diff options
Diffstat (limited to 'net')
61 files changed, 1254 insertions, 937 deletions
diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c index f044ae56a313..2a9a60733594 100644 --- a/net/8021q/vlan_dev.c +++ b/net/8021q/vlan_dev.c @@ -370,10 +370,12 @@ static int vlan_dev_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd) ifrr.ifr_ifru = ifr->ifr_ifru; switch (cmd) { + case SIOCSHWTSTAMP: + if (!net_eq(dev_net(dev), &init_net)) + break; case SIOCGMIIPHY: case SIOCGMIIREG: case SIOCSMIIREG: - case SIOCSHWTSTAMP: case SIOCGHWTSTAMP: if (netif_device_present(real_dev) && ops->ndo_do_ioctl) err = ops->ndo_do_ioctl(real_dev, &ifrr, cmd); diff --git a/net/batman-adv/main.c b/net/batman-adv/main.c index dabcaff87e34..4a89177def64 100644 --- a/net/batman-adv/main.c +++ b/net/batman-adv/main.c @@ -168,6 +168,7 @@ int batadv_mesh_init(struct net_device *soft_iface) spin_lock_init(&bat_priv->tt.commit_lock); spin_lock_init(&bat_priv->gw.list_lock); #ifdef CONFIG_BATMAN_ADV_MCAST + spin_lock_init(&bat_priv->mcast.mla_lock); spin_lock_init(&bat_priv->mcast.want_lists_lock); #endif spin_lock_init(&bat_priv->tvlv.container_list_lock); diff --git a/net/batman-adv/main.h b/net/batman-adv/main.h index f827e441025f..c59afcba31e0 100644 --- a/net/batman-adv/main.h +++ b/net/batman-adv/main.h @@ -13,7 +13,7 @@ #define BATADV_DRIVER_DEVICE "batman-adv" #ifndef BATADV_SOURCE_VERSION -#define BATADV_SOURCE_VERSION "2019.1" +#define BATADV_SOURCE_VERSION "2019.2" #endif /* B.A.T.M.A.N. parameters */ diff --git a/net/batman-adv/multicast.c b/net/batman-adv/multicast.c index 3feb9435b715..ec54e236e345 100644 --- a/net/batman-adv/multicast.c +++ b/net/batman-adv/multicast.c @@ -314,8 +314,6 @@ static void batadv_mcast_mla_list_free(struct hlist_head *mcast_list) * translation table except the ones listed in the given mcast_list. * * If mcast_list is NULL then all are retracted. - * - * Do not call outside of the mcast worker! (or cancel mcast worker first) */ static void batadv_mcast_mla_tt_retract(struct batadv_priv *bat_priv, struct hlist_head *mcast_list) @@ -323,8 +321,6 @@ static void batadv_mcast_mla_tt_retract(struct batadv_priv *bat_priv, struct batadv_hw_addr *mcast_entry; struct hlist_node *tmp; - WARN_ON(delayed_work_pending(&bat_priv->mcast.work)); - hlist_for_each_entry_safe(mcast_entry, tmp, &bat_priv->mcast.mla_list, list) { if (mcast_list && @@ -348,8 +344,6 @@ static void batadv_mcast_mla_tt_retract(struct batadv_priv *bat_priv, * * Adds multicast listener announcements from the given mcast_list to the * translation table if they have not been added yet. - * - * Do not call outside of the mcast worker! (or cancel mcast worker first) */ static void batadv_mcast_mla_tt_add(struct batadv_priv *bat_priv, struct hlist_head *mcast_list) @@ -357,8 +351,6 @@ static void batadv_mcast_mla_tt_add(struct batadv_priv *bat_priv, struct batadv_hw_addr *mcast_entry; struct hlist_node *tmp; - WARN_ON(delayed_work_pending(&bat_priv->mcast.work)); - if (!mcast_list) return; @@ -647,7 +639,10 @@ static void batadv_mcast_mla_update(struct work_struct *work) priv_mcast = container_of(delayed_work, struct batadv_priv_mcast, work); bat_priv = container_of(priv_mcast, struct batadv_priv, mcast); + spin_lock(&bat_priv->mcast.mla_lock); __batadv_mcast_mla_update(bat_priv); + spin_unlock(&bat_priv->mcast.mla_lock); + batadv_mcast_start_timer(bat_priv); } diff --git a/net/batman-adv/types.h b/net/batman-adv/types.h index 357ca119329a..74b644738a36 100644 --- a/net/batman-adv/types.h +++ b/net/batman-adv/types.h @@ -1212,6 +1212,11 @@ struct batadv_priv_mcast { unsigned char bridged:1; /** + * @mla_lock: a lock protecting mla_list and mla_flags + */ + spinlock_t mla_lock; + + /** * @num_want_all_unsnoopables: number of nodes wanting unsnoopable IP * traffic */ diff --git a/net/bridge/br_if.c b/net/bridge/br_if.c index 4a9aaa3fac8f..6d4a24a7534b 100644 --- a/net/bridge/br_if.c +++ b/net/bridge/br_if.c @@ -602,13 +602,15 @@ int br_add_if(struct net_bridge *br, struct net_device *dev, call_netdevice_notifiers(NETDEV_JOIN, dev); err = dev_set_allmulti(dev, 1); - if (err) - goto put_back; + if (err) { + kfree(p); /* kobject not yet init'd, manually free */ + goto err1; + } err = kobject_init_and_add(&p->kobj, &brport_ktype, &(dev->dev.kobj), SYSFS_BRIDGE_PORT_ATTR); if (err) - goto err1; + goto err2; err = br_sysfs_addif(p); if (err) @@ -700,12 +702,9 @@ err3: sysfs_remove_link(br->ifobj, p->dev->name); err2: kobject_put(&p->kobj); - p = NULL; /* kobject_put frees */ -err1: dev_set_allmulti(dev, -1); -put_back: +err1: dev_put(dev); - kfree(p); return err; } diff --git a/net/bridge/netfilter/ebtables.c b/net/bridge/netfilter/ebtables.c index 4e0091311d40..6b07e4978eb3 100644 --- a/net/bridge/netfilter/ebtables.c +++ b/net/bridge/netfilter/ebtables.c @@ -2153,7 +2153,9 @@ static int compat_copy_entries(unsigned char *data, unsigned int size_user, if (ret < 0) return ret; - WARN_ON(size_remaining); + if (size_remaining) + return -EINVAL; + return state->buf_kern_offset; } diff --git a/net/ceph/cls_lock_client.c b/net/ceph/cls_lock_client.c index 2105a6eaa66c..4cc28541281b 100644 --- a/net/ceph/cls_lock_client.c +++ b/net/ceph/cls_lock_client.c @@ -271,7 +271,7 @@ static int decode_locker(void **p, void *end, struct ceph_locker *locker) dout("%s %s%llu cookie %s addr %s\n", __func__, ENTITY_NAME(locker->id.name), locker->id.cookie, - ceph_pr_addr(&locker->info.addr.in_addr)); + ceph_pr_addr(&locker->info.addr)); return 0; } diff --git a/net/ceph/debugfs.c b/net/ceph/debugfs.c index 46f65709a6ff..63aef9915f75 100644 --- a/net/ceph/debugfs.c +++ b/net/ceph/debugfs.c @@ -46,7 +46,7 @@ static int monmap_show(struct seq_file *s, void *p) seq_printf(s, "\t%s%lld\t%s\n", ENTITY_NAME(inst->name), - ceph_pr_addr(&inst->addr.in_addr)); + ceph_pr_addr(&inst->addr)); } return 0; } @@ -82,7 +82,7 @@ static int osdmap_show(struct seq_file *s, void *p) char sb[64]; seq_printf(s, "osd%d\t%s\t%3d%%\t(%s)\t%3d%%\n", - i, ceph_pr_addr(&addr->in_addr), + i, ceph_pr_addr(addr), ((map->osd_weight[i]*100) >> 16), ceph_osdmap_state_str(sb, sizeof(sb), state), ((ceph_get_primary_affinity(map, i)*100) >> 16)); diff --git a/net/ceph/messenger.c b/net/ceph/messenger.c index 579d6a1ac7fe..cd0b094468b6 100644 --- a/net/ceph/messenger.c +++ b/net/ceph/messenger.c @@ -186,17 +186,18 @@ static atomic_t addr_str_seq = ATOMIC_INIT(0); static struct page *zero_page; /* used in certain error cases */ -const char *ceph_pr_addr(const struct sockaddr_storage *ss) +const char *ceph_pr_addr(const struct ceph_entity_addr *addr) { int i; char *s; - struct sockaddr_in *in4 = (struct sockaddr_in *) ss; - struct sockaddr_in6 *in6 = (struct sockaddr_in6 *) ss; + struct sockaddr_storage ss = addr->in_addr; /* align */ + struct sockaddr_in *in4 = (struct sockaddr_in *)&ss; + struct sockaddr_in6 *in6 = (struct sockaddr_in6 *)&ss; i = atomic_inc_return(&addr_str_seq) & ADDR_STR_COUNT_MASK; s = addr_str[i]; - switch (ss->ss_family) { + switch (ss.ss_family) { case AF_INET: snprintf(s, MAX_ADDR_STR_LEN, "%pI4:%hu", &in4->sin_addr, ntohs(in4->sin_port)); @@ -209,7 +210,7 @@ const char *ceph_pr_addr(const struct sockaddr_storage *ss) default: snprintf(s, MAX_ADDR_STR_LEN, "(unknown sockaddr family %hu)", - ss->ss_family); + ss.ss_family); } return s; @@ -449,7 +450,7 @@ static void set_sock_callbacks(struct socket *sock, */ static int ceph_tcp_connect(struct ceph_connection *con) { - struct sockaddr_storage *paddr = &con->peer_addr.in_addr; + struct sockaddr_storage ss = con->peer_addr.in_addr; /* align */ struct socket *sock; unsigned int noio_flag; int ret; @@ -458,7 +459,7 @@ static int ceph_tcp_connect(struct ceph_connection *con) /* sock_create_kern() allocates with GFP_KERNEL */ noio_flag = memalloc_noio_save(); - ret = sock_create_kern(read_pnet(&con->msgr->net), paddr->ss_family, + ret = sock_create_kern(read_pnet(&con->msgr->net), ss.ss_family, SOCK_STREAM, IPPROTO_TCP, &sock); memalloc_noio_restore(noio_flag); if (ret) @@ -471,18 +472,18 @@ static int ceph_tcp_connect(struct ceph_connection *con) set_sock_callbacks(sock, con); - dout("connect %s\n", ceph_pr_addr(&con->peer_addr.in_addr)); + dout("connect %s\n", ceph_pr_addr(&con->peer_addr)); con_sock_state_connecting(con); - ret = sock->ops->connect(sock, (struct sockaddr *)paddr, sizeof(*paddr), + ret = sock->ops->connect(sock, (struct sockaddr *)&ss, sizeof(ss), O_NONBLOCK); if (ret == -EINPROGRESS) { dout("connect %s EINPROGRESS sk_state = %u\n", - ceph_pr_addr(&con->peer_addr.in_addr), + ceph_pr_addr(&con->peer_addr), sock->sk->sk_state); } else if (ret < 0) { pr_err("connect %s error %d\n", - ceph_pr_addr(&con->peer_addr.in_addr), ret); + ceph_pr_addr(&con->peer_addr), ret); sock_release(sock); return ret; } @@ -669,8 +670,7 @@ static void reset_connection(struct ceph_connection *con) void ceph_con_close(struct ceph_connection *con) { mutex_lock(&con->mutex); - dout("con_close %p peer %s\n", con, - ceph_pr_addr(&con->peer_addr.in_addr)); + dout("con_close %p peer %s\n", con, ceph_pr_addr(&con->peer_addr)); con->state = CON_STATE_CLOSED; con_flag_clear(con, CON_FLAG_LOSSYTX); /* so we retry next connect */ @@ -694,7 +694,7 @@ void ceph_con_open(struct ceph_connection *con, struct ceph_entity_addr *addr) { mutex_lock(&con->mutex); - dout("con_open %p %s\n", con, ceph_pr_addr(&addr->in_addr)); + dout("con_open %p %s\n", con, ceph_pr_addr(addr)); WARN_ON(con->state != CON_STATE_CLOSED); con->state = CON_STATE_PREOPEN; @@ -1788,21 +1788,22 @@ static int verify_hello(struct ceph_connection *con) { if (memcmp(con->in_banner, CEPH_BANNER, strlen(CEPH_BANNER))) { pr_err("connect to %s got bad banner\n", - ceph_pr_addr(&con->peer_addr.in_addr)); + ceph_pr_addr(&con->peer_addr)); con->error_msg = "protocol error, bad banner"; return -1; } return 0; } -static bool addr_is_blank(struct sockaddr_storage *ss) +static bool addr_is_blank(struct ceph_entity_addr *addr) { - struct in_addr *addr = &((struct sockaddr_in *)ss)->sin_addr; - struct in6_addr *addr6 = &((struct sockaddr_in6 *)ss)->sin6_addr; + struct sockaddr_storage ss = addr->in_addr; /* align */ + struct in_addr *addr4 = &((struct sockaddr_in *)&ss)->sin_addr; + struct in6_addr *addr6 = &((struct sockaddr_in6 *)&ss)->sin6_addr; - switch (ss->ss_family) { + switch (ss.ss_family) { case AF_INET: - return addr->s_addr == htonl(INADDR_ANY); + return addr4->s_addr == htonl(INADDR_ANY); case AF_INET6: return ipv6_addr_any(addr6); default: @@ -1810,25 +1811,25 @@ static bool addr_is_blank(struct sockaddr_storage *ss) } } -static int addr_port(struct sockaddr_storage *ss) +static int addr_port(struct ceph_entity_addr *addr) { - switch (ss->ss_family) { + switch (get_unaligned(&addr->in_addr.ss_family)) { case AF_INET: - return ntohs(((struct sockaddr_in *)ss)->sin_port); + return ntohs(get_unaligned(&((struct sockaddr_in *)&addr->in_addr)->sin_port)); case AF_INET6: - return ntohs(((struct sockaddr_in6 *)ss)->sin6_port); + return ntohs(get_unaligned(&((struct sockaddr_in6 *)&addr->in_addr)->sin6_port)); } return 0; } -static void addr_set_port(struct sockaddr_storage *ss, int p) +static void addr_set_port(struct ceph_entity_addr *addr, int p) { - switch (ss->ss_family) { + switch (get_unaligned(&addr->in_addr.ss_family)) { case AF_INET: - ((struct sockaddr_in *)ss)->sin_port = htons(p); + put_unaligned(htons(p), &((struct sockaddr_in *)&addr->in_addr)->sin_port); break; case AF_INET6: - ((struct sockaddr_in6 *)ss)->sin6_port = htons(p); + put_unaligned(htons(p), &((struct sockaddr_in6 *)&addr->in_addr)->sin6_port); break; } } @@ -1836,21 +1837,18 @@ static void addr_set_port(struct sockaddr_storage *ss, int p) /* * Unlike other *_pton function semantics, zero indicates success. */ -static int ceph_pton(const char *str, size_t len, struct sockaddr_storage *ss, +static int ceph_pton(const char *str, size_t len, struct ceph_entity_addr *addr, char delim, const char **ipend) { - struct sockaddr_in *in4 = (struct sockaddr_in *) ss; - struct sockaddr_in6 *in6 = (struct sockaddr_in6 *) ss; + memset(&addr->in_addr, 0, sizeof(addr->in_addr)); - memset(ss, 0, sizeof(*ss)); - - if (in4_pton(str, len, (u8 *)&in4->sin_addr.s_addr, delim, ipend)) { - ss->ss_family = AF_INET; + if (in4_pton(str, len, (u8 *)&((struct sockaddr_in *)&addr->in_addr)->sin_addr.s_addr, delim, ipend)) { + put_unaligned(AF_INET, &addr->in_addr.ss_family); return 0; } - if (in6_pton(str, len, (u8 *)&in6->sin6_addr.s6_addr, delim, ipend)) { - ss->ss_family = AF_INET6; + if (in6_pton(str, len, (u8 *)&((struct sockaddr_in6 *)&addr->in_addr)->sin6_addr.s6_addr, delim, ipend)) { + put_unaligned(AF_INET6, &addr->in_addr.ss_family); return 0; } @@ -1862,7 +1860,7 @@ static int ceph_pton(const char *str, size_t len, struct sockaddr_storage *ss, */ #ifdef CONFIG_CEPH_LIB_USE_DNS_RESOLVER static int ceph_dns_resolve_name(const char *name, size_t namelen, - struct sockaddr_storage *ss, char delim, const char **ipend) + struct ceph_entity_addr *addr, char delim, const char **ipend) { const char *end, *delim_p; char *colon_p, *ip_addr = NULL; @@ -1891,7 +1889,7 @@ static int ceph_dns_resolve_name(const char *name, size_t namelen, /* do dns_resolve upcall */ ip_len = dns_query(NULL, name, end - name, NULL, &ip_addr, NULL, false); if (ip_len > 0) - ret = ceph_pton(ip_addr, ip_len, ss, -1, NULL); + ret = ceph_pton(ip_addr, ip_len, addr, -1, NULL); else ret = -ESRCH; @@ -1900,13 +1898,13 @@ static int ceph_dns_resolve_name(const char *name, size_t namelen, *ipend = end; pr_info("resolve '%.*s' (ret=%d): %s\n", (int)(end - name), name, - ret, ret ? "failed" : ceph_pr_addr(ss)); + ret, ret ? "failed" : ceph_pr_addr(addr)); return ret; } #else static inline int ceph_dns_resolve_name(const char *name, size_t namelen, - struct sockaddr_storage *ss, char delim, const char **ipend) + struct ceph_entity_addr *addr, char delim, const char **ipend) { return -EINVAL; } @@ -1917,13 +1915,13 @@ static inline int ceph_dns_resolve_name(const char *name, size_t namelen, * then try to extract a hostname to resolve using userspace DNS upcall. */ static int ceph_parse_server_name(const char *name, size_t namelen, - struct sockaddr_storage *ss, char delim, const char **ipend) + struct ceph_entity_addr *addr, char delim, const char **ipend) { int ret; - ret = ceph_pton(name, namelen, ss, delim, ipend); + ret = ceph_pton(name, namelen, addr, delim, ipend); if (ret) - ret = ceph_dns_resolve_name(name, namelen, ss, delim, ipend); + ret = ceph_dns_resolve_name(name, namelen, addr, delim, ipend); return ret; } @@ -1942,7 +1940,6 @@ int ceph_parse_ips(const char *c, const char *end, dout("parse_ips on '%.*s'\n", (int)(end-c), c); for (i = 0; i < max_count; i++) { const char *ipend; - struct sockaddr_storage *ss = &addr[i].in_addr; int port; char delim = ','; @@ -1951,7 +1948,7 @@ int ceph_parse_ips(const char *c, const char *end, p++; } - ret = ceph_parse_server_name(p, end - p, ss, delim, &ipend); + ret = ceph_parse_server_name(p, end - p, &addr[i], delim, &ipend); if (ret) goto bad; ret = -EINVAL; @@ -1982,9 +1979,9 @@ int ceph_parse_ips(const char *c, const char *end, port = CEPH_MON_PORT; } - addr_set_port(ss, port); + addr_set_port(&addr[i], port); - dout("parse_ips got %s\n", ceph_pr_addr(ss)); + dout("parse_ips got %s\n", ceph_pr_addr(&addr[i])); if (p == end) break; @@ -2023,12 +2020,12 @@ static int process_banner(struct ceph_connection *con) */ if (memcmp(&con->peer_addr, &con->actual_peer_addr, sizeof(con->peer_addr)) != 0 && - !(addr_is_blank(&con->actual_peer_addr.in_addr) && + !(addr_is_blank(&con->actual_peer_addr) && con->actual_peer_addr.nonce == con->peer_addr.nonce)) { pr_warn("wrong peer, want %s/%d, got %s/%d\n", - ceph_pr_addr(&con->peer_addr.in_addr), + ceph_pr_addr(&con->peer_addr), (int)le32_to_cpu(con->peer_addr.nonce), - ceph_pr_addr(&con->actual_peer_addr.in_addr), + ceph_pr_addr(&con->actual_peer_addr), (int)le32_to_cpu(con->actual_peer_addr.nonce)); con->error_msg = "wrong peer at address"; return -1; @@ -2037,16 +2034,16 @@ static int process_banner(struct ceph_connection *con) /* * did we learn our address? */ - if (addr_is_blank(&con->msgr->inst.addr.in_addr)) { - int port = addr_port(&con->msgr->inst.addr.in_addr); + if (addr_is_blank(&con->msgr->inst.addr)) { + int port = addr_port(&con->msgr->inst.addr); memcpy(&con->msgr->inst.addr.in_addr, &con->peer_addr_for_me.in_addr, sizeof(con->peer_addr_for_me.in_addr)); - addr_set_port(&con->msgr->inst.addr.in_addr, port); + addr_set_port(&con->msgr->inst.addr, port); encode_my_addr(con->msgr); dout("process_banner learned my addr is %s\n", - ceph_pr_addr(&con->msgr->inst.addr.in_addr)); + ceph_pr_addr(&con->msgr->inst.addr)); } return 0; @@ -2097,7 +2094,7 @@ static int process_connect(struct ceph_connection *con) pr_err("%s%lld %s feature set mismatch," " my %llx < server's %llx, missing %llx\n", ENTITY_NAME(con->peer_name), - ceph_pr_addr(&con->peer_addr.in_addr), + ceph_pr_addr(&con->peer_addr), sup_feat, server_feat, server_feat & ~sup_feat); con->error_msg = "missing required protocol features"; reset_connection(con); @@ -2107,7 +2104,7 @@ static int process_connect(struct ceph_connection *con) pr_err("%s%lld %s protocol version mismatch," " my %d != server's %d\n", ENTITY_NAME(con->peer_name), - ceph_pr_addr(&con->peer_addr.in_addr), + ceph_pr_addr(&con->peer_addr), le32_to_cpu(con->out_connect.protocol_version), le32_to_cpu(con->in_reply.protocol_version)); con->error_msg = "protocol version mismatch"; @@ -2141,7 +2138,7 @@ static int process_connect(struct ceph_connection *con) le32_to_cpu(con->in_reply.connect_seq)); pr_err("%s%lld %s connection reset\n", ENTITY_NAME(con->peer_name), - ceph_pr_addr(&con->peer_addr.in_addr)); + ceph_pr_addr(&con->peer_addr)); reset_connection(con); con_out_kvec_reset(con); ret = prepare_write_connect(con); @@ -2198,7 +2195,7 @@ static int process_connect(struct ceph_connection *con) pr_err("%s%lld %s protocol feature mismatch," " my required %llx > server's %llx, need %llx\n", ENTITY_NAME(con->peer_name), - ceph_pr_addr(&con->peer_addr.in_addr), + ceph_pr_addr(&con->peer_addr), req_feat, server_feat, req_feat & ~server_feat); con->error_msg = "missing required protocol features"; reset_connection(con); @@ -2405,7 +2402,7 @@ static int read_partial_message(struct ceph_connection *con) if ((s64)seq - (s64)con->in_seq < 1) { pr_info("skipping %s%lld %s seq %lld expected %lld\n", ENTITY_NAME(con->peer_name), - ceph_pr_addr(&con->peer_addr.in_addr), + ceph_pr_addr(&con->peer_addr), seq, con->in_seq + 1); con->in_base_pos = -front_len - middle_len - data_len - sizeof_footer(con); @@ -2984,10 +2981,10 @@ static void ceph_con_workfn(struct work_struct *work) static void con_fault(struct ceph_connection *con) { dout("fault %p state %lu to peer %s\n", - con, con->state, ceph_pr_addr(&con->peer_addr.in_addr)); + con, con->state, ceph_pr_addr(&con->peer_addr)); pr_warn("%s%lld %s %s\n", ENTITY_NAME(con->peer_name), - ceph_pr_addr(&con->peer_addr.in_addr), con->error_msg); + ceph_pr_addr(&con->peer_addr), con->error_msg); con->error_msg = NULL; WARN_ON(con->state != CON_STATE_CONNECTING && diff --git a/net/ceph/mon_client.c b/net/ceph/mon_client.c index a53e4fbb6319..895679d3529b 100644 --- a/net/ceph/mon_client.c +++ b/net/ceph/mon_client.c @@ -76,7 +76,7 @@ struct ceph_monmap *ceph_monmap_decode(void *p, void *end) m->num_mon); for (i = 0; i < m->num_mon; i++) dout("monmap_decode mon%d is %s\n", i, - ceph_pr_addr(&m->mon_inst[i].addr.in_addr)); + ceph_pr_addr(&m->mon_inst[i].addr)); return m; bad: @@ -203,7 +203,7 @@ static void reopen_session(struct ceph_mon_client *monc) { if (!monc->hunting) pr_info("mon%d %s session lost, hunting for new mon\n", - monc->cur_mon, ceph_pr_addr(&monc->con.peer_addr.in_addr)); + monc->cur_mon, ceph_pr_addr(&monc->con.peer_addr)); __close_session(monc); __open_session(monc); @@ -1178,7 +1178,7 @@ static void handle_auth_reply(struct ceph_mon_client *monc, __resend_generic_request(monc); pr_info("mon%d %s session established\n", monc->cur_mon, - ceph_pr_addr(&monc->con.peer_addr.in_addr)); + ceph_pr_addr(&monc->con.peer_addr)); } out: diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c index 6f739de28918..9a8eca5eda65 100644 --- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -4926,7 +4926,7 @@ static int decode_watcher(void **p, void *end, struct ceph_watch_item *item) dout("%s %s%llu cookie %llu addr %s\n", __func__, ENTITY_NAME(item->name), item->cookie, - ceph_pr_addr(&item->addr.in_addr)); + ceph_pr_addr(&item->addr)); return 0; } diff --git a/net/ceph/pagevec.c b/net/ceph/pagevec.c index d3736f5bffec..74cafc0142ea 100644 --- a/net/ceph/pagevec.c +++ b/net/ceph/pagevec.c @@ -27,7 +27,7 @@ struct page **ceph_get_direct_page_vector(const void __user *data, while (got < num_pages) { rc = get_user_pages_fast( (unsigned long)data + ((unsigned long)got * PAGE_SIZE), - num_pages - got, write_page, pages + got); + num_pages - got, write_page ? FOLL_WRITE : 0, pages + got); if (rc < 0) break; BUG_ON(rc == 0); diff --git a/net/core/fib_rules.c b/net/core/fib_rules.c index 18f8dd8329ed..43f0115cce9c 100644 --- a/net/core/fib_rules.c +++ b/net/core/fib_rules.c @@ -757,9 +757,9 @@ int fib_nl_newrule(struct sk_buff *skb, struct nlmsghdr *nlh, if (err) goto errout; - if ((nlh->nlmsg_flags & NLM_F_EXCL) && - rule_exists(ops, frh, tb, rule)) { - err = -EEXIST; + if (rule_exists(ops, frh, tb, rule)) { + if (nlh->nlmsg_flags & NLM_F_EXCL) + err = -EEXIST; goto errout_free; } diff --git a/net/core/flow_dissector.c b/net/core/flow_dissector.c index 9ca784c592ac..548f39dde307 100644 --- a/net/core/flow_dissector.c +++ b/net/core/flow_dissector.c @@ -734,7 +734,9 @@ bool bpf_flow_dissect(struct bpf_prog *prog, struct bpf_flow_dissector *ctx, flow_keys->nhoff = nhoff; flow_keys->thoff = flow_keys->nhoff; + preempt_disable(); result = BPF_PROG_RUN(prog, ctx); + preempt_enable(); flow_keys->nhoff = clamp_t(u16, flow_keys->nhoff, nhoff, hlen); flow_keys->thoff = clamp_t(u16, flow_keys->thoff, diff --git a/net/dccp/proto.c b/net/dccp/proto.c index 0e2f71ab8367..5dd85ec51bfe 100644 --- a/net/dccp/proto.c +++ b/net/dccp/proto.c @@ -263,7 +263,6 @@ int dccp_disconnect(struct sock *sk, int flags) struct inet_connection_sock *icsk = inet_csk(sk); struct inet_sock *inet = inet_sk(sk); struct dccp_sock *dp = dccp_sk(sk); - int err = 0; const int old_state = sk->sk_state; if (old_state != DCCP_CLOSED) @@ -307,7 +306,7 @@ int dccp_disconnect(struct sock *sk, int flags) WARN_ON(inet->inet_num && !icsk->icsk_bind_hash); sk->sk_error_report(sk); - return err; + return 0; } EXPORT_SYMBOL_GPL(dccp_disconnect); diff --git a/net/dsa/slave.c b/net/dsa/slave.c index fe7b6a62e8f1..9892ca1f6859 100644 --- a/net/dsa/slave.c +++ b/net/dsa/slave.c @@ -463,6 +463,8 @@ static netdev_tx_t dsa_slave_xmit(struct sk_buff *skb, struct net_device *dev) s->tx_bytes += skb->len; u64_stats_update_end(&s->syncp); + DSA_SKB_CB(skb)->deferred_xmit = false; + /* Identify PTP protocol packets, clone them, and pass them to the * switch driver */ diff --git a/net/dsa/tag_brcm.c b/net/dsa/tag_brcm.c index d52db5f2c721..9c3114179690 100644 --- a/net/dsa/tag_brcm.c +++ b/net/dsa/tag_brcm.c @@ -206,10 +206,10 @@ static const struct dsa_device_ops brcm_prepend_netdev_ops = { .rcv = brcm_tag_rcv_prepend, .overhead = BRCM_TAG_LEN, }; -#endif DSA_TAG_DRIVER(brcm_prepend_netdev_ops); MODULE_ALIAS_DSA_TAG_DRIVER(DSA_TAG_PROTO_BRCM_PREPEND); +#endif static struct dsa_tag_driver *dsa_tag_driver_array[] = { #if IS_ENABLED(CONFIG_NET_DSA_TAG_BRCM) diff --git a/net/ipv4/raw.c b/net/ipv4/raw.c index c55a5432cf37..dc91c27bb788 100644 --- a/net/ipv4/raw.c +++ b/net/ipv4/raw.c @@ -173,6 +173,7 @@ static int icmp_filter(const struct sock *sk, const struct sk_buff *skb) static int raw_v4_input(struct sk_buff *skb, const struct iphdr *iph, int hash) { int sdif = inet_sdif(skb); + int dif = inet_iif(skb); struct sock *sk; struct hlist_head *head; int delivered = 0; @@ -185,8 +186,7 @@ static int raw_v4_input(struct sk_buff *skb, const struct iphdr *iph, int hash) net = dev_net(skb->dev); sk = __raw_v4_lookup(net, __sk_head(head), iph->protocol, - iph->saddr, iph->daddr, - skb->dev->ifindex, sdif); + iph->saddr, iph->daddr, dif, sdif); while (sk) { delivered = 1; diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 077d9abdfcf5..20f6fac5882e 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -77,7 +77,7 @@ #include <asm/unaligned.h> #include <linux/errqueue.h> #include <trace/events/tcp.h> -#include <linux/static_key.h> +#include <linux/jump_label_ratelimit.h> #include <net/busy_poll.h> int sysctl_tcp_max_orphans __read_mostly = NR_FILE; @@ -113,22 +113,28 @@ int sysctl_tcp_max_orphans __read_mostly = NR_FILE; #define REXMIT_NEW 2 /* FRTO-style transmit of unsent/new packets */ #if IS_ENABLED(CONFIG_TLS_DEVICE) -static DEFINE_STATIC_KEY_FALSE(clean_acked_data_enabled); +static DEFINE_STATIC_KEY_DEFERRED_FALSE(clean_acked_data_enabled, HZ); void clean_acked_data_enable(struct inet_connection_sock *icsk, void (*cad)(struct sock *sk, u32 ack_seq)) { icsk->icsk_clean_acked = cad; - static_branch_inc(&clean_acked_data_enabled); + static_branch_inc(&clean_acked_data_enabled.key); } EXPORT_SYMBOL_GPL(clean_acked_data_enable); void clean_acked_data_disable(struct inet_connection_sock *icsk) { - static_branch_dec(&clean_acked_data_enabled); + static_branch_slow_dec_deferred(&clean_acked_data_enabled); icsk->icsk_clean_acked = NULL; } EXPORT_SYMBOL_GPL(clean_acked_data_disable); + +void clean_acked_data_flush(void) +{ + static_key_deferred_flush(&clean_acked_data_enabled); +} +EXPORT_SYMBOL_GPL(clean_acked_data_flush); #endif static void tcp_gro_dev_warn(struct sock *sk, const struct sk_buff *skb, @@ -3598,7 +3604,7 @@ static int tcp_ack(struct sock *sk, const struct sk_buff *skb, int flag) icsk->icsk_retransmits = 0; #if IS_ENABLED(CONFIG_TLS_DEVICE) - if (static_branch_unlikely(&clean_acked_data_enabled)) + if (static_branch_unlikely(&clean_acked_data_enabled.key)) if (icsk->icsk_clean_acked) icsk->icsk_clean_acked(sk, ack); #endif diff --git a/net/netfilter/core.c b/net/netfilter/core.c index 71f06900473e..b96fd3f54705 100644 --- a/net/netfilter/core.c +++ b/net/netfilter/core.c @@ -163,7 +163,7 @@ nf_hook_entries_grow(const struct nf_hook_entries *old, static void hooks_validate(const struct nf_hook_entries *hooks) { -#ifdef CONFIG_DEBUG_KERNEL +#ifdef CONFIG_DEBUG_MISC struct nf_hook_ops **orig_ops; int prio = INT_MIN; size_t i = 0; diff --git a/net/netfilter/nf_conntrack_h323_asn1.c b/net/netfilter/nf_conntrack_h323_asn1.c index 1601275efe2d..4c2ef42e189c 100644 --- a/net/netfilter/nf_conntrack_h323_asn1.c +++ b/net/netfilter/nf_conntrack_h323_asn1.c @@ -172,7 +172,7 @@ static int nf_h323_error_boundary(struct bitstr *bs, size_t bytes, size_t bits) if (bits % BITS_PER_BYTE > 0) bytes++; - if (*bs->cur + bytes > *bs->end) + if (bs->cur + bytes > bs->end) return 1; return 0; diff --git a/net/netfilter/nf_conntrack_h323_main.c b/net/netfilter/nf_conntrack_h323_main.c index 005589c6d0f6..12de40390e97 100644 --- a/net/netfilter/nf_conntrack_h323_main.c +++ b/net/netfilter/nf_conntrack_h323_main.c @@ -748,24 +748,19 @@ static int callforward_do_filter(struct net *net, } break; } -#if IS_ENABLED(CONFIG_NF_CONNTRACK_IPV6) +#if IS_ENABLED(CONFIG_IPV6) case AF_INET6: { - const struct nf_ipv6_ops *v6ops; struct rt6_info *rt1, *rt2; struct flowi6 fl1, fl2; - v6ops = nf_get_ipv6_ops(); - if (!v6ops) - return 0; - memset(&fl1, 0, sizeof(fl1)); fl1.daddr = src->in6; memset(&fl2, 0, sizeof(fl2)); fl2.daddr = dst->in6; - if (!v6ops->route(net, (struct dst_entry **)&rt1, + if (!nf_ip6_route(net, (struct dst_entry **)&rt1, flowi6_to_flowi(&fl1), false)) { - if (!v6ops->route(net, (struct dst_entry **)&rt2, + if (!nf_ip6_route(net, (struct dst_entry **)&rt2, flowi6_to_flowi(&fl2), false)) { if (ipv6_addr_equal(rt6_nexthop(rt1, &fl1.daddr), rt6_nexthop(rt2, &fl2.daddr)) && diff --git a/net/netfilter/nf_conntrack_netlink.c b/net/netfilter/nf_conntrack_netlink.c index 8dcc064d518d..7db79c1b8084 100644 --- a/net/netfilter/nf_conntrack_netlink.c +++ b/net/netfilter/nf_conntrack_netlink.c @@ -1256,7 +1256,7 @@ static int ctnetlink_del_conntrack(struct net *net, struct sock *ctnl, struct nf_conntrack_tuple tuple; struct nf_conn *ct; struct nfgenmsg *nfmsg = nlmsg_data(nlh); - u_int8_t u3 = nfmsg->nfgen_family; + u_int8_t u3 = nfmsg->version ? nfmsg->nfgen_family : AF_UNSPEC; struct nf_conntrack_zone zone; int err; diff --git a/net/netfilter/nf_flow_table_core.c b/net/netfilter/nf_flow_table_core.c index 7aabfd4b1e50..4469519a4879 100644 --- a/net/netfilter/nf_flow_table_core.c +++ b/net/netfilter/nf_flow_table_core.c @@ -185,14 +185,25 @@ static const struct rhashtable_params nf_flow_offload_rhash_params = { int flow_offload_add(struct nf_flowtable *flow_table, struct flow_offload *flow) { - flow->timeout = (u32)jiffies; + int err; - rhashtable_insert_fast(&flow_table->rhashtable, - &flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].node, - nf_flow_offload_rhash_params); - rhashtable_insert_fast(&flow_table->rhashtable, - &flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].node, - nf_flow_offload_rhash_params); + err = rhashtable_insert_fast(&flow_table->rhashtable, + &flow->tuplehash[0].node, + nf_flow_offload_rhash_params); + if (err < 0) + return err; + + err = rhashtable_insert_fast(&flow_table->rhashtable, + &flow->tuplehash[1].node, + nf_flow_offload_rhash_params); + if (err < 0) { + rhashtable_remove_fast(&flow_table->rhashtable, + &flow->tuplehash[0].node, + nf_flow_offload_rhash_params); + return err; + } + + flow->timeout = (u32)jiffies; return 0; } EXPORT_SYMBOL_GPL(flow_offload_add); @@ -232,6 +243,7 @@ flow_offload_lookup(struct nf_flowtable *flow_table, { struct flow_offload_tuple_rhash *tuplehash; struct flow_offload *flow; + struct flow_offload_entry *e; int dir; tuplehash = rhashtable_lookup(&flow_table->rhashtable, tuple, @@ -244,6 +256,10 @@ flow_offload_lookup(struct nf_flowtable *flow_table, if (flow->flags & (FLOW_OFFLOAD_DYING | FLOW_OFFLOAD_TEARDOWN)) return NULL; + e = container_of(flow, struct flow_offload_entry, flow); + if (unlikely(nf_ct_is_dying(e->ct))) + return NULL; + return tuplehash; } EXPORT_SYMBOL_GPL(flow_offload_lookup); @@ -290,8 +306,10 @@ static inline bool nf_flow_has_expired(const struct flow_offload *flow) static void nf_flow_offload_gc_step(struct flow_offload *flow, void *data) { struct nf_flowtable *flow_table = data; + struct flow_offload_entry *e; - if (nf_flow_has_expired(flow) || + e = container_of(flow, struct flow_offload_entry, flow); + if (nf_flow_has_expired(flow) || nf_ct_is_dying(e->ct) || (flow->flags & (FLOW_OFFLOAD_DYING | FLOW_OFFLOAD_TEARDOWN))) flow_offload_del(flow_table, flow); } diff --git a/net/netfilter/nf_flow_table_ip.c b/net/netfilter/nf_flow_table_ip.c index 6452550d187f..0d603e20b519 100644 --- a/net/netfilter/nf_flow_table_ip.c +++ b/net/netfilter/nf_flow_table_ip.c @@ -181,6 +181,9 @@ static int nf_flow_tuple_ip(struct sk_buff *skb, const struct net_device *dev, iph->protocol != IPPROTO_UDP) return -1; + if (iph->ttl <= 1) + return -1; + thoff = iph->ihl * 4; if (!pskb_may_pull(skb, thoff + sizeof(*ports))) return -1; @@ -408,6 +411,9 @@ static int nf_flow_tuple_ipv6(struct sk_buff *skb, const struct net_device *dev, ip6h->nexthdr != IPPROTO_UDP) return -1; + if (ip6h->hop_limit <= 1) + return -1; + thoff = sizeof(*ip6h); if (!pskb_may_pull(skb, thoff + sizeof(*ports))) return -1; diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c index d98416e83d4e..28241e82fd15 100644 --- a/net/netfilter/nf_tables_api.c +++ b/net/netfilter/nf_tables_api.c @@ -213,33 +213,33 @@ static int nft_deltable(struct nft_ctx *ctx) return err; } -static int nft_trans_chain_add(struct nft_ctx *ctx, int msg_type) +static struct nft_trans *nft_trans_chain_add(struct nft_ctx *ctx, int msg_type) { struct nft_trans *trans; trans = nft_trans_alloc(ctx, msg_type, sizeof(struct nft_trans_chain)); if (trans == NULL) - return -ENOMEM; + return ERR_PTR(-ENOMEM); if (msg_type == NFT_MSG_NEWCHAIN) nft_activate_next(ctx->net, ctx->chain); list_add_tail(&trans->list, &ctx->net->nft.commit_list); - return 0; + return trans; } static int nft_delchain(struct nft_ctx *ctx) { - int err; + struct nft_trans *trans; - err = nft_trans_chain_add(ctx, NFT_MSG_DELCHAIN); - if (err < 0) - return err; + trans = nft_trans_chain_add(ctx, NFT_MSG_DELCHAIN); + if (IS_ERR(trans)) + return PTR_ERR(trans); ctx->table->use--; nft_deactivate_next(ctx->net, ctx->chain); - return err; + return 0; } static void nft_rule_expr_activate(const struct nft_ctx *ctx, @@ -1189,6 +1189,9 @@ static int nft_dump_stats(struct sk_buff *skb, struct nft_stats __percpu *stats) u64 pkts, bytes; int cpu; + if (!stats) + return 0; + memset(&total, 0, sizeof(total)); for_each_possible_cpu(cpu) { cpu_stats = per_cpu_ptr(stats, cpu); @@ -1246,6 +1249,7 @@ static int nf_tables_fill_chain_info(struct sk_buff *skb, struct net *net, if (nft_is_base_chain(chain)) { const struct nft_base_chain *basechain = nft_base_chain(chain); const struct nf_hook_ops *ops = &basechain->ops; + struct nft_stats __percpu *stats; struct nlattr *nest; nest = nla_nest_start_noflag(skb, NFTA_CHAIN_HOOK); @@ -1267,8 +1271,9 @@ static int nf_tables_fill_chain_info(struct sk_buff *skb, struct net *net, if (nla_put_string(skb, NFTA_CHAIN_TYPE, basechain->type->name)) goto nla_put_failure; - if (rcu_access_pointer(basechain->stats) && - nft_dump_stats(skb, rcu_dereference(basechain->stats))) + stats = rcu_dereference_check(basechain->stats, + lockdep_commit_lock_is_held(net)); + if (nft_dump_stats(skb, stats)) goto nla_put_failure; } @@ -1615,6 +1620,7 @@ static int nf_tables_addchain(struct nft_ctx *ctx, u8 family, u8 genmask, struct nft_base_chain *basechain; struct nft_stats __percpu *stats; struct net *net = ctx->net; + struct nft_trans *trans; struct nft_chain *chain; struct nft_rule **rules; int err; @@ -1662,7 +1668,7 @@ static int nf_tables_addchain(struct nft_ctx *ctx, u8 family, u8 genmask, ops->dev = hook.dev; chain->flags |= NFT_BASE_CHAIN; - basechain->policy = policy; + basechain->policy = NF_ACCEPT; } else { chain = kzalloc(sizeof(*chain), GFP_KERNEL); if (chain == NULL) @@ -1698,13 +1704,18 @@ static int nf_tables_addchain(struct nft_ctx *ctx, u8 family, u8 genmask, if (err) goto err2; - err = nft_trans_chain_add(ctx, NFT_MSG_NEWCHAIN); - if (err < 0) { + trans = nft_trans_chain_add(ctx, NFT_MSG_NEWCHAIN); + if (IS_ERR(trans)) { + err = PTR_ERR(trans); rhltable_remove(&table->chains_ht, &chain->rhlhead, nft_chain_ht_params); goto err2; } + nft_trans_chain_policy(trans) = -1; + if (nft_is_base_chain(chain)) + nft_trans_chain_policy(trans) = policy; + table->use++; list_add_tail_rcu(&chain->list, &table->chains); @@ -6310,6 +6321,27 @@ static int nf_tables_validate(struct net *net) return 0; } +/* a drop policy has to be deferred until all rules have been activated, + * otherwise a large ruleset that contains a drop-policy base chain will + * cause all packets to get dropped until the full transaction has been + * processed. + * + * We defer the drop policy until the transaction has been finalized. + */ +static void nft_chain_commit_drop_policy(struct nft_trans *trans) +{ + struct nft_base_chain *basechain; + + if (nft_trans_chain_policy(trans) != NF_DROP) + return; + + if (!nft_is_base_chain(trans->ctx.chain)) + return; + + basechain = nft_base_chain(trans->ctx.chain); + basechain->policy = NF_DROP; +} + static void nft_chain_commit_update(struct nft_trans *trans) { struct nft_base_chain *basechain; @@ -6631,6 +6663,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb) nf_tables_chain_notify(&trans->ctx, NFT_MSG_NEWCHAIN); /* trans destroyed after rcu grace period */ } else { + nft_chain_commit_drop_policy(trans); nft_clear(net, trans->ctx.chain); nf_tables_chain_notify(&trans->ctx, NFT_MSG_NEWCHAIN); nft_trans_destroy(trans); diff --git a/net/netfilter/nft_flow_offload.c b/net/netfilter/nft_flow_offload.c index 6e6b9adf7d38..69d7a8439c7a 100644 --- a/net/netfilter/nft_flow_offload.c +++ b/net/netfilter/nft_flow_offload.c @@ -94,8 +94,7 @@ static void nft_flow_offload_eval(const struct nft_expr *expr, if (help) goto out; - if (ctinfo == IP_CT_NEW || - ctinfo == IP_CT_RELATED) + if (!nf_ct_is_confirmed(ct)) goto out; if (test_and_set_bit(IPS_OFFLOAD_BIT, &ct->status)) @@ -113,6 +112,7 @@ static void nft_flow_offload_eval(const struct nft_expr *expr, if (ret < 0) goto err_flow_add; + dst_release(route.tuple[!dir].dst); return; err_flow_add: diff --git a/net/openvswitch/conntrack.c b/net/openvswitch/conntrack.c index 333ec5f298fe..4c597a0bb168 100644 --- a/net/openvswitch/conntrack.c +++ b/net/openvswitch/conntrack.c @@ -1322,7 +1322,7 @@ static int ovs_ct_add_helper(struct ovs_conntrack_info *info, const char *name, return -ENOMEM; } -#ifdef CONFIG_NF_NAT_NEEDED +#if IS_ENABLED(CONFIG_NF_NAT) if (info->nat) { ret = nf_nat_helper_try_module_get(name, info->family, key->ip.proto); @@ -1811,7 +1811,7 @@ void ovs_ct_free_action(const struct nlattr *a) static void __ovs_ct_free_action(struct ovs_conntrack_info *ct_info) { if (ct_info->helper) { -#ifdef CONFIG_NF_NAT_NEEDED +#if IS_ENABLED(CONFIG_NF_NAT) if (ct_info->nat) nf_nat_helper_put(ct_info->helper); #endif diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c index 90d4e3ce00e5..fbc775fbf712 100644 --- a/net/packet/af_packet.c +++ b/net/packet/af_packet.c @@ -4598,14 +4598,29 @@ static void __exit packet_exit(void) static int __init packet_init(void) { - int rc = proto_register(&packet_proto, 0); + int rc; - if (rc != 0) + rc = proto_register(&packet_proto, 0); + if (rc) goto out; + rc = sock_register(&packet_family_ops); + if (rc) + goto out_proto; + rc = register_pernet_subsys(&packet_net_ops); + if (rc) + goto out_sock; + rc = register_netdevice_notifier(&packet_netdev_notifier); + if (rc) + goto out_pernet; - sock_register(&packet_family_ops); - register_pernet_subsys(&packet_net_ops); - register_netdevice_notifier(&packet_netdev_notifier); + return 0; + +out_pernet: + unregister_pernet_subsys(&packet_net_ops); +out_sock: + sock_unregister(PF_PACKET); +out_proto: + proto_unregister(&packet_proto); out: return rc; } diff --git a/net/qrtr/qrtr.c b/net/qrtr/qrtr.c index dd0e97f4f6c0..801872a2e7aa 100644 --- a/net/qrtr/qrtr.c +++ b/net/qrtr/qrtr.c @@ -728,12 +728,13 @@ static int qrtr_sendmsg(struct socket *sock, struct msghdr *msg, size_t len) DECLARE_SOCKADDR(struct sockaddr_qrtr *, addr, msg->msg_name); int (*enqueue_fn)(struct qrtr_node *, struct sk_buff *, int, struct sockaddr_qrtr *, struct sockaddr_qrtr *); + __le32 qrtr_type = cpu_to_le32(QRTR_TYPE_DATA); struct qrtr_sock *ipc = qrtr_sk(sock->sk); struct sock *sk = sock->sk; struct qrtr_node *node; struct sk_buff *skb; + u32 type = 0; size_t plen; - u32 type = QRTR_TYPE_DATA; int rc; if (msg->msg_flags & ~(MSG_DONTWAIT)) @@ -807,8 +808,8 @@ static int qrtr_sendmsg(struct socket *sock, struct msghdr *msg, size_t len) } /* control messages already require the type as 'command' */ - skb_copy_bits(skb, 0, &type, 4); - type = le32_to_cpu(type); + skb_copy_bits(skb, 0, &qrtr_type, 4); + type = le32_to_cpu(qrtr_type); } rc = enqueue_fn(node, skb, type, &ipc->us, addr); diff --git a/net/rds/info.c b/net/rds/info.c index e367a97a18c8..03f6fd56d237 100644 --- a/net/rds/info.c +++ b/net/rds/info.c @@ -193,7 +193,7 @@ int rds_info_getsockopt(struct socket *sock, int optname, char __user *optval, ret = -ENOMEM; goto out; } - ret = get_user_pages_fast(start, nr_pages, 1, pages); + ret = get_user_pages_fast(start, nr_pages, FOLL_WRITE, pages); if (ret != nr_pages) { if (ret > 0) nr_pages = ret; diff --git a/net/rds/rdma.c b/net/rds/rdma.c index 182ab8430594..b340ed4fc43a 100644 --- a/net/rds/rdma.c +++ b/net/rds/rdma.c @@ -158,7 +158,8 @@ static int rds_pin_pages(unsigned long user_addr, unsigned int nr_pages, { int ret; - ret = get_user_pages_fast(user_addr, nr_pages, write, pages); + ret = get_user_pages_fast(user_addr, nr_pages, write ? FOLL_WRITE : 0, + pages); if (ret >= 0 && ret < nr_pages) { while (ret--) diff --git a/net/sched/cls_matchall.c b/net/sched/cls_matchall.c index 1e98a517fb0b..db42d97a2006 100644 --- a/net/sched/cls_matchall.c +++ b/net/sched/cls_matchall.c @@ -308,6 +308,7 @@ static int mall_reoffload(struct tcf_proto *tp, bool add, tc_setup_cb_t *cb, NL_SET_ERR_MSG_MOD(extack, "Failed to setup flow action"); return err; } + return 0; } err = cb(TC_SETUP_CLSMATCHALL, &cls_mall, cb_priv); diff --git a/net/smc/smc_ib.c b/net/smc/smc_ib.c index 53f429c04843..d14ca4af6f94 100644 --- a/net/smc/smc_ib.c +++ b/net/smc/smc_ib.c @@ -146,18 +146,13 @@ out: static int smc_ib_fill_mac(struct smc_ib_device *smcibdev, u8 ibport) { const struct ib_gid_attr *attr; - int rc = 0; + int rc; attr = rdma_get_gid_attr(smcibdev->ibdev, ibport, 0); if (IS_ERR(attr)) return -ENODEV; - if (attr->ndev) - memcpy(smcibdev->mac[ibport - 1], attr->ndev->dev_addr, - ETH_ALEN); - else - rc = -ENODEV; - + rc = rdma_read_gid_l2_fields(attr, NULL, smcibdev->mac[ibport - 1]); rdma_put_gid_attr(attr); return rc; } @@ -185,6 +180,7 @@ int smc_ib_determine_gid(struct smc_ib_device *smcibdev, u8 ibport, unsigned short vlan_id, u8 gid[], u8 *sgid_index) { const struct ib_gid_attr *attr; + const struct net_device *ndev; int i; for (i = 0; i < smcibdev->pattr[ibport - 1].gid_tbl_len; i++) { @@ -192,11 +188,14 @@ int smc_ib_determine_gid(struct smc_ib_device *smcibdev, u8 ibport, if (IS_ERR(attr)) continue; - if (attr->ndev && + rcu_read_lock(); + ndev = rdma_read_gid_attr_ndev_rcu(attr); + if (!IS_ERR(ndev) && ((!vlan_id && !is_vlan_dev(attr->ndev)) || (vlan_id && is_vlan_dev(attr->ndev) && vlan_dev_vlan_id(attr->ndev) == vlan_id)) && attr->gid_type == IB_GID_TYPE_ROCE) { + rcu_read_unlock(); if (gid) memcpy(gid, &attr->gid, SMC_GID_SIZE); if (sgid_index) @@ -204,6 +203,7 @@ int smc_ib_determine_gid(struct smc_ib_device *smcibdev, u8 ibport, rdma_put_gid_attr(attr); return 0; } + rcu_read_unlock(); rdma_put_gid_attr(attr); } return -ENODEV; diff --git a/net/sunrpc/auth_gss/auth_gss.c b/net/sunrpc/auth_gss/auth_gss.c index 3fd56c0c90ae..4ce42c62458e 100644 --- a/net/sunrpc/auth_gss/auth_gss.c +++ b/net/sunrpc/auth_gss/auth_gss.c @@ -269,6 +269,7 @@ err: struct gss_upcall_msg { refcount_t count; kuid_t uid; + const char *service_name; struct rpc_pipe_msg msg; struct list_head list; struct gss_auth *auth; @@ -316,6 +317,7 @@ gss_release_msg(struct gss_upcall_msg *gss_msg) gss_put_ctx(gss_msg->ctx); rpc_destroy_wait_queue(&gss_msg->rpc_waitqueue); gss_put_auth(gss_msg->auth); + kfree_const(gss_msg->service_name); kfree(gss_msg); } @@ -410,9 +412,12 @@ gss_upcall_callback(struct rpc_task *task) gss_release_msg(gss_msg); } -static void gss_encode_v0_msg(struct gss_upcall_msg *gss_msg) +static void gss_encode_v0_msg(struct gss_upcall_msg *gss_msg, + const struct cred *cred) { - uid_t uid = from_kuid(&init_user_ns, gss_msg->uid); + struct user_namespace *userns = cred->user_ns; + + uid_t uid = from_kuid_munged(userns, gss_msg->uid); memcpy(gss_msg->databuf, &uid, sizeof(uid)); gss_msg->msg.data = gss_msg->databuf; gss_msg->msg.len = sizeof(uid); @@ -420,17 +425,31 @@ static void gss_encode_v0_msg(struct gss_upcall_msg *gss_msg) BUILD_BUG_ON(sizeof(uid) > sizeof(gss_msg->databuf)); } +static ssize_t +gss_v0_upcall(struct file *file, struct rpc_pipe_msg *msg, + char __user *buf, size_t buflen) +{ + struct gss_upcall_msg *gss_msg = container_of(msg, + struct gss_upcall_msg, + msg); + if (msg->copied == 0) + gss_encode_v0_msg(gss_msg, file->f_cred); + return rpc_pipe_generic_upcall(file, msg, buf, buflen); +} + static int gss_encode_v1_msg(struct gss_upcall_msg *gss_msg, const char *service_name, - const char *target_name) + const char *target_name, + const struct cred *cred) { + struct user_namespace *userns = cred->user_ns; struct gss_api_mech *mech = gss_msg->auth->mech; char *p = gss_msg->databuf; size_t buflen = sizeof(gss_msg->databuf); int len; len = scnprintf(p, buflen, "mech=%s uid=%d", mech->gm_name, - from_kuid(&init_user_ns, gss_msg->uid)); + from_kuid_munged(userns, gss_msg->uid)); buflen -= len; p += len; gss_msg->msg.len = len; @@ -491,6 +510,25 @@ out_overflow: return -ENOMEM; } +static ssize_t +gss_v1_upcall(struct file *file, struct rpc_pipe_msg *msg, + char __user *buf, size_t buflen) +{ + struct gss_upcall_msg *gss_msg = container_of(msg, + struct gss_upcall_msg, + msg); + int err; + if (msg->copied == 0) { + err = gss_encode_v1_msg(gss_msg, + gss_msg->service_name, + gss_msg->auth->target_name, + file->f_cred); + if (err) + return err; + } + return rpc_pipe_generic_upcall(file, msg, buf, buflen); +} + static struct gss_upcall_msg * gss_alloc_msg(struct gss_auth *gss_auth, kuid_t uid, const char *service_name) @@ -513,16 +551,14 @@ gss_alloc_msg(struct gss_auth *gss_auth, refcount_set(&gss_msg->count, 1); gss_msg->uid = uid; gss_msg->auth = gss_auth; - switch (vers) { - case 0: - gss_encode_v0_msg(gss_msg); - break; - default: - err = gss_encode_v1_msg(gss_msg, service_name, gss_auth->target_name); - if (err) + kref_get(&gss_auth->kref); + if (service_name) { + gss_msg->service_name = kstrdup_const(service_name, GFP_NOFS); + if (!gss_msg->service_name) { + err = -ENOMEM; goto err_put_pipe_version; + } } - kref_get(&gss_auth->kref); return gss_msg; err_put_pipe_version: put_pipe_version(gss_auth->net); @@ -581,8 +617,8 @@ gss_refresh_upcall(struct rpc_task *task) /* XXX: warning on the first, under the assumption we * shouldn't normally hit this case on a refresh. */ warn_gssd(); - task->tk_timeout = 15*HZ; - rpc_sleep_on(&pipe_version_rpc_waitqueue, task, NULL); + rpc_sleep_on_timeout(&pipe_version_rpc_waitqueue, + task, NULL, jiffies + (15 * HZ)); err = -EAGAIN; goto out; } @@ -595,7 +631,6 @@ gss_refresh_upcall(struct rpc_task *task) if (gss_cred->gc_upcall != NULL) rpc_sleep_on(&gss_cred->gc_upcall->rpc_waitqueue, task, NULL); else if (gss_msg->ctx == NULL && gss_msg->msg.errno >= 0) { - task->tk_timeout = 0; gss_cred->gc_upcall = gss_msg; /* gss_upcall_callback will release the reference to gss_upcall_msg */ refcount_inc(&gss_msg->count); @@ -707,7 +742,7 @@ gss_pipe_downcall(struct file *filp, const char __user *src, size_t mlen) goto err; } - uid = make_kuid(&init_user_ns, id); + uid = make_kuid(current_user_ns(), id); if (!uid_valid(uid)) { err = -EINVAL; goto err; @@ -2116,7 +2151,7 @@ static const struct rpc_credops gss_nullops = { }; static const struct rpc_pipe_ops gss_upcall_ops_v0 = { - .upcall = rpc_pipe_generic_upcall, + .upcall = gss_v0_upcall, .downcall = gss_pipe_downcall, .destroy_msg = gss_pipe_destroy_msg, .open_pipe = gss_pipe_open_v0, @@ -2124,7 +2159,7 @@ static const struct rpc_pipe_ops gss_upcall_ops_v0 = { }; static const struct rpc_pipe_ops gss_upcall_ops_v1 = { - .upcall = rpc_pipe_generic_upcall, + .upcall = gss_v1_upcall, .downcall = gss_pipe_downcall, .destroy_msg = gss_pipe_destroy_msg, .open_pipe = gss_pipe_open_v1, diff --git a/net/sunrpc/auth_gss/svcauth_gss.c b/net/sunrpc/auth_gss/svcauth_gss.c index 0c5d7896d6dd..8be2f209982b 100644 --- a/net/sunrpc/auth_gss/svcauth_gss.c +++ b/net/sunrpc/auth_gss/svcauth_gss.c @@ -474,12 +474,12 @@ static int rsc_parse(struct cache_detail *cd, * treatment so are checked for validity here.) */ /* uid */ - rsci.cred.cr_uid = make_kuid(&init_user_ns, id); + rsci.cred.cr_uid = make_kuid(current_user_ns(), id); /* gid */ if (get_int(&mesg, &id)) goto out; - rsci.cred.cr_gid = make_kgid(&init_user_ns, id); + rsci.cred.cr_gid = make_kgid(current_user_ns(), id); /* number of additional gid's */ if (get_int(&mesg, &N)) @@ -497,7 +497,7 @@ static int rsc_parse(struct cache_detail *cd, kgid_t kgid; if (get_int(&mesg, &id)) goto out; - kgid = make_kgid(&init_user_ns, id); + kgid = make_kgid(current_user_ns(), id); if (!gid_valid(kgid)) goto out; rsci.cred.cr_group_info->gid[i] = kgid; diff --git a/net/sunrpc/auth_unix.c b/net/sunrpc/auth_unix.c index d4018e5a24c5..e7df1f782b2e 100644 --- a/net/sunrpc/auth_unix.c +++ b/net/sunrpc/auth_unix.c @@ -107,6 +107,8 @@ unx_marshal(struct rpc_task *task, struct xdr_stream *xdr) __be32 *p, *cred_len, *gidarr_len; int i; struct group_info *gi = cred->cr_cred->group_info; + struct user_namespace *userns = clnt->cl_cred ? + clnt->cl_cred->user_ns : &init_user_ns; /* Credential */ @@ -122,14 +124,13 @@ unx_marshal(struct rpc_task *task, struct xdr_stream *xdr) p = xdr_reserve_space(xdr, 3 * sizeof(*p)); if (!p) goto marshal_failed; - *p++ = cpu_to_be32(from_kuid(&init_user_ns, cred->cr_cred->fsuid)); - *p++ = cpu_to_be32(from_kgid(&init_user_ns, cred->cr_cred->fsgid)); + *p++ = cpu_to_be32(from_kuid_munged(userns, cred->cr_cred->fsuid)); + *p++ = cpu_to_be32(from_kgid_munged(userns, cred->cr_cred->fsgid)); gidarr_len = p++; if (gi) for (i = 0; i < UNX_NGROUPS && i < gi->ngroups; i++) - *p++ = cpu_to_be32(from_kgid(&init_user_ns, - gi->gid[i])); + *p++ = cpu_to_be32(from_kgid_munged(userns, gi->gid[i])); *gidarr_len = cpu_to_be32(p - gidarr_len - 1); *cred_len = cpu_to_be32((p - cred_len - 1) << 2); p = xdr_reserve_space(xdr, (p - gidarr_len - 1) << 2); diff --git a/net/sunrpc/cache.c b/net/sunrpc/cache.c index 261131dfa1f1..d22328984853 100644 --- a/net/sunrpc/cache.c +++ b/net/sunrpc/cache.c @@ -40,6 +40,7 @@ static bool cache_defer_req(struct cache_req *req, struct cache_head *item); static void cache_revisit_request(struct cache_head *item); +static bool cache_listeners_exist(struct cache_detail *detail); static void cache_init(struct cache_head *h, struct cache_detail *detail) { @@ -306,7 +307,8 @@ int cache_check(struct cache_detail *detail, cache_fresh_unlocked(h, detail); break; } - } + } else if (!cache_listeners_exist(detail)) + rv = try_to_negate_entry(detail, h); } if (rv == -EAGAIN) { diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index 8ff11dc98d7f..c1f1afabd024 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -394,6 +394,7 @@ static struct rpc_clnt * rpc_new_client(const struct rpc_create_args *args, if (err) goto out_no_clid; + clnt->cl_cred = get_cred(args->cred); clnt->cl_procinfo = version->procs; clnt->cl_maxproc = version->nrprocs; clnt->cl_prog = args->prognumber ? : program->number; @@ -439,6 +440,7 @@ static struct rpc_clnt * rpc_new_client(const struct rpc_create_args *args, out_no_path: rpc_free_iostats(clnt->cl_metrics); out_no_stats: + put_cred(clnt->cl_cred); rpc_free_clid(clnt); out_no_clid: kfree(clnt); @@ -484,8 +486,11 @@ static struct rpc_clnt *rpc_create_xprt(struct rpc_create_args *args, } clnt->cl_softrtry = 1; - if (args->flags & RPC_CLNT_CREATE_HARDRTRY) + if (args->flags & (RPC_CLNT_CREATE_HARDRTRY|RPC_CLNT_CREATE_SOFTERR)) { clnt->cl_softrtry = 0; + if (args->flags & RPC_CLNT_CREATE_SOFTERR) + clnt->cl_softerr = 1; + } if (args->flags & RPC_CLNT_CREATE_AUTOBIND) clnt->cl_autobind = 1; @@ -623,10 +628,12 @@ static struct rpc_clnt *__rpc_clone_client(struct rpc_create_args *args, /* Turn off autobind on clones */ new->cl_autobind = 0; new->cl_softrtry = clnt->cl_softrtry; + new->cl_softerr = clnt->cl_softerr; new->cl_noretranstimeo = clnt->cl_noretranstimeo; new->cl_discrtry = clnt->cl_discrtry; new->cl_chatty = clnt->cl_chatty; new->cl_principal = clnt->cl_principal; + new->cl_cred = get_cred(clnt->cl_cred); return new; out_err: @@ -648,6 +655,7 @@ struct rpc_clnt *rpc_clone_client(struct rpc_clnt *clnt) .prognumber = clnt->cl_prog, .version = clnt->cl_vers, .authflavor = clnt->cl_auth->au_flavor, + .cred = clnt->cl_cred, }; return __rpc_clone_client(&args, clnt); } @@ -669,6 +677,7 @@ rpc_clone_client_set_auth(struct rpc_clnt *clnt, rpc_authflavor_t flavor) .prognumber = clnt->cl_prog, .version = clnt->cl_vers, .authflavor = flavor, + .cred = clnt->cl_cred, }; return __rpc_clone_client(&args, clnt); } @@ -827,14 +836,8 @@ void rpc_killall_tasks(struct rpc_clnt *clnt) * Spin lock all_tasks to prevent changes... */ spin_lock(&clnt->cl_lock); - list_for_each_entry(rovr, &clnt->cl_tasks, tk_task) { - if (!RPC_IS_ACTIVATED(rovr)) - continue; - if (!(rovr->tk_flags & RPC_TASK_KILLED)) { - rovr->tk_flags |= RPC_TASK_KILLED; - rpc_exit(rovr, -EIO); - } - } + list_for_each_entry(rovr, &clnt->cl_tasks, tk_task) + rpc_signal_task(rovr); spin_unlock(&clnt->cl_lock); } EXPORT_SYMBOL_GPL(rpc_killall_tasks); @@ -882,6 +885,7 @@ rpc_free_client(struct rpc_clnt *clnt) xprt_put(rcu_dereference_raw(clnt->cl_xprt)); xprt_iter_destroy(&clnt->cl_xpi); rpciod_down(); + put_cred(clnt->cl_cred); rpc_free_clid(clnt); kfree(clnt); return parent; @@ -946,6 +950,7 @@ struct rpc_clnt *rpc_bind_new_program(struct rpc_clnt *old, .prognumber = program->number, .version = vers, .authflavor = old->cl_auth->au_flavor, + .cred = old->cl_cred, }; struct rpc_clnt *clnt; int err; @@ -1007,6 +1012,8 @@ void rpc_task_set_client(struct rpc_task *task, struct rpc_clnt *clnt) atomic_inc(&clnt->cl_count); if (clnt->cl_softrtry) task->tk_flags |= RPC_TASK_SOFT; + if (clnt->cl_softerr) + task->tk_flags |= RPC_TASK_TIMEOUT; if (clnt->cl_noretranstimeo) task->tk_flags |= RPC_TASK_NO_RETRANS_TIMEOUT; if (atomic_read(&clnt->cl_swapper)) @@ -1470,22 +1477,14 @@ void rpc_force_rebind(struct rpc_clnt *clnt) } EXPORT_SYMBOL_GPL(rpc_force_rebind); -/* - * Restart an (async) RPC call from the call_prepare state. - * Usually called from within the exit handler. - */ -int -rpc_restart_call_prepare(struct rpc_task *task) +static int +__rpc_restart_call(struct rpc_task *task, void (*action)(struct rpc_task *)) { - if (RPC_ASSASSINATED(task)) - return 0; - task->tk_action = call_start; task->tk_status = 0; - if (task->tk_ops->rpc_call_prepare != NULL) - task->tk_action = rpc_prepare_task; + task->tk_rpc_status = 0; + task->tk_action = action; return 1; } -EXPORT_SYMBOL_GPL(rpc_restart_call_prepare); /* * Restart an (async) RPC call. Usually called from within the @@ -1494,14 +1493,23 @@ EXPORT_SYMBOL_GPL(rpc_restart_call_prepare); int rpc_restart_call(struct rpc_task *task) { - if (RPC_ASSASSINATED(task)) - return 0; - task->tk_action = call_start; - task->tk_status = 0; - return 1; + return __rpc_restart_call(task, call_start); } EXPORT_SYMBOL_GPL(rpc_restart_call); +/* + * Restart an (async) RPC call from the call_prepare state. + * Usually called from within the exit handler. + */ +int +rpc_restart_call_prepare(struct rpc_task *task) +{ + if (task->tk_ops->rpc_call_prepare != NULL) + return __rpc_restart_call(task, rpc_prepare_task); + return rpc_restart_call(task); +} +EXPORT_SYMBOL_GPL(rpc_restart_call_prepare); + const char *rpc_proc_name(const struct rpc_task *task) { @@ -1516,6 +1524,19 @@ const char return "no proc"; } +static void +__rpc_call_rpcerror(struct rpc_task *task, int tk_status, int rpc_status) +{ + task->tk_rpc_status = rpc_status; + rpc_exit(task, tk_status); +} + +static void +rpc_call_rpcerror(struct rpc_task *task, int status) +{ + __rpc_call_rpcerror(task, status, status); +} + /* * 0. Initial state * @@ -1580,7 +1601,7 @@ call_reserveresult(struct rpc_task *task) printk(KERN_ERR "%s: status=%d, but no request slot, exiting\n", __func__, status); - rpc_exit(task, -EIO); + rpc_call_rpcerror(task, -EIO); return; } @@ -1608,7 +1629,7 @@ call_reserveresult(struct rpc_task *task) __func__, status); break; } - rpc_exit(task, status); + rpc_call_rpcerror(task, status); } /* @@ -1676,7 +1697,7 @@ call_refreshresult(struct rpc_task *task) } dprintk("RPC: %5u %s: refresh creds failed with error %d\n", task->tk_pid, __func__, status); - rpc_exit(task, status); + rpc_call_rpcerror(task, status); } /* @@ -1727,7 +1748,7 @@ call_allocate(struct rpc_task *task) if (status == 0) return; if (status != -ENOMEM) { - rpc_exit(task, status); + rpc_call_rpcerror(task, status); return; } @@ -1793,10 +1814,17 @@ call_encode(struct rpc_task *task) rpc_delay(task, HZ >> 4); break; case -EKEYEXPIRED: - task->tk_action = call_refresh; + if (!task->tk_cred_retry) { + rpc_exit(task, task->tk_status); + } else { + task->tk_action = call_refresh; + task->tk_cred_retry--; + dprintk("RPC: %5u %s: retry refresh creds\n", + task->tk_pid, __func__); + } break; default: - rpc_exit(task, task->tk_status); + rpc_call_rpcerror(task, task->tk_status); } return; } else { @@ -1857,7 +1885,6 @@ call_bind(struct rpc_task *task) if (!xprt_prepare_transmit(task)) return; - task->tk_timeout = xprt->bind_timeout; xprt->ops->rpcbind(task); } @@ -1938,7 +1965,7 @@ call_bind_status(struct rpc_task *task) task->tk_pid, -task->tk_status); } - rpc_exit(task, status); + rpc_call_rpcerror(task, status); return; retry_timeout: @@ -1973,7 +2000,7 @@ call_connect(struct rpc_task *task) if (task->tk_status < 0) return; if (task->tk_flags & RPC_TASK_NOCONNECT) { - rpc_exit(task, -ENOTCONN); + rpc_call_rpcerror(task, -ENOTCONN); return; } if (!xprt_prepare_transmit(task)) @@ -2033,7 +2060,7 @@ call_connect_status(struct rpc_task *task) task->tk_action = call_transmit; return; } - rpc_exit(task, status); + rpc_call_rpcerror(task, status); return; out_retry: /* Check for timeouts before looping back to call_bind */ @@ -2118,7 +2145,7 @@ call_transmit_status(struct rpc_task *task) if (!task->tk_msg.rpc_proc->p_proc) trace_xprt_ping(task->tk_xprt, task->tk_status); - rpc_exit(task, task->tk_status); + rpc_call_rpcerror(task, task->tk_status); return; } /* fall through */ @@ -2282,7 +2309,7 @@ call_status(struct rpc_task *task) rpc_check_timeout(task); return; out_exit: - rpc_exit(task, status); + rpc_call_rpcerror(task, status); } static bool @@ -2306,29 +2333,40 @@ rpc_check_timeout(struct rpc_task *task) task->tk_timeouts++; if (RPC_IS_SOFTCONN(task) && !rpc_check_connected(task->tk_rqstp)) { - rpc_exit(task, -ETIMEDOUT); + rpc_call_rpcerror(task, -ETIMEDOUT); return; } if (RPC_IS_SOFT(task)) { + /* + * Once a "no retrans timeout" soft tasks (a.k.a NFSv4) has + * been sent, it should time out only if the transport + * connection gets terminally broken. + */ + if ((task->tk_flags & RPC_TASK_NO_RETRANS_TIMEOUT) && + rpc_check_connected(task->tk_rqstp)) + return; + if (clnt->cl_chatty) { - printk(KERN_NOTICE "%s: server %s not responding, timed out\n", + pr_notice_ratelimited( + "%s: server %s not responding, timed out\n", clnt->cl_program->name, task->tk_xprt->servername); } if (task->tk_flags & RPC_TASK_TIMEOUT) - rpc_exit(task, -ETIMEDOUT); + rpc_call_rpcerror(task, -ETIMEDOUT); else - rpc_exit(task, -EIO); + __rpc_call_rpcerror(task, -EIO, -ETIMEDOUT); return; } if (!