From 302fc6bbcba4beee6ff5e73c2fcc257e62667d4d Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Thu, 17 Oct 2024 11:31:37 -0700 Subject: [PATCH 01/16] phonet: Pass ifindex to fill_route(). We will convert route_doit() to RCU. route_doit() will call rtm_phonet_notify() outside of RCU due to GFP_KERNEL, so dev will not be available in fill_route(). Let's pass ifindex directly to fill_route(). Signed-off-by: Kuniyuki Iwashima Reviewed-by: Eric Dumazet Signed-off-by: Paolo Abeni --- net/phonet/pn_netlink.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/net/phonet/pn_netlink.c b/net/phonet/pn_netlink.c index 14928fa04675..c9a4215ec560 100644 --- a/net/phonet/pn_netlink.c +++ b/net/phonet/pn_netlink.c @@ -170,8 +170,8 @@ out: /* Routes handling */ -static int fill_route(struct sk_buff *skb, struct net_device *dev, u8 dst, - u32 portid, u32 seq, int event) +static int fill_route(struct sk_buff *skb, u32 ifindex, u8 dst, + u32 portid, u32 seq, int event) { struct rtmsg *rtm; struct nlmsghdr *nlh; @@ -190,8 +190,7 @@ static int fill_route(struct sk_buff *skb, struct net_device *dev, u8 dst, rtm->rtm_scope = RT_SCOPE_UNIVERSE; rtm->rtm_type = RTN_UNICAST; rtm->rtm_flags = 0; - if (nla_put_u8(skb, RTA_DST, dst) || - nla_put_u32(skb, RTA_OIF, READ_ONCE(dev->ifindex))) + if (nla_put_u8(skb, RTA_DST, dst) || nla_put_u32(skb, RTA_OIF, ifindex)) goto nla_put_failure; nlmsg_end(skb, nlh); return 0; @@ -210,7 +209,8 @@ void rtm_phonet_notify(int event, struct net_device *dev, u8 dst) nla_total_size(1) + nla_total_size(4), GFP_KERNEL); if (skb == NULL) goto errout; - err = fill_route(skb, dev, dst, 0, 0, event); + + err = fill_route(skb, dev->ifindex, dst, 0, 0, event); if (err < 0) { WARN_ON(err == -EMSGSIZE); kfree_skb(skb); @@ -286,7 +286,7 @@ static int route_dumpit(struct sk_buff *skb, struct netlink_callback *cb) if (!dev) continue; - err = fill_route(skb, dev, addr << 2, + err = fill_route(skb, READ_ONCE(dev->ifindex), addr << 2, NETLINK_CB(cb->skb).portid, cb->nlh->nlmsg_seq, RTM_NEWROUTE); if (err < 0) -- 2.51.0 From de51ad08b1177bbbb8b60cb7dd4c3c5dd50d262f Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Thu, 17 Oct 2024 11:31:38 -0700 Subject: [PATCH 02/16] phonet: Pass net and ifindex to rtm_phonet_notify(). Currently, rtm_phonet_notify() fetches netns and ifindex from dev. Once route_doit() is converted to RCU, rtm_phonet_notify() will be called outside of RCU due to GFP_KERNEL, and dev will be unavailable there. Let's pass net and ifindex to rtm_phonet_notify(). Signed-off-by: Kuniyuki Iwashima Reviewed-by: Eric Dumazet Signed-off-by: Paolo Abeni --- include/net/phonet/pn_dev.h | 2 +- net/phonet/pn_dev.c | 10 +++++++--- net/phonet/pn_netlink.c | 16 +++++++++------- 3 files changed, 17 insertions(+), 11 deletions(-) diff --git a/include/net/phonet/pn_dev.h b/include/net/phonet/pn_dev.h index ac0331d83a81..021e524fd20a 100644 --- a/include/net/phonet/pn_dev.h +++ b/include/net/phonet/pn_dev.h @@ -43,7 +43,7 @@ void phonet_address_notify(struct net *net, int event, u32 ifindex, u8 addr); int phonet_route_add(struct net_device *dev, u8 daddr); int phonet_route_del(struct net_device *dev, u8 daddr); -void rtm_phonet_notify(int event, struct net_device *dev, u8 dst); +void rtm_phonet_notify(struct net *net, int event, u32 ifindex, u8 dst); struct net_device *phonet_route_get_rcu(struct net *net, u8 daddr); struct net_device *phonet_route_output(struct net *net, u8 daddr); diff --git a/net/phonet/pn_dev.c b/net/phonet/pn_dev.c index 545279ef5910..6ded0d347b9f 100644 --- a/net/phonet/pn_dev.c +++ b/net/phonet/pn_dev.c @@ -263,9 +263,13 @@ static int phonet_device_autoconf(struct net_device *dev) static void phonet_route_autodel(struct net_device *dev) { - struct phonet_net *pnn = phonet_pernet(dev_net(dev)); - unsigned int i; + struct net *net = dev_net(dev); DECLARE_BITMAP(deleted, 64); + u32 ifindex = dev->ifindex; + struct phonet_net *pnn; + unsigned int i; + + pnn = phonet_pernet(net); /* Remove left-over Phonet routes */ bitmap_zero(deleted, 64); @@ -281,7 +285,7 @@ static void phonet_route_autodel(struct net_device *dev) return; /* short-circuit RCU */ synchronize_rcu(); for_each_set_bit(i, deleted, 64) { - rtm_phonet_notify(RTM_DELROUTE, dev, i); + rtm_phonet_notify(net, RTM_DELROUTE, ifindex, i); dev_put(dev); } } diff --git a/net/phonet/pn_netlink.c b/net/phonet/pn_netlink.c index c9a4215ec560..bfec5bd639b6 100644 --- a/net/phonet/pn_netlink.c +++ b/net/phonet/pn_netlink.c @@ -200,7 +200,7 @@ nla_put_failure: return -EMSGSIZE; } -void rtm_phonet_notify(int event, struct net_device *dev, u8 dst) +void rtm_phonet_notify(struct net *net, int event, u32 ifindex, u8 dst) { struct sk_buff *skb; int err = -ENOBUFS; @@ -210,17 +210,17 @@ void rtm_phonet_notify(int event, struct net_device *dev, u8 dst) if (skb == NULL) goto errout; - err = fill_route(skb, dev->ifindex, dst, 0, 0, event); + err = fill_route(skb, ifindex, dst, 0, 0, event); if (err < 0) { WARN_ON(err == -EMSGSIZE); kfree_skb(skb); goto