From 3c10ffddc61f8a1a59e29a110ba70b47e679206a Mon Sep 17 00:00:00 2001 From: Pavel Skripkin Date: Thu, 2 Sep 2021 22:04:00 +0300 Subject: net: xfrm: fix shift-out-of-bounds in xfrm_get_default Syzbot hit shift-out-of-bounds in xfrm_get_default. The problem was in missing validation check for user data. up->dirmask comes from user-space, so we need to check if this value is less than XFRM_USERPOLICY_DIRMASK_MAX to avoid shift-out-of-bounds bugs. Fixes: 2d151d39073a ("xfrm: Add possibility to set the default to block if we have no policy") Reported-and-tested-by: syzbot+b2be9dd8ca6f6c73ee2d@syzkaller.appspotmail.com Signed-off-by: Pavel Skripkin Signed-off-by: Steffen Klassert --- net/xfrm/xfrm_user.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'net') diff --git a/net/xfrm/xfrm_user.c b/net/xfrm/xfrm_user.c index 03b66d154b2b..4719a6d54aa6 100644 --- a/net/xfrm/xfrm_user.c +++ b/net/xfrm/xfrm_user.c @@ -2005,6 +2005,11 @@ static int xfrm_get_default(struct sk_buff *skb, struct nlmsghdr *nlh, return -EMSGSIZE; } + if (up->dirmask >= XFRM_USERPOLICY_DIRMASK_MAX) { + kfree_skb(r_skb); + return -EINVAL; + } + r_up = nlmsg_data(r_nlh); r_up->action = ((net->xfrm.policy_default & (1 << up->dirmask)) >> up->dirmask); -- cgit v1.2.3 From f8d858e607b2a36808ac6d4218f5f5203d7a7d63 Mon Sep 17 00:00:00 2001 From: Nicolas Dichtel Date: Tue, 14 Sep 2021 16:46:33 +0200 Subject: xfrm: make user policy API complete >From a userland POV, this API was based on some magic values: - dirmask and action were bitfields but meaning of bits (XFRM_POL_DEFAULT_*) are not exported; - action is confusing, if a bit is set, does it mean drop or accept? Let's try to simplify this uapi by using explicit field and macros. Fixes: 2d151d39073a ("xfrm: Add possibility to set the default to block if we have no policy") Signed-off-by: Nicolas Dichtel Signed-off-by: Steffen Klassert --- include/uapi/linux/xfrm.h | 9 ++++++--- net/xfrm/xfrm_user.c | 36 +++++++++++++++++++----------------- 2 files changed, 25 insertions(+), 20 deletions(-) (limited to 'net') diff --git a/include/uapi/linux/xfrm.h b/include/uapi/linux/xfrm.h index 26f456b1f33e..eda0426ec4c2 100644 --- a/include/uapi/linux/xfrm.h +++ b/include/uapi/linux/xfrm.h @@ -514,9 +514,12 @@ struct xfrm_user_offload { #define XFRM_OFFLOAD_INBOUND 2 struct xfrm_userpolicy_default { -#define XFRM_USERPOLICY_DIRMASK_MAX (sizeof(__u8) * 8) - __u8 dirmask; - __u8 action; +#define XFRM_USERPOLICY_UNSPEC 0 +#define XFRM_USERPOLICY_BLOCK 1 +#define XFRM_USERPOLICY_ACCEPT 2 + __u8 in; + __u8 fwd; + __u8 out; }; #ifndef __KERNEL__ diff --git a/net/xfrm/xfrm_user.c b/net/xfrm/xfrm_user.c index 4719a6d54aa6..90c88390f1fe 100644 --- a/net/xfrm/xfrm_user.c +++ b/net/xfrm/xfrm_user.c @@ -1966,16 +1966,21 @@ static int xfrm_set_default(struct sk_buff *skb, struct nlmsghdr *nlh, { struct net *net = sock_net(skb->sk); struct xfrm_userpolicy_default *up = nlmsg_data(nlh); - u8 dirmask; - u8 old_default = net->xfrm.policy_default; - if (up->dirmask >= XFRM_USERPOLICY_DIRMASK_MAX) - return -EINVAL; + if (up->in == XFRM_USERPOLICY_BLOCK) + net->xfrm.policy_default |= XFRM_POL_DEFAULT_IN; + else if (up->in == XFRM_USERPOLICY_ACCEPT) + net->xfrm.policy_default &= ~XFRM_POL_DEFAULT_IN; - dirmask = (1 << up->dirmask) & XFRM_POL_DEFAULT_MASK; + if (up->fwd == XFRM_USERPOLICY_BLOCK) + net->xfrm.policy_default |= XFRM_POL_DEFAULT_FWD; + else if (up->fwd == XFRM_USERPOLICY_ACCEPT) + net->xfrm.policy_default &= ~XFRM_POL_DEFAULT_FWD; - net->xfrm.policy_default = (old_default & (0xff ^ dirmask)) - | (up->action << up->dirmask); + if (up->out == XFRM_USERPOLICY_BLOCK) + net->xfrm.policy_default |= XFRM_POL_DEFAULT_OUT; + else if (up->out == XFRM_USERPOLICY_ACCEPT) + net->xfrm.policy_default &= ~XFRM_POL_DEFAULT_OUT; rt_genid_bump_all(net); @@ -1988,13 +1993,11 @@ static int xfrm_get_default(struct sk_buff *skb, struct nlmsghdr *nlh, struct sk_buff *r_skb; struct nlmsghdr *r_nlh; struct net *net = sock_net(skb->sk); - struct xfrm_userpolicy_default *r_up, *up; + struct xfrm_userpolicy_default *r_up; int len = NLMSG_ALIGN(sizeof(struct xfrm_userpolicy_default)); u32 portid = NETLINK_CB(skb).portid; u32 seq = nlh->nlmsg_seq; - up = nlmsg_data(nlh); - r_skb = nlmsg_new(len, GFP_ATOMIC); if (!r_skb) return -ENOMEM; @@ -2005,15 +2008,14 @@ static int xfrm_get_default(struct sk_buff *skb, struct nlmsghdr *nlh, return -EMSGSIZE; } - if (up->dirmask >= XFRM_USERPOLICY_DIRMASK_MAX) { - kfree_skb(r_skb); - return -EINVAL; - } - r_up = nlmsg_data(r_nlh); - r_up->action = ((net->xfrm.policy_default & (1 << up->dirmask)) >> up->dirmask); - r_up->dirmask = up->dirmask; + r_up->in = net->xfrm.policy_default & XFRM_POL_DEFAULT_IN ? + XFRM_USERPOLICY_BLOCK : XFRM_USERPOLICY_ACCEPT; + r_up->fwd = net->xfrm.policy_default & XFRM_POL_DEFAULT_FWD ? + XFRM_USERPOLICY_BLOCK : XFRM_USERPOLICY_ACCEPT; + r_up->out = net->xfrm.policy_default & XFRM_POL_DEFAULT_OUT ? + XFRM_USERPOLICY_BLOCK : XFRM_USERPOLICY_ACCEPT; nlmsg_end(r_skb, r_nlh); return nlmsg_unicast(net->xfrm.nlsk, r_skb, portid); -- cgit v1.2.3 From 88d0adb5f13b1c52fbb7d755f6f79db18c2f0c2c Mon Sep 17 00:00:00 2001 From: Nicolas Dichtel Date: Tue, 14 Sep 2021 16:46:34 +0200 Subject: xfrm: notify default policy on update This configuration knob is very sensible, it should be notified when changing. Fixes: 2d151d39073a ("xfrm: Add possibility to set the default to block if we have no policy") Signed-off-by: Nicolas Dichtel Signed-off-by: Steffen Klassert --- net/xfrm/xfrm_user.c | 31 +++++++++++++++++++++++++++++++ 1 file changed, 31 insertions(+) (limited to 'net') diff --git a/net/xfrm/xfrm_user.c b/net/xfrm/xfrm_user.c index 90c88390f1fe..0eba0c27c665 100644 --- a/net/xfrm/xfrm_user.c +++ b/net/xfrm/xfrm_user.c @@ -1961,6 +1961,36 @@ static struct sk_buff *xfrm_policy_netlink(struct sk_buff *in_skb, return skb; } +static int xfrm_notify_userpolicy(struct net *net) +{ + struct xfrm_userpolicy_default *up; + int len = NLMSG_ALIGN(sizeof(*up)); + struct nlmsghdr *nlh; + struct sk_buff *skb; + + skb = nlmsg_new(len, GFP_ATOMIC); + if (skb == NULL) + return -ENOMEM; + + nlh = nlmsg_put(skb, 0, 0, XFRM_MSG_GETDEFAULT, sizeof(*up), 0); + if (nlh == NULL) { + kfree_skb(skb); + return -EMSGSIZE; + } + + up = nlmsg_data(nlh); + up->in = net->xfrm.policy_default & XFRM_POL_DEFAULT_IN ? + XFRM_USERPOLICY_BLOCK : XFRM_USERPOLICY_ACCEPT; + up->fwd = net->xfrm.policy_default & XFRM_POL_DEFAULT_FWD ? + XFRM_USERPOLICY_BLOCK : XFRM_USERPOLICY_ACCEPT; + up->out = net->xfrm.policy_default & XFRM_POL_DEFAULT_OUT ? + XFRM_USERPOLICY_BLOCK : XFRM_USERPOLICY_ACCEPT; + + nlmsg_end(skb, nlh); + + return xfrm_nlmsg_multicast(net, skb, 0, XFRMNLGRP_POLICY); +} + static int xfrm_set_default(struct sk_buff *skb, struct nlmsghdr *nlh, struct nlattr **attrs) { @@ -1984,6 +2014,7 @@ static int xfrm_set_default(struct sk_buff *skb, struct nlmsghdr *nlh, rt_genid_bump_all(net); + xfrm_notify_userpolicy(net); return 0; } -- cgit v1.2.3 From 93ec1320b0170d7a207eda2d119c669b673401ed Mon Sep 17 00:00:00 2001 From: Nicolas Dichtel Date: Wed, 22 Sep 2021 10:50:06 +0200 Subject: xfrm: fix rcu lock in xfrm_notify_userpolicy() As stated in the comment above xfrm_nlmsg_multicast(), rcu read lock must be held before calling this function. Reported-by: syzbot+3d9866419b4aa8f985d6@syzkaller.appspotmail.com Fixes: 703b94b93c19 ("xfrm: notify default policy on update") Signed-off-by: Nicolas Dichtel Signed-off-by: Steffen Klassert --- net/xfrm/xfrm_user.