(task->tk_flags & RPC_CALL_MAJORSEEN)) { task->tk_flags |= RPC_CALL_MAJORSEEN; if (clnt->cl_chatty) { - printk(KERN_NOTICE "%s: server %s not responding, still trying\n", - clnt->cl_program->name, - task->tk_xprt->servername); + pr_notice_ratelimited( + "%s: server %s not responding, still trying\n", + clnt->cl_program->name, + task->tk_xprt->servername); } } rpc_force_rebind(clnt); @@ -2358,7 +2396,7 @@ call_decode(struct rpc_task *task) if (task->tk_flags & RPC_CALL_MAJORSEEN) { if (clnt->cl_chatty) { - printk(KERN_NOTICE "%s: server %s OK\n", + pr_notice_ratelimited("%s: server %s OK\n", clnt->cl_program->name, task->tk_xprt->servername); } @@ -2881,7 +2919,7 @@ static void rpc_show_task(const struct rpc_clnt *clnt, printk(KERN_INFO "%5u %04x %6d %8p %8p %8ld %8p %sv%u %s a:%ps q:%s\n", task->tk_pid, task->tk_flags, task->tk_status, - clnt, task->tk_rqstp, task->tk_timeout, task->tk_ops, + clnt, task->tk_rqstp, rpc_task_timeout(task), task->tk_ops, clnt->cl_program->name, clnt->cl_vers, rpc_proc_name(task), task->tk_action, rpc_waitq); } diff --git a/net/sunrpc/debugfs.c b/net/sunrpc/debugfs.c index 19bb356230ed..95ebd76b132d 100644 --- a/net/sunrpc/debugfs.c +++ b/net/sunrpc/debugfs.c @@ -33,7 +33,7 @@ tasks_show(struct seq_file *f, void *v) seq_printf(f, "%5u %04x %6d 0x%x 0x%x %8ld %ps %sv%u %s a:%ps q:%s\n", task->tk_pid, task->tk_flags, task->tk_status, - clnt->cl_clid, xid, task->tk_timeout, task->tk_ops, + clnt->cl_clid, xid, rpc_task_timeout(task), task->tk_ops, clnt->cl_program->name, clnt->cl_vers, rpc_proc_name(task), task->tk_action, rpc_waitq); return 0; diff --git a/net/sunrpc/rpcb_clnt.c b/net/sunrpc/rpcb_clnt.c index 41a971ac1c63..2277b7cdad27 100644 --- a/net/sunrpc/rpcb_clnt.c +++ b/net/sunrpc/rpcb_clnt.c @@ -240,6 +240,7 @@ static int rpcb_create_local_unix(struct net *net) .program = &rpcb_program, .version = RPCBVERS_2, .authflavor = RPC_AUTH_NULL, + .cred = current_cred(), /* * We turn off the idle timeout to prevent the kernel * from automatically disconnecting the socket. @@ -299,6 +300,7 @@ static int rpcb_create_local_net(struct net *net) .program = &rpcb_program, .version = RPCBVERS_2, .authflavor = RPC_AUTH_UNIX, + .cred = current_cred(), .flags = RPC_CLNT_CREATE_NOPING, }; struct rpc_clnt *clnt, *clnt4; @@ -358,7 +360,8 @@ out: static struct rpc_clnt *rpcb_create(struct net *net, const char *nodename, const char *hostname, struct sockaddr *srvaddr, size_t salen, - int proto, u32 version) + int proto, u32 version, + const struct cred *cred) { struct rpc_create_args args = { .net = net, @@ -370,6 +373,7 @@ static struct rpc_clnt *rpcb_create(struct net *net, const char *nodename, .program = &rpcb_program, .version = version, .authflavor = RPC_AUTH_UNIX, + .cred = cred, .flags = (RPC_CLNT_CREATE_NOPING | RPC_CLNT_CREATE_NONPRIVPORT), }; @@ -694,7 +698,8 @@ void rpcb_getport_async(struct rpc_task *task) /* Put self on the wait queue to ensure we get notified if * some other task is already attempting to bind the port */ - rpc_sleep_on(&xprt->binding, task, NULL); + rpc_sleep_on_timeout(&xprt->binding, task, + NULL, jiffies + xprt->bind_timeout); if (xprt_test_and_set_binding(xprt)) { dprintk("RPC: %5u %s: waiting for another binder\n", @@ -744,7 +749,8 @@ void rpcb_getport_async(struct rpc_task *task) rpcb_clnt = rpcb_create(xprt->xprt_net, clnt->cl_nodename, xprt->servername, sap, salen, - xprt->prot, bind_version); + xprt->prot, bind_version, + clnt->cl_cred); if (IS_ERR(rpcb_clnt)) { status = PTR_ERR(rpcb_clnt); dprintk("RPC: %5u %s: rpcb_create failed, error %ld\n", diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c index 28956c70100a..1a12fb03e611 100644 --- a/net/sunrpc/sched.c +++ b/net/sunrpc/sched.c @@ -58,6 +58,20 @@ static struct rpc_wait_queue delay_queue; struct workqueue_struct *rpciod_workqueue __read_mostly; struct workqueue_struct *xprtiod_workqueue __read_mostly; +unsigned long +rpc_task_timeout(const struct rpc_task *task) +{ + unsigned long timeout = READ_ONCE(task->tk_timeout); + + if (timeout != 0) { + unsigned long now = jiffies; + if (time_before(now, timeout)) + return timeout - now; + } + return 0; +} +EXPORT_SYMBOL_GPL(rpc_task_timeout); + /* * Disable the timer for a given RPC task. Should be called with * queue->lock and bh_disabled in order to avoid races within @@ -66,7 +80,7 @@ struct workqueue_struct *xprtiod_workqueue __read_mostly; static void __rpc_disable_timer(struct rpc_wait_queue *queue, struct rpc_task *task) { - if (task->tk_timeout == 0) + if (list_empty(&task->u.tk_wait.timer_list)) return; dprintk("RPC: %5u disabling timer\n", task->tk_pid); task->tk_timeout = 0; @@ -78,25 +92,21 @@ __rpc_disable_timer(struct rpc_wait_queue *queue, struct rpc_task *task) static void rpc_set_queue_timer(struct rpc_wait_queue *queue, unsigned long expires) { - queue->timer_list.expires = expires; - mod_timer(&queue->timer_list.timer, expires); + timer_reduce(&queue->timer_list.timer, expires); } /* * Set up a timer for the current task. */ static void -__rpc_add_timer(struct rpc_wait_queue *queue, struct rpc_task *task) +__rpc_add_timer(struct rpc_wait_queue *queue, struct rpc_task *task, + unsigned long timeout) { - if (!task->tk_timeout) - return; - dprintk("RPC: %5u setting alarm for %u ms\n", - task->tk_pid, jiffies_to_msecs(task->tk_timeout)); + task->tk_pid, jiffies_to_msecs(timeout - jiffies)); - task->u.tk_wait.expires = jiffies + task->tk_timeout; - if (list_empty(&queue->timer_list.list) || time_before(task->u.tk_wait.expires, queue->timer_list.expires)) - rpc_set_queue_timer(queue, task->u.tk_wait.expires); + task->tk_timeout = timeout; + rpc_set_queue_timer(queue, timeout); list_add(&task->u.tk_wait.timer_list, &queue->timer_list.list); } @@ -188,6 +198,7 @@ static void __rpc_add_wait_queue(struct rpc_wait_queue *queue, if (RPC_IS_QUEUED(task)) return; + INIT_LIST_HEAD(&task->u.tk_wait.timer_list); if (RPC_IS_PRIORITY(queue)) __rpc_add_wait_queue_priority(queue, task, queue_priority); else if (RPC_IS_SWAPPER(task)) @@ -238,7 +249,9 @@ static void __rpc_init_priority_wait_queue(struct rpc_wait_queue *queue, const c queue->maxpriority = nr_queues - 1; rpc_reset_waitqueue_priority(queue); queue->qlen = 0; - timer_setup(&queue->timer_list.timer, __rpc_queue_timer_fn, 0); + timer_setup(&queue->timer_list.timer, + __rpc_queue_timer_fn, + TIMER_DEFERRABLE); INIT_LIST_HEAD(&queue->timer_list.list); rpc_assign_waitqueue_name(queue, qname); } @@ -362,7 +375,6 @@ static void rpc_make_runnable(struct workqueue_struct *wq, */ static void __rpc_sleep_on_priority(struct rpc_wait_queue *q, struct rpc_task *task, - rpc_action action, unsigned char queue_priority) { dprintk("RPC: %5u sleep_on(queue \"%s\" time %lu)\n", @@ -372,47 +384,100 @@ static void __rpc_sleep_on_priority(struct rpc_wait_queue *q, __rpc_add_wait_queue(q, task, queue_priority); - WARN_ON_ONCE(task->tk_callback != NULL); - task->tk_callback = action; - __rpc_add_timer(q, task); } -void rpc_sleep_on(struct rpc_wait_queue *q, struct rpc_task *task, - rpc_action action) +static void __rpc_sleep_on_priority_timeout(struct rpc_wait_queue *q, + struct rpc_task *task, unsigned long timeout, + unsigned char queue_priority) +{ + if (time_is_after_jiffies(timeout)) { + __rpc_sleep_on_priority(q, task, queue_priority); + __rpc_add_timer(q, task, timeout); + } else + task->tk_status = -ETIMEDOUT; +} + +static void rpc_set_tk_callback(struct rpc_task *task, rpc_action action) +{ + if (action && !WARN_ON_ONCE(task->tk_callback != NULL)) + task->tk_callback = action; +} + +static bool rpc_sleep_check_activated(struct rpc_task *task) { /* We shouldn't ever put an inactive task to sleep */ - WARN_ON_ONCE(!RPC_IS_ACTIVATED(task)); - if (!RPC_IS_ACTIVATED(task)) { + if (WARN_ON_ONCE(!RPC_IS_ACTIVATED(task))) { task->tk_status = -EIO; rpc_put_task_async(task); - return; + return false; } + return true; +} + +void rpc_sleep_on_timeout(struct rpc_wait_queue *q, struct rpc_task *task, + rpc_action action, unsigned long timeout) +{ + if (!rpc_sleep_check_activated(task)) + return; + + rpc_set_tk_callback(task, action); + + /* + * Protect the queue operations. + */ + spin_lock_bh(&q->lock); + __rpc_sleep_on_priority_timeout(q, task, timeout, task->tk_priority); + spin_unlock_bh(&q->lock); +} +EXPORT_SYMBOL_GPL(rpc_sleep_on_timeout); +void rpc_sleep_on(struct rpc_wait_queue *q, struct rpc_task *task, + rpc_action action) +{ + if (!rpc_sleep_check_activated(task)) + return; + + rpc_set_tk_callback(task, action); + + WARN_ON_ONCE(task->tk_timeout != 0); /* * Protect the queue operations. */ spin_lock_bh(&q->lock); - __rpc_sleep_on_priority(q, task, action, task->tk_priority); + __rpc_sleep_on_priority(q, task, task->tk_priority); spin_unlock_bh(&q->lock); } EXPORT_SYMBOL_GPL(rpc_sleep_on); +void rpc_sleep_on_priority_timeout(struct rpc_wait_queue *q, + struct rpc_task *task, unsigned long timeout, int priority) +{ + if (!rpc_sleep_check_activated(task)) + return; + + priority -= RPC_PRIORITY_LOW; + /* + * Protect the queue operations. + */ + spin_lock_bh(&q->lock); + __rpc_sleep_on_priority_timeout(q, task, timeout, priority); + spin_unlock_bh(&q->lock); +} +EXPORT_SYMBOL_GPL(rpc_sleep_on_priority_timeout); + void rpc_sleep_on_priority(struct rpc_wait_queue *q, struct rpc_task *task, - rpc_action action, int priority) + int priority) { - /* We shouldn't ever put an inactive task to sleep */ - WARN_ON_ONCE(!RPC_IS_ACTIVATED(task)); - if (!RPC_IS_ACTIVATED(task)) { - task->tk_status = -EIO; - rpc_put_task_async(task); + if (!rpc_sleep_check_activated(task)) return; - } + WARN_ON_ONCE(task->tk_timeout != 0); + priority -= RPC_PRIORITY_LOW; /* * Protect the queue operations. */ spin_lock_bh(&q->lock); - __rpc_sleep_on_priority(q, task, action, priority - RPC_PRIORITY_LOW); + __rpc_sleep_on_priority(q, task, priority); spin_unlock_bh(&q->lock); } EXPORT_SYMBOL_GPL(rpc_sleep_on_priority); @@ -704,7 +769,7 @@ static void __rpc_queue_timer_fn(struct timer_list *t) spin_lock(&queue->lock); expires = now = jiffies; list_for_each_entry_safe(task, n, &queue->timer_list.list, u.tk_wait.timer_list) { - timeo = task->u.tk_wait.expires; + timeo = task->tk_timeout; if (time_after_eq(now, timeo)) { dprintk("RPC: %5u timeout\n", task->tk_pid); task->tk_status = -ETIMEDOUT; @@ -730,8 +795,7 @@ static void __rpc_atrun(struct rpc_task *task) */ void rpc_delay(struct rpc_task *task, unsigned long delay) { - task->tk_timeout = delay; - rpc_sleep_on(&delay_queue, task, __rpc_atrun); + rpc_sleep_on_timeout(&delay_queue, task, __rpc_atrun, jiffies + delay); } EXPORT_SYMBOL_GPL(rpc_delay); @@ -759,8 +823,7 @@ static void rpc_reset_task_statistics(struct rpc_task *task) { task->tk_timeouts = 0; - task->tk_flags &= ~(RPC_CALL_MAJORSEEN|RPC_TASK_KILLED|RPC_TASK_SENT); - + task->tk_flags &= ~(RPC_CALL_MAJORSEEN|RPC_TASK_SENT); rpc_init_task_statistics(task); } @@ -773,7 +836,6 @@ void rpc_exit_task(struct rpc_task *task) if (task->tk_ops->rpc_call_done != NULL) { task->tk_ops->rpc_call_done(task, task->tk_calldata); if (task->tk_action != NULL) { - WARN_ON(RPC_ASSASSINATED(task)); /* Always release the RPC slot and buffer memory */ xprt_release(task); rpc_reset_task_statistics(task); @@ -781,6 +843,19 @@ void rpc_exit_task(struct rpc_task *task) } } +void rpc_signal_task(struct rpc_task *task) +{ + struct rpc_wait_queue *queue; + + if (!RPC_IS_ACTIVATED(task)) + return; + set_bit(RPC_TASK_SIGNALLED, &task->tk_runstate); + smp_mb__after_atomic(); + queue = READ_ONCE(task->tk_waitqueue); + if (queue) + rpc_wake_up_queued_task_set_status(queue, task, -ERESTARTSYS); +} + void rpc_exit(struct rpc_task *task, int status) { task->tk_status = status; @@ -836,6 +911,13 @@ static void __rpc_execute(struct rpc_task *task) */ if (!RPC_IS_QUEUED(task)) continue; + + /* + * Signalled tasks should exit rather than sleep. + */ + if (RPC_SIGNALLED(task)) + rpc_exit(task, -ERESTARTSYS); + /* * The queue->lock protects against races with * rpc_make_runnable(). @@ -861,7 +943,7 @@ static void __rpc_execute(struct rpc_task *task) status = out_of_line_wait_on_bit(&task->tk_runstate, RPC_TASK_QUEUED, rpc_wait_bit_killable, TASK_KILLABLE); - if (status == -ERESTARTSYS) { + if (status < 0) { /* * When a sync task receives a signal, it exits with * -ERESTARTSYS. In order to catch any callbacks that @@ -869,7 +951,7 @@ static void __rpc_execute(struct rpc_task *task) * break the loop here, but go around once more. */ dprintk("RPC: %5u got signal\n", task->tk_pid); - task->tk_flags |= RPC_TASK_KILLED; + set_bit(RPC_TASK_SIGNALLED, &task->tk_runstate); rpc_exit(task, -ERESTARTSYS); } dprintk("RPC: %5u sync task resuming\n", task->tk_pid); diff --git a/net/sunrpc/socklib.c b/net/sunrpc/socklib.c index 7e55cfc69697..9faea12624a6 100644 --- a/net/sunrpc/socklib.c +++ b/net/sunrpc/socklib.c @@ -106,7 +106,7 @@ xdr_partial_copy_from_skb(struct xdr_buf *xdr, unsigned int base, struct xdr_skb /* ACL likes to be lazy in allocating pages - ACLs * are small by default but can get huge. */ if ((xdr->flags & XDRBUF_SPARSE_PAGES) && *ppage == NULL) { - *ppage = alloc_page(GFP_ATOMIC); + *ppage = alloc_page(GFP_NOWAIT | __GFP_NOWARN); if (unlikely(*ppage == NULL)) { if (copied == 0) copied = -ENOMEM; diff --git a/net/sunrpc/svc.c b/net/sunrpc/svc.c index dbd19697ee38..2be827820247 100644 --- a/net/sunrpc/svc.c +++ b/net/sunrpc/svc.c @@ -993,6 +993,58 @@ static int __svc_register(struct net *net, const char *progname, return error; } +int svc_rpcbind_set_version(struct net *net, + const struct svc_program *progp, + u32 version, int family, + unsigned short proto, + unsigned short port) +{ + dprintk("svc: svc_register(%sv%d, %s, %u, %u)\n", + progp->pg_name, version, + proto == IPPROTO_UDP? "udp" : "tcp", + port, family); + + return __svc_register(net, progp->pg_name, progp->pg_prog, + version, family, proto, port); + +} +EXPORT_SYMBOL_GPL(svc_rpcbind_set_version); + +int svc_generic_rpcbind_set(struct net *net, + const struct svc_program *progp, + u32 version, int family, + unsigned short proto, + unsigned short port) +{ + const struct svc_version *vers = progp->pg_vers[version]; + int error; + + if (vers == NULL) + return 0; + + if (vers->vs_hidden) { + dprintk("svc: svc_register(%sv%d, %s, %u, %u)" + " (but not telling portmap)\n", + progp->pg_name, version, + proto == IPPROTO_UDP? "udp" : "tcp", + port, family); + return 0; + } + + /* + * Don't register a UDP port if we need congestion + * control. + */ + if (vers->vs_need_cong_ctrl && proto == IPPROTO_UDP) + return 0; + + error = svc_rpcbind_set_version(net, progp, version, + family, proto, port); + + return (vers->vs_rpcb_optnl) ? 0 : error; +} +EXPORT_SYMBOL_GPL(svc_generic_rpcbind_set); + /** * svc_register - register an RPC service with the local portmapper * @serv: svc_serv struct for the service to register @@ -1008,7 +1060,6 @@ int svc_register(const struct svc_serv *serv, struct net *net, const unsigned short port) { struct svc_program *progp; - const struct svc_version *vers; unsigned int i; int error = 0; @@ -1018,37 +1069,9 @@ int svc_register(const struct svc_serv *serv, struct net *net, for (progp = serv->sv_program; progp; progp = progp->pg_next) { for (i = 0; i < progp->pg_nvers; i++) { - vers = progp->pg_vers[i]; - if (vers == NULL) - continue; - - dprintk("svc: svc_register(%sv%d, %s, %u, %u)%s\n", - progp->pg_name, - i, - proto == IPPROTO_UDP? "udp" : "tcp", - port, - family, - vers->vs_hidden ? - " (but not telling portmap)" : ""); - - if (vers->vs_hidden) - continue; - - /* - * Don't register a UDP port if we need congestion - * control. - */ - if (vers->vs_need_cong_ctrl && proto == IPPROTO_UDP) - continue; - - error = __svc_register(net, progp->pg_name, progp->pg_prog, - i, family, proto, port); - - if (vers->vs_rpcb_optnl) { - error = 0; - continue; - } + error = progp->pg_rpcbind_set(net, progp, i, + family, proto, port); if (error < 0) { printk(KERN_WARNING "svc: failed to register " "%sv%u RPC service (errno %d).\n", @@ -1144,6 +1167,114 @@ void svc_printk(struct svc_rqst *rqstp, const char *fmt, ...) static __printf(2,3) void svc_printk(struct svc_rqst *rqstp, const char *fmt, ...) {} #endif +__be32 +svc_return_autherr(struct svc_rqst *rqstp, __be32 auth_err) +{ + set_bit(RQ_AUTHERR, &rqstp->rq_flags); + return auth_err; +} +EXPORT_SYMBOL_GPL(svc_return_autherr); + +static __be32 +svc_get_autherr(struct svc_rqst *rqstp, __be32 *statp) +{ + if (test_and_clear_bit(RQ_AUTHERR, &rqstp->rq_flags)) + return *statp; + return rpc_auth_ok; +} + +static int +svc_generic_dispatch(struct svc_rqst *rqstp, __be32 *statp) +{ + struct kvec *argv = &rqstp->rq_arg.head[0]; + struct kvec *resv = &rqstp->rq_res.head[0]; + const struct svc_procedure *procp = rqstp->rq_procinfo; + + /* + * Decode arguments + * XXX: why do we ignore the return value? + */ + if (procp->pc_decode && + !procp->pc_decode(rqstp, argv->iov_base)) { + *statp = rpc_garbage_args; + return 1; + } + + *statp = procp->pc_func(rqstp); + + if (*statp == rpc_drop_reply || + test_bit(RQ_DROPME, &rqstp->rq_flags)) + return 0; + + if (test_bit(RQ_AUTHERR, &rqstp->rq_flags)) + return 1; + + if (*statp != rpc_success) + return 1; + + /* Encode reply */ + if (procp->pc_encode && + !procp->pc_encode(rqstp, resv->iov_base + resv->iov_len)) { + dprintk("svc: failed to encode reply\n"); + /* serv->sv_stats->rpcsystemerr++; */ + *statp = rpc_system_err; + } + return 1; +} + +__be32 +svc_generic_init_request(struct svc_rqst *rqstp, + const struct svc_program *progp, + struct svc_process_info *ret) +{ + const struct svc_version *versp = NULL; /* compiler food */ + const struct svc_procedure *procp = NULL; + + if (rqstp->rq_vers >= progp->pg_nvers ) + goto err_bad_vers; + versp = progp->pg_vers[rqstp->rq_vers]; + if (!versp) + goto err_bad_vers; + + /* + * Some protocol versions (namely NFSv4) require some form of + * congestion control. (See RFC 7530 section 3.1 paragraph 2) + * In other words, UDP is not allowed. We mark those when setting + * up the svc_xprt, and verify that here. + * + * The spec is not very clear about what error should be returned + * when someone tries to access a server that is listening on UDP + * for lower versions. RPC_PROG_MISMATCH seems to be the closest + * fit. + */ + if (versp->vs_need_cong_ctrl && rqstp->rq_xprt && + !test_bit(XPT_CONG_CTRL, &rqstp->rq_xprt->xpt_flags)) + goto err_bad_vers; + + if (rqstp->rq_proc >= versp->vs_nproc) + goto err_bad_proc; + rqstp->rq_procinfo = procp = &versp->vs_proc[rqstp->rq_proc]; + if (!procp) + goto err_bad_proc; + + /* Initialize storage for argp and resp */ + memset(rqstp->rq_argp, 0, procp->pc_argsize); + memset(rqstp->rq_resp, 0, procp->pc_ressize); + + /* Bump per-procedure stats counter */ + versp->vs_count[rqstp->rq_proc]++; + + ret->dispatch = versp->vs_dispatch; + return rpc_success; +err_bad_vers: + ret->mismatch.lovers = progp->pg_lovers; + ret->mismatch.hivers = progp->pg_hivers; + return rpc_prog_mismatch; +err_bad_proc: + return rpc_proc_unavail; +} +EXPORT_SYMBOL_GPL(svc_generic_init_request); + /* * Common routine for processing the RPC request. */ @@ -1151,11 +1282,11 @@ static int svc_process_common(struct svc_rqst *rqstp, struct kvec *argv, struct kvec *resv) { struct svc_program *progp; - const struct svc_version *versp = NULL; /* compiler food */ const struct svc_procedure *procp = NULL; struct svc_serv *serv = rqstp->rq_server; + struct svc_process_info process; __be32 *statp; - u32 prog, vers, proc; + u32 prog, vers; __be32 auth_stat, rpc_stat; int auth_res; __be32 *reply_statp; @@ -1187,8 +1318,8 @@ svc_process_common(struct svc_rqst *rqstp, struct kvec *argv, struct kvec *resv) svc_putnl(resv, 0); /* ACCEPT */ rqstp->rq_prog = prog = svc_getnl(argv); /* program number */ - rqstp->rq_vers = vers = svc_getnl(argv); /* version number */ - rqstp->rq_proc = proc = svc_getnl(argv); /* procedure number */ + rqstp->rq_vers = svc_getnl(argv); /* version number */ + rqstp->rq_proc = svc_getnl(argv); /* procedure number */ for (progp = serv->sv_program; progp; progp = progp->pg_next) if (prog == progp->pg_prog) @@ -1226,29 +1357,22 @@ svc_process_common(struct svc_rqst *rqstp, struct kvec *argv, struct kvec *resv) if (progp == NULL) goto err_bad_prog; - if (vers >= progp->pg_nvers || - !