errout; } - rtnl_notify(skb, dev_net(dev), 0, - RTNLGRP_PHONET_ROUTE, NULL, GFP_KERNEL); + + rtnl_notify(skb, net, 0, RTNLGRP_PHONET_ROUTE, NULL, GFP_KERNEL); return; errout: - rtnl_set_sk_err(dev_net(dev), RTNLGRP_PHONET_ROUTE, err); + rtnl_set_sk_err(net, RTNLGRP_PHONET_ROUTE, err); } static const struct nla_policy rtm_phonet_policy[RTA_MAX+1] = { @@ -235,6 +235,7 @@ static int route_doit(struct sk_buff *skb, struct nlmsghdr *nlh, struct nlattr *tb[RTA_MAX+1]; struct net_device *dev; struct rtmsg *rtm; + u32 ifindex; int err; u8 dst; @@ -260,7 +261,8 @@ static int route_doit(struct sk_buff *skb, struct nlmsghdr *nlh, if (dst & 3) /* Phonet addresses only have 6 high-order bits */ return -EINVAL; - dev = __dev_get_by_index(net, nla_get_u32(tb[RTA_OIF])); + ifindex = nla_get_u32(tb[RTA_OIF]); + dev = __dev_get_by_index(net, ifindex); if (dev == NULL) return -ENODEV; @@ -269,7 +271,7 @@ static int route_doit(struct sk_buff *skb, struct nlmsghdr *nlh, else err = phonet_route_del(dev, dst); if (!err) - rtm_phonet_notify(nlh->nlmsg_type, dev, dst); + rtm_phonet_notify(net, nlh->nlmsg_type, ifindex, dst); return err; } -- 2.51.0 From 3deec3b4afb4c767007eae1eeedbcf3da599395b Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Thu, 17 Oct 2024 11:31:39 -0700 Subject: [PATCH 03/16] phonet: Convert phonet_routes.lock to spinlock_t. route_doit() calls phonet_route_add() or phonet_route_del() for RTM_NEWROUTE or RTM_DELROUTE, respectively. Both functions only touch phonet_pernet(dev_net(dev))->routes, which is currently protected by RTNL and its dedicated mutex, phonet_routes.lock. We will convert route_doit() to RCU and cannot use mutex inside RCU. Let's convert the mutex to spinlock_t. Signed-off-by: Kuniyuki Iwashima Reviewed-by: Eric Dumazet Signed-off-by: Paolo Abeni --- include/net/phonet/pn_dev.h | 1 - net/phonet/pn_dev.c | 23 ++++++++++++++--------- 2 files changed, 14 insertions(+), 10 deletions(-) diff --git a/include/net/phonet/pn_dev.h b/include/net/phonet/pn_dev.h index 021e524fd20a..37a3e83531c6 100644 --- a/include/net/phonet/pn_dev.h +++ b/include/net/phonet/pn_dev.h @@ -11,7 +11,6 @@ #define PN_DEV_H #include -#include #include struct net; diff --git a/net/phonet/pn_dev.c b/net/phonet/pn_dev.c index 6ded0d347b9f..19234d664c4f 100644 --- a/net/phonet/pn_dev.c +++ b/net/phonet/pn_dev.c @@ -22,7 +22,7 @@ #include struct phonet_routes { - struct mutex lock; + spinlock_t lock; struct net_device __rcu *table[64]; }; @@ -273,13 +273,15 @@ static void phonet_route_autodel(struct net_device *dev) /* Remove left-over Phonet routes */ bitmap_zero(deleted, 64); - mutex_lock(&pnn->routes.lock); - for (i = 0; i < 64; i++) + + spin_lock(&pnn->routes.lock); + for (i = 0; i < 64; i++) { if (rcu_access_pointer(pnn->routes.table[i]) == dev) { RCU_INIT_POINTER(pnn->routes.table[i], NULL); set_bit(i, deleted); } - mutex_unlock(&pnn->routes.lock); + } + spin_unlock(&pnn->routes.lock); if (bitmap_empty(deleted, 64)) return; /* short-circuit RCU */ @@ -326,7 +328,7 @@ static int __net_init phonet_init_net(struct net *net) INIT_LIST_HEAD(&pnn->pndevs.list); spin_lock_init(&pnn->pndevs.lock); - mutex_init(&pnn->routes.lock); + spin_lock_init(&pnn->routes.lock); return 0; } @@ -376,13 +378,15 @@ int phonet_route_add(struct net_device *dev, u8 daddr) int err = -EEXIST; daddr = daddr >> 2; - mutex_lock(&routes->lock); + + spin_lock(&routes->lock); if (routes->table[daddr] == NULL) { rcu_assign_pointer(routes->table[daddr], dev); dev_hold(dev); err = 0; } - mutex_unlock(&routes->lock); + spin_unlock(&routes->lock); + return err; } @@ -392,12 +396,13 @@ int phonet_route_del(struct net_device *dev, u8 daddr) struct phonet_routes *routes = &pnn->routes; daddr = daddr >> 2; - mutex_lock(&routes->lock); + + spin_lock(&routes->lock); if (rcu_access_pointer(routes->table[daddr]) == dev) RCU_INIT_POINTER(routes->table[daddr], NULL); else dev = NULL; - mutex_unlock(&routes->lock); + spin_unlock(&routes->lock); if (!dev) return -ENOENT; -- 2.51.0 From 17a1ac0018ae1cee0b2c2235ce54e91ecbbed7be Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Thu, 17 Oct 2024 11:31:40 -0700 Subject: [PATCH 04/16] phonet: Don't hold RTNL for route_doit(). Now only __dev_get_by_index() depends on RTNL in route_doit(). Let's use dev_get_by_index_rcu() and register route_doit() with RTNL_FLAG_DOIT_UNLOCKED. Signed-off-by: Kuniyuki Iwashima Reviewed-by: Eric Dumazet Signed-off-by: Paolo Abeni --- net/phonet/pn_netlink.c | 19 +++++++++++++------ 1 file changed, 13 insertions(+), 6 deletions(-) diff --git a/net/phonet/pn_netlink.c b/net/phonet/pn_netlink.c index bfec5bd639b6..ca1f04e4a2d9 100644 --- a/net/phonet/pn_netlink.c +++ b/net/phonet/pn_netlink.c @@ -245,8 +245,6 @@ static int route_doit(struct sk_buff *skb, struct nlmsghdr *nlh, if (!netlink_capable(skb, CAP_SYS_ADMIN)) return -EPERM; - ASSERT_RTNL(); - err = nlmsg_parse_deprecated(nlh, sizeof(*rtm), tb, RTA_MAX, rtm_phonet_policy, extack); if (err < 0) @@ -262,16 +260,25 @@ static int route_doit(struct sk_buff *skb, struct nlmsghdr *nlh, return -EINVAL; ifindex = nla_get_u32(tb[RTA_OIF]); - dev = __dev_get_by_index(net, ifindex); - if (dev == NULL) + + rcu_read_lock(); + + dev = dev_get_by_index_rcu(net, ifindex); + if (!dev) { + rcu_read_unlock(); return -ENODEV; + } if (nlh->nlmsg_type == RTM_NEWROUTE) err = phonet_route_add(dev, dst); else err = phonet_route_del(dev, dst); + + rcu_read_unlock(); + if (!err) rtm_phonet_notify(net, nlh->nlmsg_type, ifindex, dst); + return err; } @@ -308,9 +315,9 @@ static const struct rtnl_msg_handler phonet_rtnl_msg_handlers[] __initdata_or_mo {.owner = THIS_MODULE, .protocol = PF_PHONET, .msgtype = RTM_GETADDR, .dumpit = getaddr_dumpit, .flags = RTNL_FLAG_DUMP_UNLOCKED}, {.owner = THIS_MODULE, .protocol = PF_PHONET, .msgtype = RTM_NEWROUTE, - .doit = route_doit}, + .doit = route_doit, .flags = RTNL_FLAG_DOIT_UNLOCKED}, {.owner = THIS_MODULE, .protocol = PF_PHONET, .msgtype = RTM_DELROUTE, - .doit = route_doit}, + .doit = route_doit, .flags = RTNL_FLAG_DOIT_UNLOCKED}, {.owner = THIS_MODULE, .protocol = PF_PHONET, .msgtype = RTM_GETROUTE, .dumpit = route_dumpit, .flags = RTNL_FLAG_DUMP_UNLOCKED}, }; -- 2.51.0 From 81bc949f640f78b507c7523de7c750bcc87c1bb8 Mon Sep 17 00:00:00 2001 From: Sabrina Dubroca Date: Fri, 18 Oct 2024 12:55:58 +0200 Subject: [PATCH 05/16] selftests: tls: add a selftest for wrapping rec_seq Set the initial rec_seq to 0xffffffffffffffff so that it wraps immediately. The send() call should fail with EBADMSG. A bug in this code was fixed in commit cfaa80c91f6f ("net/tls: do not free tls_rec on async operation in bpf_exec_tx_verdict()"). Signed-off-by: Sabrina Dubroca Reviewed-by: Simon Horman Link: https://patch.msgid.link/20775fcfd0371422921ee60a42de170c0398ac10.1729244987.git.sd@queasysnail.net Signed-off-by: Paolo Abeni --- tools/testing/selftests/net/tls.c | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) diff --git a/tools/testing/selftests/net/tls.c b/tools/testing/selftests/net/tls.c index f27a12d2a2c9..1a706d03bb6b 100644 --- a/tools/testing/selftests/net/tls.c +++ b/tools/testing/selftests/net/tls.c @@ -266,6 +266,25 @@ TEST_F(tls_basic, bad_cipher) EXPECT_EQ(setsockopt(self->fd, SOL_TLS, TLS_TX, &tls12, sizeof(struct tls12_crypto_info_aes_gcm_128)), -1); } +TEST_F(tls_basic, recseq_wrap) +{ + struct tls_crypto_info_keys tls12; + char const *test_str = "test_read"; + int send_len = 10; + + if (self->notls) + SKIP(return, "no TLS support"); + + tls_crypto_info_init(TLS_1_2_VERSION, TLS_CIPHER_AES_GCM_128, &tls12); + memset(&tls12.aes128.rec_seq, 0xff, sizeof(tls12.aes128.rec_seq)); + + ASSERT_EQ(setsockopt(self->fd, SOL_TLS, TLS_TX, &tls12, tls12.len), 0); + ASSERT_EQ(setsockopt(self->cfd, SOL_TLS, TLS_RX, &tls12, tls12.len), 0); + + EXPECT_EQ(send(self->fd, test_str, send_len, 0), -1); + EXPECT_EQ(errno, EBADMSG); +} + FIXTURE(tls) { int fd, cfd; -- 2.51.0 From f009e946c15540cdff2974771fb979f40b794153 Mon Sep 17 00:00:00 2001 From: Dominique Martinet Date: Thu, 24 Oct 2024 08:29:19 +0900 Subject: [PATCH 06/16] Revert "9p: Enable multipage folios" This reverts commit 1325e4a91a405f88f1b18626904d37860a4f9069. using multipage folios apparently break some madvise operations like MADV_PAGEOUT which do not reliably unload the specified page anymore, Revert the patch until that is figured out. Reported-by: Andrii Nakryiko Fixes: 1325e4a91a40 ("9p: Enable multipage folios") Signed-off-by: Dominique Martinet Acked-by: Andrii Nakryiko Signed-off-by: Linus Torvalds --- fs/9p/vfs_inode.c | 1 - 1 file changed, 1 deletion(-) diff --git a/fs/9p/vfs_inode.c b/fs/9p/vfs_inode.c index effb3aa1f3ed..fd72fc38c8f5 100644 --- a/fs/9p/vfs_inode.c +++ b/fs/9p/vfs_inode.c @@ -295,7 +295,6 @@ int v9fs_init_inode(struct v9fs_session_info *v9ses, inode->i_op = &v9fs_file_inode_operations; inode->i_fop = &v9fs_file_operations; } - mapping_set_large_folios(inode->i_mapping); break; case S_IFLNK: -- 2.51.0 From 788d5d655bc94ded8a2b550caa03ba4c255c55eb Mon Sep 17 00:00:00 2001 From: Guillaume Nault Date: Fri, 18 Oct 2024 15:35:28 +0200 Subject: [PATCH 07/16] bareudp: Use pcpu stats to update rx_dropped counter. Use the core_stats rx_dropped counter to avoid the cost of atomic increments. Signed-off-by: Guillaume Nault Reviewed-by: Simon Horman Signed-off-by: David S. Miller --- drivers/net/bareudp.