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) (limited to 'net') diff --git a/net/xfrm/xfrm_user.c b/net/xfrm/xfrm_user.c index 0eba0c27c665..3a3cb09eec12 100644 --- a/net/xfrm/xfrm_user.c +++ b/net/xfrm/xfrm_user.c @@ -1967,6 +1967,7 @@ static int xfrm_notify_userpolicy(struct net *net) int len = NLMSG_ALIGN(sizeof(*up)); struct nlmsghdr *nlh; struct sk_buff *skb; + int err; skb = nlmsg_new(len, GFP_ATOMIC); if (skb == NULL) @@ -1988,7 +1989,11 @@ static int xfrm_notify_userpolicy(struct net *net) nlmsg_end(skb, nlh); - return xfrm_nlmsg_multicast(net, skb, 0, XFRMNLGRP_POLICY); + rcu_read_lock(); + err = xfrm_nlmsg_multicast(net, skb, 0, XFRMNLGRP_POLICY); + rcu_read_unlock(); + + return err; } static int xfrm_set_default(struct sk_buff *skb, struct nlmsghdr *nlh, -- cgit v1.2.3 From 339031bafe6b281cf2dcb8364217288b9fdab555 Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Thu, 23 Sep 2021 16:44:34 +0200 Subject: netfilter: conntrack: fix boot failure with nf_conntrack.enable_hooks=1 This is a revert of 7b1957b049 ("netfilter: nf_defrag_ipv4: use net_generic infra") and a partial revert of 8b0adbe3e3 ("netfilter: nf_defrag_ipv6: use net_generic infra"). If conntrack is builtin and kernel is booted with: nf_conntrack.enable_hooks=1 .... kernel will fail to boot due to a NULL deref in nf_defrag_ipv4_enable(): Its called before the ipv4 defrag initcall is made, so net_generic() returns NULL. To resolve this, move the user refcount back to struct net so calls to those functions are possible even before their initcalls have run. Fixes: 7b1957b04956 ("netfilter: nf_defrag_ipv4: use net_generic infra") Fixes: 8b0adbe3e38d ("netfilter: nf_defrag_ipv6: use net_generic infra"). Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso --- include/net/netfilter/ipv6/nf_defrag_ipv6.h | 1 - include/net/netns/netfilter.h | 6 ++++++ net/ipv4/netfilter/nf_defrag_ipv4.c | 30 +++++++++-------------------- net/ipv6/netfilter/nf_conntrack_reasm.c | 2 +- net/ipv6/netfilter/nf_defrag_ipv6_hooks.c | 25 +++++++++--------------- 5 files changed, 25 insertions(+), 39 deletions(-) (limited to 'net') diff --git a/include/net/netfilter/ipv6/nf_defrag_ipv6.h b/include/net/netfilter/ipv6/nf_defrag_ipv6.h index 0fd8a4159662..ceadf8ba25a4 100644 --- a/include/net/netfilter/ipv6/nf_defrag_ipv6.h +++ b/include/net/netfilter/ipv6/nf_defrag_ipv6.h @@ -17,7 +17,6 @@ struct inet_frags_ctl; struct nft_ct_frag6_pernet { struct ctl_table_header *nf_frag_frags_hdr; struct fqdir *fqdir; - unsigned int users; }; #endif /* _NF_DEFRAG_IPV6_H */ diff --git a/include/net/netns/netfilter.h b/include/net/netns/netfilter.h index 986a2a9cfdfa..b593f95e9991 100644 --- a/include/net/netns/netfilter.h +++ b/include/net/netns/netfilter.h @@ -27,5 +27,11 @@ struct netns_nf { #if IS_ENABLED(CONFIG_DECNET) struct nf_hook_entries __rcu *hooks_decnet[NF_DN_NUMHOOKS]; #endif +#if IS_ENABLED(CONFIG_NF_DEFRAG_IPV4) + unsigned int defrag_ipv4_users; +#endif +#if IS_ENABLED(CONFIG_NF_DEFRAG_IPV6) + unsigned int defrag_ipv6_users; +#endif }; #endif diff --git a/net/ipv4/netfilter/nf_defrag_ipv4.c b/net/ipv4/netfilter/nf_defrag_ipv4.c index 613432a36f0a..e61ea428ea18 100644 --- a/net/ipv4/netfilter/nf_defrag_ipv4.c +++ b/net/ipv4/netfilter/nf_defrag_ipv4.c @@ -20,13 +20,8 @@ #endif #include -static unsigned int defrag4_pernet_id __read_mostly; static DEFINE_MUTEX(defrag4_mutex); -struct defrag4_pernet { - unsigned int users; -}; - static int nf_ct_ipv4_gather_frags(struct net *net, struct sk_buff *skb, u_int32_t user) { @@ -111,19 +106,15 @@ static const struct nf_hook_ops ipv4_defrag_ops[] = { static void __net_exit