(versp = progp->pg_vers[vers])) - goto err_bad_vers; - - /* - * Some protocol versions (namely NFSv4) require some form of - * congestion control. (See RFC 7530 section 3.1 paragraph 2) - * In other words, UDP is not allowed. We mark those when setting - * up the svc_xprt, and verify that here. - * - * The spec is not very clear about what error should be returned - * when someone tries to access a server that is listening on UDP - * for lower versions. RPC_PROG_MISMATCH seems to be the closest - * fit. - */ - if (versp->vs_need_cong_ctrl && rqstp->rq_xprt && - !test_bit(XPT_CONG_CTRL, &rqstp->rq_xprt->xpt_flags)) + rpc_stat = progp->pg_init_request(rqstp, progp, &process); + switch (rpc_stat) { + case rpc_success: + break; + case rpc_prog_unavail: + goto err_bad_prog; + case rpc_prog_mismatch: goto err_bad_vers; + case rpc_proc_unavail: + goto err_bad_proc; + } - procp = versp->vs_proc + proc; - if (proc >= versp->vs_nproc || !procp->pc_func) + procp = rqstp->rq_procinfo; + /* Should this check go into the dispatcher? */ + if (!procp || !procp->pc_func) goto err_bad_proc; - rqstp->rq_procinfo = procp; /* Syntactic check complete */ serv->sv_stats->rpccnt++; @@ -1258,13 +1382,6 @@ svc_process_common(struct svc_rqst *rqstp, struct kvec *argv, struct kvec *resv) statp = resv->iov_base +resv->iov_len; svc_putnl(resv, RPC_SUCCESS); - /* Bump per-procedure stats counter */ - versp->vs_count[proc]++; - - /* Initialize storage for argp and resp */ - memset(rqstp->rq_argp, 0, procp->pc_argsize); - memset(rqstp->rq_resp, 0, procp->pc_ressize); - /* un-reserve some of the out-queue now that we have a * better idea of reply size */ @@ -1272,43 +1389,18 @@ svc_process_common(struct svc_rqst *rqstp, struct kvec *argv, struct kvec *resv) svc_reserve_auth(rqstp, procp->pc_xdrressize<<2); /* Call the function that processes the request. */ - if (!versp->vs_dispatch) { - /* - * Decode arguments - * XXX: why do we ignore the return value? - */ - if (procp->pc_decode && - !procp->pc_decode(rqstp, argv->iov_base)) + if (!process.dispatch) { + if (!svc_generic_dispatch(rqstp, statp)) + goto release_dropit; + if (*statp == rpc_garbage_args) goto err_garbage; - - *statp = procp->pc_func(rqstp); - - /* Encode reply */ - if (*statp == rpc_drop_reply || - test_bit(RQ_DROPME, &rqstp->rq_flags)) { - if (procp->pc_release) - procp->pc_release(rqstp); - goto dropit; - } - if (*statp == rpc_autherr_badcred) { - if (procp->pc_release) - procp->pc_release(rqstp); - goto err_bad_auth; - } - if (*statp == rpc_success && procp->pc_encode && - !procp->pc_encode(rqstp, resv->iov_base + resv->iov_len)) { - dprintk("svc: failed to encode reply\n"); - /* serv->sv_stats->rpcsystemerr++; */ - *statp = rpc_system_err; - } + auth_stat = svc_get_autherr(rqstp, statp); + if (auth_stat != rpc_auth_ok) + goto err_release_bad_auth; } else { dprintk("svc: calling dispatcher\n"); - if (!versp->vs_dispatch(rqstp, statp)) { - /* Release reply info */ - if (procp->pc_release) - procp->pc_release(rqstp); - goto dropit; - } + if (!process.dispatch(rqstp, statp)) + goto release_dropit; /* Release reply info */ } /* Check RPC status result */ @@ -1327,6 +1419,9 @@ svc_process_common(struct svc_rqst *rqstp, struct kvec *argv, struct kvec *resv) goto close; return 1; /* Caller can now send it */ +release_dropit: + if (procp->pc_release) + procp->pc_release(rqstp); dropit: svc_authorise(rqstp); /* doesn't hurt to call this twice */ dprintk("svc: svc_process dropit\n"); @@ -1351,6 +1446,9 @@ err_bad_rpc: svc_putnl(resv, 2); goto sendit; +err_release_bad_auth: + if (procp->pc_release) + procp->pc_release(rqstp); err_bad_auth: dprintk("svc: authentication failed (%d)\n", ntohl(auth_stat)); serv->sv_stats->rpcbadauth++; @@ -1369,16 +1467,16 @@ err_bad_prog: err_bad_vers: svc_printk(rqstp, "unknown version (%d for prog %d, %s)\n", - vers, prog, progp->pg_name); + rqstp->rq_vers, rqstp->rq_prog, progp->pg_name); serv->sv_stats->rpcbadfmt++; svc_putnl(resv, RPC_PROG_MISMATCH); - svc_putnl(resv, progp->pg_lovers); - svc_putnl(resv, progp->pg_hivers); + svc_putnl(resv, process.mismatch.lovers); + svc_putnl(resv, process.mismatch.hivers); goto sendit; err_bad_proc: - svc_printk(rqstp, "unknown procedure (%d)\n", proc); + svc_printk(rqstp, "unknown procedure (%d)\n", rqstp->rq_proc); serv->sv_stats->rpcbadfmt++; svc_putnl(resv, RPC_PROC_UNAVAIL); diff --git a/net/sunrpc/svc_xprt.c b/net/sunrpc/svc_xprt.c index 61530b1b7754..9429b28e9ba0 100644 --- a/net/sunrpc/svc_xprt.c +++ b/net/sunrpc/svc_xprt.c @@ -136,6 +136,7 @@ static void svc_xprt_free(struct kref *kref) struct module *owner = xprt->xpt_class->xcl_owner; if (test_bit(XPT_CACHE_AUTH, &xprt->xpt_flags)) svcauth_unix_info_release(xprt); + put_cred(xprt->xpt_cred); put_net(xprt->xpt_net); /* See comment on corresponding get in xs_setup_bc_tcp(): */ if (xprt->xpt_bc_xprt) @@ -252,7 +253,8 @@ void svc_add_new_perm_xprt(struct svc_serv *serv, struct svc_xprt *new) static int _svc_create_xprt(struct svc_serv *serv, const char *xprt_name, struct net *net, const int family, - const unsigned short port, int flags) + const unsigned short port, int flags, + const struct cred *cred) { struct svc_xprt_class *xcl; @@ -273,6 +275,7 @@ static int _svc_create_xprt(struct svc_serv *serv, const char *xprt_name, module_put(xcl->xcl_owner); return PTR_ERR(newxprt); } + newxprt->xpt_cred = get_cred(cred); svc_add_new_perm_xprt(serv, newxprt); newport = svc_xprt_local_port(newxprt); return newport; @@ -286,15 +289,16 @@ static int _svc_create_xprt(struct svc_serv *serv, const char *xprt_name, int svc_create_xprt(struct svc_serv *serv, const char *xprt_name, struct net *net, const int family, - const unsigned short port, int flags) + const unsigned short port, int flags, + const struct cred *cred) { int err; dprintk("svc: creating transport %s[%d]\n", xprt_name, port); - err = _svc_create_xprt(serv, xprt_name, net, family, port, flags); + err = _svc_create_xprt(serv, xprt_name, net, family, port, flags, cred); if (err == -EPROTONOSUPPORT) { request_module("svc%s", xprt_name); - err = _svc_create_xprt(serv, xprt_name, net, family, port, flags); + err = _svc_create_xprt(serv, xprt_name, net, family, port, flags, cred); } if (err < 0) dprintk("svc: transport %s not found, err %d\n", @@ -782,9 +786,10 @@ static int svc_handle_xprt(struct svc_rqst *rqstp, struct svc_xprt *xprt) __module_get(xprt->xpt_class->xcl_owner); svc_check_conn_limits(xprt->xpt_server); newxpt = xprt->xpt_ops->xpo_accept(xprt); - if (newxpt) + if (newxpt) { + newxpt->xpt_cred = get_cred(xprt->xpt_cred); svc_add_new_temp_xprt(serv, newxpt); - else + } else module_put(xprt->xpt_class->xcl_owner); } else if (svc_xprt_reserve_slot(rqstp, xprt)) { /* XPT_DATA|XPT_DEFERRED case: */ diff --git a/net/sunrpc/svcauth_unix.c b/net/sunrpc/svcauth_unix.c index fb9041b92f72..f92ef79c8ea5 100644 --- a/net/sunrpc/svcauth_unix.c +++ b/net/sunrpc/svcauth_unix.c @@ -500,7 +500,7 @@ static int unix_gid_parse(struct cache_detail *cd, rv = get_int(&mesg, &id); if (rv) return -EINVAL; - uid = make_kuid(&init_user_ns, id); + uid = make_kuid(current_user_ns(), id); ug.uid = uid; expiry = get_expiry(&mesg); @@ -522,7 +522,7 @@ static int unix_gid_parse(struct cache_detail *cd, err = -EINVAL; if (rv) goto out; - kgid = make_kgid(&init_user_ns, gid); + kgid = make_kgid(current_user_ns(), gid); if (!gid_valid(kgid)) goto out; ug.gi->gid[i] = kgid; @@ -555,7 +555,7 @@ static int unix_gid_show(struct seq_file *m, struct cache_detail *cd, struct cache_head *h) { - struct user_namespace *user_ns = &init_user_ns; + struct user_namespace *user_ns = m->file->f_cred->user_ns; struct unix_gid *ug; int i; int glen; @@ -796,6 +796,7 @@ svcauth_unix_accept(struct svc_rqst *rqstp, __be32 *authp) struct kvec *argv = &rqstp->rq_arg.head[0]; struct kvec *resv = &rqstp->rq_res.head[0]; struct svc_cred *cred = &rqstp->rq_cred; + struct user_namespace *userns; u32 slen, i; int len = argv->iov_len; @@ -816,8 +817,10 @@ svcauth_unix_accept(struct svc_rqst *rqstp, __be32 *authp) * (export-specific) anonymous id by nfsd_setuser. * Supplementary gid's will be left alone. */ - cred->cr_uid = make_kuid(&init_user_ns, svc_getnl(argv)); /* uid */ - cred->cr_gid = make_kgid(&init_user_ns, svc_getnl(argv)); /* gid */ + userns = (rqstp->rq_xprt && rqstp->rq_xprt->xpt_cred) ? + rqstp->rq_xprt->xpt_cred->user_ns : &init_user_ns; + cred->cr_uid = make_kuid(userns, svc_getnl(argv)); /* uid */ + cred->cr_gid = make_kgid(userns, svc_getnl(argv)); /* gid */ slen = svc_getnl(argv); /* gids length */ if (slen > UNX_NGROUPS || (len -= (slen + 2)*4) < 0) goto badcred; @@ -825,7 +828,7 @@ svcauth_unix_accept(struct svc_rqst *rqstp, __be32 *authp) if (cred->cr_group_info == NULL) return SVC_CLOSE; for (i = 0; i < slen; i++) { - kgid_t kgid = make_kgid(&init_user_ns, svc_getnl(argv)); + kgid_t kgid = make_kgid(userns, svc_getnl(argv)); cred->cr_group_info->gid[i] = kgid; } groups_sort(cred->cr_group_info); diff --git a/net/sunrpc/svcsock.c b/net/sunrpc/svcsock.c index 43590a968b73..540fde2804d0 100644 --- a/net/sunrpc/svcsock.c +++ b/net/sunrpc/svcsock.c @@ -1332,13 +1332,14 @@ EXPORT_SYMBOL_GPL(svc_alien_sock); * @fd: file descriptor of the new listener * @name_return: pointer to buffer to fill in with name of listener * @len: size of the buffer + * @cred: credential * * Fills in socket name and returns positive length of name if successful. * Name is terminated with '\n'. On error, returns a negative errno * value. */ int svc_addsock(struct svc_serv *serv, const int fd, char *name_return, - const size_t len) + const size_t len, const struct cred *cred) { int err = 0; struct socket *so = sockfd_lookup(fd, &err); @@ -1371,6 +1372,7 @@ int svc_addsock(struct svc_serv *serv, const int fd, char *name_return, salen = kernel_getsockname(svsk->sk_sock, sin); if (salen >= 0) svc_xprt_set_local(&svsk->sk_xprt, sin, salen); + svsk->sk_xprt.xpt_cred = get_cred(cred); svc_add_new_perm_xprt(serv, &svsk->sk_xprt); return svc_one_sock_name(svsk, name_return, len); out: diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c index d7117d241460..a9d40bc7ebed 100644 --- a/net/sunrpc/xprt.c +++ b/net/sunrpc/xprt.c @@ -73,6 +73,15 @@ static void xprt_destroy(struct rpc_xprt *xprt); static DEFINE_SPINLOCK(xprt_list_lock); static LIST_HEAD(xprt_list); +static unsigned long xprt_request_timeout(const struct rpc_rqst *req) +{ + unsigned long timeout = jiffies + req->rq_timeout; + + if (time_before(timeout, req->rq_majortimeo)) + return timeout; + return req->rq_majortimeo; +} + /** * xprt_register_transport - register a transport implementation * @transport: transport to register @@ -209,9 +218,12 @@ out_unlock: out_sleep: dprintk("RPC: %5u failed to lock transport %p\n", task->tk_pid, xprt); - task->tk_timeout = RPC_IS_SOFT(task) ? req->rq_timeout : 0; task->tk_status = -EAGAIN; - rpc_sleep_on(&xprt->sending, task, NULL); + if (RPC_IS_SOFT(task)) + rpc_sleep_on_timeout(&xprt->sending, task, NULL, + xprt_request_timeout(req)); + else + rpc_sleep_on(&xprt->sending, task, NULL); return 0; } EXPORT_SYMBOL_GPL(xprt_reserve_xprt); @@ -273,9 +285,12 @@ out_unlock: xprt_clear_locked(xprt); out_sleep: dprintk("RPC: %5u failed to lock transport %p\n", task->tk_pid, xprt); - task->tk_timeout = RPC_IS_SOFT(task) ? req->rq_timeout : 0; task->tk_status = -EAGAIN; - rpc_sleep_on(&xprt->sending, task, NULL); + if (RPC_IS_SOFT(task)) + rpc_sleep_on_timeout(&xprt->sending, task, NULL, + xprt_request_timeout(req)); + else + rpc_sleep_on(&xprt->sending, task, NULL); return 0; } EXPORT_SYMBOL_GPL(xprt_reserve_xprt_cong); @@ -554,53 +569,44 @@ bool xprt_write_space(struct rpc_xprt *xprt) } EXPORT_SYMBOL_GPL(xprt_write_space); -/** - * xprt_set_retrans_timeout_def - set a request's retransmit timeout - * @task: task whose timeout is to be set - * - * Set a request's retransmit timeout based on the transport's - * default timeout parameters. Used by transports that don't adjust - * the retransmit timeout based on round-trip time estimation. - */ -void xprt_set_retrans_timeout_def(struct rpc_task *task) +static unsigned long xprt_abs_ktime_to_jiffies(ktime_t abstime) { - task->tk_timeout = task->tk_rqstp->rq_timeout; + s64 delta = ktime_to_ns(ktime_get() - abstime); + return likely(delta >= 0) ? + jiffies - nsecs_to_jiffies(delta) : + jiffies + nsecs_to_jiffies(-delta); } -EXPORT_SYMBOL_GPL(xprt_set_retrans_timeout_def); -/** - * xprt_set_retrans_timeout_rtt - set a request's retransmit timeout - * @task: task whose timeout is to be set - * - * Set a request's retransmit timeout using the RTT estimator. - */ -void xprt_set_retrans_timeout_rtt(struct rpc_task *task) +static unsigned long xprt_calc_majortimeo(struct rpc_rqst *req) { - int timer = task->tk_msg.rpc_proc->p_timer; - struct rpc_clnt *clnt = task->tk_client; - struct rpc_rtt *rtt = clnt->cl_rtt; - struct rpc_rqst *req = task->tk_rqstp; - unsigned long max_timeout = clnt->cl_timeout->to_maxval; + const struct rpc_timeout *to = req->rq_task->tk_client->cl_timeout; + unsigned long majortimeo = req->rq_timeout; - task->tk_timeout = rpc_calc_rto(rtt, timer); - task->tk_timeout <<= rpc_ntimeo(rtt, timer) + req->rq_retries; - if (task->tk_timeout > max_timeout || task->tk_timeout == 0) - task->tk_timeout = max_timeout; + if (to->to_exponential) + majortimeo <<= to->to_retries; + else + majortimeo += to->to_increment * to->to_retries; + if (majortimeo > to->to_maxval || majortimeo == 0) + majortimeo = to->to_maxval; + return majortimeo; } -EXPORT_SYMBOL_GPL(xprt_set_retrans_timeout_rtt); static void xprt_reset_majortimeo(struct rpc_rqst *req) { - const struct rpc_timeout *to = req->rq_task->tk_client->cl_timeout; + req->rq_majortimeo += xprt_calc_majortimeo(req); +} - req->rq_majortimeo = req->rq_timeout; - if (to->to_exponential) - req->rq_majortimeo <<= to->to_retries; +static void xprt_init_majortimeo(struct rpc_task *task, struct rpc_rqst *req) +{ + unsigned long time_init; + struct rpc_xprt *xprt = req->rq_xprt; + + if (likely(xprt && xprt_connected(xprt))) + time_init = jiffies; else - req->rq_majortimeo += to->to_increment * to->to_retries; - if (req->rq_majortimeo > to->to_maxval || req->rq_majortimeo == 0) - req->rq_majortimeo = to->to_maxval; - req->rq_majortimeo += jiffies; + time_init = xprt_abs_ktime_to_jiffies(task->tk_start); + req->rq_timeout = task->tk_client->cl_timeout->to_initval; + req->rq_majortimeo = time_init + xprt_calc_majortimeo(req); } /** @@ -822,9 +828,9 @@ void xprt_connect(struct rpc_task *task) xprt->ops->close(xprt); if (!xprt_connected(xprt)) { - task->tk_timeout = task->tk_rqstp->rq_timeout; task->tk_rqstp->rq_connect_cookie = xprt->connect_cookie; - rpc_sleep_on(&xprt->pending, task, NULL); + rpc_sleep_on_timeout(&xprt->pending, task, NULL, + xprt_request_timeout(task->tk_rqstp)); if (test_bit(XPRT_CLOSING, &xprt->state)) return; @@ -949,7 +955,7 @@ xprt_is_pinned_rqst(struct rpc_rqst *req) * @req: Request to pin * * Caller must ensure this is atomic with the call to xprt_lookup_rqst() - * so should be holding the xprt receive lock. + * so should be holding xprt->queue_lock. */ void xprt_pin_rqst(struct rpc_rqst *req) { @@ -961,7 +967,7 @@ EXPORT_SYMBOL_GPL(xprt_pin_rqst); * xprt_unpin_rqst - Unpin a request on the transport receive list * @req: Request to pin * - * Caller should be holding the xprt receive lock. + * Caller should be holding xprt->queue_lock. */ void xprt_unpin_rqst(struct rpc_rqst *req) { @@ -1017,7 +1023,6 @@ xprt_request_enqueue_receive(struct rpc_task *task) set_bit(RPC_TASK_NEED_RECV, &task->tk_runstate); spin_unlock(&xprt->queue_lock); - xprt_reset_majortimeo(req); /* Turn off autodisconnect */ del_singleshot_timer_sync(&xprt->timer); } @@ -1103,6 +1108,49 @@ static void xprt_timer(struct rpc_task *task) } /** + * xprt_wait_for_reply_request_def - wait for reply + * @task: pointer to rpc_task + * + * Set a request's retransmit timeout based on the transport's + * default timeout parameters. Used by transports that don't adjust + * the retransmit timeout based on round-trip time estimation, + * and put the task to sleep on the pending queue. + */ +void xprt_wait_for_reply_request_def(struct rpc_task *task) +{ + struct rpc_rqst *req = task->tk_rqstp; + + rpc_sleep_on_timeout(&req->rq_xprt->pending, task, xprt_timer, + xprt_request_timeout(req)); +} +EXPORT_SYMBOL_GPL(xprt_wait_for_reply_request_def); + +/** + * xprt_wait_for_reply_request_rtt - wait for reply using RTT estimator + * @task: pointer to rpc_task + * + * Set a request's retransmit timeout using the RTT estimator, + * and put the task to sleep on the pending queue. + */ +void xprt_wait_for_reply_request_rtt(struct rpc_task *task) +{ + int timer = task->tk_msg.rpc_proc->p_timer; + struct rpc_clnt *clnt = task->tk_client; + struct rpc_rtt *rtt = clnt->cl_rtt; + struct rpc_rqst *req = task->tk_rqstp; + unsigned long max_timeout = clnt->cl_timeout->to_maxval; + unsigned long timeout; + + timeout = rpc_calc_rto(rtt, timer); + timeout <<= rpc_ntimeo(rtt, timer) + req->rq_retries; + if (timeout > max_timeout || timeout == 0) + timeout = max_timeout; + rpc_sleep_on_timeout(&req->rq_xprt->pending, task, xprt_timer, + jiffies + timeout); +} +EXPORT_SYMBOL_GPL(xprt_wait_for_reply_request_rtt); + +/** * xprt_request_wait_receive - wait for the reply to an RPC request * @task: RPC task about to send a request * @@ -1121,8 +1169,7 @@ void xprt_request_wait_receive(struct rpc_task *task) */ spin_lock(&xprt->queue_lock); if (test_bit(RPC_TASK_NEED_RECV, &task->tk_runstate)) { - xprt->ops->set_retrans_timeout(task); - rpc_sleep_on(&xprt->pending, task, xprt_timer); + xprt->ops->wait_for_reply_request(task); /* * Send an extra queue wakeup call if the * connection was dropped in case the call to @@ -1337,6 +1384,10 @@ xprt_request_transmit(struct rpc_rqst *req, struct rpc_task *snd_task) if (status < 0) goto out_dequeue; } + if (RPC_SIGNALLED(task)) { + status = -ERESTARTSYS; + goto out_dequeue; + } } /* @@ -1605,7 +1656,6 @@ xprt_request_init(struct rpc_task *task) struct rpc_xprt *xprt = task->tk_xprt; struct rpc_rqst *req = task->tk_rqstp; - req->rq_timeout = task->tk_client->cl_timeout->to_initval; req->rq_task = task; req->rq_xprt = xprt; req->rq_buffer = NULL; @@ -1618,7 +1668,7 @@ xprt_request_init(struct rpc_task *task) req->rq_snd_buf.bvec = NULL; req->rq_rcv_buf.bvec = NULL; req->rq_release_snd_buf = NULL; - xprt_reset_majortimeo(req); + xprt_init_majortimeo(task, req); dprintk("RPC: %5u reserved req %p xid %08x\n", task->tk_pid, req, ntohl(req->rq_xid)); } @@ -1647,7 +1697,6 @@ void xprt_reserve(struct rpc_task *task) if (task->tk_rqstp != NULL) return; - task->tk_timeout = 0; task->tk_status = -EAGAIN; if (!xprt_throttle_congested(xprt, task)) xprt_do_reserve(xprt, task); @@ -1670,7 +1719,6 @@ void xprt_retry_reserve(struct rpc_task *task) if (task->tk_rqstp != NULL) return; - task->tk_timeout = 0; task->tk_status = -EAGAIN; xprt_do_reserve(xprt, task); } @@ -1827,7 +1875,9 @@ found: xprt->idle_timeout = 0; INIT_WORK(&xprt->task_cleanup, xprt_autoclose); if (xprt_has_timer(xprt)) - timer_setup(&xprt->timer, xprt_init_autodisconnect, 0); + timer_setup(&xprt->timer, + xprt_init_autodisconnect, + TIMER_DEFERRABLE); else timer_setup(&xprt->timer, NULL, 0); diff --git a/net/sunrpc/xprtrdma/backchannel.c b/net/sunrpc/xprtrdma/backchannel.c index d79b18c1f4cd..ce986591f213 100644 --- a/net/sunrpc/xprtrdma/backchannel.c +++ b/net/sunrpc/xprtrdma/backchannel.c @@ -19,45 +19,6 @@ #undef RPCRDMA_BACKCHANNEL_DEBUG -static int rpcrdma_bc_setup_reqs(struct rpcrdma_xprt *r_xprt, - unsigned int count) -{ - struct rpc_xprt *xprt = &r_xprt->rx_xprt; - struct rpcrdma_req *req; - struct rpc_rqst *rqst; - unsigned int i; - - for (i = 0; i < (count << 1); i++) { - struct rpcrdma_regbuf *rb; - size_t size; - - req = rpcrdma_create_req(r_xprt); - if (IS_ERR(req)) - return PTR_ERR(req); - rqst = &req->rl_slot; - - rqst->rq_xprt = xprt; - INIT_LIST_HEAD(&rqst->rq_bc_list); - __set_bit(RPC_BC_PA_IN_USE, &rqst->rq_bc_pa_state); - spin_lock(&xprt->bc_pa_lock); - list_add(&rqst->rq_bc_pa_list, &xprt->bc_pa_list); - spin_unlock(&xprt->bc_pa_lock); - - size = r_xprt->rx_data.inline_rsize; - rb = rpcrdma_alloc_regbuf(size, DMA_TO_DEVICE, GFP_KERNEL); - if (IS_ERR(rb)) - goto out_fail; - req->rl_sendbuf = rb; - xdr_buf_init(&rqst->rq_snd_buf, rb->rg_base, - min_t(size_t, size, PAGE_SIZE)); - } - return 0; - -out_fail: - rpcrdma_req_destroy(req); - return -ENOMEM; -} - /** * xprt_rdma_bc_setup - Pre-allocate resources for handling backchannel requests * @xprt: transport associated with these backchannel resources @@ -68,34 +29,10 @@ out_fail: int xprt_rdma_bc_setup(struct rpc_xprt *xprt, unsigned int reqs) { struct rpcrdma_xprt *r_xprt = rpcx_to_rdmax(xprt); - int rc; - /* The backchannel reply path returns each rpc_rqst to the - * bc_pa_list _after_ the reply is sent. If the server is - * faster than the client, it can send another backward - * direction request before the rpc_rqst is returned to the - * list. The client rejects the request in this case. - * - * Twice as many rpc_rqsts are prepared to ensure there is - * always an rpc_rqst available as soon as a reply is sent. - */ - if (reqs > RPCRDMA_BACKWARD_WRS >> 1) - goto out_err; - - rc = rpcrdma_bc_setup_reqs(r_xprt, reqs); - if (rc) - goto out_free; - - r_xprt->rx_buf.rb_bc_srv_max_requests = reqs; + r_xprt->rx_buf.rb_bc_srv_max_requests = RPCRDMA_BACKWARD_WRS >> 1; trace_xprtrdma_cb_setup(r_xprt, reqs); return 0; - -out_free: - xprt_rdma_bc_destroy(xprt, reqs); - -out_err: - pr_err("RPC: %s: setup backchannel transport failed\n", __func__); - return -ENOMEM; } /** @@ -107,10 +44,10 @@ out_err: size_t xprt_rdma_bc_maxpayload(struct rpc_xprt *xprt) { struct rpcrdma_xprt *r_xprt = rpcx_to_rdmax(xprt); - struct rpcrdma_create_data_internal *cdata = &r_xprt->rx_data; + struct rpcrdma_ep *ep = &r_xprt->rx_ep; size_t maxmsg; - maxmsg = min_t(unsigned int, cdata->inline_rsize, cdata->inline_wsize); + maxmsg = min_t(unsigned int, ep->rep_inline_send, ep->rep_inline_recv); maxmsg = min_t(unsigned int, maxmsg, PAGE_SIZE); return maxmsg - RPCRDMA_HDRLEN_MIN; } @@ -123,7 +60,7 @@ static int rpcrdma_bc_marshal_reply(struct rpc_rqst *rqst) rpcrdma_set_xdrlen(&req->rl_hdrbuf, 0); xdr_init_encode(&req->rl_stream, &req->rl_hdrbuf, - req->rl_rdmabuf->rg_base, rqst); + rdmab_data(req->rl_rdmabuf), rqst); p = xdr_reserve_space(&req->rl_stream, 28); if (unlikely(!p)) @@ -223,6 +160,43 @@ void xprt_rdma_bc_free_rqst(struct rpc_rqst *rqst) spin_unlock(&xprt->bc_pa_lock); } +static struct rpc_rqst *rpcrdma_bc_rqst_get(struct rpcrdma_xprt *r_xprt) +{ + struct rpc_xprt *xprt = &r_xprt->rx_xprt; + struct rpcrdma_req *req; + struct rpc_rqst *rqst; + size_t size; + + spin_lock(&xprt->bc_pa_lock); + rqst = list_first_entry_or_null(&xprt->bc_pa_list, struct rpc_rqst, + rq_bc_pa_list); + if (!rqst) + goto create_req; + list_del(&rqst->rq_bc_pa_list); + spin_unlock(&xprt->bc_pa_lock); + return rqst; + +create_req: + spin_unlock(&xprt->bc_pa_lock); + + /* Set a limit to prevent a remote from overrunning our resources. + */ + if (xprt->bc_alloc_count >= RPCRDMA_BACKWARD_WRS) + return NULL; + + size = min_t(size_t, r_xprt->rx_ep.rep_inline_recv, PAGE_SIZE); + req = rpcrdma_req_create(r_xprt, size, GFP_KERNEL); + if (!req) + return NULL; + + xprt->bc_alloc_count++; + rqst = &req->rl_slot; + rqst->rq_xprt = xprt; + __set_bit(RPC_BC_PA_IN_USE, &rqst->rq_bc_pa_state); + xdr_buf_init(&rqst->rq_snd_buf, rdmab_data(req->rl_sendbuf), size); + return rqst; +} + /** * rpcrdma_bc_receive_call - Handle a backward direction call * @r_xprt: transport receiving the call @@ -254,18 +228,10 @@ void rpcrdma_bc_receive_call(struct rpcrdma_xprt *r_xprt, pr_info("RPC: %s: %*ph\n", __func__, size, p); #endif - /* Grab a free bc rqst */ - spin_lock(&xprt->bc_pa_lock); - if (list_empty(&xprt->bc_pa_list)) { - spin_unlock(&xprt->bc_pa_lock); + rqst = rpcrdma_bc_rqst_get(r_xprt); + if (!rqst) goto out_overflow; - } - rqst = list_first_entry(&xprt->bc_pa_list, - struct rpc_rqst, rq_bc_pa_list); - list_del(&rqst->rq_bc_pa_list); - spin_unlock(&xprt->bc_pa_lock); - /* Prepare rqst */ rqst->rq_reply_bytes_recvd = 0; rqst->rq_xid = *p; diff --git a/net/sunrpc/xprtrdma/frwr_ops.c b/net/sunrpc/xprtrdma/frwr_ops.c index 52cb6c1b0c2b..794ba4ca0994 100644 --- a/net/sunrpc/xprtrdma/frwr_ops.c +++ b/net/sunrpc/xprtrdma/frwr_ops.c @@ -82,13 +82,13 @@ /** * frwr_is_supported - Check if device supports FRWR - * @ia: interface adapter to check + * @device: interface adapter to check * * Returns true if device supports FRWR, otherwise false */ -bool frwr_is_supported(struct rpcrdma_ia *ia) +bool frwr_is_supported(struct ib_device *device) { - struct ib_device_attr *attrs = &ia->ri_device->attrs; + struct ib_device_attr *attrs = &device->attrs; if (!