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/drivers/net/bareudp.c b/drivers/net/bareudp.c index fa2dd76ba3d9..a2abfade82dd 100644 --- a/drivers/net/bareudp.c +++ b/drivers/net/bareudp.c @@ -84,7 +84,7 @@ static int bareudp_udp_encap_recv(struct sock *sk, struct sk_buff *skb) if (skb_copy_bits(skb, BAREUDP_BASE_HLEN, &ipversion, sizeof(ipversion))) { - DEV_STATS_INC(bareudp->dev, rx_dropped); + dev_core_stats_rx_dropped_inc(bareudp->dev); goto drop; } ipversion >>= 4; @@ -94,7 +94,7 @@ static int bareudp_udp_encap_recv(struct sock *sk, struct sk_buff *skb) } else if (ipversion == 6 && bareudp->multi_proto_mode) { proto = htons(ETH_P_IPV6); } else { - DEV_STATS_INC(bareudp->dev, rx_dropped); + dev_core_stats_rx_dropped_inc(bareudp->dev); goto drop; } } else if (bareudp->ethertype == htons(ETH_P_MPLS_UC)) { @@ -108,7 +108,7 @@ static int bareudp_udp_encap_recv(struct sock *sk, struct sk_buff *skb) ipv4_is_multicast(tunnel_hdr->daddr)) { proto = htons(ETH_P_MPLS_MC); } else { - DEV_STATS_INC(bareudp->dev, rx_dropped); + dev_core_stats_rx_dropped_inc(bareudp->dev); goto drop; } } else { @@ -124,7 +124,7 @@ static int bareudp_udp_encap_recv(struct sock *sk, struct sk_buff *skb) (addr_type & IPV6_ADDR_MULTICAST)) { proto = htons(ETH_P_MPLS_MC); } else { - DEV_STATS_INC(bareudp->dev, rx_dropped); + dev_core_stats_rx_dropped_inc(bareudp->dev); goto drop; } } @@ -136,7 +136,7 @@ static int bareudp_udp_encap_recv(struct sock *sk, struct sk_buff *skb) proto, !net_eq(bareudp->net, dev_net(bareudp->dev)))) { - DEV_STATS_INC(bareudp->dev, rx_dropped); + dev_core_stats_rx_dropped_inc(bareudp->dev); goto drop; } @@ -144,7 +144,7 @@ static int bareudp_udp_encap_recv(struct sock *sk, struct sk_buff *skb) tun_dst = udp_tun_rx_dst(skb, family, key, 0, 0); if (!tun_dst) { - DEV_STATS_INC(bareudp->dev, rx_dropped); + dev_core_stats_rx_dropped_inc(bareudp->dev); goto drop; } skb_dst_set(skb, &tun_dst->dst); -- 2.51.0 From b8ee7a11c75436b85fa1641aa5f970de0f8a575c Mon Sep 17 00:00:00 2001 From: Javier Carrasco Date: Sat, 19 Oct 2024 22:16:49 +0200 Subject: [PATCH 08/16] net: dsa: mv88e6xxx: fix unreleased fwnode_handle in setup_port() 'ports_fwnode' is initialized via device_get_named_child_node(), which requires a call to fwnode_handle_put() when the variable is no longer required to avoid leaking memory. Add the missing fwnode_handle_put() after 'ports_fwnode' has been used and is no longer required. Fixes: 94a2a84f5e9e ("net: dsa: mv88e6xxx: Support LED control") Signed-off-by: Javier Carrasco Reviewed-by: Andrew Lunn Reviewed-by: Linus Walleij Signed-off-by: David S. Miller --- drivers/net/dsa/mv88e6xxx/chip.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/dsa/mv88e6xxx/chip.c b/drivers/net/dsa/mv88e6xxx/chip.c index 4f5193d86e65..c75005b4d86e 100644 --- a/drivers/net/dsa/mv88e6xxx/chip.c +++ b/drivers/net/dsa/mv88e6xxx/chip.c @@ -3379,6 +3379,7 @@ static int mv88e6xxx_setup_port(struct mv88e6xxx_chip *chip, int port) break; } } + fwnode_handle_put(ports_fwnode); } else { dev_dbg(chip->dev, "no ethernet ports node defined for the device\n"); } -- 2.51.0 From a3e4bf7f9675b11d970bdbc9ccb24434d448b2c2 Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Tue, 15 Oct 2024 18:11:44 -0700 Subject: [PATCH 09/16] configs/debug: make sure PROVE_RCU_LIST=y takes effect Commit 0aaa8977acbf ("configs: introduce debug.config for CI-like setup") added CONFIG_PROVE_RCU_LIST=y to the common CI config, but RCU_EXPERT is not set, and it's a dependency for CONFIG_PROVE_RCU_LIST=y. Make sure CIs take advantage of CONFIG_PROVE_RCU_LIST=y, recent fixes in networking indicate that it does catch bugs. Reviewed-by: Joel Fernandes (Google) Acked-by: Matthieu Baerts (NGI0) Acked-by: Paul E. McKenney Reviewed-by: Simon Horman Link: https://patch.msgid.link/20241016011144.3058445-1-kuba@kernel.org Signed-off-by: Jakub Kicinski --- kernel/configs/debug.config | 1 + 1 file changed, 1 insertion(+) diff --git a/kernel/configs/debug.config b/kernel/configs/debug.config index 509ee703de15..20552f163930 100644 --- a/kernel/configs/debug.config +++ b/kernel/configs/debug.config @@ -103,6 +103,7 @@ CONFIG_BUG_ON_DATA_CORRUPTION=y # # RCU Debugging # +CONFIG_RCU_EXPERT=y CONFIG_PROVE_RCU=y CONFIG_PROVE_RCU_LIST=y # -- 2.51.0 From 05c9afb9bfa358ae8885b9d6eaa11603395e3c63 Mon Sep 17 00:00:00 2001 From: Danila Tikhonov Date: Sun, 20 Oct 2024 23:56:09 +0300 Subject: [PATCH 10/16] dt-bindings: nfc: nxp,nci: Document PN553 compatible The PN553 is another NFC chip from NXP, document the compatible in the bindings. Signed-off-by: Danila Tikhonov Acked-by: Krzysztof Kozlowski Acked-by: Rob Herring (Arm) Link: https://patch.msgid.link/20241020205615.211256-2-danila@jiaxyga.com Signed-off-by: Jakub Kicinski --- Documentation/devicetree/bindings/net/nfc/nxp,nci.yaml | 1 + 1 file changed, 1 insertion(+) diff --git a/Documentation/devicetree/bindings/net/nfc/nxp,nci.yaml b/Documentation/devicetree/bindings/net/nfc/nxp,nci.yaml index 6924aff0b2c5..364b36151180 100644 --- a/Documentation/devicetree/bindings/net/nfc/nxp,nci.yaml +++ b/Documentation/devicetree/bindings/net/nfc/nxp,nci.yaml @@ -17,6 +17,7 @@ properties: - enum: - nxp,nq310 - nxp,pn547 + - nxp,pn553 - const: nxp,nxp-nci-i2c enable-gpios: -- 2.51.0 From 9f6cb31979739e46b4f863cf13736f8298cbd21e Mon Sep 17 00:00:00 2001 From: Yazen Ghannam Date: Mon, 21 Oct 2024 15:38:25 +0000 Subject: [PATCH 11/16] net: amd8111e: Remove duplicate definition of PCI_VENDOR_ID_AMD The AMD PCI vendor ID is already defined in . Remove this local definition as it is not needed. Signed-off-by: Yazen Ghannam Reviewed-by: Kalesh AP Link: https://patch.msgid.link/20241021153825.2536819-1-yazen.ghannam@amd.