_queue *nd_q; while (!tx_ring->is_xdp && tx_ring->rd_p != tx_ring->wr_p) { @@ -1635,6 +1824,7 @@ static int nfp_net_rx(struct nfp_net_rx_ring *rx_ring, int budget) struct nfp_net_rx_buf *rxbuf; struct nfp_net_rx_desc *rxd; struct nfp_meta_parsed meta; + bool redir_egress = false; struct net_device *netdev; dma_addr_t new_dma_addr; u32 meta_len_xdp = 0; @@ -1770,13 +1960,16 @@ static int nfp_net_rx(struct nfp_net_rx_ring *rx_ring, int budget) struct nfp_net *nn; nn = netdev_priv(dp->netdev); - netdev = nfp_app_repr_get(nn->app, meta.portid); + netdev = nfp_app_dev_get(nn->app, meta.portid, + &redir_egress); if (unlikely(!netdev)) { nfp_net_rx_drop(dp, r_vec, rx_ring, rxbuf, NULL); continue; } - nfp_repr_inc_rx_stats(netdev, pkt_len); + + if (nfp_netdev_is_nfp_repr(netdev)) + nfp_repr_inc_rx_stats(netdev, pkt_len); } skb = build_skb(rxbuf->frag, true_bufsz); @@ -1805,13 +1998,29 @@ static int nfp_net_rx(struct nfp_net_rx_ring *rx_ring, int budget) nfp_net_rx_csum(dp, r_vec, rxd, &meta, skb); +#ifdef CONFIG_TLS_DEVICE + if (rxd->rxd.flags & PCIE_DESC_RX_DECRYPTED) { + skb->decrypted = true; + u64_stats_update_begin(&r_vec->rx_sync); + r_vec->hw_tls_rx++; + u64_stats_update_end(&r_vec->rx_sync); + } +#endif + if (rxd->rxd.flags & PCIE_DESC_RX_VLAN) __vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q), le16_to_cpu(rxd->rxd.vlan)); if (meta_len_xdp) skb_metadata_set(skb, meta_len_xdp); - napi_gro_receive(&rx_ring->r_vec->napi, skb); + if (likely(!redir_egress)) { + napi_gro_receive(&rx_ring->r_vec->napi, skb); + } else { + skb->dev = netdev; + skb_reset_network_header(skb); + __skb_push(skb, ETH_HLEN); + dev_queue_xmit(skb); + } } if (xdp_prog) { @@ -3111,7 +3320,9 @@ static int nfp_net_change_mtu(struct net_device *netdev, int new_mtu) static int nfp_net_vlan_rx_add_vid(struct net_device *netdev, __be16 proto, u16 vid) { + const u32 cmd = NFP_NET_CFG_MBOX_CMD_CTAG_FILTER_ADD; struct nfp_net *nn = netdev_priv(netdev); + int err; /* Priority tagged packets with vlan id 0 are processed by the * NFP as untagged packets @@ -3119,17 +3330,23 @@ nfp_net_vlan_rx_add_vid(struct net_device *netdev, __be16 proto, u16 vid) if (!vid) return 0; + err = nfp_net_mbox_lock(nn, NFP_NET_CFG_VLAN_FILTER_SZ); + if (err) + return err; + nn_writew(nn, nn->tlv_caps.mbox_off + NFP_NET_CFG_VLAN_FILTER_VID, vid); nn_writew(nn, nn->tlv_caps.mbox_off + NFP_NET_CFG_VLAN_FILTER_PROTO, ETH_P_8021Q); - return nfp_net_reconfig_mbox(nn, NFP_NET_CFG_MBOX_CMD_CTAG_FILTER_ADD); + return nfp_net_mbox_reconfig_and_unlock(nn, cmd); } static int nfp_net_vlan_rx_kill_vid(struct net_device *netdev, __be16 proto, u16 vid) { + const u32 cmd = NFP_NET_CFG_MBOX_CMD_CTAG_FILTER_KILL; struct nfp_net *nn = netdev_priv(netdev); + int err; /* Priority tagged packets with vlan id 0 are processed by the * NFP as untagged packets @@ -3137,11 +3354,15 @@ nfp_net_vlan_rx_kill_vid(struct net_device *netdev, __be16 proto, u16 vid) if (!vid) return 0; + err = nfp_net_mbox_lock(nn, NFP_NET_CFG_VLAN_FILTER_SZ); + if (err) + return err; + nn_writew(nn, nn->tlv_caps.mbox_off + NFP_NET_CFG_VLAN_FILTER_VID, vid); nn_writew(nn, nn->tlv_caps.mbox_off + NFP_NET_CFG_VLAN_FILTER_PROTO, ETH_P_8021Q); - return nfp_net_reconfig_mbox(nn, NFP_NET_CFG_MBOX_CMD_CTAG_FILTER_KILL); + return nfp_net_mbox_reconfig_and_unlock(nn, cmd); } static void nfp_net_stat64(struct net_device *netdev, @@ -3324,8 +3545,11 @@ nfp_net_get_phys_port_name(struct net_device *netdev, char *name, size_t len) struct nfp_net *nn = netdev_priv(netdev); int n; + /* If port is defined, devlink_port is registered and devlink core + * is taking care of name formatting. + */ if (nn->port) - return nfp_port_get_phys_port_name(netdev, name, len); + return -EOPNOTSUPP; if (nn->dp.is_vf || nn->vnic_no_name) return -EOPNOTSUPP; @@ -3517,6 +3741,7 @@ const struct net_device_ops nfp_net_netdev_ops = { .ndo_set_vf_mac = nfp_app_set_vf_mac, .ndo_set_vf_vlan = nfp_app_set_vf_vlan, .ndo_set_vf_spoofchk = nfp_app_set_vf_spoofchk, + .ndo_set_vf_trust = nfp_app_set_vf_trust, .ndo_get_vf_config = nfp_app_get_vf_config, .ndo_set_vf_link_state = nfp_app_set_vf_link_state, .ndo_setup_tc = nfp_port_setup_tc, @@ -3530,8 +3755,7 @@ const struct net_device_ops nfp_net_netdev_ops = { .ndo_udp_tunnel_add = nfp_net_add_vxlan_port, .ndo_udp_tunnel_del = nfp_net_del_vxlan_port, .ndo_bpf = nfp_net_xdp, - .ndo_get_port_parent_id = nfp_port_get_port_parent_id, - .ndo_get_devlink = nfp_devlink_get_devlink, + .ndo_get_devlink_port = nfp_devlink_get_devlink_port, }; /** @@ -3548,7 +3772,7 @@ void nfp_net_info(struct nfp_net *nn) nn->fw_ver.resv, nn->fw_ver.class, nn->fw_ver.major, nn->fw_ver.minor, nn->max_mtu); - nn_info(nn, "CAP: %#x %s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s\n", + nn_info(nn, "CAP: %#x %s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s\n", nn->cap, nn->cap & NFP_NET_CFG_CTRL_PROMISC ? "PROMISC " : "", nn->cap & NFP_NET_CFG_CTRL_L2BC ? "L2BCFILT " : "", @@ -3564,7 +3788,6 @@ void nfp_net_info(struct nfp_net *nn) nn->cap & NFP_NET_CFG_CTRL_RSS ? "RSS1 " : "", nn->cap & NFP_NET_CFG_CTRL_RSS2 ? "RSS2 " : "", nn->cap & NFP_NET_CFG_CTRL_CTAG_FILTER ? "CTAG_FILTER " : "", - nn->cap & NFP_NET_CFG_CTRL_L2SWITCH ? "L2SWITCH " : "", nn->cap & NFP_NET_CFG_CTRL_MSIXAUTO ? "AUTOMASK " : "", nn->cap & NFP_NET_CFG_CTRL_IRQMOD ? "IRQMOD " : "", nn->cap & NFP_NET_CFG_CTRL_VXLAN ? "VXLAN " : "", @@ -3632,6 +3855,8 @@ nfp_net_alloc(struct pci_dev *pdev, void __iomem *ctrl_bar, bool needs_netdev, nn->dp.txd_cnt = NFP_NET_TX_DESCS_DEFAULT; nn->dp.rxd_cnt = NFP_NET_RX_DESCS_DEFAULT; + sema_init(&nn->bar_lock, 1); + spin_lock_init(&nn->reconfig_lock); spin_lock_init(&nn->link_status_lock); @@ -3642,6 +3867,10 @@ nfp_net_alloc(struct pci_dev *pdev, void __iomem *ctrl_bar, bool needs_netdev, if (err) goto err_free_nn; + err = nfp_ccm_mbox_alloc(nn); + if (err) + goto err_free_nn; + return nn; err_free_nn: @@ -3659,6 +3888,8 @@ err_free_nn: void nfp_net_free(struct nfp_net *nn) { WARN_ON(timer_pending(&nn->reconfig_timer) || nn->reconfig_posted); + nfp_ccm_mbox_free(nn); + if (nn->dp.netdev) free_netdev(nn->dp.netdev); else @@ -3886,14 +4117,7 @@ int nfp_net_init(struct nfp_net *nn) /* Set default MTU and Freelist buffer size */ if (!nfp_net_is_data_vnic(nn) && nn->app->ctrl_mtu) { - if (nn->app->ctrl_mtu <= nn->max_mtu) { - nn->dp.mtu = nn->app->ctrl_mtu; - } else { - if (nn->app->ctrl_mtu != NFP_APP_CTRL_MTU_MAX) - nn_warn(nn, "app requested MTU above max supported %u > %u\n", - nn->app->ctrl_mtu, nn->max_mtu); - nn->dp.mtu = nn->max_mtu; - } + nn->dp.mtu = min(nn->app->ctrl_mtu, nn->max_mtu); } else if (nn->max_mtu < NFP_NET_DEFAULT_MTU) { nn->dp.mtu = nn->max_mtu; } else { @@ -3920,9 +4144,6 @@ int nfp_net_init(struct nfp_net *nn) nn->dp.ctrl |= NFP_NET_CFG_CTRL_IRQMOD; } - if (nn->dp.netdev) - nfp_net_netdev_init(nn); - /* Stash the re-configuration queue away. First odd queue in TX Bar */ nn->qcp_cfg = nn->tx_bar + NFP_QCP_QUEUE_ADDR_SZ; @@ -3935,11 +4156,27 @@ int nfp_net_init(struct nfp_net *nn) if (err) return err; + if (nn->dp.netdev) { + nfp_net_netdev_init(nn); + + err = nfp_ccm_mbox_init(nn); + if (err) + return err; + + err = nfp_net_tls_init(nn); + if (err) + goto err_clean_mbox; + } + nfp_net_vecs_init(nn); if (!nn->dp.netdev) return 0; return register_netdev(nn->dp.netdev); + +err_clean_mbox: + nfp_ccm_mbox_clean(nn); + return err; } /** @@ -3952,5 +4189,6 @@ void nfp_net_clean(struct nfp_net *nn) return; unregister_netdev(nn->dp.netdev); + nfp_ccm_mbox_clean(nn); nfp_net_reconfig_wait_posted(nn); } diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_ctrl.c b/drivers/net/ethernet/netronome/nfp/nfp_net_ctrl.c index 6d5213b5bcb0..d835c14b7257 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_net_ctrl.c +++ b/drivers/net/ethernet/netronome/nfp/nfp_net_ctrl.c @@ -99,6 +99,21 @@ int nfp_net_tlv_caps_parse(struct device *dev, u8 __iomem *ctrl_mem, caps->repr_cap = readl(data); break; + case NFP_NET_CFG_TLV_TYPE_MBOX_CMSG_TYPES: + if (length >= 4) + caps->mbox_cmsg_types = readl(data); + break; + case NFP_NET_CFG_TLV_TYPE_CRYPTO_OPS: + if (length < 32) { + dev_err(dev, + "CRYPTO OPS TLV should be at least 32B, is %dB offset:%u\n", + length, offset); + return -EINVAL; + } + + caps->crypto_ops = readl(data); + caps->crypto_enable_off = data - ctrl_mem + 16; + break; default: if (!FIELD_GET(NFP_NET_CFG_TLV_HEADER_REQUIRED, hdr)) break; diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_ctrl.h b/drivers/net/ethernet/netronome/nfp/nfp_net_ctrl.h index 372adea10e14..ee6b24e4eacd 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_net_ctrl.h +++ b/drivers/net/ethernet/netronome/nfp/nfp_net_ctrl.h @@ -44,6 +44,7 @@ #define NFP_NET_META_MARK 2 #define NFP_NET_META_PORTID 5 #define NFP_NET_META_CSUM 6 /* checksum complete type */ +#define NFP_NET_META_CONN_HANDLE 7 #define NFP_META_PORT_ID_CTRL ~0U @@ -104,8 +105,6 @@ #define NFP_NET_CFG_CTRL_RINGPRIO (0x1 << 19) /* Ring priorities */ #define NFP_NET_CFG_CTRL_MSIXAUTO (0x1 << 20) /* MSI-X auto-masking */ #define NFP_NET_CFG_CTRL_TXRWB (0x1 << 21) /* Write-back of TX ring*/ -#define NFP_NET_CFG_CTRL_L2SWITCH (0x1 << 22) /* L2 Switch */ -#define NFP_NET_CFG_CTRL_L2SWITCH_LOCAL (0x1 << 23) /* Switch to local */ #define NFP_NET_CFG_CTRL_VXLAN (0x1 << 24) /* VXLAN tunnel support */ #define NFP_NET_CFG_CTRL_NVGRE (0x1 << 25) /* NVGRE tunnel support */ #define NFP_NET_CFG_CTRL_BPF (0x1 << 27) /* BPF offload capable */ @@ -130,7 +129,6 @@ #define NFP_NET_CFG_UPDATE_TXRPRIO (0x1 << 3) /* TX Ring prio change */ #define NFP_NET_CFG_UPDATE_RXRPRIO (0x1 << 4) /* RX Ring prio change */ #define NFP_NET_CFG_UPDATE_MSIX (0x1 << 5) /* MSI-X change */ -#define NFP_NET_CFG_UPDATE_L2SWITCH (0x1 << 6) /* Switch changes */ #define NFP_NET_CFG_UPDATE_RESET (0x1 << 7) /* Update due to FLR */ #define NFP_NET_CFG_UPDATE_IRQMOD (0x1 << 8) /* IRQ mod change */ #define NFP_NET_CFG_UPDATE_VXLAN (0x1 << 9) /* VXLAN port change */ @@ -138,6 +136,7 @@ #define NFP_NET_CFG_UPDATE_MACADDR (0x1 << 11) /* MAC address change */ #define NFP_NET_CFG_UPDATE_MBOX (0x1 << 12) /* Mailbox update */ #define NFP_NET_CFG_UPDATE_VF (0x1 << 13) /* VF settings change */ +#define NFP_NET_CFG_UPDATE_CRYPTO (0x1 << 14) /* Crypto on/off */ #define NFP_NET_CFG_UPDATE_ERR (0x1 << 31) /* A error occurred */ #define NFP_NET_CFG_TXRS_ENABLE 0x0008 #define NFP_NET_CFG_RXRS_ENABLE 0x0010 @@ -392,12 +391,12 @@ #define NFP_NET_CFG_MBOX_SIMPLE_CMD 0x0 #define NFP_NET_CFG_MBOX_SIMPLE_RET 0x4 #define NFP_NET_CFG_MBOX_SIMPLE_VAL 0x8 -#define NFP_NET_CFG_MBOX_SIMPLE_LEN 12 #define NFP_NET_CFG_MBOX_CMD_CTAG_FILTER_ADD 1 #define NFP_NET_CFG_MBOX_CMD_CTAG_FILTER_KILL 2 #define NFP_NET_CFG_MBOX_CMD_PCI_DSCP_PRIOMAP_SET 5 +#define NFP_NET_CFG_MBOX_CMD_TLV_CMSG 6 /** * VLAN filtering using general use mailbox @@ -470,6 +469,16 @@ * %NFP_NET_CFG_TLV_TYPE_REPR_CAP: * Single word, equivalent of %NFP_NET_CFG_CAP for representors, features which * can be used on representors. + * + * %NFP_NET_CFG_TLV_TYPE_MBOX_CMSG_TYPES: + * Variable, bitmap of control message types supported by the mailbox handler. + * Bit 0 corresponds to message type 0, bit 1 to 1, etc. Control messages are + * encapsulated into simple TLVs, with an end TLV and written to the Mailbox. + * + * %NFP_NET_CFG_TLV_TYPE_CRYPTO_OPS: + * 8 words, bitmaps of supported and enabled crypto operations. + * First 16B (4 words) contains a bitmap of supported crypto operations, + * and next 16B contain the enabled operations. */ #define NFP_NET_CFG_TLV_TYPE_UNKNOWN 0 #define NFP_NET_CFG_TLV_TYPE_RESERVED 1 @@ -479,6 +488,8 @@ #define NFP_NET_CFG_TLV_TYPE_EXPERIMENTAL0 5 #define NFP_NET_CFG_TLV_TYPE_EXPERIMENTAL1 6 #define NFP_NET_CFG_TLV_TYPE_REPR_CAP 7 +#define NFP_NET_CFG_TLV_TYPE_MBOX_CMSG_TYPES 10 +#define NFP_NET_CFG_TLV_TYPE_CRYPTO_OPS 11 /* see crypto/fw.h */ struct device; @@ -488,20 +499,20 @@ struct device; * @mbox_off: vNIC mailbox area offset * @mbox_len: vNIC mailbox area length * @repr_cap: capabilities for representors + * @mbox_cmsg_types: cmsgs which can be passed through the mailbox + * @crypto_ops: supported crypto operations + * @crypto_enable_off: offset of crypto ops enable region */ struct nfp_net_tlv_caps { u32 me_freq_mhz; unsigned int mbox_off; unsigned int mbox_len; u32 repr_cap; + u32 mbox_cmsg_types; + u32 crypto_ops; + unsigned int crypto_enable_off; }; int nfp_net_tlv_caps_parse(struct device *dev, u8 __iomem *ctrl_mem, struct nfp_net_tlv_caps *caps); - -static inline bool nfp_net_has_mbox(struct nfp_net_tlv_caps *caps) -{ - return caps->mbox_len >= NFP_NET_CFG_MBOX_SIMPLE_LEN; -} - #endif /* _NFP_NET_CTRL_H_ */ diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_debugfs.c b/drivers/net/ethernet/netronome/nfp/nfp_net_debugfs.c index ab7f2498e1c4..553c708694e8 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_net_debugfs.c +++ b/drivers/net/ethernet/netronome/nfp/nfp_net_debugfs.c @@ -159,19 +159,13 @@ void nfp_net_debugfs_vnic_add(struct nfp_net *nn, struct dentry *ddir) else strcpy(name, "ctrl-vnic"); nn->debugfs_dir = debugfs_create_dir(name, ddir); - if (IS_ERR_OR_NULL(nn->debugfs_dir)) - return; /* Create queue debugging sub-tree */ queues = debugfs_create_dir("queue", nn->debugfs_dir); - if (IS_ERR_OR_NULL(queues)) - return; rx = debugfs_create_dir("rx", queues); tx = debugfs_create_dir("tx", queues); xdp = debugfs_create_dir("xdp", queues); - if (IS_ERR_OR_NULL(rx) || IS_ERR_OR_NULL(tx) || IS_ERR_OR_NULL(xdp)) - return; for (i = 0; i < min(nn->max_rx_rings, nn->max_r_vecs); i++) { sprintf(name, "%d", i); @@ -190,16 +184,7 @@ void nfp_net_debugfs_vnic_add(struct nfp_net *nn, struct dentry *ddir) struct dentry *nfp_net_debugfs_device_add(struct pci_dev *pdev) { - struct dentry *dev_dir; - - if (IS_ERR_OR_NULL(nfp_dir)) - return NULL; - - dev_dir = debugfs_create_dir(pci_name(pdev), nfp_dir); - if (IS_ERR_OR_NULL(dev_dir)) - return NULL; - - return dev_dir; + return debugfs_create_dir(pci_name(pdev), nfp_dir); } void nfp_net_debugfs_dir_clean(struct dentry **dir) diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c b/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c index 690b62718dbb..1b840ee47339 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c +++ b/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c @@ -18,6 +18,7 @@ #include <linux/pci.h> #include <linux/ethtool.h> #include <linux/firmware.h> +#include <linux/sfp.h> #include "nfpcore/nfp.h" #include "nfpcore/nfp_nsp.h" @@ -149,8 +150,11 @@ static const struct nfp_et_stat nfp_mac_et_stats[] = { #define NN_ET_GLOBAL_STATS_LEN ARRAY_SIZE(nfp_net_et_stats) #define NN_ET_SWITCH_STATS_LEN 9 -#define NN_RVEC_GATHER_STATS 9 +#define NN_RVEC_GATHER_STATS 13 #define NN_RVEC_PER_Q_STATS 3 +#define NN_CTRL_PATH_STATS 1 + +#define SFP_SFF_REV_COMPLIANCE 1 static void nfp_net_get_nspinfo(struct nfp_app *app, char *version) { @@ -420,7 +424,8 @@ static unsigned int nfp_vnic_get_sw_stats_count(struct net_device *netdev) { struct nfp_net *nn = netdev_priv(netdev); - return NN_RVEC_GATHER_STATS + nn->max_r_vecs * NN_RVEC_PER_Q_STATS; + return NN_RVEC_GATHER_STATS + nn->max_r_vecs * NN_RVEC_PER_Q_STATS + + NN_CTRL_PATH_STATS; } static u8 *nfp_vnic_get_sw_stats_strings(struct net_device *netdev, u8 *data) @@ -439,10 +444,16 @@ static u8 *nfp_vnic_get_sw_stats_strings(struct net_device *netdev, u8 *data) data = nfp_pr_et(data, "hw_rx_csum_complete"); data = nfp_pr_et(data, "hw_rx_csum_err"); data = nfp_pr_et(data, "rx_replace_buf_alloc_fail"); + data = nfp_pr_et(data, "rx_tls_decrypted_packets"); data = nfp_pr_et(data, "hw_tx_csum"); data = nfp_pr_et(data, "hw_tx_inner_csum"); data = nfp_pr_et(data, "tx_gather"); data = nfp_pr_et(data, "tx_lso"); + data = nfp_pr_et(data, "tx_tls_encrypted_packets"); + data = nfp_pr_et(data, "tx_tls_ooo"); + data = nfp_pr_et(data, "tx_tls_drop_no_sync_data"); + + data = nfp_pr_et(data, "hw_tls_no_space"); return data; } @@ -465,16 +476,20 @@ static u64 *nfp_vnic_get_sw_stats(struct net_device *netdev, u64 *data) tmp[2] = nn->r_vecs[i].hw_csum_rx_complete; tmp[3] = nn->r_vecs[i].hw_csum_rx_error; tmp[4] = nn->r_vecs[i].rx_replace_buf_alloc_fail; + tmp[5] = nn->r_vecs[i].hw_tls_rx; } while (u64_stats_fetch_retry(&nn->r_vecs[i].rx_sync, start)); do { start = u64_stats_fetch_begin(&nn->r_vecs[i].tx_sync); data[1] = nn->r_vecs[i].tx_pkts; data[2] = nn->r_vecs[i].tx_busy; - tmp[5] = nn->r_vecs[i].hw_csum_tx; - tmp[6] = nn->r_vecs[i].hw_csum_tx_inner; - tmp[7] = nn->r_vecs[i].tx_gather; - tmp[8] = nn->r_vecs[i].tx_lso; + tmp[6] = nn->r_vecs[i].hw_csum_tx; + tmp[7] = nn->r_vecs[i].hw_csum_tx_inner; + tmp[8] = nn->r_vecs[i].tx_gather; + tmp[9] = nn->r_vecs[i].tx_lso; + tmp[10] = nn->r_vecs[i].hw_tls_tx; + tmp[11] = nn->r_vecs[i].tls_tx_fallback; + tmp[12] = nn->r_vecs[i].tls_tx_no_fallback; } while (u64_stats_fetch_retry(&nn->r_vecs[i].tx_sync, start)); data += NN_RVEC_PER_Q_STATS; @@ -486,6 +501,8 @@ static u64 *nfp_vnic_get_sw_stats(struct net_device *netdev, u64 *data) for (j = 0; j < NN_RVEC_GATHER_STATS; j++) *data++ = gathered_stats[j]; + *data++ = atomic_read(&nn->ktls_no_space); + return data; } @@ -1096,6 +1113,130 @@ nfp_app_get_dump_data(struct net_device *netdev, struct ethtool_dump *dump, buffer); } +static int +nfp_port_get_module_info(struct net_device *netdev, + struct ethtool_modinfo *modinfo) +{ + struct nfp_eth_table_port *eth_port; + struct nfp_port *port; + unsigned int read_len; + struct nfp_nsp *nsp; + int err = 0; + u8 data; + + port = nfp_port_from_netdev(netdev); + eth_port = nfp_port_get_eth_port(port); + if (!eth_port) + return -EOPNOTSUPP; + + nsp = nfp_nsp_open(port->app->cpp); + if (IS_ERR(nsp)) { + err = PTR_ERR(nsp); + netdev_err(netdev, "Failed to access the NSP: %d\n", err); + return err; + } + + if (!nfp_nsp_has_read_module_eeprom(nsp)) { + netdev_info(netdev, "reading module EEPROM not supported. Please update flash\n"); + err = -EOPNOTSUPP; + goto exit_close_nsp; + } + + switch (eth_port->interface) { + case NFP_INTERFACE_SFP: + case NFP_INTERFACE_SFP28: + err = nfp_nsp_read_module_eeprom(nsp, eth_port->eth_index, + SFP_SFF8472_COMPLIANCE, &data, + 1, &read_len); + if (err < 0) + goto exit_close_nsp; + + if (!data) { + modinfo->type = ETH_MODULE_SFF_8079; + modinfo->eeprom_len = ETH_MODULE_SFF_8079_LEN; + } else { + modinfo->type = ETH_MODULE_SFF_8472; + modinfo->eeprom_len = ETH_MODULE_SFF_8472_LEN; + } + break; + case NFP_INTERFACE_QSFP: + err = nfp_nsp_read_module_eeprom(nsp, eth_port->eth_index, + SFP_SFF_REV_COMPLIANCE, &data, + 1, &read_len); + if (err < 0) + goto exit_close_nsp; + + if (data < 0x3) { + modinfo->type = ETH_MODULE_SFF_8436; + modinfo->eeprom_len = ETH_MODULE_SFF_8436_LEN; + } else { + modinfo->type = ETH_MODULE_SFF_8636; + modinfo->eeprom_len = ETH_MODULE_SFF_8636_LEN; + } + break; + case NFP_INTERFACE_QSFP28: + modinfo->type = ETH_MODULE_SFF_8636; + modinfo->eeprom_len = ETH_MODULE_SFF_8636_LEN; + break; + default: + netdev_err(netdev, "Unsupported module 0x%x detected\n", + eth_port->interface); + err = -EINVAL; + } + +exit_close_nsp: + nfp_nsp_close(nsp); + return err; +} + +static int +nfp_port_get_module_eeprom(struct net_device *netdev, + struct ethtool_eeprom *eeprom, u8 *data) +{ + struct nfp_eth_table_port *eth_port; + struct nfp_port *port; + struct nfp_nsp *nsp; + int err; + + port = nfp_port_from_netdev(netdev); + eth_port = __nfp_port_get_eth_port(port); + if (!eth_port) + return -EOPNOTSUPP; + + nsp = nfp_nsp_open(port->app->cpp); + if (IS_ERR(nsp)) { + err = PTR_ERR(nsp); + netdev_err(netdev, "Failed to access the NSP: %d\n", err); + return err; + } + + if (!nfp_nsp_has_read_module_eeprom(nsp)) { + netdev_info(netdev, "reading module EEPROM not supported. Please update flash\n"); + err = -EOPNOTSUPP; + goto exit_close_nsp; + } + + err = nfp_nsp_read_module_eeprom(nsp, eth_port->eth_index, + eeprom->offset, data, eeprom->len, + &eeprom->len); + if (err < 0) { + if (eeprom->len) { + netdev_warn(netdev, + "Incomplete read from module EEPROM: %d\n", + err); + err = 0; + } else { + netdev_err(netdev, + "Reading from module EEPROM failed: %d\n", + err); + } + } + +exit_close_nsp: + nfp_nsp_close(nsp); + return err; +} + static int nfp_net_set_coalesce(struct net_device *netdev, struct ethtool_coalesce *ec) { @@ -1253,6 +1394,8 @@ static const struct ethtool_ops nfp_net_ethtool_ops = { .set_dump = nfp_app_set_dump, .get_dump_flag = nfp_app_get_dump_flag, .get_dump_data = nfp_app_get_dump_data, + .get_module_info = nfp_port_get_module_info, + .get_module_eeprom = nfp_port_get_module_eeprom, .get_coalesce = nfp_net_get_coalesce, .set_coalesce = nfp_net_set_coalesce, .get_channels = nfp_net_get_channels, @@ -1272,6 +1415,8 @@ const struct ethtool_ops nfp_port_ethtool_ops = { .set_dump = nfp_app_set_dump, .get_dump_flag = nfp_app_get_dump_flag, .get_dump_data = nfp_app_get_dump_data, + .get_module_info = nfp_port_get_module_info, + .get_module_eeprom = nfp_port_get_module_eeprom, .get_link_ksettings = nfp_net_get_link_ksettings, .set_link_ksettings = nfp_net_set_link_ksettings, .get_fecparam = nfp_port_get_fecparam, diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_main.c b/drivers/net/ethernet/netronome/nfp/nfp_net_main.c index 08f5fdbd8e41..921db40047d7 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_net_main.c +++ b/drivers/net/ethernet/netronome/nfp/nfp_net_main.c @@ -150,34 +150,39 @@ nfp_net_pf_init_vnic(struct nfp_pf *pf, struct nfp_net *nn, unsigned int id) nn->id = id; + if (nn->port) { + err = nfp_devlink_port_register(pf->app, nn->port); + if (err) + return err; + } + err = nfp_net_init(nn); if (err) - return err; + goto err_devlink_port_clean; nfp_net_debugfs_vnic_add(nn, pf->ddir); - if (nn->port) { - err = nfp_devlink_port_register(pf->app, nn->port); - if (err) - goto err_dfs_clean; - } + if (nn->port) + nfp_devlink_port_type_eth_set(nn->port); nfp_net_info(nn); if (nfp_net_is_data_vnic(nn)) { err = nfp_app_vnic_init(pf->app, nn); if (err) - goto err_devlink_port_clean; + goto err_devlink_port_type_clean; } return 0; -err_devlink_port_clean: +err_devlink_port_type_clean: if (nn->port) - nfp_devlink_port_unregister(nn->port); -err_dfs_clean: + nfp_devlink_port_type_clear(nn->port); nfp_net_debugfs_dir_clean(&nn->debugfs_dir); nfp_net_clean(nn); +err_devlink_port_clean: + if (nn->port) + nfp_devlink_port_unregister(nn->port); return err; } @@ -200,10 +205,8 @@ nfp_net_pf_alloc_vnics(struct nfp_pf *pf, void __iomem *ctrl_bar, ctrl_bar += NFP_PF_CSR_SLICE_SIZE; /* Kill the vNIC if app init marked it as invalid */ - if (nn->port && nn->port->type == NFP_PORT_INVALID) { + if (nn->port && nn->port->type == NFP_PORT_INVALID) nfp_net_pf_free_vnic(pf, nn); - continue; - } } if (list_empty(&pf->vnics)) @@ -221,9 +224,11 @@ static void nfp_net_pf_clean_vnic(struct nfp_pf *pf, struct nfp_net *nn) if (nfp_net_is_data_vnic(nn)) nfp_app_vnic_clean(pf->app, nn); if (nn->port) - nfp_devlink_port_unregister(nn->port); + nfp_devlink_port_type_clear(nn->port); nfp_net_debugfs_dir_clean(&nn->debugfs_dir); nfp_net_clean(nn); + if (nn->port) + nfp_devlink_port_unregister(nn->port); } static int nfp_net_pf_alloc_irqs(struct nfp_pf *pf) @@ -704,6 +709,10 @@ int nfp_net_pci_probe(struct nfp_pf *pf) if (err) goto err_devlink_unreg; + err = nfp_devlink_params_register(pf); + if (err) + goto err_shared_buf_unreg; + mutex_lock(&pf->lock); pf->ddir = nfp_net_debugfs_device_add(pf->pdev); @@ -737,6 +746,8 @@ err_free_vnics: err_clean_ddir: nfp_net_debugfs_dir_clean(&pf->ddir); mutex_unlock(&pf->lock); + nfp_devlink_params_unregister(pf); +err_shared_buf_unreg: nfp_shared_buf_unregister(pf); err_devlink_unreg: cancel_work_sync(&pf->port_refresh_work); @@ -766,6 +777,7 @@ void nfp_net_pci_remove(struct nfp_pf *pf) mutex_unlock(&pf->lock); + nfp_devlink_params_unregister(pf); nfp_shared_buf_unregister(pf); devlink_unregister(priv_to_devlink(pf)); diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_repr.c b/drivers/net/ethernet/netronome/nfp/nfp_net_repr.c index 94d228c04496..79d72c88bbef 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_net_repr.c +++ b/drivers/net/ethernet/netronome/nfp/nfp_net_repr.c @@ -267,13 +267,14 @@ const struct net_device_ops nfp_repr_netdev_ops = { .ndo_set_vf_mac = nfp_app_set_vf_mac, .ndo_set_vf_vlan = nfp_app_set_vf_vlan, .ndo_set_vf_spoofchk = nfp_app_set_vf_spoofchk, + .ndo_set_vf_trust = nfp_app_set_vf_trust, .ndo_get_vf_config = nfp_app_get_vf_config, .ndo_set_vf_link_state = nfp_app_set_vf_link_state, .ndo_fix_features = nfp_repr_fix_features, .ndo_set_features = nfp_port_set_features, .ndo_set_mac_address = eth_mac_addr, .ndo_get_port_parent_id = nfp_port_get_port_parent_id, - .ndo_get_devlink = nfp_devlink_get_devlink, + .ndo_get_devlink_port = nfp_devlink_get_devlink_port, }; void @@ -298,22 +299,6 @@ static void nfp_repr_clean(struct nfp_repr *repr) nfp_port_free(repr->port); } -static struct lock_class_key nfp_repr_netdev_xmit_lock_key; -static struct lock_class_key nfp_repr_netdev_addr_lock_key; - -static void nfp_repr_set_lockdep_class_one(struct net_device *dev, - struct netdev_queue *txq, - void *_unused) -{ - lockdep_set_class(&txq->_xmit_lock, &nfp_repr_netdev_xmit_lock_key); -} - -static void nfp_repr_set_lockdep_class(struct net_device *dev) -{ - lockdep_set_class(&dev->addr_list_lock, &nfp_repr_netdev_addr_lock_key); - netdev_for_each_tx_queue(dev, nfp_repr_set_lockdep_class_one, NULL); -} - int nfp_repr_init(struct nfp_app *app, struct net_device *netdev, u32 cmsg_port_id, struct nfp_port *port, struct net_device *pf_netdev) @@ -323,8 +308,6 @@ int nfp_repr_init(struct nfp_app *app, struct net_device *netdev, u32 repr_cap = nn->tlv_caps.repr_cap; int err; - nfp_repr_set_lockdep_class(netdev); - repr->port = port; repr->dst = metadata_dst_alloc(0, METADATA_HW_PORT_MUX, GFP_KERNEL); if (!repr->dst) diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_sriov.c b/drivers/net/ethernet/netronome/nfp/nfp_net_sriov.c index b6ec46ed0540..3fdaaf8ed2ba 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_net_sriov.c +++ b/drivers/net/ethernet/netronome/nfp/nfp_net_sriov.c @@ -1,5 +1,5 @@ // SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) -/* Copyright (C) 2017 Netronome Systems, Inc. */ +/* Copyright (C) 2017-2019 Netronome Systems, Inc. */ #include <linux/bitfield.h> #include <linux/errno.h> @@ -146,6 +146,30 @@ int nfp_app_set_vf_spoofchk(struct net_device *netdev, int vf, bool enable) "spoofchk"); } +int nfp_app_set_vf_trust(struct net_device *netdev, int vf, bool enable) +{ + struct nfp_app *app = nfp_app_from_netdev(netdev); + unsigned int vf_offset; + u8 vf_ctrl; + int err; + + err = nfp_net_sriov_check(app, vf, NFP_NET_VF_CFG_MB_CAP_TRUST, + "trust"); + if (err) + return err; + + /* Write trust control bit to VF entry in VF config symbol */ + vf_offset = NFP_NET_VF_CFG_MB_SZ + vf * NFP_NET_VF_CFG_SZ + + NFP_NET_VF_CFG_CTRL; + vf_ctrl = readb(app->pf->vfcfg_tbl2 + vf_offset); + vf_ctrl &= ~NFP_NET_VF_CFG_CTRL_TRUST; + vf_ctrl |= FIELD_PREP(NFP_NET_VF_CFG_CTRL_TRUST, enable); + writeb(vf_ctrl, app->pf->vfcfg_tbl2 + vf_offset); + + return nfp_net_sriov_update(app, vf, NFP_NET_VF_CFG_MB_UPD_TRUST, + "trust"); +} + int nfp_app_set_vf_link_state(struct net_device *netdev, int vf, int link_state) { @@ -213,6 +237,7 @@ int nfp_app_get_vf_config(struct net_device *netdev, int vf, ivi->qos = FIELD_GET(NFP_NET_VF_CFG_VLAN_QOS, vlan_tci); ivi->spoofchk = FIELD_GET(NFP_NET_VF_CFG_CTRL_SPOOF, flags); + ivi->trusted = FIELD_GET(NFP_NET_VF_CFG_CTRL_TRUST, flags); ivi->linkstate = FIELD_GET(NFP_NET_VF_CFG_CTRL_LINK_STATE, flags); return 0; diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_sriov.h b/drivers/net/ethernet/netronome/nfp/nfp_net_sriov.h index c9f09c5bb5ee..a3db0cbf6425 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_net_sriov.h +++ b/drivers/net/ethernet/netronome/nfp/nfp_net_sriov.h @@ -1,5 +1,5 @@ /* SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) */ -/* Copyright (C) 2017 Netronome Systems, Inc. */ +/* Copyright (C) 2017-2019 Netronome Systems, Inc. */ #ifndef _NFP_NET_SRIOV_H_ #define _NFP_NET_SRIOV_H_ @@ -19,12 +19,14 @@ #define NFP_NET_VF_CFG_MB_CAP_VLAN (0x1 << 1) #define NFP_NET_VF_CFG_MB_CAP_SPOOF (0x1 << 2) #define NFP_NET_VF_CFG_MB_CAP_LINK_STATE (0x1 << 3) +#define NFP_NET_VF_CFG_MB_CAP_TRUST (0x1 << 4) #define NFP_NET_VF_CFG_MB_RET 0x2 #define NFP_NET_VF_CFG_MB_UPD 0x4 #define NFP_NET_VF_CFG_MB_UPD_MAC (0x1 << 0) #define NFP_NET_VF_CFG_MB_UPD_VLAN (0x1 << 1) #define NFP_NET_VF_CFG_MB_UPD_SPOOF (0x1 << 2) #define NFP_NET_VF_CFG_MB_UPD_LINK_STATE (0x1 << 3) +#define NFP_NET_VF_CFG_MB_UPD_TRUST (0x1 << 4) #define NFP_NET_VF_CFG_MB_VF_NUM 0x7 /* VF config entry @@ -35,6 +37,7 @@ #define NFP_NET_VF_CFG_MAC_HI 0x0 #define NFP_NET_VF_CFG_MAC_LO 0x6 #define NFP_NET_VF_CFG_CTRL 0x4 +#define NFP_NET_VF_CFG_CTRL_TRUST 0x8 #define NFP_NET_VF_CFG_CTRL_SPOOF 0x4 #define NFP_NET_VF_CFG_CTRL_LINK_STATE 0x3 #define NFP_NET_VF_CFG_LS_MODE_AUTO 0 @@ -48,6 +51,7 @@ int nfp_app_set_vf_mac(struct net_device *netdev, int vf, u8 *mac); int nfp_app_set_vf_vlan(struct net_device *netdev, int vf, u16 vlan, u8 qos, __be16 vlan_proto); int nfp_app_set_vf_spoofchk(struct net_device *netdev, int vf, bool setting); +int nfp_app_set_vf_trust(struct net_device *netdev, int vf, bool setting); int nfp_app_set_vf_link_state(struct net_device *netdev, int vf, int link_state); int nfp_app_get_vf_config(struct net_device *netdev, int vf, diff --git a/drivers/net/ethernet/netronome/nfp/nfp_netvf_main.c b/drivers/net/ethernet/netronome/nfp/nfp_netvf_main.c index 1145849ca7ba..e4977cdf7678 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_netvf_main.c +++ b/drivers/net/ethernet/netronome/nfp/nfp_netvf_main.c @@ -282,8 +282,14 @@ err_free_vf: static void nfp_netvf_pci_remove(struct pci_dev *pdev) { - struct nfp_net_vf *vf = pci_get_drvdata(pdev); - struct nfp_net *nn = vf->nn; + struct nfp_net_vf *vf; + struct nfp_net *nn; + + vf = pci_get_drvdata(pdev); + if (!vf) + return; + + nn = vf->nn; /* Note, the order is slightly different from above as we need * to keep the nn pointer around till we have freed everything. @@ -317,4 +323,5 @@ struct pci_driver nfp_netvf_pci_driver = { .id_table = nfp_netvf_pci_device_ids, .probe = nfp_netvf_pci_probe, .remove = nfp_netvf_pci_remove, + .shutdown = nfp_netvf_pci_remove, }; diff --git a/drivers/net/ethernet/netronome/nfp/nfp_port.h b/drivers/net/ethernet/netronome/nfp/nfp_port.h index 90ae053f5c07..d7fd203bb180 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_port.h +++ b/drivers/net/ethernet/netronome/nfp/nfp_port.h @@ -131,6 +131,8 @@ int nfp_net_refresh_port_table_sync(struct nfp_pf *pf); int nfp_devlink_port_register(struct nfp_app *app, struct nfp_port *port); void nfp_devlink_port_unregister(struct nfp_port *port); +void nfp_devlink_port_type_eth_set(struct nfp_port *port); +void nfp_devlink_port_type_clear(struct nfp_port *port); /** * Mac stats (0x0000 - 0x0200) diff --git a/drivers/net/ethernet/netronome/nfp/nfpcore/Makefile b/drivers/net/ethernet/netronome/nfp/nfpcore/Makefile deleted file mode 100644 index 805fa28f391a..000000000000 --- a/drivers/net/ethernet/netronome/nfp/nfpcore/Makefile +++ /dev/null @@ -1,2 +0,0 @@ -# SPDX-License-Identifier: GPL-2.0 -# kbuild requires Makefile in a directory to build individual objects diff --git a/drivers/net/ethernet/netronome/nfp/nfpcore/nfp6000/Makefile b/drivers/net/ethernet/netronome/nfp/nfpcore/nfp6000/Makefile deleted file mode 100644 index 805fa28f391a..000000000000 --- a/drivers/net/ethernet/netronome/nfp/nfpcore/nfp6000/Makefile +++ /dev/null @@ -1,2 +0,0 @@ -# SPDX-License-Identifier: GPL-2.0 -# kbuild requires Makefile in a directory to build individual objects diff --git a/drivers/net/ethernet/netronome/nfp/nfpcore/nfp_cpplib.c b/drivers/net/ethernet/netronome/nfp/nfpcore/nfp_cpplib.c index 3cfecf105bde..85734c6badf5 100644 --- a/drivers/net/ethernet/netronome/nfp/nfpcore/nfp_cpplib.c +++ b/drivers/net/ethernet/netronome/nfp/nfpcore/nfp_cpplib.c @@ -24,8 +24,9 @@ /* NFP6000 PL */ #define NFP_PL_DEVICE_ID 0x00000004 #define NFP_PL_DEVICE_ID_MASK GENMASK(7, 0) - -#define NFP6000_ARM_GCSR_SOFTMODEL0 0x00400144 +#define NFP_PL_DEVICE_PART_MASK GENMASK(31, 16) +#define NFP_PL_DEVICE_MODEL_MASK (NFP_PL_DEVICE_PART_MASK | \ + NFP_PL_DEVICE_ID_MASK) /** * nfp_cpp_readl() - Read a u32 word from a CPP location @@ -120,22 +121,17 @@ int nfp_cpp_writeq(struct nfp_cpp *cpp, u32 cpp_id, */ int nfp_cpp_model_autodetect(struct nfp_cpp *cpp, u32 *model) { - const u32 arm_id = NFP_CPP_ID(NFP_CPP_TARGET_ARM, 0, 0); u32 reg; int err; - err = nfp_cpp_readl(cpp, arm_id, NFP6000_ARM_GCSR_SOFTMODEL0, model); - if (err < 0) - return err; - - /* The PL's PluDeviceID revision code is authoratative */ - *model &= ~0xff; err = nfp_xpb_readl(cpp, NFP_XPB_DEVICE(1, 1, 16) + NFP_PL_DEVICE_ID, ®); if (err < 0) return err; - *model |= (NFP_PL_DEVICE_ID_MASK & reg) - 0x10; + *model = reg & NFP_PL_DEVICE_MODEL_MASK; + if (*model & NFP_PL_DEVICE_ID_MASK) + *model -= 0x10; return 0; } diff --git a/drivers/net/ethernet/netronome/nfp/nfpcore/nfp_nsp.c b/drivers/net/ethernet/netronome/nfp/nfpcore/nfp_nsp.c index 3a4e224a64b7..f18e787fa9ad 100644 --- a/drivers/net/ethernet/netronome/nfp/nfpcore/nfp_nsp.c +++ b/drivers/net/ethernet/netronome/nfp/nfpcore/nfp_nsp.c @@ -79,6 +79,8 @@ #define NFP_VERSIONS_NCSI_OFF 22 #define NFP_VERSIONS_CFGR_OFF 26 +#define NSP_SFF_EEPROM_BLOCK_LEN 8 + enum nfp_nsp_cmd { SPCODE_NOOP = 0, /* No operation */ SPCODE_SOFT_RESET = 1, /* Soft reset the NFP */ @@ -94,7 +96,10 @@ enum nfp_nsp_cmd { SPCODE_NSP_IDENTIFY = 13, /* Read NSP version */ SPCODE_FW_STORED = 16, /* If no FW loaded, load flash app FW */ SPCODE_HWINFO_LOOKUP = 17, /* Lookup HWinfo with overwrites etc. */ + SPCODE_HWINFO_SET = 18, /* Set HWinfo entry */ + SPCODE_FW_LOADED = 19, /* Is application firmware loaded */ SPCODE_VERSIONS = 21, /* Report FW versions */ + SPCODE_READ_SFF_EEPROM = 22, /* Read module EEPROM */ }; struct nfp_nsp_dma_buf { @@ -140,6 +145,8 @@ struct nfp_nsp { * @option: NFP SP Command Argument * @buf: NFP SP Buffer Address * @error_cb: Callback for interpreting option if error occurred + * @error_quiet:Don't print command error/warning. Protocol errors are still + * logged. */ struct nfp_nsp_command_arg { u16 code; @@ -148,6 +155,7 @@ struct nfp_nsp_command_arg { u32 option; u64 buf; void (*error_cb)(struct nfp_nsp *state, u32 ret_val); + bool error_quiet; }; /** @@ -238,11 +246,16 @@ static int nfp_nsp_check(struct nfp_nsp *state) state->ver.major = FIELD_GET(NSP_STATUS_MAJOR, reg); state->ver.minor = FIELD_GET(NSP_STATUS_MINOR, reg); - if (state->ver.major != NSP_MAJOR || state->ver.minor < NSP_MINOR) { + if (state->ver.major != NSP_MAJOR) { nfp_err(cpp, "Unsupported ABI %hu.%hu\n", state->ver.major, state->ver.minor); return -EINVAL; } + if (state->ver.minor < NSP_MINOR) { + nfp_err(cpp, "ABI too old to support NIC operation (%u.%hu < %u.%u), please update the management FW on the flash\n", + NSP_MAJOR, state->ver.minor, NSP_MAJOR, NSP_MINOR); + return -EINVAL; + } if (reg & NSP_STATUS_BUSY) { nfp_err(cpp, "Service processor busy!\n"); @@ -397,8 +410,10 @@ __nfp_nsp_command(struct nfp_nsp *state, const struct nfp_nsp_command_arg *arg) err = FIELD_GET(NSP_STATUS_RESULT, reg); if (err) { - nfp_warn(cpp, "Result (error) code set: %d (%d) command: %d\n", - -err, (int)ret_val, arg->code); + if (!arg->error_quiet) + nfp_warn(cpp, "Result (error) code set: %d (%d) command: %d\n", + -err, (int)ret_val, arg->code); + if (arg->error_cb) arg->error_cb(state, ret_val); else @@ -883,12 +898,14 @@ int nfp_nsp_load_stored_fw(struct nfp_nsp *state) } static int -__nfp_nsp_hwinfo_lookup(struct nfp_nsp *state, void *buf, unsigned int size) +__nfp_nsp_hwinfo_lookup(struct nfp_nsp *state, void *buf, unsigned int size, + bool optional) { struct nfp_nsp_command_buf_arg hwinfo_lookup = { { .code = SPCODE_HWINFO_LOOKUP, .option = size, + .error_quiet = optional, }, .in_buf = buf, .in_size = size, @@ -905,7 +922,7 @@ int nfp_nsp_hwinfo_lookup(struct nfp_nsp *state, void *buf, unsigned int size) size = min_t(u32, size, NFP_HWINFO_LOOKUP_SIZE); - err = __nfp_nsp_hwinfo_lookup(state, buf, size); + err = __nfp_nsp_hwinfo_lookup(state, buf, size, false); if (err) return err; @@ -917,6 +934,66 @@ int nfp_nsp_hwinfo_lookup(struct nfp_nsp *state, void *buf, unsigned int size) return 0; } +int nfp_nsp_hwinfo_lookup_optional(struct nfp_nsp *state, void *buf, + unsigned int size, const char *default_val) +{ + int err; + + /* Ensure that the default value is usable irrespective of whether + * it is actually going to be used. + */ + if (strnlen(default_val, size) == size) + return -EINVAL; + + if (!nfp_nsp_has_hwinfo_lookup(state)) { + strcpy(buf, default_val); + return 0; + } + + size = min_t(u32, size, NFP_HWINFO_LOOKUP_SIZE); + + err = __nfp_nsp_hwinfo_lookup(state, buf, size, true); + if (err) { + if (err == -ENOENT) { + strcpy(buf, default_val); + return 0; + } + + nfp_err(state->cpp, "NSP HWinfo lookup failed: %d\n", err); + return err; + } + + if (strnlen(buf, size) == size) { + nfp_err(state->cpp, "NSP HWinfo value not NULL-terminated\n"); + return -EINVAL; + } + + return 0; +} + +int nfp_nsp_hwinfo_set(struct nfp_nsp *state, void *buf, unsigned int size) +{ + struct nfp_nsp_command_buf_arg hwinfo_set = { + { + .code = SPCODE_HWINFO_SET, + .option = size, + }, + .in_buf = buf, + .in_size = size, + }; + + return nfp_nsp_command_buf(state, &hwinfo_set); +} + +int nfp_nsp_fw_loaded(struct nfp_nsp *state) +{ + const struct nfp_nsp_command_arg arg = { + .code = SPCODE_FW_LOADED, + }; + + return __nfp_nsp_command(state, &arg); +} + int nfp_nsp_versions(struct nfp_nsp *state, void *buf, unsigned int size) { struct nfp_nsp_command_buf_arg versions = { @@ -965,3 +1042,62 @@ const char *nfp_nsp_versions_get(enum nfp_nsp_versions id, bool flash, return (const char *)&buf[buf_off]; } + +static int +__nfp_nsp_module_eeprom(struct nfp_nsp *state, void *buf, unsigned int size) +{ + struct nfp_nsp_command_buf_arg module_eeprom = { + { + .code = SPCODE_READ_SFF_EEPROM, + .option = size, + }, + .in_buf = buf, + .in_size = size, + .out_buf = buf, + .out_size = size, + }; + + return nfp_nsp_command_buf(state, &module_eeprom); +} + +int nfp_nsp_read_module_eeprom(struct nfp_nsp *state, int eth_index, + unsigned int offset, void *data, + unsigned int len, unsigned int *read_len) +{ + struct eeprom_buf { + u8 metalen; + __le16 length; + __le16 offset; + __le16 readlen; + u8 eth_index; + u8 data[0]; + } __packed *buf; + int bufsz, ret; + + BUILD_BUG_ON(offsetof(struct eeprom_buf, data) % 8); + + /* Buffer must be large enough and rounded to the next block size. */ + bufsz = struct_size(buf, data, round_up(len, NSP_SFF_EEPROM_BLOCK_LEN)); + buf = kzalloc(bufsz, GFP_KERNEL); + if (!buf) + return -ENOMEM; + + buf->metalen = + offsetof(struct eeprom_buf, data) / NSP_SFF_EEPROM_BLOCK_LEN; + buf->length = cpu_to_le16(len); + buf->offset = cpu_to_le16(offset); + buf->eth_index = eth_index; + + ret = __nfp_nsp_module_eeprom(state, buf, bufsz); + + *read_len = min_t(unsigned int, len, le16_to_cpu(buf->readlen)); + if (*read_len) + memcpy(data, buf->data, *read_len); + + if (!ret && *read_len < len) + ret = -EIO; + + kfree(buf); + + return ret; +} diff --git a/drivers/net/ethernet/netronome/nfp/nfpcore/nfp_nsp.h b/drivers/net/ethernet/netronome/nfp/nfpcore/nfp_nsp.h index bd9c358c646f..1531c1870020 100644 --- a/drivers/net/ethernet/netronome/nfp/nfpcore/nfp_nsp.h +++ b/drivers/net/ethernet/netronome/nfp/nfpcore/nfp_nsp.h @@ -22,6 +22,13 @@ int nfp_nsp_write_flash(struct nfp_nsp *state, const struct firmware *fw); int nfp_nsp_mac_reinit(struct nfp_nsp *state); int nfp_nsp_load_stored_fw(struct nfp_nsp *state); int nfp_nsp_hwinfo_lookup(struct nfp_nsp *state, void *buf, unsigned int size); +int nfp_nsp_hwinfo_lookup_optional(struct nfp_nsp *state, void *buf, + unsigned int size, const char *default_val); +int nfp_nsp_hwinfo_set(struct nfp_nsp *state, void *buf, unsigned int size); +int nfp_nsp_fw_loaded(struct nfp_nsp *state); +int nfp_nsp_read_module_eeprom(struct nfp_nsp *state, int eth_index, + unsigned int offset, void *data, + unsigned int len, unsigned int *read_len); static inline bool nfp_nsp_has_mac_reinit(struct nfp_nsp *state) { @@ -38,11 +45,26 @@ static inline bool nfp_nsp_has_hwinfo_lookup(struct nfp_nsp *state) return nfp_nsp_get_abi_ver_minor(state) > 24; } +static inline bool nfp_nsp_has_hwinfo_set(struct nfp_nsp *state) +{ + return nfp_nsp_get_abi_ver_minor(state) > 25; +} + +static inline bool nfp_nsp_has_fw_loaded(struct nfp_nsp *state) +{ + return nfp_nsp_get_abi_ver_minor(state) > 25; +} + static inline bool nfp_nsp_has_versions(struct nfp_nsp *state) { return nfp_nsp_get_abi_ver_minor(state) > 27; } +static inline bool nfp_nsp_has_read_module_eeprom(struct nfp_nsp *state) +{ + return nfp_nsp_get_abi_ver_minor(state) > 28; +} + enum nfp_eth_interface { NFP_INTERFACE_NONE = 0, NFP_INTERFACE_SFP = 1, @@ -80,6 +102,21 @@ enum nfp_eth_fec { #define NFP_FEC_REED_SOLOMON BIT(NFP_FEC_REED_SOLOMON_BIT) #define NFP_FEC_DISABLED BIT(NFP_FEC_DISABLED_BIT) +/* Defines the valid values of the 'abi_drv_reset' hwinfo key */ +#define NFP_NSP_DRV_RESET_DISK 0 +#define NFP_NSP_DRV_RESET_ALWAYS 1 +#define NFP_NSP_DRV_RESET_NEVER 2 +#define NFP_NSP_DRV_RESET_DEFAULT "0" + +/* Defines the valid values of the 'app_fw_from_flash' hwinfo key */ +#define NFP_NSP_APP_FW_LOAD_DISK 0 +#define NFP_NSP_APP_FW_LOAD_FLASH 1 +#define NFP_NSP_APP_FW_LOAD_PREF 2 +#define NFP_NSP_APP_FW_LOAD_DEFAULT "2" + +/* Define the default value for the 'abi_drv_load_ifc' key */ +#define NFP_NSP_DRV_LOAD_IFC_DEFAULT "0x10ff" + /** * struct nfp_eth_table - ETH table information * @count: number of table entries diff --git a/drivers/net/ethernet/netronome/nfp/nic/Makefile b/drivers/net/ethernet/netronome/nfp/nic/Makefile deleted file mode 100644 index 805fa28f391a..000000000000 --- a/drivers/net/ethernet/netronome/nfp/nic/Makefile +++ /dev/null @@ -1,2 +0,0 @@ -# SPDX-License-Identifier: GPL-2.0 -# kbuild requires Makefile in a directory to build individual objects |