defrag4_net_exit(struct net *net) { - struct defrag4_pernet *nf_defrag = net_generic(net, defrag4_pernet_id); - - if (nf_defrag->users) { + if (net->nf.defrag_ipv4_users) { nf_unregister_net_hooks(net, ipv4_defrag_ops, ARRAY_SIZE(ipv4_defrag_ops)); - nf_defrag->users = 0; + net->nf.defrag_ipv4_users = 0; } } static struct pernet_operations defrag4_net_ops = { .exit = defrag4_net_exit, - .id = &defrag4_pernet_id, - .size = sizeof(struct defrag4_pernet), }; static int __init nf_defrag_init(void) @@ -138,24 +129,23 @@ static void __exit nf_defrag_fini(void) int nf_defrag_ipv4_enable(struct net *net) { - struct defrag4_pernet *nf_defrag = net_generic(net, defrag4_pernet_id); int err = 0; mutex_lock(&defrag4_mutex); - if (nf_defrag->users == UINT_MAX) { + if (net->nf.defrag_ipv4_users == UINT_MAX) { err = -EOVERFLOW; goto out_unlock; } - if (nf_defrag->users) { - nf_defrag->users++; + if (net->nf.defrag_ipv4_users) { + net->nf.defrag_ipv4_users++; goto out_unlock; } err = nf_register_net_hooks(net, ipv4_defrag_ops, ARRAY_SIZE(ipv4_defrag_ops)); if (err == 0) - nf_defrag->users = 1; + net->nf.defrag_ipv4_users = 1; out_unlock: mutex_unlock(&defrag4_mutex); @@ -165,12 +155,10 @@ EXPORT_SYMBOL_GPL(nf_defrag_ipv4_enable); void nf_defrag_ipv4_disable(struct net *net) { - struct defrag4_pernet *nf_defrag = net_generic(net, defrag4_pernet_id); - mutex_lock(&defrag4_mutex); - if (nf_defrag->users) { - nf_defrag->users--; - if (nf_defrag->users == 0) + if (net->nf.defrag_ipv4_users) { + net->nf.defrag_ipv4_users--; + if (net->nf.defrag_ipv4_users == 0) nf_unregister_net_hooks(net, ipv4_defrag_ops, ARRAY_SIZE(ipv4_defrag_ops)); } diff --git a/net/ipv6/netfilter/nf_conntrack_reasm.c b/net/ipv6/netfilter/nf_conntrack_reasm.c index a0108415275f..5c47be29b9ee 100644 --- a/net/ipv6/netfilter/nf_conntrack_reasm.c +++ b/net/ipv6/netfilter/nf_conntrack_reasm.c @@ -33,7 +33,7 @@ static const char nf_frags_cache_name[] = "nf-frags"; -unsigned int nf_frag_pernet_id __read_mostly; +static unsigned int nf_frag_pernet_id __read_mostly; static struct inet_frags nf_frags; static struct nft_ct_frag6_pernet *nf_frag_pernet(struct net *net) diff --git a/net/ipv6/netfilter/nf_defrag_ipv6_hooks.c b/net/ipv6/netfilter/nf_defrag_ipv6_hooks.c index e8a59d8bf2ad..cb4eb1d2c620 100644 --- a/net/ipv6/netfilter/nf_defrag_ipv6_hooks.c +++ b/net/ipv6/netfilter/nf_defrag_ipv6_hooks.c @@ -25,8 +25,6 @@ #include #include -extern unsigned int nf_frag_pernet_id; - static DEFINE_MUTEX(defrag6_mutex); static enum ip6_defrag_users nf_ct6_defrag_user(unsigned int hooknum, @@ -91,12 +89,10 @@ static const struct nf_hook_ops ipv6_defrag_ops[] = { static void __net_exit defrag6_net_exit(struct net *net) { - struct nft_ct_frag6_pernet *nf_frag = net_generic(net, nf_frag_pernet_id); - - if (nf_frag->users) { + if (net->nf.defrag_ipv6_users) { nf_unregister_net_hooks(net, ipv6_defrag_ops, ARRAY_SIZE(ipv6_defrag_ops)); - nf_frag->users = 0; + net->nf.defrag_ipv6_users = 0; } } @@ -134,24 +130,23 @@ static void __exit nf_defrag_fini(void) int nf_defrag_ipv6_enable(struct net *net) { - struct nft_ct_frag6_pernet *nf_frag = net_generic(net, nf_frag_pernet_id); int err = 0; mutex_lock(&defrag6_mutex); - if (nf_frag->users == UINT_MAX) { + if (net->nf.defrag_ipv6_users == UINT_MAX) { err = -EOVERFLOW; goto out_unlock; } - if (nf_frag->users) { - nf_frag->users++; + if (net->nf.defrag_ipv6_users) { + net->nf.defrag_ipv6_users++; goto out_unlock; } err = nf_register_net_hooks(net, ipv6_defrag_ops, ARRAY_SIZE(ipv6_defrag_ops)); if (err == 0) - nf_frag->users = 1; + net->nf.defrag_ipv6_users = 1; out_unlock: mutex_unlock(&defrag6_mutex); @@ -161,12 +156,10 @@ EXPORT_SYMBOL_GPL(nf_defrag_ipv6_enable); void