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/amd/amd8111e.h | 1 - 1 file changed, 1 deletion(-) diff --git a/drivers/net/ethernet/amd/amd8111e.h b/drivers/net/ethernet/amd/amd8111e.h index 305232f5476d..e4ee4c28800c 100644 --- a/drivers/net/ethernet/amd/amd8111e.h +++ b/drivers/net/ethernet/amd/amd8111e.h @@ -550,7 +550,6 @@ typedef enum { /* Driver definitions */ -#define PCI_VENDOR_ID_AMD 0x1022 #define PCI_DEVICE_ID_AMD8111E_7462 0x7462 #define MAX_UNITS 8 /* Maximum number of devices possible */ -- 2.51.0 From 47e99f30730c0167cd32c9a2fd4a74f0a024cb2b Mon Sep 17 00:00:00 2001 From: Leo Stone Date: Mon, 21 Oct 2024 10:46:44 -0700 Subject: [PATCH 12/16] selftest/tcp-ao: Add filter tests Add tests that check if getsockopt(TCP_AO_GET_KEYS) returns the right keys when using different filters. Sample output: > # ok 114 filter keys: by sndid, rcvid, address > # ok 115 filter keys: by is_current > # ok 116 filter keys: by is_rnext > # ok 117 filter keys: by sndid, rcvid > # ok 118 filter keys: correct nkeys when in.nkeys < matches Acked-by: Dmitry Safonov <0x7f454c46@gmail.com> Signed-off-by: Leo Stone Reviewed-by: Simon Horman Link: https://patch.msgid.link/20241021174652.6949-1-leocstone@gmail.com Signed-off-by: Jakub Kicinski --- .../selftests/net/tcp_ao/setsockopt-closed.c | 186 +++++++++++++++++- 1 file changed, 181 insertions(+), 5 deletions(-) diff --git a/tools/testing/selftests/net/tcp_ao/setsockopt-closed.c b/tools/testing/selftests/net/tcp_ao/setsockopt-closed.c index 084db4ecdff6..0abb9807d742 100644 --- a/tools/testing/selftests/net/tcp_ao/setsockopt-closed.c +++ b/tools/testing/selftests/net/tcp_ao/setsockopt-closed.c @@ -6,6 +6,8 @@ static union tcp_addr tcp_md5_client; +#define FILTER_TEST_NKEYS 16 + static int test_port = 7788; static void make_listen(int sk) { @@ -813,23 +815,197 @@ static void duplicate_tests(void) setsockopt_checked(sk, TCP_AO_ADD_KEY, &ao, EEXIST, "duplicate: SendID differs"); } +static void fetch_all_keys(int sk, struct tcp_ao_getsockopt *keys) +{ + socklen_t optlen = sizeof(struct tcp_ao_getsockopt); + + memset(keys, 0, sizeof(struct tcp_ao_getsockopt) * FILTER_TEST_NKEYS); + keys[0].get_all = 1; + keys[0].nkeys = FILTER_TEST_NKEYS; + if (getsockopt(sk, IPPROTO_TCP, TCP_AO_GET_KEYS, &keys[0], &optlen)) + test_error("getsockopt"); +} + +static int prepare_test_keys(struct tcp_ao_getsockopt *keys) +{ + const char *test_password = "Test password number "; + struct tcp_ao_add test_ao[FILTER_TEST_NKEYS]; + char test_password_scratch[64] = {}; + u8 rcvid = 100, sndid = 100; + int sk; + + sk = socket(test_family, SOCK_STREAM, IPPROTO_TCP); + if (sk < 0) + test_error("socket()"); + + for (int i = 0; i < FILTER_TEST_NKEYS; i++) { + snprintf(test_password_scratch, 64, "%s %d", test_password, i); + test_prepare_key(&test_ao[i], DEFAULT_TEST_ALGO, this_ip_dest, + false, false, DEFAULT_TEST_PREFIX, 0, sndid++, + rcvid++, 0, 0, strlen(test_password_scratch), + test_password_scratch); + } + test_ao[0].set_current = 1; + test_ao[1].set_rnext = 1; + /* One key with a different addr and overlapping sndid, rcvid */ + tcp_addr_to_sockaddr_in(&test_ao[2].addr, &this_ip_addr, 0); + test_ao[2].sndid = 100; + test_ao[2].rcvid = 100; + + /* Add keys in a random order */ + for (int i = 0; i < FILTER_TEST_NKEYS; i++) { + int randidx = rand() % (FILTER_TEST_NKEYS - i); + + if (setsockopt(sk, IPPROTO_TCP, TCP_AO_ADD_KEY, + &test_ao[randidx], sizeof(struct tcp_ao_add))) + test_error("setsockopt()"); + memcpy(&test_ao[randidx], &test_ao[FILTER_TEST_NKEYS - 1 - i], + sizeof(struct tcp_ao_add)); + } + + fetch_all_keys(sk, keys); + + return sk; +} + +/* Assumes passwords are unique */ +static int compare_mkts(struct tcp_ao_getsockopt *expected, int nexpected, + struct tcp_ao_getsockopt *actual, int nactual) +{ + int matches = 0; + + for (int i = 0; i < nexpected; i++) { + for (int j = 0; j < nactual; j++) { + if (memcmp(expected[i].key, actual[j].key, + TCP_AO_MAXKEYLEN) == 0) + matches++; + } + } + return nexpected - matches; +} + +static void filter_keys_checked(int sk, struct tcp_ao_getsockopt *filter, + struct tcp_ao_getsockopt *expected, + unsigned int nexpected, const char *tst) +{ + struct tcp_ao_getsockopt filtered_keys[FILTER_TEST_NKEYS] = {}; + struct tcp_ao_getsockopt all_keys[FILTER_TEST_NKEYS] = {}; + socklen_t len = sizeof(struct tcp_ao_getsockopt); + + fetch_all_keys(sk, all_keys); + memcpy(&filtered_keys[0], filter, sizeof(struct tcp_ao_getsockopt)); + filtered_keys[0].nkeys = FILTER_TEST_NKEYS; + if (getsockopt(sk, IPPROTO_TCP, TCP_AO_GET_KEYS, filtered_keys, &len)) + test_error("getsockopt"); + if (filtered_keys[0].nkeys != nexpected) { + test_fail("wrong nr of keys, expected %u got %u", nexpected, + filtered_keys[0].nkeys); + goto out_close; + } + if (compare_mkts(expected, nexpected, filtered_keys, + filtered_keys[0].nkeys)) { + test_fail("got wrong keys back"); + goto out_close; + } + test_ok("filter keys: %s", tst); + +out_close: + close(sk); + memset(filter, 0, sizeof(struct