(attrs->device_cap_flags & IB_DEVICE_MEM_MGT_EXTENSIONS)) goto out_not_supported; @@ -98,7 +98,7 @@ bool frwr_is_supported(struct rpcrdma_ia *ia) out_not_supported: pr_info("rpcrdma: 'frwr' mode is not supported by device %s\n", - ia->ri_device->name); + device->name); return false; } @@ -131,7 +131,7 @@ frwr_mr_recycle_worker(struct work_struct *work) if (mr->mr_dir != DMA_NONE) { trace_xprtrdma_mr_unmap(mr); - ib_dma_unmap_sg(r_xprt->rx_ia.ri_device, + ib_dma_unmap_sg(r_xprt->rx_ia.ri_id->device, mr->mr_sg, mr->mr_nents, mr->mr_dir); mr->mr_dir = DMA_NONE; } @@ -194,12 +194,11 @@ out_list_err: * frwr_open - Prepare an endpoint for use with FRWR * @ia: interface adapter this endpoint will use * @ep: endpoint to prepare - * @cdata: transport parameters * * On success, sets: * ep->rep_attr.cap.max_send_wr * ep->rep_attr.cap.max_recv_wr - * cdata->max_requests + * ep->rep_max_requests * ia->ri_max_segs * * And these FRWR-related fields: @@ -208,10 +207,9 @@ out_list_err: * * On failure, a negative errno is returned. */ -int frwr_open(struct rpcrdma_ia *ia, struct rpcrdma_ep *ep, - struct rpcrdma_create_data_internal *cdata) +int frwr_open(struct rpcrdma_ia *ia, struct rpcrdma_ep *ep) { - struct ib_device_attr *attrs = &ia->ri_device->attrs; + struct ib_device_attr *attrs = &ia->ri_id->device->attrs; int max_qp_wr, depth, delta; ia->ri_mrtype = IB_MR_TYPE_MEM_REG; @@ -253,24 +251,23 @@ int frwr_open(struct rpcrdma_ia *ia, struct rpcrdma_ep *ep, } while (delta > 0); } - max_qp_wr = ia->ri_device->attrs.max_qp_wr; + max_qp_wr = ia->ri_id->device->attrs.max_qp_wr; max_qp_wr -= RPCRDMA_BACKWARD_WRS; max_qp_wr -= 1; if (max_qp_wr < RPCRDMA_MIN_SLOT_TABLE) return -ENOMEM; - if (cdata->max_requests > max_qp_wr) - cdata->max_requests = max_qp_wr; - ep->rep_attr.cap.max_send_wr = cdata->max_requests * depth; + if (ep->rep_max_requests > max_qp_wr) + ep->rep_max_requests = max_qp_wr; + ep->rep_attr.cap.max_send_wr = ep->rep_max_requests * depth; if (ep->rep_attr.cap.max_send_wr > max_qp_wr) { - cdata->max_requests = max_qp_wr / depth; - if (!cdata->max_requests) + ep->rep_max_requests = max_qp_wr / depth; + if (!ep->rep_max_requests) return -EINVAL; - ep->rep_attr.cap.max_send_wr = cdata->max_requests * - depth; + ep->rep_attr.cap.max_send_wr = ep->rep_max_requests * depth; } ep->rep_attr.cap.max_send_wr += RPCRDMA_BACKWARD_WRS; ep->rep_attr.cap.max_send_wr += 1; /* for ib_drain_sq */ - ep->rep_attr.cap.max_recv_wr = cdata->max_requests; + ep->rep_attr.cap.max_recv_wr = ep->rep_max_requests; ep->rep_attr.cap.max_recv_wr += RPCRDMA_BACKWARD_WRS; ep->rep_attr.cap.max_recv_wr += 1; /* for ib_drain_rq */ @@ -300,15 +297,6 @@ size_t frwr_maxpages(struct rpcrdma_xprt *r_xprt) (ia->ri_max_segs - 2) * ia->ri_max_frwr_depth); } -static void -__frwr_sendcompletion_flush(struct ib_wc *wc, const char *wr) -{ - if (wc->status != IB_WC_WR_FLUSH_ERR) - pr_err("rpcrdma: %s: %s (%u/0x%x)\n", - wr, ib_wc_status_msg(wc->status), - wc->status, wc->vendor_err); -} - /** * frwr_wc_fastreg - Invoked by RDMA provider for a flushed FastReg WC * @cq: completion queue (ignored) @@ -323,10 +311,8 @@ frwr_wc_fastreg(struct ib_cq *cq, struct ib_wc *wc) container_of(cqe, struct rpcrdma_frwr, fr_cqe); /* WARNING: Only wr_cqe and status are reliable at this point */ - if (wc->status != IB_WC_SUCCESS) { + if (wc->status != IB_WC_SUCCESS) frwr->fr_state = FRWR_FLUSHED_FR; - __frwr_sendcompletion_flush(wc, "fastreg"); - } trace_xprtrdma_wc_fastreg(wc, frwr); } @@ -344,10 +330,8 @@ frwr_wc_localinv(struct ib_cq *cq, struct ib_wc *wc) fr_cqe); /* WARNING: Only wr_cqe and status are reliable at this point */ - if (wc->status != IB_WC_SUCCESS) { + if (wc->status != IB_WC_SUCCESS) frwr->fr_state = FRWR_FLUSHED_LI; - __frwr_sendcompletion_flush(wc, "localinv"); - } trace_xprtrdma_wc_li(wc, frwr); } @@ -366,12 +350,10 @@ frwr_wc_localinv_wake(struct ib_cq *cq, struct ib_wc *wc) fr_cqe); /* WARNING: Only wr_cqe and status are reliable at this point */ - if (wc->status != IB_WC_SUCCESS) { + if (wc->status != IB_WC_SUCCESS) frwr->fr_state = FRWR_FLUSHED_LI; - __frwr_sendcompletion_flush(wc, "localinv"); - } - complete(&frwr->fr_linv_done); trace_xprtrdma_wc_li_wake(wc, frwr); + complete(&frwr->fr_linv_done); } /** @@ -436,7 +418,8 @@ struct rpcrdma_mr_seg *frwr_map(struct rpcrdma_xprt *r_xprt, } mr->mr_dir = rpcrdma_data_dir(writing); - mr->mr_nents = ib_dma_map_sg(ia->ri_device, mr->mr_sg, i, mr->mr_dir); + mr->mr_nents = + ib_dma_map_sg(ia->ri_id->device, mr->mr_sg, i, mr->mr_dir); if (!mr->mr_nents) goto out_dmamap_err; @@ -466,7 +449,7 @@ struct rpcrdma_mr_seg *frwr_map(struct rpcrdma_xprt *r_xprt, return seg; out_dmamap_err: - frwr->fr_state = FRWR_IS_INVALID; + mr->mr_dir = DMA_NONE; trace_xprtrdma_frwr_sgerr(mr, i); rpcrdma_mr_put(mr); return ERR_PTR(-EIO); diff --git a/net/sunrpc/xprtrdma/rpc_rdma.c b/net/sunrpc/xprtrdma/rpc_rdma.c index 6c1fb270f127..85115a2e2639 100644 --- a/net/sunrpc/xprtrdma/rpc_rdma.c +++ b/net/sunrpc/xprtrdma/rpc_rdma.c @@ -105,16 +105,23 @@ static unsigned int rpcrdma_max_reply_header_size(unsigned int maxsegs) return size; } +/** + * rpcrdma_set_max_header_sizes - Initialize inline payload sizes + * @r_xprt: transport instance to initialize + * + * The max_inline fields contain the maximum size of an RPC message + * so the marshaling code doesn't have to repeat this calculation + * for every RPC. + */ void rpcrdma_set_max_header_sizes(struct rpcrdma_xprt *r_xprt) { - struct rpcrdma_create_data_internal *cdata = &r_xprt->rx_data; - struct rpcrdma_ia *ia = &r_xprt->rx_ia; - unsigned int maxsegs = ia->ri_max_segs; - - ia->ri_max_inline_write = cdata->inline_wsize - - rpcrdma_max_call_header_size(maxsegs); - ia->ri_max_inline_read = cdata->inline_rsize - - rpcrdma_max_reply_header_size(maxsegs); + unsigned int maxsegs = r_xprt->rx_ia.ri_max_segs; + struct rpcrdma_ep *ep = &r_xprt->rx_ep; + + ep->rep_max_inline_send = + ep->rep_inline_send - rpcrdma_max_call_header_size(maxsegs); + ep->rep_max_inline_recv = + ep->rep_inline_recv - rpcrdma_max_reply_header_size(maxsegs); } /* The client can send a request inline as long as the RPCRDMA header @@ -131,7 +138,7 @@ static bool rpcrdma_args_inline(struct rpcrdma_xprt *r_xprt, struct xdr_buf *xdr = &rqst->rq_snd_buf; unsigned int count, remaining, offset; - if (xdr->len > r_xprt->rx_ia.ri_max_inline_write) + if (xdr->len > r_xprt->rx_ep.rep_max_inline_send) return false; if (xdr->page_len) { @@ -159,9 +166,7 @@ static bool rpcrdma_args_inline(struct rpcrdma_xprt *r_xprt, static bool rpcrdma_results_inline(struct rpcrdma_xprt *r_xprt, struct rpc_rqst *rqst) { - struct rpcrdma_ia *ia = &r_xprt->rx_ia; - - return rqst->rq_rcv_buf.buflen <= ia->ri_max_inline_read; + return rqst->rq_rcv_buf.buflen <= r_xprt->rx_ep.rep_max_inline_recv; } /* The client is required to provide a Reply chunk if the maximum @@ -173,10 +178,9 @@ rpcrdma_nonpayload_inline(const struct rpcrdma_xprt *r_xprt, const struct rpc_rqst *rqst) { const struct xdr_buf *buf = &rqst->rq_rcv_buf; - const struct rpcrdma_ia *ia = &r_xprt->rx_ia; - return buf->head[0].iov_len + buf->tail[0].iov_len < - ia->ri_max_inline_read; + return (buf->head[0].iov_len + buf->tail[0].iov_len) < + r_xprt->rx_ep.rep_max_inline_recv; } /* Split @vec on page boundaries into SGEs. FMR registers pages, not @@ -238,7 +242,7 @@ rpcrdma_convert_iovs(struct rpcrdma_xprt *r_xprt, struct xdr_buf *xdrbuf, */ if (unlikely(xdrbuf->flags & XDRBUF_SPARSE_PAGES)) { if (!*ppages) - *ppages = alloc_page(GFP_ATOMIC); + *ppages = alloc_page(GFP_NOWAIT | __GFP_NOWARN); if (!*ppages) return -ENOBUFS; } @@ -508,50 +512,45 @@ rpcrdma_encode_reply_chunk(struct rpcrdma_xprt *r_xprt, struct rpcrdma_req *req, } /** - * rpcrdma_unmap_sendctx - DMA-unmap Send buffers + * rpcrdma_sendctx_unmap - DMA-unmap Send buffer * @sc: sendctx containing SGEs to unmap * */ -void -rpcrdma_unmap_sendctx(struct rpcrdma_sendctx *sc) +void rpcrdma_sendctx_unmap(struct rpcrdma_sendctx *sc) { - struct rpcrdma_ia *ia = &sc->sc_xprt->rx_ia; struct ib_sge *sge; - unsigned int count; /* The first two SGEs contain the transport header and * the inline buffer. These are always left mapped so * they can be cheaply re-used. */ - sge = &sc->sc_sges[2]; - for (count = sc->sc_unmap_count; count; ++sge, --count) - ib_dma_unmap_page(ia->ri_device, - sge->addr, sge->length, DMA_TO_DEVICE); + for (sge = &sc->sc_sges[2]; sc->sc_unmap_count; + ++sge, --sc->sc_unmap_count) + ib_dma_unmap_page(sc->sc_device, sge->addr, sge->length, + DMA_TO_DEVICE); - if (test_and_clear_bit(RPCRDMA_REQ_F_TX_RESOURCES, &sc->sc_req->rl_flags)) { - smp_mb__after_atomic(); + if (test_and_clear_bit(RPCRDMA_REQ_F_TX_RESOURCES, + &sc->sc_req->rl_flags)) wake_up_bit(&sc->sc_req->rl_flags, RPCRDMA_REQ_F_TX_RESOURCES); - } } /* Prepare an SGE for the RPC-over-RDMA transport header. */ -static bool -rpcrdma_prepare_hdr_sge(struct rpcrdma_ia *ia, struct rpcrdma_req *req, - u32 len) +static bool rpcrdma_prepare_hdr_sge(struct rpcrdma_xprt *r_xprt, + struct rpcrdma_req *req, u32 len) { struct rpcrdma_sendctx *sc = req->rl_sendctx; struct rpcrdma_regbuf *rb = req->rl_rdmabuf; struct ib_sge *sge = sc->sc_sges; - if (!rpcrdma_dma_map_regbuf(ia, rb)) + if (!rpcrdma_regbuf_dma_map(r_xprt, rb)) goto out_regbuf; sge->addr = rdmab_addr(rb); sge->length = len; sge->lkey = rdmab_lkey(rb); - ib_dma_sync_single_for_device(rdmab_device(rb), sge->addr, - sge->length, DMA_TO_DEVICE); + ib_dma_sync_single_for_device(rdmab_device(rb), sge->addr, sge->length, + DMA_TO_DEVICE); sc->sc_wr.num_sge++; return true; @@ -563,23 +562,23 @@ out_regbuf: /* Prepare the Send SGEs. The head and tail iovec, and each entry * in the page list, gets its own SGE. */ -static bool -rpcrdma_prepare_msg_sges(struct rpcrdma_ia *ia, struct rpcrdma_req *req, - struct xdr_buf *xdr, enum rpcrdma_chunktype rtype) +static bool rpcrdma_prepare_msg_sges(struct rpcrdma_xprt *r_xprt, + struct rpcrdma_req *req, + struct xdr_buf *xdr, + enum rpcrdma_chunktype rtype) { struct rpcrdma_sendctx *sc = req->rl_sendctx; unsigned int sge_no, page_base, len, remaining; struct rpcrdma_regbuf *rb = req->rl_sendbuf; - struct ib_device *device = ia->ri_device; struct ib_sge *sge = sc->sc_sges; - u32 lkey = ia->ri_pd->local_dma_lkey; struct page *page, **ppages; /* The head iovec is straightforward, as it is already * DMA-mapped. Sync the content that has changed. */ - if (!rpcrdma_dma_map_regbuf(ia, rb)) + if (!rpcrdma_regbuf_dma_map(r_xprt, rb)) goto out_regbuf; + sc->sc_device = rdmab_device(rb); sge_no = 1; sge[sge_no].addr = rdmab_addr(rb); sge[sge_no].length = xdr->head[0].iov_len; @@ -626,13 +625,14 @@ rpcrdma_prepare_msg_sges(struct rpcrdma_ia *ia, struct rpcrdma_req *req, goto out_mapping_overflow; len = min_t(u32, PAGE_SIZE - page_base, remaining); - sge[sge_no].addr = ib_dma_map_page(device, *ppages, - page_base, len, - DMA_TO_DEVICE); - if (ib_dma_mapping_error(device, sge[sge_no].addr)) + sge[sge_no].addr = + ib_dma_map_page(rdmab_device(rb), *ppages, + page_base, len, DMA_TO_DEVICE); + if (ib_dma_mapping_error(rdmab_device(rb), + sge[sge_no].addr)) goto out_mapping_err; sge[sge_no].length = len; - sge[sge_no].lkey = lkey; + sge[sge_no].lkey = rdmab_lkey(rb); sc->sc_unmap_count++; ppages++; @@ -653,13 +653,13 @@ rpcrdma_prepare_msg_sges(struct rpcrdma_ia *ia, struct rpcrdma_req *req, map_tail: sge_no++; - sge[sge_no].addr = ib_dma_map_page(device, page, - page_base, len, - DMA_TO_DEVICE); - if (ib_dma_mapping_error(device, sge[sge_no].addr)) + sge[sge_no].addr = + ib_dma_map_page(rdmab_device(rb), page, page_base, len, + DMA_TO_DEVICE); + if (ib_dma_mapping_error(rdmab_device(rb), sge[sge_no].addr)) goto out_mapping_err; sge[sge_no].length = len; - sge[sge_no].lkey = lkey; + sge[sge_no].lkey = rdmab_lkey(rb); sc->sc_unmap_count++; } @@ -674,12 +674,12 @@ out_regbuf: return false; out_mapping_overflow: - rpcrdma_unmap_sendctx(sc); + rpcrdma_sendctx_unmap(sc); pr_err("rpcrdma: too many Send SGEs (%u)\n", sge_no); return false; out_mapping_err: - rpcrdma_unmap_sendctx(sc); + rpcrdma_sendctx_unmap(sc); trace_xprtrdma_dma_maperr(sge[sge_no].addr); return false; } @@ -699,7 +699,7 @@ rpcrdma_prepare_send_sges(struct rpcrdma_xprt *r_xprt, struct rpcrdma_req *req, u32 hdrlen, struct xdr_buf *xdr, enum rpcrdma_chunktype rtype) { - req->rl_sendctx = rpcrdma_sendctx_get_locked(&r_xprt->rx_buf); + req->rl_sendctx = rpcrdma_sendctx_get_locked(r_xprt); if (!req->rl_sendctx) return -EAGAIN; req->rl_sendctx->sc_wr.num_sge = 0; @@ -707,11 +707,11 @@ rpcrdma_prepare_send_sges(struct rpcrdma_xprt *r_xprt, req->rl_sendctx->sc_req = req; __clear_bit(RPCRDMA_REQ_F_TX_RESOURCES, &req->rl_flags); - if (!rpcrdma_prepare_hdr_sge(&r_xprt->rx_ia, req, hdrlen)) + if (!rpcrdma_prepare_hdr_sge(r_xprt, req, hdrlen)) return -EIO; if (rtype != rpcrdma_areadch) - if (!rpcrdma_prepare_msg_sges(&r_xprt->rx_ia, req, xdr, rtype)) + if (!rpcrdma_prepare_msg_sges(r_xprt, req, xdr, rtype)) return -EIO; return 0; @@ -747,8 +747,8 @@ rpcrdma_marshal_req(struct rpcrdma_xprt *r_xprt, struct rpc_rqst *rqst) int ret; rpcrdma_set_xdrlen(&req->rl_hdrbuf, 0); - xdr_init_encode(xdr, &req->rl_hdrbuf, - req->rl_rdmabuf->rg_base, rqst); + xdr_init_encode(xdr, &req->rl_hdrbuf, rdmab_data(req->rl_rdmabuf), + rqst); /* Fixed header fields */ ret = -EMSGSIZE; @@ -876,6 +876,7 @@ rpcrdma_marshal_req(struct rpcrdma_xprt *r_xprt, struct rpc_rqst *rqst) return 0; out_err: + trace_xprtrdma_marshal_failed(rqst, ret); switch (ret) { case -EAGAIN: xprt_wait_for_buffer_space(rqst->rq_xprt); diff --git a/net/sunrpc/xprtrdma/svc_rdma_backchannel.c b/net/sunrpc/xprtrdma/svc_rdma_backchannel.c index 907464c2a9f0..bed57d8b5c19 100644 --- a/net/sunrpc/xprtrdma/svc_rdma_backchannel.c +++ b/net/sunrpc/xprtrdma/svc_rdma_backchannel.c @@ -261,7 +261,7 @@ static const struct rpc_xprt_ops xprt_rdma_bc_procs = { .buf_alloc = xprt_rdma_bc_allocate, .buf_free = xprt_rdma_bc_free, .send_request = xprt_rdma_bc_send_request, - .set_retrans_timeout = xprt_set_retrans_timeout_def, + .wait_for_reply_request = xprt_wait_for_reply_request_def, .close = xprt_rdma_bc_close, .destroy = xprt_rdma_bc_put, .print_stats = xprt_rdma_print_stats diff --git a/net/sunrpc/xprtrdma/transport.c b/net/sunrpc/xprtrdma/transport.c index 5d261353bd90..1f73a6a7e43c 100644 --- a/net/sunrpc/xprtrdma/transport.c +++ b/net/sunrpc/xprtrdma/transport.c @@ -68,9 +68,9 @@ * tunables */ -static unsigned int xprt_rdma_slot_table_entries = RPCRDMA_DEF_SLOT_TABLE; +unsigned int xprt_rdma_slot_table_entries = RPCRDMA_DEF_SLOT_TABLE; unsigned int xprt_rdma_max_inline_read = RPCRDMA_DEF_INLINE; -static unsigned int xprt_rdma_max_inline_write = RPCRDMA_DEF_INLINE; +unsigned int xprt_rdma_max_inline_write = RPCRDMA_DEF_INLINE; unsigned int xprt_rdma_memreg_strategy = RPCRDMA_FRWR; int xprt_rdma_pad_optimize; @@ -288,7 +288,7 @@ xprt_rdma_destroy(struct rpc_xprt *xprt) cancel_delayed_work_sync(&r_xprt->rx_connect_worker); - rpcrdma_ep_destroy(&r_xprt->rx_ep, &r_xprt->rx_ia); + rpcrdma_ep_destroy(r_xprt); rpcrdma_buffer_destroy(&r_xprt->rx_buf); rpcrdma_ia_close(&r_xprt->rx_ia); @@ -311,10 +311,8 @@ static const struct rpc_timeout xprt_rdma_default_timeout = { static struct rpc_xprt * xprt_setup_rdma(struct xprt_create *args) { - struct rpcrdma_create_data_internal cdata; struct rpc_xprt *xprt; struct rpcrdma_xprt *new_xprt; - struct rpcrdma_ep *new_ep; struct sockaddr *sap; int rc; @@ -349,40 +347,12 @@ xprt_setup_rdma(struct xprt_create *args) xprt_set_bound(xprt); xprt_rdma_format_addresses(xprt, sap); - cdata.max_requests = xprt_rdma_slot_table_entries; - - cdata.rsize = RPCRDMA_MAX_SEGS * PAGE_SIZE; /* RDMA write max */ - cdata.wsize = RPCRDMA_MAX_SEGS * PAGE_SIZE; /* RDMA read max */ - - cdata.inline_wsize = xprt_rdma_max_inline_write; - if (cdata.inline_wsize > cdata.wsize) - cdata.inline_wsize = cdata.wsize; - - cdata.inline_rsize = xprt_rdma_max_inline_read; - if (cdata.inline_rsize > cdata.rsize) - cdata.inline_rsize = cdata.rsize; - - /* - * Create new transport instance, which includes initialized - * o ia - * o endpoint - * o buffers - */ - new_xprt = rpcx_to_rdmax(xprt); - rc = rpcrdma_ia_open(new_xprt); if (rc) goto out1; - /* - * initialize and create ep - */ - new_xprt->rx_data = cdata; - new_ep = &new_xprt->rx_ep; - - rc = rpcrdma_ep_create(&new_xprt->rx_ep, - &new_xprt->rx_ia, &new_xprt->rx_data); + rc = rpcrdma_ep_create(new_xprt); if (rc) goto out2; @@ -413,7 +383,7 @@ out4: rpcrdma_buffer_destroy(&new_xprt->rx_buf); rc = -ENODEV; out3: - rpcrdma_ep_destroy(new_ep, &new_xprt->rx_ia); + rpcrdma_ep_destroy(new_xprt); out2: rpcrdma_ia_close(&new_xprt->rx_ia); out1: @@ -585,52 +555,15 @@ xprt_rdma_free_slot(struct rpc_xprt *xprt, struct rpc_rqst *rqst) rpc_wake_up_next(&xprt->backlog); } -static bool -rpcrdma_get_sendbuf(struct rpcrdma_xprt *r_xprt, struct rpcrdma_req *req, - size_t size, gfp_t flags) +static bool rpcrdma_check_regbuf(struct rpcrdma_xprt *r_xprt, + struct rpcrdma_regbuf *rb, size_t size, + gfp_t flags) { - struct rpcrdma_regbuf *rb; - - if (req->rl_sendbuf && rdmab_length(req->rl_sendbuf) >= size) - return true; - - rb = rpcrdma_alloc_regbuf(size, DMA_TO_DEVICE, flags); - if (IS_ERR(rb)) - return false; - - rpcrdma_free_regbuf(req->rl_sendbuf); - r_xprt->rx_stats.hardway_register_count += size; - req->rl_sendbuf = rb; - return true; -} - -/* The