nf_defrag_ipv6_disable(struct net *net) { - struct nft_ct_frag6_pernet *nf_frag = net_generic(net, nf_frag_pernet_id); - mutex_lock(&defrag6_mutex); - if (nf_frag->users) { - nf_frag->users--; - if (nf_frag->users == 0) + if (net->nf.defrag_ipv6_users) { + net->nf.defrag_ipv6_users--; + if (net->nf.defrag_ipv6_users == 0) nf_unregister_net_hooks(net, ipv6_defrag_ops, ARRAY_SIZE(ipv6_defrag_ops)); } -- cgit v1.2.3 From e189ae161dd784aa5d454b0832f818cacc0e131b Mon Sep 17 00:00:00 2001 From: Pablo Neira Ayuso Date: Tue, 21 Sep 2021 00:37:03 +0200 Subject: netfilter: nf_tables: add position handle in event notification Add position handle to allow to identify the rule location from netlink events. Otherwise, userspace cannot incrementally update a userspace cache through monitoring events. Skip handle dump if the rule has been either inserted (at the beginning of the ruleset) or appended (at the end of the ruleset), the NLM_F_APPEND netlink flag is sufficient in these two cases. Handle NLM_F_REPLACE as NLM_F_APPEND since the rule replacement expansion appends it after the specified rule handle. Fixes: 96518518cc41 ("netfilter: add nftables") Signed-off-by: Pablo Neira Ayuso --- net/netfilter/nf_tables_api.c | 34 +++++++++++++++++++++++++--------- 1 file changed, 25 insertions(+), 9 deletions(-) (limited to 'net') diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c index b9546defdc28..085783b14075 100644 --- a/net/netfilter/nf_tables_api.c +++ b/net/netfilter/nf_tables_api.c @@ -2866,8 +2866,7 @@ static int nf_tables_fill_rule_info(struct sk_buff *skb, struct net *net, u32 flags, int family, const struct nft_table *table, const struct nft_chain *chain, - const struct nft_rule *rule, - const struct nft_rule *prule) + const struct nft_rule *rule, u64 handle) { struct nlmsghdr *nlh; const struct nft_expr *expr, *next; @@ -2887,9 +2886,8 @@ static int nf_tables_fill_rule_info(struct sk_buff *skb, struct net *net, NFTA_RULE_PAD)) goto nla_put_failure; - if (event != NFT_MSG_DELRULE && prule) { - if (nla_put_be64(skb, NFTA_RULE_POSITION, - cpu_to_be64(prule->handle), + if (event != NFT_MSG_DELRULE && handle) { + if (nla_put_be64(skb, NFTA_RULE_POSITION, cpu_to_be64(handle), NFTA_RULE_PAD)) goto nla_put_failure; } @@ -2925,7 +2923,10 @@ static void nf_tables_rule_notify(const struct nft_ctx *ctx, const struct nft_rule *rule, int event) { struct nftables_pernet *nft_net = nft_pernet(ctx->net); + const struct nft_rule *prule; struct sk_buff *skb; + u64 handle = 0; + u16 flags = 0; int err; if (!ctx->report && @@ -2936,9 +2937,18 @@ static void nf_tables_rule_notify(const struct nft_ctx *ctx, if (skb == NULL) goto err; + if (event == NFT_MSG_NEWRULE && + !list_is_first(&rule->list, &ctx->chain->rules) && + !list_is_last(&rule->list, &ctx->chain->rules)) { + prule = list_prev_entry(rule, list); + handle = prule->handle; + } + if (ctx->flags & (NLM_F_APPEND | NLM_F_REPLACE)) + flags |= NLM_F_APPEND; + err = nf_tables_fill_rule_info(skb, ctx->net, ctx->portid, ctx->seq, - event, 0, ctx->family, ctx->table, - ctx->chain, rule, NULL); + event, flags, ctx->family, ctx->table, + ctx->chain, rule, handle); if (err < 0) { kfree_skb(skb); goto err; @@ -2964,6 +2974,7 @@ static int __nf_tables_dump_rules(struct sk_buff *skb, struct net *net = sock_net(skb->sk); const struct nft_rule *rule, *prule; unsigned int s_idx = cb->args[0]; + u64 handle; prule = NULL; list_for_each_entry_rcu(rule, &chain->rules, list) { @@ -2975,12 +2986,17 @@ static int __nf_tables_dump_rules(struct sk_buff *skb, memset(&cb->args[1], 0, sizeof(cb->args) - sizeof(cb->args[0])); } + if (prule) + handle = prule->handle; + else + handle = 0; + if (nf_tables_fill_rule_info(skb, net, NETLINK_CB(cb->skb).portid, cb->nlh->nlmsg_seq, NFT_MSG_NEWRULE, NLM_F_MULTI | NLM_F_APPEND, table->family, - table, chain, rule, prule) < 0) + table, chain, rule, handle) < 0) return 1; nl_dump_check_consistent(cb, nlmsg_hdr(skb)); @@ -3143,7 +3159,7 @@ static int nf_tables_getrule(struct sk_buff *skb, const struct nfnl_info *info, err = nf_tables_fill_rule_info(skb2, net, NETLINK_CB(skb).portid, info->nlh->nlmsg_seq, NFT_MSG_NEWRULE, 0, - family, table, chain, rule, NULL); + family, table, chain, rule, 0); if (err < 0) goto err_fill_rule_info; -- cgit v1.2.3 From 2c964c558641a3bddaee5719c9e6d8805f777812 Mon Sep 17 00:00:00 2001 From: Pablo Neira Ayuso Date: Sat, 25 Sep 2021 00:27:38 +0200 Subject: netfilter: nf_tables: reverse order in rule replacement expansion Deactivate old rule first, then append the new rule, so rule replacement notification via netlink first reports the deletion of the old rule with handle X in first place, then it adds the new rule (reusing the handle X of the replaced old rule). Note that the abort path releases the transaction that has been created by nft_delrule() on error. Fixes: ca08987885a1 ("netfilter: nf_tables: deactivate expressions in rule replecement routine") Signed-off-by: Pablo Neira Ayuso --- net/netfilter/nf_tables_api.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) (limited to 'net') diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c index 085783b14075..c8acd26c7201 100644 --- a/net/netfilter/nf_tables_api.c +++ b/net/netfilter/nf_tables_api.c @@ -3419,17 +3419,15 @@ static int nf_tables_newrule(struct sk_buff *skb, const struct nfnl_info *info, } if (info->nlh->nlmsg_flags & NLM_F_REPLACE) { + err = nft_delrule(&ctx, old_rule); + if (err < 0) + goto err_destroy_flow_rule; + trans = nft_trans_rule_add(&ctx, NFT_MSG_NEWRULE, rule); if (trans == NULL) { err = -ENOMEM; goto err_destroy_flow_rule; } - err = nft_delrule(&ctx, old_rule); - if (err < 0) { - nft_trans_destroy(trans); - goto err_destroy_flow_rule; - } - list_add_tail_rcu(&rule->list, &old_rule->list); } else { trans = nft_trans_rule_add(&ctx, NFT_MSG_NEWRULE, rule); -- cgit v1.2.3 From 2ba5acfb34957e8a7fe47cd78c77ca88e9cc2b03 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Fri, 1 Oct 2021 09:59:21 -0400 Subject: SUNRPC: fix sign error causing rpcsec_gss drops If sd_max is unsigned, then sd_max - GSS_SEQ_WIN is a very large number whenever sd_max is less than GSS_SEQ_WIN, and the comparison: seq_num <= sd->sd_max - GSS_SEQ_WIN in gss_check_seq_num is pretty much always true, even when that's clearly not what was intended. This was causing pynfs to hang when using krb5, because pynfs uses zero as the initial gss sequence number. That's perfectly legal, but this logic error causes knfsd to drop the rpc in that case. Out-of-order sequence IDs in the first GSS_SEQ_WIN (128) calls will also cause this. Fixes: 10b9d99a3dbb ("SUNRPC: Augment server-side rpcgss tracepoints") Cc: stable@vger.kernel.org Signed-off-by: J. Bruce Fields Signed-off-by: Chuck Lever --- net/sunrpc/auth_gss/svcauth_gss.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/sunrpc/auth_gss/svcauth_gss.c b/net/sunrpc/auth_gss/svcauth_gss.c index 475a965b7b43..1ea82a2e63af 100644 --- a/net/sunrpc/auth_gss/svcauth_gss.c +++ b/net/sunrpc/auth_gss/svcauth_gss.c @@ -645,7 +645,7 @@ static bool gss_check_seq_num(const struct svc_rqst *rqstp, struct rsc *rsci, } __set_bit(seq_num % GSS_SEQ_WIN, sd->sd_win); goto ok; - } else if (seq_num <= sd->sd_max - GSS_SEQ_WIN) { + } else if (seq_num + GSS_SEQ_WIN <= sd->sd_max) { goto toolow; } if (__test_and_set_bit(seq_num % GSS_SEQ_WIN, sd->sd_win)) -- cgit v1.2.3 From 