tcp_ao_getsockopt)); +} + +static void filter_tests(void) +{ + struct tcp_ao_getsockopt original_keys[FILTER_TEST_NKEYS]; + struct tcp_ao_getsockopt expected_keys[FILTER_TEST_NKEYS]; + struct tcp_ao_getsockopt filter = {}; + int sk, f, nmatches; + socklen_t len; + + f = 2; + sk = prepare_test_keys(original_keys); + filter.rcvid = original_keys[f].rcvid; + filter.sndid = original_keys[f].sndid; + memcpy(&filter.addr, &original_keys[f].addr, + sizeof(original_keys[f].addr)); + filter.prefix = original_keys[f].prefix; + filter_keys_checked(sk, &filter, &original_keys[f], 1, + "by sndid, rcvid, address"); + + f = -1; + sk = prepare_test_keys(original_keys); + for (int i = 0; i < original_keys[0].nkeys; i++) { + if (original_keys[i].is_current) { + f = i; + break; + } + } + if (f < 0) + test_error("No current key after adding one"); + filter.is_current = 1; + filter_keys_checked(sk, &filter, &original_keys[f], 1, "by is_current"); + + f = -1; + sk = prepare_test_keys(original_keys); + for (int i = 0; i < original_keys[0].nkeys; i++) { + if (original_keys[i].is_rnext) { + f = i; + break; + } + } + if (f < 0) + test_error("No rnext key after adding one"); + filter.is_rnext = 1; + filter_keys_checked(sk, &filter, &original_keys[f], 1, "by is_rnext"); + + f = -1; + nmatches = 0; + sk = prepare_test_keys(original_keys); + for (int i = 0; i < original_keys[0].nkeys; i++) { + if (original_keys[i].sndid == 100) { + f = i; + memcpy(&expected_keys[nmatches], &original_keys[i], + sizeof(struct tcp_ao_getsockopt)); + nmatches++; + } + } + if (f < 0) + test_error("No key for sndid 100"); + if (nmatches != 2) + test_error("Should have 2 keys with sndid 100"); + filter.rcvid = original_keys[f].rcvid; + filter.sndid = original_keys[f].sndid; + filter.addr.ss_family = test_family; + filter_keys_checked(sk, &filter, expected_keys, nmatches, + "by sndid, rcvid"); + + sk = prepare_test_keys(original_keys); + filter.get_all = 1; + filter.nkeys = FILTER_TEST_NKEYS / 2; + len = sizeof(struct tcp_ao_getsockopt); + if (getsockopt(sk, IPPROTO_TCP, TCP_AO_GET_KEYS, &filter, &len)) + test_error("getsockopt"); + if (filter.nkeys == FILTER_TEST_NKEYS) + test_ok("filter keys: correct nkeys when in.nkeys < matches"); + else + test_fail("filter keys: wrong nkeys, expected %u got %u", + FILTER_TEST_NKEYS, filter.nkeys); +} + static void *client_fn(void *arg) { if (inet_pton(TEST_FAMILY, __TEST_CLIENT_IP(2), &tcp_md5_client) != 1) test_error("Can't convert ip address"); extend_tests(); einval_tests(); + filter_tests(); duplicate_tests(); - /* - * TODO: check getsockopt(TCP_AO_GET_KEYS) with different filters - * returning proper nr & keys; - */ return NULL; } int main(int argc, char *argv[]) { - test_init(121, client_fn, NULL); + test_init(126, client_fn, NULL); return 0; } -- 2.51.0 From 890bde75a2360d80712f394a31982fbd93fa0891 Mon Sep 17 00:00:00 2001 From: Florian Fainelli Date: Mon, 21 Oct 2024 10:49:34 -0700 Subject: [PATCH 13/16] net: systemport: Remove unused txchk accessors Vladimir reported the following warning with clang-16 and W=1: warning: unused function 'txchk_readl' [-Wunused-function] BCM_SYSPORT_IO_MACRO(txchk, SYS_PORT_TXCHK_OFFSET); note: expanded from macro 'BCM_SYSPORT_IO_MACRO' warning: unused function 'txchk_writel' [-Wunused-function] note: expanded from macro 'BCM_SYSPORT_IO_MACRO' warning: unused function 'tbuf_readl' [-Wunused-function] BCM_SYSPORT_IO_MACRO(tbuf, SYS_PORT_TBUF_OFFSET); note: expanded from macro 'BCM_SYSPORT_IO_MACRO' warning: unused function 'tbuf_writel' [-Wunused-function] note: expanded from macro 'BCM_SYSPORT_IO_MACRO' The TXCHK and RBUF blocks are not being accessed, remove the IO macros used to access those blocks. No functional impact. Reported-by: Vladimir Oltean Signed-off-by: Florian Fainelli Reviewed-by: Simon Horman Link: https://patch.msgid.link/20241021174935.57658-2-florian.fainelli@broadcom.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/broadcom/bcmsysport.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/drivers/net/ethernet/broadcom/bcmsysport.c b/drivers/net/ethernet/broadcom/bcmsysport.c index 0b7088ca4822..9e42b5db721e 100644 --- a/drivers/net/ethernet/broadcom/bcmsysport.c +++ b/drivers/net/ethernet/broadcom/bcmsysport.c @@ -46,9 +46,7 @@ BCM_SYSPORT_IO_MACRO(umac, SYS_PORT_UMAC_OFFSET); BCM_SYSPORT_IO_MACRO(gib, SYS_PORT_GIB_OFFSET); BCM_SYSPORT_IO_MACRO(tdma, SYS_PORT_TDMA_OFFSET); BCM_SYSPORT_IO_MACRO(rxchk, SYS_PORT_RXCHK_OFFSET); -BCM_SYSPORT_IO_MACRO(txchk, SYS_PORT_TXCHK_OFFSET); BCM_SYSPORT_IO_MACRO(rbuf, SYS_PORT_RBUF_OFFSET); -BCM_SYSPORT_IO_MACRO(tbuf, SYS_PORT_TBUF_OFFSET); BCM_SYSPORT_IO_MACRO(topctrl, SYS_PORT_TOPCTRL_OFFSET); /* On SYSTEMPORT Lite, any register after RDMA_STATUS has the exact -- 2.51.0 From e69fbd287d5a8b1038e50612545d7e3fc6db2b8a Mon Sep 17 00:00:00 2001 From: Florian Fainelli Date: Mon, 21 Oct 2024 10:49:35 -0700 Subject: [PATCH 14/16] net: systemport: Move IO macros to header file Move the BCM_SYSPORT_IO_MACRO() definition and its use to bcmsysport.h where it is more appropriate and where static inline helpers are acceptable. While at it, make sure that the macro 'offset' argument does not trigger a checkpatch warning due to possible argument re-use. Suggested-by: Vladimir Oltean Signed-off-by: Florian Fainelli Reviewed-by: Simon Horman Link: https://patch.msgid.link/20241021174935.57658-3-florian.fainelli@broadcom.