rq_rcv_buf is used only if a Reply chunk is necessary. - * The decision to use a Reply chunk is made later in - * rpcrdma_marshal_req. This buffer is registered at that time. - * - * Otherwise, the associated RPC Reply arrives in a separate - * Receive buffer, arbitrarily chosen by the HCA. The buffer - * allocated here for the RPC Reply is not utilized in that - * case. See rpcrdma_inline_fixup. - * - * A regbuf is used here to remember the buffer size. - */ -static bool -rpcrdma_get_recvbuf(struct rpcrdma_xprt *r_xprt, struct rpcrdma_req *req, - size_t size, gfp_t flags) -{ - struct rpcrdma_regbuf *rb; - - if (req->rl_recvbuf && rdmab_length(req->rl_recvbuf) >= size) - return true; - - rb = rpcrdma_alloc_regbuf(size, DMA_NONE, flags); - if (IS_ERR(rb)) - return false; - - rpcrdma_free_regbuf(req->rl_recvbuf); - r_xprt->rx_stats.hardway_register_count += size; - req->rl_recvbuf = rb; + if (unlikely(rdmab_length(rb) < size)) { + if (!rpcrdma_regbuf_realloc(rb, size, flags)) + return false; + r_xprt->rx_stats.hardway_register_count += size; + } return true; } @@ -655,13 +588,15 @@ xprt_rdma_allocate(struct rpc_task *task) if (RPC_IS_SWAPPER(task)) flags = __GFP_MEMALLOC | GFP_NOWAIT | __GFP_NOWARN; - if (!rpcrdma_get_sendbuf(r_xprt, req, rqst->rq_callsize, flags)) + if (!rpcrdma_check_regbuf(r_xprt, req->rl_sendbuf, rqst->rq_callsize, + flags)) goto out_fail; - if (!rpcrdma_get_recvbuf(r_xprt, req, rqst->rq_rcvsize, flags)) + if (!rpcrdma_check_regbuf(r_xprt, req->rl_recvbuf, rqst->rq_rcvsize, + flags)) goto out_fail; - rqst->rq_buffer = req->rl_sendbuf->rg_base; - rqst->rq_rbuffer = req->rl_recvbuf->rg_base; + rqst->rq_buffer = rdmab_data(req->rl_sendbuf); + rqst->rq_rbuffer = rdmab_data(req->rl_recvbuf); trace_xprtrdma_op_allocate(task, req); return 0; @@ -815,7 +750,7 @@ static const struct rpc_xprt_ops xprt_rdma_procs = { .alloc_slot = xprt_rdma_alloc_slot, .free_slot = xprt_rdma_free_slot, .release_request = xprt_release_rqst_cong, /* ditto */ - .set_retrans_timeout = xprt_set_retrans_timeout_def, /* ditto */ + .wait_for_reply_request = xprt_wait_for_reply_request_def, /* ditto */ .timer = xprt_rdma_timer, .rpcbind = rpcb_getport_async, /* sunrpc/rpcb_clnt.c */ .set_port = xprt_rdma_set_port, diff --git a/net/sunrpc/xprtrdma/verbs.c b/net/sunrpc/xprtrdma/verbs.c index 30cfc0efe699..bef5eac8ab38 100644 --- a/net/sunrpc/xprtrdma/verbs.c +++ b/net/sunrpc/xprtrdma/verbs.c @@ -76,11 +76,16 @@ static void rpcrdma_sendctx_put_locked(struct rpcrdma_sendctx *sc); static void rpcrdma_mrs_create(struct rpcrdma_xprt *r_xprt); static void rpcrdma_mrs_destroy(struct rpcrdma_buffer *buf); -static int rpcrdma_create_rep(struct rpcrdma_xprt *r_xprt, bool temp); -static void rpcrdma_dma_unmap_regbuf(struct rpcrdma_regbuf *rb); +static struct rpcrdma_regbuf * +rpcrdma_regbuf_alloc(size_t size, enum dma_data_direction direction, + gfp_t flags); +static void rpcrdma_regbuf_dma_unmap(struct rpcrdma_regbuf *rb); +static void rpcrdma_regbuf_free(struct rpcrdma_regbuf *rb); static void rpcrdma_post_recvs(struct rpcrdma_xprt *r_xprt, bool temp); -/* Wait for outstanding transport work to finish. +/* Wait for outstanding transport work to finish. ib_drain_qp + * handles the drains in the wrong order for us, so open code + * them here. */ static void rpcrdma_xprt_drain(struct rpcrdma_xprt *r_xprt) { @@ -132,11 +137,6 @@ rpcrdma_wc_send(struct ib_cq *cq, struct ib_wc *wc) /* WARNING: Only wr_cqe and status are reliable at this point */ trace_xprtrdma_wc_send(sc, wc); - if (wc->status != IB_WC_SUCCESS && wc->status != IB_WC_WR_FLUSH_ERR) - pr_err("rpcrdma: Send: %s (%u/0x%x)\n", - ib_wc_status_msg(wc->status), - wc->status, wc->vendor_err); - rpcrdma_sendctx_put_locked(sc); } @@ -174,10 +174,6 @@ rpcrdma_wc_receive(struct ib_cq *cq, struct ib_wc *wc) return; out_flushed: - if (wc->status != IB_WC_WR_FLUSH_ERR) - pr_err("rpcrdma: Recv: %s (%u/0x%x)\n", - ib_wc_status_msg(wc->status), - wc->status, wc->vendor_err); rpcrdma_recv_buffer_put(rep); } @@ -185,7 +181,6 @@ static void rpcrdma_update_connect_private(struct rpcrdma_xprt *r_xprt, struct rdma_conn_param *param) { - struct rpcrdma_create_data_internal *cdata = &r_xprt->rx_data; const struct rpcrdma_connect_private *pmsg = param->private_data; unsigned int rsize, wsize; @@ -202,12 +197,13 @@ rpcrdma_update_connect_private(struct rpcrdma_xprt *r_xprt, wsize = rpcrdma_decode_buffer_size(pmsg->cp_recv_size); } - if (rsize < cdata->inline_rsize) - cdata->inline_rsize = rsize; - if (wsize < cdata->inline_wsize) - cdata->inline_wsize = wsize; - dprintk("RPC: %s: max send %u, max recv %u\n", - __func__, cdata->inline_wsize, cdata->inline_rsize); + if (rsize < r_xprt->rx_ep.rep_inline_recv) + r_xprt->rx_ep.rep_inline_recv = rsize; + if (wsize < r_xprt->rx_ep.rep_inline_send) + r_xprt->rx_ep.rep_inline_send = wsize; + dprintk("RPC: %s: max send %u, max recv %u\n", __func__, + r_xprt->rx_ep.rep_inline_send, + r_xprt->rx_ep.rep_inline_recv); rpcrdma_set_max_header_sizes(r_xprt); } @@ -247,7 +243,7 @@ rpcrdma_cm_event_handler(struct rdma_cm_id *id, struct rdma_cm_event *event) case RDMA_CM_EVENT_DEVICE_REMOVAL: #if IS_ENABLED(CONFIG_SUNRPC_DEBUG) pr_info("rpcrdma: removing device %s for %s:%s\n", - ia->ri_device->name, + ia->ri_id->device->name, rpcrdma_addrstr(r_xprt), rpcrdma_portstr(r_xprt)); #endif set_bit(RPCRDMA_IAF_REMOVING, &ia->ri_flags); @@ -256,7 +252,6 @@ rpcrdma_cm_event_handler(struct rdma_cm_id *id, struct rdma_cm_event *event) wait_for_completion(&ia->ri_remove_done); ia->ri_id = NULL; - ia->ri_device = NULL; /* Return 1 to ensure the core destroys the id. */ return 1; case RDMA_CM_EVENT_ESTABLISHED: @@ -291,7 +286,7 @@ disconnected: dprintk("RPC: %s: %s:%s on %s/frwr: %s\n", __func__, rpcrdma_addrstr(r_xprt), rpcrdma_portstr(r_xprt), - ia->ri_device->name, rdma_event_msg(event->event)); + ia->ri_id->device->name, rdma_event_msg(event->event)); return 0; } @@ -370,9 +365,8 @@ rpcrdma_ia_open(struct rpcrdma_xprt *xprt) rc = PTR_ERR(ia->ri_id); goto out_err; } - ia->ri_device = ia->ri_id->device; - ia->ri_pd = ib_alloc_pd(ia->ri_device, 0); + ia->ri_pd = ib_alloc_pd(ia->ri_id->device, 0); if (IS_ERR(ia->ri_pd)) { rc = PTR_ERR(ia->ri_pd); pr_err("rpcrdma: ib_alloc_pd() returned %d\n", rc); @@ -381,12 +375,12 @@ rpcrdma_ia_open(struct rpcrdma_xprt *xprt) switch (xprt_rdma_memreg_strategy) { case RPCRDMA_FRWR: - if (frwr_is_supported(ia)) + if (frwr_is_supported(ia->ri_id->device)) break; /*FALLTHROUGH*/ default: pr_err("rpcrdma: Device %s does not support memreg mode %d\n", - ia->ri_device->name, xprt_rdma_memreg_strategy); + ia->ri_id->device->name, xprt_rdma_memreg_strategy); rc = -EINVAL; goto out_err; } @@ -438,11 +432,11 @@ rpcrdma_ia_remove(struct rpcrdma_ia *ia) * mappings and MRs are gone. */ list_for_each_entry(rep, &buf->rb_recv_bufs, rr_list) - rpcrdma_dma_unmap_regbuf(rep->rr_rdmabuf); + rpcrdma_regbuf_dma_unmap(rep->rr_rdmabuf); list_for_each_entry(req, &buf->rb_allreqs, rl_all) { - rpcrdma_dma_unmap_regbuf(req->rl_rdmabuf); - rpcrdma_dma_unmap_regbuf(req->rl_sendbuf); - rpcrdma_dma_unmap_regbuf(req->rl_recvbuf); + rpcrdma_regbuf_dma_unmap(req->rl_rdmabuf); + rpcrdma_regbuf_dma_unmap(req->rl_sendbuf); + rpcrdma_regbuf_dma_unmap(req->rl_recvbuf); } rpcrdma_mrs_destroy(buf); ib_dealloc_pd(ia->ri_pd); @@ -468,7 +462,6 @@ rpcrdma_ia_close(struct rpcrdma_ia *ia) rdma_destroy_id(ia->ri_id); } ia->ri_id = NULL; - ia->ri_device = NULL; /* If the pd is still busy, xprtrdma missed freeing a resource */ if (ia->ri_pd && !IS_ERR(ia->ri_pd)) @@ -476,19 +469,26 @@ rpcrdma_ia_close(struct rpcrdma_ia *ia) ia->ri_pd = NULL; } -/* - * Create unconnected endpoint. +/** + * rpcrdma_ep_create - Create unconnected endpoint + * @r_xprt: transport to instantiate + * + * Returns zero on success, or a negative errno. */ -int -rpcrdma_ep_create(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia, - struct rpcrdma_create_data_internal *cdata) +int rpcrdma_ep_create(struct rpcrdma_xprt *r_xprt) { + struct rpcrdma_ep *ep = &r_xprt->rx_ep; + struct rpcrdma_ia *ia = &r_xprt->rx_ia; struct rpcrdma_connect_private *pmsg = &ep->rep_cm_private; struct ib_cq *sendcq, *recvcq; unsigned int max_sge; int rc; - max_sge = min_t(unsigned int, ia->ri_device->attrs.max_send_sge, + ep->rep_max_requests = xprt_rdma_slot_table_entries; + ep->rep_inline_send = xprt_rdma_max_inline_write; + ep->rep_inline_recv = xprt_rdma_max_inline_read; + + max_sge = min_t(unsigned int, ia->ri_id->device->attrs.max_send_sge, RPCRDMA_MAX_SEND_SGES); if (max_sge < RPCRDMA_MIN_SEND_SGES) { pr_warn("rpcrdma: HCA provides only %d send SGEs\n", max_sge); @@ -496,7 +496,7 @@ rpcrdma_ep_create(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia, } ia->ri_max_send_sges = max_sge; - rc = frwr_open(ia, ep, cdata); + rc = frwr_open(ia, ep); if (rc) return rc; @@ -518,23 +518,21 @@ rpcrdma_ep_create(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia, ep->rep_attr.cap.max_send_sge, ep->rep_attr.cap.max_recv_sge); - /* set trigger for requesting send completion */ - ep->rep_send_batch = min_t(unsigned int, RPCRDMA_MAX_SEND_BATCH, - cdata->max_requests >> 2); + ep->rep_send_batch = ep->rep_max_requests >> 3; ep->rep_send_count = ep->rep_send_batch; init_waitqueue_head(&ep->rep_connect_wait); ep->rep_receive_count = 0; - sendcq = ib_alloc_cq(ia->ri_device, NULL, + sendcq = ib_alloc_cq(ia->ri_id->device, NULL, ep->rep_attr.cap.max_send_wr + 1, - ia->ri_device->num_comp_vectors > 1 ? 1 : 0, + ia->ri_id->device->num_comp_vectors > 1 ? 1 : 0, IB_POLL_WORKQUEUE); if (IS_ERR(sendcq)) { rc = PTR_ERR(sendcq); goto out1; } - recvcq = ib_alloc_cq(ia->ri_device, NULL, + recvcq = ib_alloc_cq(ia->ri_id->device, NULL, ep->rep_attr.cap.max_recv_wr + 1, 0, IB_POLL_WORKQUEUE); if (IS_ERR(recvcq)) { @@ -552,15 +550,15 @@ rpcrdma_ep_create(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia, pmsg->cp_magic = rpcrdma_cmp_magic; pmsg->cp_version = RPCRDMA_CMP_VERSION; pmsg->cp_flags |= RPCRDMA_CMP_F_SND_W_INV_OK; - pmsg->cp_send_size = rpcrdma_encode_buffer_size(cdata->inline_wsize); - pmsg->cp_recv_size = rpcrdma_encode_buffer_size(cdata->inline_rsize); + pmsg->cp_send_size = rpcrdma_encode_buffer_size(ep->rep_inline_send); + pmsg->cp_recv_size = rpcrdma_encode_buffer_size(ep->rep_inline_recv); ep->rep_remote_cma.private_data = pmsg; ep->rep_remote_cma.private_data_len = sizeof(*pmsg); /* Client offers RDMA Read but does not initiate */ ep->rep_remote_cma.initiator_depth = 0; ep->rep_remote_cma.responder_resources = - min_t(int, U8_MAX, ia->ri_device->attrs.max_qp_rd_atom); + min_t(int, U8_MAX, ia->ri_id->device->attrs.max_qp_rd_atom); /* Limit transport retries so client can detect server * GID changes quickly. RPC layer handles re-establishing @@ -583,16 +581,16 @@ out1: return rc; } -/* - * rpcrdma_ep_destroy +/** + * rpcrdma_ep_destroy - Disconnect and destroy endpoint. + * @r_xprt: transport instance to shut down * - * Disconnect and destroy endpoint. After this, the only - * valid operations on the ep are to free it (if dynamically - * allocated) or re-create it. */ -void -rpcrdma_ep_destroy(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia) +void rpcrdma_ep_destroy(struct rpcrdma_xprt *r_xprt) { + struct rpcrdma_ep *ep = &r_xprt->rx_ep; + struct rpcrdma_ia *ia = &r_xprt->rx_ia; + if (ia->ri_id && ia->ri_id->qp) { rpcrdma_ep_disconnect(ep, ia); rdma_destroy_qp(ia->ri_id); @@ -622,7 +620,7 @@ rpcrdma_ep_recreate_xprt(struct rpcrdma_xprt *r_xprt, goto out1; rc = -ENOMEM; - err = rpcrdma_ep_create(ep, ia, &r_xprt->rx_data); + err = rpcrdma_ep_create(r_xprt); if (err) { pr_err("rpcrdma: rpcrdma_ep_create returned %d\n", err); goto out2; @@ -639,7 +637,7 @@ rpcrdma_ep_recreate_xprt(struct rpcrdma_xprt *r_xprt, return 0; out3: - rpcrdma_ep_destroy(ep, ia); + rpcrdma_ep_destroy(r_xprt); out2: rpcrdma_ia_close(ia); out1: @@ -672,7 +670,7 @@ rpcrdma_ep_reconnect(struct rpcrdma_xprt *r_xprt, struct rpcrdma_ep *ep, */ old = id; rc = -ENETUNREACH; - if (ia->ri_device != id->device) { + if (ia->ri_id->device != id->device) { pr_err("rpcrdma: can't reconnect on different device!\n"); goto out_destroy; } @@ -796,8 +794,8 @@ rpcrdma_ep_disconnect(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia) */ /* rpcrdma_sendctxs_destroy() assumes caller has already quiesced - * queue activity, and ib_drain_qp has flushed all remaining Send - * requests. + * queue activity, and rpcrdma_xprt_drain has flushed all remaining + * Send requests. */ static void rpcrdma_sendctxs_destroy(struct rpcrdma_buffer *buf) { @@ -867,20 +865,20 @@ static unsigned long rpcrdma_sendctx_next(struct rpcrdma_buffer *buf, /** * rpcrdma_sendctx_get_locked - Acquire a send context - * @buf: transport buffers from which to acquire an unused context + * @r_xprt: controlling transport instance * * Returns pointer to a free send completion context; or NULL if * the queue is empty. * * Usage: Called to acquire an SGE array before preparing a Send WR. * - * The caller serializes calls to this function (per rpcrdma_buffer), - * and provides an effective memory barrier that flushes the new value + * The caller serializes calls to this function (per transport), and + * provides an effective memory barrier that flushes the new value * of rb_sc_head. */ -struct rpcrdma_sendctx *rpcrdma_sendctx_get_locked(struct rpcrdma_buffer *buf) +struct rpcrdma_sendctx *rpcrdma_sendctx_get_locked(struct rpcrdma_xprt *r_xprt) { - struct rpcrdma_xprt *r_xprt; + struct rpcrdma_buffer *buf = &r_xprt->rx_buf; struct rpcrdma_sendctx *sc; unsigned long next_head; @@ -905,7 +903,6 @@ out_emptyq: * backing up. Cause the caller to pause and try again. */ set_bit(RPCRDMA_BUF_F_EMPTY_SCQ, &buf->rb_flags); - r_xprt = container_of(buf, struct rpcrdma_xprt, rx_buf); r_xprt->rx_stats.empty_sendctx_q++; return NULL; } @@ -917,7 +914,7 @@ out_emptyq: * Usage: Called from Send completion to return a sendctxt * to the queue. * - * The caller serializes calls to this function (per rpcrdma_buffer). + * The caller serializes calls to this function (per transport). */ static void rpcrdma_sendctx_put_locked(struct rpcrdma_sendctx *sc) @@ -925,7 +922,7 @@ rpcrdma_sendctx_put_locked(struct rpcrdma_sendctx *sc) struct rpcrdma_buffer *buf = &sc->sc_xprt->rx_buf; unsigned long next_tail; - /* Unmap SGEs of previously completed by unsignaled + /* Unmap SGEs of previously completed but unsignaled * Sends by walking up the queue until @sc is found. */ next_tail = buf->rb_sc_tail; @@ -933,7 +930,7 @@ rpcrdma_sendctx_put_locked(struct rpcrdma_sendctx *sc) next_tail = rpcrdma_sendctx_next(buf, next_tail); /* ORDER: item must be accessed _before_ tail is updated */ - rpcrdma_unmap_sendctx(buf->rb_sc_ctxs[next_tail]); + rpcrdma_sendctx_unmap(buf->rb_sc_ctxs[next_tail]); } while (buf->rb_sc_ctxs[next_tail] != sc); @@ -996,54 +993,70 @@ rpcrdma_mr_refresh_worker(struct work_struct *work) rpcrdma_mrs_create(r_xprt); } -struct rpcrdma_req * -rpcrdma_create_req(struct rpcrdma_xprt *r_xprt) +/** + * rpcrdma_req_create - Allocate an rpcrdma_req object + * @r_xprt: controlling r_xprt + * @size: initial size, in bytes, of send and receive buffers + * @flags: GFP flags passed to memory allocators + * + * Returns an allocated and fully initialized rpcrdma_req or NULL. + */ +struct rpcrdma_req *rpcrdma_req_create(struct rpcrdma_xprt *r_xprt, size_t size, + gfp_t flags) { struct rpcrdma_buffer *buffer = &r_xprt->rx_buf; struct rpcrdma_regbuf *rb; struct rpcrdma_req *req; - req = kzalloc(sizeof(*req), GFP_KERNEL); + req = kzalloc(sizeof(*req), flags); if (req == NULL) - return ERR_PTR(-ENOMEM); + goto out1; - rb = rpcrdma_alloc_regbuf(RPCRDMA_HDRBUF_SIZE, - DMA_TO_DEVICE, GFP_KERNEL); - if (IS_ERR(rb)) { - kfree(req); - return ERR_PTR(-ENOMEM); - } + rb = rpcrdma_regbuf_alloc(RPCRDMA_HDRBUF_SIZE, DMA_TO_DEVICE, flags); + if (!rb) + goto out2; req->rl_rdmabuf = rb; - xdr_buf_init(&req->rl_hdrbuf, rb->rg_base, rdmab_length(rb)); + xdr_buf_init(&req->rl_hdrbuf, rdmab_data(rb), rdmab_length(rb)); + + req->rl_sendbuf = rpcrdma_regbuf_alloc(size, DMA_TO_DEVICE, flags); + if (!req->rl_sendbuf) + goto out3; + + req->rl_recvbuf = rpcrdma_regbuf_alloc(size, DMA_NONE, flags); + if (!req->rl_recvbuf) + goto out4; + req->rl_buffer = buffer; INIT_LIST_HEAD(&req->rl_registered); - spin_lock(&buffer->rb_lock); list_add(&req->rl_all, &buffer->rb_allreqs); spin_unlock(&buffer->rb_lock); return req; + +out4: + kfree(req->rl_sendbuf); +out3: + kfree(req->rl_rdmabuf); +out2: + kfree(req); +out1: + return NULL; } -static int -rpcrdma_create_rep(struct rpcrdma_xprt *r_xprt, bool temp) +static bool rpcrdma_rep_create(struct rpcrdma_xprt *r_xprt, bool temp) { - struct rpcrdma_create_data_internal *cdata = &r_xprt->rx_data; struct rpcrdma_buffer *buf = &r_xprt->rx_buf; struct rpcrdma_rep *rep; - int rc; - rc = -ENOMEM; rep = kzalloc(sizeof(*rep), GFP_KERNEL); if (rep == NULL) goto out; - rep->rr_rdmabuf = rpcrdma_alloc_regbuf(cdata->inline_rsize, + rep->rr_rdmabuf = rpcrdma_regbuf_alloc(r_xprt->rx_ep.rep_inline_recv, DMA_FROM_DEVICE, GFP_KERNEL); - if (IS_ERR(rep->rr_rdmabuf)) { - rc = PTR_ERR(rep->rr_rdmabuf); + if (!rep->rr_rdmabuf) goto out_free; - } - xdr_buf_init(&rep->rr_hdrbuf, rep->rr_rdmabuf->rg_base, + xdr_buf_init(&rep->rr_hdrbuf, rdmab_data(rep->rr_rdmabuf), rdmab_length(rep->rr_rdmabuf)); rep->rr_cqe.done = rpcrdma_wc_receive; @@ -1058,22 +1071,27 @@ rpcrdma_create_rep(struct rpcrdma_xprt *r_xprt, bool temp) spin_lock(&buf->rb_lock); list_add(&rep->rr_list, &buf->rb_recv_bufs); spin_unlock(&buf->rb_lock); - return 0; + return true; out_free: kfree(rep); out: - return rc; + return false; } -int -rpcrdma_buffer_create(struct rpcrdma_xprt *r_xprt) +/** + * rpcrdma_buffer_create - Create initial set of req/rep objects + * @r_xprt: transport instance to (re)initialize + * + * Returns zero on success, otherwise a negative errno. + */ +int rpcrdma_buffer_create(struct rpcrdma_xprt *r_xprt) { struct rpcrdma_buffer *buf = &r_xprt->rx_buf; int i, rc; buf->rb_flags = 0; - buf->rb_max_requests = r_xprt->rx_data.max_requests; + buf->rb_max_requests = r_xprt->rx_ep.rep_max_requests; buf->rb_bc_srv_max_requests = 0; spin_lock_init(&buf->rb_mrlock); spin_lock_init(&buf->rb_lock); @@ -1086,16 +1104,15 @@ rpcrdma_buffer_create(struct rpcrdma_xprt *r_xprt) INIT_LIST_HEAD(&buf->rb_send_bufs); INIT_LIST_HEAD(&buf->rb_allreqs); + + rc = -ENOMEM; for (i = 0; i < buf->rb_max_requests; i++) { struct rpcrdma_req *req; - req = rpcrdma_create_req(r_xprt); - if (IS_ERR(req)) { - dprintk("RPC: %s: request buffer %d alloc" - " failed\n", __func__, i); - rc = PTR_ERR(req); + req = rpcrdma_req_create(r_xprt, RPCRDMA_V1_DEF_INLINE_SIZE, + GFP_KERNEL); + if (!req) goto out; - } list_add(&req->rl_list, &buf->rb_send_bufs); } @@ -1121,10 +1138,9 @@ out: return rc; } -static void -rpcrdma_destroy_rep(struct rpcrdma_rep *rep) +static void rpcrdma_rep_destroy(struct rpcrdma_rep *rep) { - rpcrdma_free_regbuf(rep->rr_rdmabuf); + rpcrdma_regbuf_free(rep->rr_rdmabuf); kfree(rep); } @@ -1140,9 +1156,9 @@ rpcrdma_req_destroy(struct rpcrdma_req *req) { list_del(&req->rl_all); - rpcrdma_free_regbuf(req->rl_recvbuf); - rpcrdma_free_regbuf(req->rl_sendbuf); - rpcrdma_free_regbuf(req->rl_rdmabuf); + rpcrdma_regbuf_free(req->rl_recvbuf); + rpcrdma_regbuf_free(req->rl_sendbuf); + rpcrdma_regbuf_free(req->rl_rdmabuf); kfree(req); } @@ -1180,7 +1196,7 @@ rpcrdma_mrs_destroy(struct rpcrdma_buffer *buf) * rpcrdma_buffer_destroy - Release all hw resources * @buf: root control block for resources * - * ORDERING: relies on a prior ib_drain_qp : + * ORDERING: relies on a prior rpcrdma_xprt_drain : * - No more Send or Receive completions can occur * - All MRs, reps, and reqs are returned to their free lists */ @@ -1202,7 +1218,7 @@ rpcrdma_buffer_destroy(struct rpcrdma_buffer *buf) rep = list_first_entry(&buf->rb_recv_bufs, struct rpcrdma_rep, rr_list); list_del(&rep->rr_list); - rpcrdma_destroy_rep(rep); + rpcrdma_rep_destroy(rep); } while (!list_empty(&buf->rb_send_bufs)) { @@ -1281,7 +1297,7 @@ rpcrdma_mr_unmap_and_put(struct rpcrdma_mr *mr) if (mr->mr_dir != DMA_NONE) { trace_xprtrdma_mr_unmap(mr); - ib_dma_unmap_sg(r_xprt->rx_ia.ri_device, + ib_dma_unmap_sg(r_xprt->rx_ia.ri_id->device, mr->mr_sg, mr->mr_nents, mr->mr_dir); mr->mr_dir = DMA_NONE; } @@ -1331,7 +1347,7 @@ rpcrdma_buffer_put(struct rpcrdma_req *req) } spin_unlock(&buffers->rb_lock); if (rep) - rpcrdma_destroy_rep(rep); + rpcrdma_rep_destroy(rep); } /* @@ -1348,69 +1364,90 @@ rpcrdma_recv_buffer_put(struct rpcrdma_rep *rep) list_add(&rep->rr_list, &buffers->rb_recv_bufs); spin_unlock(&buffers->rb_lock); } else { - rpcrdma_destroy_rep(rep); + rpcrdma_rep_destroy(rep); } } -/** - * rpcrdma_alloc_regbuf - allocate and DMA-map memory for SEND/RECV buffers - * @size: size of buffer to be allocated, in bytes - * @direction: direction of data movement - * @flags: GFP flags - * - * Returns an ERR_PTR, or a pointer to a regbuf, a buffer that - * can be persistently DMA-mapped for I/O. +/* Returns a pointer to a rpcrdma_regbuf object, or NULL. * * xprtrdma uses a regbuf for posting an outgoing RDMA SEND, or for * receiving the payload of RDMA RECV operations. During Long Calls * or Replies they may be registered externally via frwr_map. */ -struct rpcrdma_regbuf * -rpcrdma_alloc_regbuf(size_t size, enum dma_data_direction direction, +static struct rpcrdma_regbuf * +rpcrdma_regbuf_alloc(size_t size, enum dma_data_direction direction, gfp_t flags) { struct rpcrdma_regbuf *rb; - rb = kmalloc(sizeof(*rb) + size, flags); - if (rb == NULL) - return ERR_PTR(-ENOMEM); + rb = kmalloc(sizeof(*rb), flags); + if (!rb) + return NULL; + rb->rg_data = kmalloc(size, flags); + if (!rb->rg_data) { + kfree(rb); + return NULL; + } rb->rg_device = NULL; rb->rg_direction = direction; rb->rg_iov.length = size; - return rb; } /** - * __rpcrdma_map_regbuf - DMA-map a regbuf - * @ia: controlling rpcrdma_ia + * rpcrdma_regbuf_realloc - re-allocate a SEND/RECV buffer + * @rb: regbuf to reallocate + * @size: size of buffer to be allocated, in bytes + * @flags: GFP flags + * + * Returns true if reallocation was successful. If false is + * returned, @rb is left untouched. + */ +bool rpcrdma_regbuf_realloc(struct rpcrdma_regbuf *rb, size_t size, gfp_t flags) +{ + void *buf; + + buf = kmalloc(size, flags); + if (!buf) + return false; + + rpcrdma_regbuf_dma_unmap(rb); + kfree(rb->rg_data); + + rb->rg_data = buf; + rb->rg_iov.length = size; + return true; +} + +/** + * __rpcrdma_regbuf_dma_map - DMA-map a regbuf + * @r_xprt: controlling transport instance * @rb: regbuf to be mapped + * + * Returns true if the buffer is now DMA mapped to @r_xprt's device */ -bool -__rpcrdma_dma_map_regbuf(struct rpcrdma_ia *ia, struct rpcrdma_regbuf *rb) +bool __rpcrdma_regbuf_dma_map(struct rpcrdma_xprt *r_xprt, + struct rpcrdma_regbuf *rb) { - struct ib_device *device = ia->ri_device; + struct ib_device *device = r_xprt->rx_ia.ri_id->device; if (rb->rg_direction == DMA_NONE) return false; - rb->rg_iov.addr = ib_dma_map_single(device, - (void *)rb->rg_base, - rdmab_length(rb), - rb->rg_direction); + rb->rg_iov.addr = ib_dma_map_single(device, rdmab_data(rb), + rdmab_length(rb), rb->rg_direction); if (ib_dma_mapping_error(device, rdmab_addr(rb))) { trace_xprtrdma_dma_maperr(rdmab_addr(rb)); return false; } rb->rg_device = device; - rb->rg_iov.lkey = ia->ri_pd->local_dma_lkey; + rb->rg_iov.lkey = r_xprt->rx_ia.ri_pd->local_dma_lkey; return true; } -static void -rpcrdma_dma_unmap_regbuf(struct rpcrdma_regbuf *rb) +static void rpcrdma_regbuf_dma_unmap(struct rpcrdma_regbuf *rb) { if (!rb) return; @@ -1418,19 +1455,16 @@ rpcrdma_dma_unmap_regbuf(struct rpcrdma_regbuf *rb) if (!rpcrdma_regbuf_is_mapped(rb)) return; - ib_dma_unmap_single(rb->rg_device, rdmab_addr(rb), - rdmab_length(rb), rb->rg_direction); + ib_dma_unmap_single(rb->rg_device, rdmab_addr(rb), rdmab_length(rb), + rb->rg_direction); rb->rg_device = NULL; } -/** - * rpcrdma_free_regbuf - deregister and free registered buffer - * @rb: regbuf to be deregistered and freed - */ -void -rpcrdma_free_regbuf(struct rpcrdma_regbuf *rb) +static void rpcrdma_regbuf_free(struct rpcrdma_regbuf *rb) { - rpcrdma_dma_unmap_regbuf(rb); + rpcrdma_regbuf_dma_unmap(rb); + if (rb) + kfree(rb->rg_data); kfree(rb); } @@ -1497,17 +1531,15 @@ rpcrdma_post_recvs(struct rpcrdma_xprt *r_xprt, bool temp) list_del(&rep->rr_list); spin_unlock(&buf->rb_lock); if (!rep) { - if (rpcrdma_create_rep(r_xprt, temp)) + if (!rpcrdma_rep_create(r_xprt, temp)) break; continue; } rb = rep->rr_rdmabuf; - if (!rpcrdma_regbuf_is_mapped(rb)) { - if (!__rpcrdma_dma_map_regbuf(&r_xprt->rx_ia, rb)) { - rpcrdma_recv_buffer_put(rep); - break; - } + if (!rpcrdma_regbuf_dma_map(r_xprt, rb)) { + rpcrdma_recv_buffer_put(rep); + break; } trace_xprtrdma_post_recv(rep->rr_recv_wr.wr_cqe); diff --git a/net/sunrpc/xprtrdma/xprt_rdma.h b/net/sunrpc/xprtrdma/xprt_rdma.h index 10f6593e1a6a..d1e0749bcbc4 100644 --- a/net/sunrpc/xprtrdma/xprt_rdma.h +++ b/net/sunrpc/xprtrdma/xprt_rdma.h @@ -66,20 +66,17 @@ * Interface Adapter -- one per transport instance */ struct rpcrdma_ia { - struct ib_device *ri_device; struct rdma_cm_id *ri_id; struct ib_pd *ri_pd; - struct completion ri_done; - struct completion ri_remove_done; int ri_async_rc; unsigned int ri_max_segs; unsigned int ri_max_frwr_depth; - unsigned int ri_max_inline_write; - unsigned int ri_max_inline_read; unsigned int ri_max_send_sges; bool ri_implicit_roundup; enum ib_mr_type ri_mrtype; unsigned long ri_flags; + struct completion ri_done; + struct completion ri_remove_done; }; enum { @@ -93,22 +90,29 @@ enum { struct rpcrdma_ep { unsigned int rep_send_count; unsigned int rep_send_batch; + unsigned int rep_max_inline_send; + unsigned int rep_max_inline_recv; int rep_connected; struct ib_qp_init_attr rep_attr; wait_queue_head_t rep_connect_wait; struct rpcrdma_connect_private rep_cm_private; struct rdma_conn_param rep_remote_cma; + unsigned int rep_max_requests; /* set by /proc */ + unsigned int rep_inline_send; /* negotiated */ + unsigned int rep_inline_recv; /* negotiated */ int rep_receive_count; }; /* Pre-allocate extra Work Requests for handling backward receives * and sends. This is a fixed value because the Work Queues are - * allocated when the forward channel is set up. + * allocated when the forward channel is set up, long before the + * backchannel is provisioned. This value is two times + * NFS4_DEF_CB_SLOT_TABLE_SIZE. */ #if defined(CONFIG_SUNRPC_BACKCHANNEL) -#define RPCRDMA_BACKWARD_WRS (8) +#define RPCRDMA_BACKWARD_WRS (32) #else -#define RPCRDMA_BACKWARD_WRS (0) +#define RPCRDMA_BACKWARD_WRS (0) #endif /* Registered buffer -- registered kmalloc'd memory for RDMA SEND/RECV @@ -121,33 +125,34 @@ struct rpcrdma_regbuf { struct ib_sge rg_iov; struct ib_device *rg_device; enum dma_data_direction rg_direction; - __be32 rg_base[0] __attribute__ ((aligned(256))); + void *rg_data; }; -static inline u64 -rdmab_addr(struct rpcrdma_regbuf *rb) +static inline u64 rdmab_addr(struct rpcrdma_regbuf *rb) { return rb->rg_iov.addr; } -static inline u32 -rdmab_length(struct rpcrdma_regbuf *rb) +static inline u32 rdmab_length(struct rpcrdma_regbuf *rb) { return rb->rg_iov.length; } -static inline u32 -rdmab_lkey(struct rpcrdma_regbuf *rb) +static inline u32 rdmab_lkey(struct rpcrdma_regbuf *rb) { return rb->rg_iov.lkey; } -static inline struct ib_device * -rdmab_device(struct rpcrdma_regbuf *rb) +static inline struct ib_device *rdmab_device(struct rpcrdma_regbuf *rb) { return rb->rg_device; } +static inline void *rdmab_data(const struct rpcrdma_regbuf *rb) +{ + return rb->rg_data; +} + #define RPCRDMA_DEF_GFP (GFP_NOIO | __GFP_NOWARN) /* To ensure a transport can always make forward progress, @@ -222,34 +227,18 @@ struct rpcrdma_xprt; struct rpcrdma_sendctx { struct ib_send_wr sc_wr; struct ib_cqe sc_cqe; + struct ib_device *sc_device; struct rpcrdma_xprt *sc_xprt; struct rpcrdma_req *sc_req; unsigned int sc_unmap_count; struct ib_sge sc_sges[]; }; -/* Limit the number of SGEs that can be unmapped during one - * Send completion. This caps the amount of work a single - * completion can do before returning to the provider. - * - * Setting this to zero disables Send completion batching. - */ -enum { - RPCRDMA_MAX_SEND_BATCH = 7, -}; - /* * struct rpcrdma_mr - external memory region metadata * * An external memory region is any buffer or page that is registered * on the fly (ie, not pre-registered). - * - * Each rpcrdma_buffer has a list of free MWs anchored in rb_mrs. During - * call_allocate, rpcrdma_buffer_get() assigns one to each segment in - * an rpcrdma_req. Then rpcrdma_register_external() grabs these to keep - * track of registration metadata while each RPC is pending. - * rpcrdma_deregister_external() uses this metadata to unmap and - * release these resources when an RPC is complete. */ enum rpcrdma_frwr_state { FRWR_IS_INVALID, /* ready to be used */ @@ -419,20 +408,6 @@ enum { }; /* - * Internal structure for transport instance creation. This - * exists primarily for modularity. - * - * This data should be set with mount options - */ -struct rpcrdma_create_data_internal { - unsigned int max_requests; /* max requests (slots) in flight */ - unsigned int rsize; /* mount rsize - max read hdr+data */ - unsigned int wsize; /* mount wsize - max write hdr+data */ - unsigned int inline_rsize; /* max non-rdma read data payload */ - unsigned int inline_wsize; /* max non-rdma write data payload */ -}; - -/* * Statistics for RPCRDMA */ struct rpcrdma_stats { @@ -476,13 +451,11 @@ struct rpcrdma_xprt { struct rpcrdma_ia rx_ia; struct rpcrdma_ep rx_ep; struct rpcrdma_buffer rx_buf; - struct rpcrdma_create_data_internal rx_data; struct delayed_work rx_connect_worker; struct rpcrdma_stats rx_stats; }; #define rpcx_to_rdmax(x) container_of(x, struct rpcrdma_xprt, rx_xprt) -#define rpcx_to_rdmad(x) (rpcx_to_rdmax(x)->rx_data) static inline const char * rpcrdma_addrstr(const struct rpcrdma_xprt *r_xprt) @@ -516,9 +489,8 @@ void rpcrdma_ia_close(struct rpcrdma_ia *); /* * Endpoint calls - xprtrdma/verbs.c */ -int rpcrdma_ep_create(struct rpcrdma_ep *, struct rpcrdma_ia *, - struct rpcrdma_create_data_internal *); -void rpcrdma_ep_destroy(struct rpcrdma_ep *, struct rpcrdma_ia *); +int rpcrdma_ep_create(struct rpcrdma_xprt *r_xprt); +void rpcrdma_ep_destroy(struct rpcrdma_xprt *r_xprt); int rpcrdma_ep_connect(struct rpcrdma_ep *, struct rpcrdma_ia *); void rpcrdma_ep_disconnect(struct rpcrdma_ep *, struct rpcrdma_ia *); @@ -528,11 +500,12 @@ int rpcrdma_ep_post(struct rpcrdma_ia *, struct rpcrdma_ep *, /* * Buffer calls - xprtrdma/verbs.c */ -struct rpcrdma_req *rpcrdma_create_req(struct rpcrdma_xprt *); +struct rpcrdma_req *rpcrdma_req_create(struct rpcrdma_xprt *r_xprt, size_t size, + gfp_t flags); void rpcrdma_req_destroy(struct rpcrdma_req *req); int rpcrdma_buffer_create(struct rpcrdma_xprt *); void rpcrdma_buffer_destroy(struct rpcrdma_buffer *); -struct rpcrdma_sendctx *rpcrdma_sendctx_get_locked(struct rpcrdma_buffer *buf); +struct rpcrdma_sendctx *rpcrdma_sendctx_get_locked(struct rpcrdma_xprt *r_xprt); struct rpcrdma_mr *rpcrdma_mr_get(struct rpcrdma_xprt *r_xprt); void rpcrdma_mr_put(struct rpcrdma_mr *mr); @@ -548,23 +521,34 @@ struct rpcrdma_req *rpcrdma_buffer_get(struct rpcrdma_buffer *); void rpcrdma_buffer_put(struct rpcrdma_req *); void rpcrdma_recv_buffer_put(struct rpcrdma_rep *); -struct rpcrdma_regbuf *rpcrdma_alloc_regbuf(size_t, enum dma_data_direction, - gfp_t); -bool __rpcrdma_dma_map_regbuf(struct rpcrdma_ia *, struct rpcrdma_regbuf *); -void rpcrdma_free_regbuf(struct rpcrdma_regbuf *); +bool rpcrdma_regbuf_realloc(struct rpcrdma_regbuf *rb, size_t size, + gfp_t flags); +bool __rpcrdma_regbuf_dma_map(struct rpcrdma_xprt *r_xprt, + struct rpcrdma_regbuf *rb); -static inline bool -rpcrdma_regbuf_is_mapped(struct rpcrdma_regbuf *rb) +/** + * rpcrdma_regbuf_is_mapped - check if buffer is DMA mapped + * + * Returns true if the buffer is now mapped to rb->rg_device. + */ +static inline bool rpcrdma_regbuf_is_mapped(struct rpcrdma_regbuf *rb) { return rb->rg_device != NULL; } -static inline bool -rpcrdma_dma_map_regbuf(struct rpcrdma_ia *ia, struct rpcrdma_regbuf *rb) +/** + * rpcrdma_regbuf_dma_map - DMA-map a regbuf + * @r_xprt: controlling transport instance + * @rb: regbuf to be mapped + * + * Returns true if the buffer is currently DMA mapped. + */ +static inline bool rpcrdma_regbuf_dma_map(struct rpcrdma_xprt *r_xprt, + struct rpcrdma_regbuf *rb) { if (likely(rpcrdma_regbuf_is_mapped(rb))) return true; - return __rpcrdma_dma_map_regbuf(ia, rb); + return __rpcrdma_regbuf_dma_map(r_xprt, rb); } /* @@ -579,9 +563,8 @@ rpcrdma_data_dir(bool writing) /* Memory registration calls xprtrdma/frwr_ops.c */ -bool frwr_is_supported(struct rpcrdma_ia *); -int frwr_open(struct rpcrdma_ia *ia, struct rpcrdma_ep *ep, - struct rpcrdma_create_data_internal *cdata); +bool frwr_is_supported(struct ib_device *device); +int frwr_open(struct rpcrdma_ia *ia, struct rpcrdma_ep *ep); int frwr_init_mr(struct rpcrdma_ia *ia, struct rpcrdma_mr *mr); void frwr_release_mr(struct rpcrdma_mr *mr); size_t frwr_maxpages(struct rpcrdma_xprt *r_xprt); @@ -610,7 +593,7 @@ int rpcrdma_prepare_send_sges(struct rpcrdma_xprt *r_xprt, struct rpcrdma_req *req, u32 hdrlen, struct xdr_buf *xdr, enum rpcrdma_chunktype rtype); -void rpcrdma_unmap_sendctx(struct rpcrdma_sendctx *sc); +void rpcrdma_sendctx_unmap(struct rpcrdma_sendctx *sc); int rpcrdma_marshal_req(struct rpcrdma_xprt *r_xprt, struct rpc_rqst *rqst); void rpcrdma_set_max_header_sizes(struct rpcrdma_xprt *); void rpcrdma_complete_rqst(struct rpcrdma_rep *rep); @@ -627,7 +610,9 @@ static inline void rpcrdma_set_xdrlen(struct xdr_buf *xdr, size_t len) /* RPC/RDMA module init - xprtrdma/transport.c */ +extern unsigned int xprt_rdma_slot_table_entries; extern unsigned int xprt_rdma_max_inline_read; +extern unsigned int xprt_rdma_max_inline_write; void xprt_rdma_format_addresses(struct rpc_xprt *xprt, struct sockaddr *sap); void xprt_rdma_free_addresses(struct rpc_xprt *xprt); void xprt_rdma_close(struct rpc_xprt *xprt); diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c index 732d4b57411a..c69951ed2ebc 100644 --- a/net/sunrpc/xprtsock.c +++ b/net/sunrpc/xprtsock.c @@ -2017,6 +2017,7 @@ static void xs_local_connect(struct rpc_xprt *xprt, struct rpc_task *task) * we'll need to figure out how to pass a namespace to * connect. */ + task->tk_rpc_status = -ENOTCONN; rpc_exit(task, -ENOTCONN); return; } @@ -2690,7 +2691,7 @@ static const struct rpc_xprt_ops xs_local_ops = { .buf_free = rpc_free, .prepare_request = xs_stream_prepare_request, .send_request = xs_local_send_request, - .set_retrans_timeout = xprt_set_retrans_timeout_def, + .wait_for_reply_request = xprt_wait_for_reply_request_def, .close = xs_close, .destroy = xs_destroy, .print_stats = xs_local_print_stats, @@ -2710,7 +2711,7 @@ static const struct rpc_xprt_ops xs_udp_ops = { .buf_alloc = rpc_malloc, .buf_free = rpc_free, .send_request = xs_udp_send_request, - .set_retrans_timeout = xprt_set_retrans_timeout_rtt, + .wait_for_reply_request = xprt_wait_for_reply_request_rtt, .timer = xs_udp_timer, .release_request = xprt_release_rqst_cong, .close = xs_close, @@ -2733,7 +2734,7 @@ static const struct rpc_xprt_ops xs_tcp_ops = { .buf_free = rpc_free, .prepare_request = xs_stream_prepare_request, .send_request = xs_tcp_send_request, - .set_retrans_timeout = xprt_set_retrans_timeout_def, + .wait_for_reply_request = xprt_wait_for_reply_request_def, .close = xs_tcp_shutdown, .destroy = xs_destroy, .set_connect_timeout = xs_tcp_set_connect_timeout, @@ -2761,7 +2762,7 @@ static const struct rpc_xprt_ops bc_tcp_ops = { .buf_alloc = bc_malloc, .buf_free = bc_free, .send_request = bc_send_request, - .set_retrans_timeout = xprt_set_retrans_timeout_def, + .wait_for_reply_request = xprt_wait_for_reply_request_def, .close = bc_close, .destroy = bc_destroy, .print_stats = xs_tcp_print_stats, diff --git a/net/tipc/socket.c b/net/tipc/socket.c index 145e4decb0c9..dd8537f988c4 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -736,11 +736,11 @@ static __poll_t tipc_poll(struct file *file, struct socket *sock, switch (sk->sk_state) { case TIPC_ESTABLISHED: - case TIPC_CONNECTING: if (!tsk->cong_link_cnt && !tsk_conn_cong(tsk)) revents |= EPOLLOUT; /* fall through */ case TIPC_LISTEN: + case TIPC_CONNECTING: if (!skb_queue_empty(&sk->sk_receive_queue)) revents |= EPOLLIN | EPOLLRDNORM; break; @@ -2043,7 +2043,7 @@ static bool tipc_sk_filter_connect(struct tipc_sock *tsk, struct sk_buff *skb) if (msg_data_sz(hdr)) return true; /* Empty ACK-, - wake up sleeping connect() and drop */ - sk->sk_data_ready(sk); + sk->sk_state_change(sk); msg_set_dest_droppable(hdr, 1); return false; } diff --git a/net/tls/tls_device.c b/net/tls/tls_device.c index e225c81e6b35..ca54a7c7ec81 100644 --- a/net/tls/tls_device.c +++ b/net/tls/tls_device.c @@ -541,14 +541,11 @@ static int tls_device_push_pending_record(struct sock *sk, int flags) void tls_device_write_space(struct sock *sk, struct tls_context *ctx) { - int rc = 0; - if (!sk->sk_write_pending && tls_is_partially_sent_record(ctx)) { gfp_t sk_allocation = sk->sk_allocation; sk->sk_allocation = GFP_ATOMIC; - rc = tls_push_partial_record(sk, ctx, - MSG_DONTWAIT | MSG_NOSIGNAL); + tls_push_partial_record(sk, ctx, MSG_DONTWAIT | MSG_NOSIGNAL); sk->sk_allocation = sk_allocation; } } @@ -1036,4 +1033,5 @@ void __exit tls_device_cleanup(void) { unregister_netdevice_notifier(&tls_dev_notifier); flush_work(&tls_device_gc_work); + clean_acked_data_flush(); } diff --git a/net/tls/tls_sw.c b/net/tls/tls_sw.c index c02293fb10e6..d93f83f77864 100644 --- a/net/tls/tls_sw.c +++ b/net/tls/tls_sw.c @@ -119,23 +119,25 @@ static int skb_nsg(struct sk_buff *skb, int offset, int len) } static int padding_length(struct tls_sw_context_rx *ctx, - struct tls_context *tls_ctx, struct sk_buff *skb) + struct tls_prot_info *prot, struct sk_buff *skb) { struct strp_msg *rxm = strp_msg(skb); int sub = 0; /* Determine zero-padding length */ - if (tls_ctx->prot_info.version == TLS_1_3_VERSION) { + if (prot->version == TLS_1_3_VERSION) { char content_type = 0; int err; int back = 17; while (content_type == 0) { - if (back > rxm->full_len) + if (back > rxm->full_len - prot->prepend_size) return -EBADMSG; err = skb_copy_bits(skb, rxm->offset + rxm->full_len - back, &content_type, 1); + if (err) + return err; if (content_type) break; sub++; @@ -170,9 +172,17 @@ static void tls_decrypt_done(struct crypto_async_request *req, int err) tls_err_abort(skb->sk, err); } else { struct strp_msg *rxm = strp_msg(skb); - rxm->full_len -= padding_length(ctx, tls_ctx, skb); - rxm->offset += prot->prepend_size; - rxm->full_len -= prot->overhead_size; + int pad; + + pad = padding_length(ctx, prot, skb); + if (pad < 0) { + ctx->async_wait.err = pad; + tls_err_abort(skb->sk, pad); + } else { + rxm->full_len -= pad; + rxm->offset += prot->prepend_size; + rxm->full_len -= prot->overhead_size; + } } /* After using skb->sk to propagate sk through crypto async callback @@ -1478,7 +1488,7 @@ static int decrypt_skb_update(struct sock *sk, struct sk_buff *skb, struct tls_prot_info *prot = &tls_ctx->prot_info; int version = prot->version; struct strp_msg *rxm = strp_msg(skb); - int err = 0; + int pad, err = 0; if (!ctx->decrypted) { #ifdef CONFIG_TLS_DEVICE @@ -1501,7 +1511,11 @@ static int decrypt_skb_update(struct sock *sk, struct sk_buff *skb, *zc = false; } - rxm->full_len -= padding_length(ctx, tls_ctx, skb); + pad = padding_length(ctx, prot, skb); + if (pad < 0) + return pad; + + rxm->full_len -= pad; rxm->offset += prot->prepend_size; rxm->full_len -= prot->overhead_size; tls_advance_record_sn(sk, &tls_ctx->rx, version); diff --git a/net/xdp/xdp_umem.c b/net/xdp/xdp_umem.c index 989e52386c35..2b18223e7eb8 100644 --- a/net/xdp/xdp_umem.c +++ b/net/xdp/xdp_umem.c @@ -253,8 +253,8 @@ static int xdp_umem_pin_pages(struct xdp_umem *umem) return -ENOMEM; down_read(¤t->mm->mmap_sem); - npgs = get_user_pages_longterm(umem->address, umem->npgs, - gup_flags, &umem->pgs[0], NULL); + npgs = get_user_pages(umem->address, umem->npgs, + gup_flags | FOLL_LONGTERM, &umem->pgs[0], NULL); up_read(¤t->mm->mmap_sem); if (npgs != umem->npgs) { |