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/broadcom/bcmsysport.c | 22 --------------------- drivers/net/ethernet/broadcom/bcmsysport.h | 23 ++++++++++++++++++++++ 2 files changed, 23 insertions(+), 22 deletions(-) diff --git a/drivers/net/ethernet/broadcom/bcmsysport.c b/drivers/net/ethernet/broadcom/bcmsysport.c index 9e42b5db721e..caff6e87a488 100644 --- a/drivers/net/ethernet/broadcom/bcmsysport.c +++ b/drivers/net/ethernet/broadcom/bcmsysport.c @@ -27,28 +27,6 @@ #include "bcmsysport.h" -/* I/O accessors register helpers */ -#define BCM_SYSPORT_IO_MACRO(name, offset) \ -static inline u32 name##_readl(struct bcm_sysport_priv *priv, u32 off) \ -{ \ - u32 reg = readl_relaxed(priv->base + offset + off); \ - return reg; \ -} \ -static inline void name##_writel(struct bcm_sysport_priv *priv, \ - u32 val, u32 off) \ -{ \ - writel_relaxed(val, priv->base + offset + off); \ -} \ - -BCM_SYSPORT_IO_MACRO(intrl2_0, SYS_PORT_INTRL2_0_OFFSET); -BCM_SYSPORT_IO_MACRO(intrl2_1, SYS_PORT_INTRL2_1_OFFSET); -BCM_SYSPORT_IO_MACRO(umac, SYS_PORT_UMAC_OFFSET); -BCM_SYSPORT_IO_MACRO(gib, SYS_PORT_GIB_OFFSET); -BCM_SYSPORT_IO_MACRO(tdma, SYS_PORT_TDMA_OFFSET); -BCM_SYSPORT_IO_MACRO(rxchk, SYS_PORT_RXCHK_OFFSET); -BCM_SYSPORT_IO_MACRO(rbuf, SYS_PORT_RBUF_OFFSET); -BCM_SYSPORT_IO_MACRO(topctrl, SYS_PORT_TOPCTRL_OFFSET); - /* On SYSTEMPORT Lite, any register after RDMA_STATUS has the exact * same layout, except it has been moved by 4 bytes up, *sigh* */ diff --git a/drivers/net/ethernet/broadcom/bcmsysport.h b/drivers/net/ethernet/broadcom/bcmsysport.h index 335cf6631db5..a34296f989f1 100644 --- a/drivers/net/ethernet/broadcom/bcmsysport.h +++ b/drivers/net/ethernet/broadcom/bcmsysport.h @@ -773,4 +773,27 @@ struct bcm_sysport_priv { struct bcm_sysport_tx_ring *ring_map[DSA_MAX_PORTS * 8]; }; + +/* I/O accessors register helpers */ +#define BCM_SYSPORT_IO_MACRO(name, offset) \ +static inline u32 name##_readl(struct bcm_sysport_priv *priv, u32 off) \ +{ \ + u32 reg = readl_relaxed(priv->base + (offset) + off); \ + return reg; \ +} \ +static inline void name##_writel(struct bcm_sysport_priv *priv, \ + u32 val, u32 off) \ +{ \ + writel_relaxed(val, priv->base + (offset) + off); \ +} \ + +BCM_SYSPORT_IO_MACRO(intrl2_0, SYS_PORT_INTRL2_0_OFFSET); +BCM_SYSPORT_IO_MACRO(intrl2_1, SYS_PORT_INTRL2_1_OFFSET); +BCM_SYSPORT_IO_MACRO(umac, SYS_PORT_UMAC_OFFSET); +BCM_SYSPORT_IO_MACRO(gib, SYS_PORT_GIB_OFFSET); +BCM_SYSPORT_IO_MACRO(tdma, SYS_PORT_TDMA_OFFSET); +BCM_SYSPORT_IO_MACRO(rxchk, SYS_PORT_RXCHK_OFFSET); +BCM_SYSPORT_IO_MACRO(rbuf, SYS_PORT_RBUF_OFFSET); +BCM_SYSPORT_IO_MACRO(topctrl, SYS_PORT_TOPCTRL_OFFSET); + #endif /* __BCM_SYSPORT_H */ -- 2.51.0 From a42f3076648e0b507de9039f8085edcc10b35fb7 Mon Sep 17 00:00:00 2001 From: "Matthieu Baerts (NGI0)" Date: Mon, 21 Oct 2024 17:14:03 +0200 Subject: [PATCH 15/16] mptcp: pm: send ACK on non-stale subflows If the subflow is considered as "staled", it is better to avoid it to send an ACK carrying an ADD_ADDR or RM_ADDR. Another subflow, if any, will then be selected. Reviewed-by: Mat Martineau Signed-off-by: Matthieu Baerts (NGI0) Link: https://patch.msgid.link/20241021-net-next-mptcp-misc-6-13-v1-1-1ef02746504a@kernel.org Signed-off-by: Jakub Kicinski --- net/mptcp/pm_netlink.c | 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 deletions(-) diff --git a/net/mptcp/pm_netlink.c b/net/mptcp/pm_netlink.c index db586a5b3866..618289aac0ab 100644 --- a/net/mptcp/pm_netlink.c +++ b/net/mptcp/pm_netlink.c @@ -781,7 +781,7 @@ bool mptcp_pm_nl_is_init_remote_addr(struct mptcp_sock *msk, void mptcp_pm_nl_addr_send_ack(struct mptcp_sock *msk) { - struct mptcp_subflow_context *subflow; + struct mptcp_subflow_context *subflow, *alt = NULL; msk_owned_by_me(msk); lockdep_assert_held(&msk->pm.lock); @@ -792,10 +792,18 @@ void mptcp_pm_nl_addr_send_ack(struct mptcp_sock *msk) mptcp_for_each_subflow(msk, subflow) { if (__mptcp_subflow_active(subflow)) { - mptcp_pm_send_ack(msk, subflow, false, false); - break; + if (!subflow->stale) { + mptcp_pm_send_ack(msk, subflow, false, false); + return; + } + + if (!alt) + alt = subflow; } } + + if (alt) + mptcp_pm_send_ack(msk, alt, false, false); } int mptcp_pm_nl_mp_prio_send_ack(struct mptcp_sock *msk, -- 2.51.0 From 581c8cbfa934aaa555daa4e843242fcecc160f05 Mon Sep 17 00:00:00 2001 From: Gang Yan Date: Mon, 21 Oct 2024 17:14:04 +0200 Subject: [PATCH 16/16] mptcp: annotate data-races around subflow->fully_established We introduce the same handling for potential data races with the 'fully_established' flag in subflow as previously done for msk->fully_established. Additionally, we make a crucial change: convert the subflow's 'fully_established' from 'bit_field' to 'bool' type. This is necessary because methods for avoiding data races don't work well with 'bit_field'. Specifically, the 'READ_ONCE' needs to know the size of the variable being accessed, which is not supported in 'bit_field'. Also, 'test_bit' expect the address of 'bit_field'. Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/516 Signed-off-by: Gang Yan Reviewed-by: Matthieu Baerts (NGI0) Signed-off-by: Matthieu Baerts (NGI0) Link: https://patch.msgid.link/20241021-net-next-mptcp-misc-6-13-v1-2-1ef02746504a@kernel.org Signed-off-by: Jakub Kicinski --- net/mptcp/diag.c | 2 +- net/mptcp/options.c | 4 ++-- net/mptcp/protocol.c | 2 +- net/mptcp/protocol.h | 6 +++--- net/mptcp/subflow.c | 4 ++-- 5 files changed, 9 insertions(+), 9 deletions(-) diff --git a/net/mptcp/diag.c b/net/mptcp/diag.c index 2d3efb405437..02205f7994d7 100644 --- a/net/mptcp/diag.c +++ b/net/mptcp/diag.c @@ -47,7 +47,7 @@ static int subflow_get_info(struct sock *sk, struct sk_buff *skb) flags |= MPTCP_SUBFLOW_FLAG_BKUP_REM; if (sf->request_bkup) flags |= MPTCP_SUBFLOW_FLAG_BKUP_LOC; - if (sf->fully_established) + if (READ_ONCE(sf->fully_established)) flags |= MPTCP_SUBFLOW_FLAG_FULLY_ESTABLISHED; if (sf->conn_finished) flags |= MPTCP_SUBFLOW_FLAG_CONNECTED; diff --git a/net/mptcp/options.c b/net/mptcp/options.c index 370c3836b771..1603b3702e22 100644 --- a/net/mptcp/options.c +++ b/net/mptcp/options.c @@ -461,7 +461,7 @@ static bool mptcp_established_options_mp(struct sock *sk, struct sk_buff *skb, return false; /* MPC/MPJ needed only on 3rd ack packet, DATA_FIN and TCP shutdown take precedence */ - if (subflow->fully_established || snd_data_fin_enable || + if (READ_ONCE(subflow->fully_established) || snd_data_fin_enable || subflow->snd_isn != TCP_SKB_CB(skb)->seq || sk->sk_state != TCP_ESTABLISHED) return false; @@ -930,7 +930,7 @@ static bool check_fully_established(struct mptcp_sock *msk, struct sock *ssk, /* here we can process OoO, in-window pkts, only in-sequence 4th ack * will make the subflow fully established */ - if (likely(subflow->fully_established)) { + if (likely(READ_ONCE(subflow->fully_established))) { /* on passive sockets, check for 3rd ack retransmission * note that msk is always set by subflow_syn_recv_sock() * for mp_join subflows diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c index 1f5c63eb21f0..a6c9661a4c45 100644 --- a/net/mptcp/protocol.c +++ b/net/mptcp/protocol.c @@ -3511,7 +3511,7 @@ static void schedule_3rdack_retransmission(struct sock *ssk) struct tcp_sock *tp = tcp_sk(ssk); unsigned long timeout; - if (mptcp_subflow_ctx(ssk)->fully_established) + if (READ_ONCE(mptcp_subflow_ctx(ssk)->fully_established)) return; /* reschedule with a timeout above RTT, as we must look only for drop */ diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h index 568a72702b08..a93e661ef5c4 100644 --- a/net/mptcp/protocol.h +++ b/net/mptcp/protocol.h @@ -513,7 +513,6 @@ struct mptcp_subflow_context { request_bkup : 1, mp_capable : 1, /* remote is MPTCP capable */ mp_join : 1, /* remote is JOINing */ - fully_established : 1, /* path validated */ pm_notified : 1, /* PM hook called for established status */ conn_finished : 1, map_valid : 1, @@ -532,10 +531,11 @@ struct mptcp_subflow_context { is_mptfo : 1, /* subflow is doing TFO */ close_event_done : 1, /* has done the post-closed part */ mpc_drop : 1, /* the MPC option has been dropped in a rtx */ - __unused : 8; + __unused : 9; bool data_avail; bool scheduled; bool pm_listener; /* a listener managed by the kernel PM? */ + bool fully_established; /* path validated */ u32 remote_nonce; u64 thmac; u32 local_nonce; @@ -780,7 +780,7 @@ static inline bool __tcp_can_send(const struct sock *ssk) static inline bool __mptcp_subflow_active(struct mptcp_subflow_context *subflow) { /* can't send if JOIN hasn't completed yet (i.e. is usable for mptcp) */ - if (subflow->request_join && !subflow->fully_established) + if (subflow->request_join && !READ_ONCE(subflow->fully_established)) return false; return __tcp_can_send(mptcp_subflow_tcp_sock(subflow)); diff --git a/net/mptcp/subflow.c b/net/mptcp/subflow.c index 6170f2fff71e..860903e06422 100644 --- a/net/mptcp/subflow.c +++ b/net/mptcp/subflow.c @@ -800,7 +800,7 @@ void __mptcp_subflow_fully_established(struct mptcp_sock *msk, const struct mptcp_options_received *mp_opt) { subflow_set_remote_key(msk, subflow, mp_opt); - subflow->fully_established = 1; + WRITE_ONCE(subflow->fully_established, true); WRITE_ONCE(msk->fully_established, true); if (subflow->is_mptfo) @@ -2062,7 +2062,7 @@ static void subflow_ulp_clone(const struct request_sock *req, } else if (subflow_req->mp_join) { new_ctx->ssn_offset = subflow_req->ssn_offset; new_ctx->mp_join = 1; - new_ctx->fully_established = 1; + WRITE_ONCE(new_ctx->fully_established, true); new_ctx->remote_key_valid = 1; new_ctx->backup = subflow_req->backup; new_ctx->request_bkup = subflow_req->request_bkup; -- 2.51.0