560ee196fe9e5037e5015e2cdb14b3aecb1cd7dc Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Thu, 30 Sep 2021 14:22:39 -0700 Subject: net_sched: fix NULL deref in fifo_set_limit() syzbot reported another NULL deref in fifo_set_limit() [1] I could repro the issue with : unshare -n tc qd add dev lo root handle 1:0 tbf limit 200000 burst 70000 rate 100Mbit tc qd replace dev lo parent 1:0 pfifo_fast tc qd change dev lo root handle 1:0 tbf limit 300000 burst 70000 rate 100Mbit pfifo_fast does not have a change() operation. Make fifo_set_limit() more robust about this. [1] BUG: kernel NULL pointer dereference, address: 0000000000000000 PGD 1cf99067 P4D 1cf99067 PUD 7ca49067 PMD 0 Oops: 0010 [#1] PREEMPT SMP KASAN CPU: 1 PID: 14443 Comm: syz-executor959 Not tainted 5.15.0-rc3-syzkaller #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 RIP: 0010:0x0 Code: Unable to access opcode bytes at RIP 0xffffffffffffffd6. RSP: 0018:ffffc9000e2f7310 EFLAGS: 00010246 RAX: dffffc0000000000 RBX: ffffffff8d6ecc00 RCX: 0000000000000000 RDX: 0000000000000000 RSI: ffff888024c27910 RDI: ffff888071e34000 RBP: ffff888071e34000 R08: 0000000000000001 R09: ffffffff8fcfb947 R10: 0000000000000001 R11: 0000000000000000 R12: ffff888024c27910 R13: ffff888071e34018 R14: 0000000000000000 R15: ffff88801ef74800 FS: 00007f321d897700(0000) GS:ffff8880b9d00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: ffffffffffffffd6 CR3: 00000000722c3000 CR4: 00000000003506e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: fifo_set_limit net/sched/sch_fifo.c:242 [inline] fifo_set_limit+0x198/0x210 net/sched/sch_fifo.c:227 tbf_change+0x6ec/0x16d0 net/sched/sch_tbf.c:418 qdisc_change net/sched/sch_api.c:1332 [inline] tc_modify_qdisc+0xd9a/0x1a60 net/sched/sch_api.c:1634 rtnetlink_rcv_msg+0x413/0xb80 net/core/rtnetlink.c:5572 netlink_rcv_skb+0x153/0x420 net/netlink/af_netlink.c:2504 netlink_unicast_kernel net/netlink/af_netlink.c:1314 [inline] netlink_unicast+0x533/0x7d0 net/netlink/af_netlink.c:1340 netlink_sendmsg+0x86d/0xdb0 net/netlink/af_netlink.c:1929 sock_sendmsg_nosec net/socket.c:704 [inline] sock_sendmsg+0xcf/0x120 net/socket.c:724 ____sys_sendmsg+0x6e8/0x810 net/socket.c:2409 ___sys_sendmsg+0xf3/0x170 net/socket.c:2463 __sys_sendmsg+0xe5/0x1b0 net/socket.c:2492 do_syscall_x64 arch/x86/entry/common.c:50 [inline] do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80 entry_SYSCALL_64_after_hwframe+0x44/0xae Fixes: fb0305ce1b03 ("net-sched: consolidate default fifo qdisc setup") Signed-off-by: Eric Dumazet Reported-by: syzbot Link: https://lore.kernel.org/r/20210930212239.3430364-1-eric.dumazet@gmail.com Signed-off-by: Jakub Kicinski --- net/sched/sch_fifo.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'net') diff --git a/net/sched/sch_fifo.c b/net/sched/sch_fifo.c index a579a4131d22..e1040421b797 100644 --- a/net/sched/sch_fifo.c +++ b/net/sched/sch_fifo.c @@ -233,6 +233,9 @@ int fifo_set_limit(struct Qdisc *q, unsigned int limit) if (strncmp(q->ops->id + 1, "fifo", 4) != 0) return 0; + if (!q->ops->change) + return 0; + nla = kmalloc(nla_attr_size(sizeof(struct tc_fifo_qopt)), GFP_KERNEL); if (nla) { nla->nla_type = RTM_NEWQDISC; -- cgit v1.2.3 From 6fb721cf781808ee2ca5e737fb0592cc68de3381 Mon Sep 17 00:00:00 2001 From: Pablo Neira Ayuso Date: Sun, 26 Sep 2021 09:59:35 +0200 Subject: netfilter: nf_tables: honor NLM_F_CREATE and NLM_F_EXCL in event notification Include the NLM_F_CREATE and NLM_F_EXCL flags in netlink event notifications, otherwise userspace cannot distiguish between create and add commands. Fixes: 96518518cc41 ("netfilter: add nftables") Signed-off-by: Pablo Neira Ayuso --- include/net/netfilter/nf_tables.h | 2 +- net/netfilter/nf_tables_api.c | 47 +++++++++++++++++++++++++++++---------- net/netfilter/nft_quota.c | 2 +- 3 files changed, 37 insertions(+), 14 deletions(-) (limited to 'net') diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h index 148f5d8ee5ab..a16171c5fd9e 100644 --- a/include/net/netfilter/nf_tables.h +++ b/include/net/netfilter/nf_tables.h @@ -1202,7 +1202,7 @@ struct nft_object *nft_obj_lookup(const struct net *net, void nft_obj_notify(struct net *net, const struct nft_table *table, struct nft_object *obj, u32 portid, u32 seq, - int event, int family, int report, gfp_t gfp); + int event, u16 flags, int family, int report, gfp_t gfp); /** * struct nft_object_type - stateful object type diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c index c8acd26c7201..c0851fec11d4 100644 --- a/net/netfilter/nf_tables_api.c +++ b/net/netfilter/nf_tables_api.c @@ -780,6 +780,7 @@ static void nf_tables_table_notify(const struct nft_ctx *ctx, int event) { struct nftables_pernet *nft_net; struct sk_buff *skb; + u16 flags = 0; int err; if (!ctx->report && @@ -790,8 +791,11 @@ static void nf_tables_table_notify(const struct nft_ctx *ctx, int event) if (skb == NULL) goto err; + if (ctx->flags & (NLM_F_CREATE | NLM_F_EXCL)) + flags |= ctx->flags & (NLM_F_CREATE | NLM_F_EXCL); + err = nf_tables_fill_table_info(skb, ctx->net, ctx->portid, ctx->seq, - event, 0, ctx->family, ctx->table); + event, flags, ctx->family, ctx->table); if (err < 0) { kfree_skb(skb); goto err; @@ -1563,6 +1567,7 @@ static void nf_tables_chain_notify(const struct nft_ctx *ctx, int event) { struct nftables_pernet *nft_net; struct sk_buff *skb; + u16 flags = 0; int err; if (!ctx->report && @@ -1573,8 +1578,11 @@ static void nf_tables_chain_notify(const struct nft_ctx *ctx, int event) if (skb == NULL) goto err; + if (ctx->flags & (NLM_F_CREATE | NLM_F_EXCL)) + flags |= ctx->flags & (NLM_F_CREATE | NLM_F_EXCL); + err = nf_tables_fill_chain_info(skb, ctx->net, ctx->portid, ctx->seq, - event, 0, ctx->family, ctx->table, + event, flags, ctx->family, ctx->table, ctx->chain); if (err < 0) { kfree_skb(skb); @@ -2945,6 +2953,8 @@ static void nf_tables_rule_notify(const struct nft_ctx *ctx, } if (ctx->flags & (NLM_F_APPEND | NLM_F_REPLACE)) flags |= NLM_F_APPEND; + if (ctx->flags & (NLM_F_CREATE | NLM_F_EXCL)) + flags |= ctx->flags & (NLM_F_CREATE | NLM_F_EXCL); err = nf_tables_fill_rule_info(skb, ctx->net, ctx->portid, ctx->seq, event, flags, ctx->family, ctx->table, @@ -3957,8 +3967,9 @@ static void nf_tables_set_notify(const struct nft_ctx *ctx, gfp_t gfp_flags) { struct nftables_pernet *nft_net = nft_pernet(ctx->net); - struct sk_buff *skb; u32 portid = ctx->portid; + struct sk_buff *skb; + u16 flags = 0; int err; if (!ctx->report && @@ -3969,7 +3980,10 @@ static void nf_tables_set_notify(const struct nft_ctx *ctx, if (skb == NULL) goto err; - err = nf_tables_fill_set(skb, ctx, set, event, 0); + if (ctx->flags & (NLM_F_CREATE | NLM_F_EXCL)) + flags |= ctx->flags & (NLM_F_CREATE | NLM_F_EXCL); + + err = nf_tables_fill_set(skb, ctx, set, event, flags); if (err < 0) { kfree_skb(skb); goto err; @@ -5245,12 +5259,13 @@ static int nf_tables_getsetelem(struct sk_buff *skb, static void nf_tables_setelem_notify(const struct nft_ctx *ctx, const struct nft_set *set, const struct nft_set_elem *elem, - int event, u16 flags) + int event) { struct nftables_pernet *nft_net; struct net *net = ctx->net; u32 portid = ctx->portid; struct sk_buff *skb; + u16 flags = 0; int err; if (!ctx->report && !nfnetlink_has_listeners(net, NFNLGRP_NFTABLES)) @@ -5260,6 +5275,9 @@ static void nf_tables_setelem_notify(const struct nft_ctx *ctx, if (skb == NULL) goto err; + if (ctx->flags & (NLM_F_CREATE | NLM_F_EXCL)) + flags |= ctx->flags & (NLM_F_CREATE | NLM_F_EXCL); + err = nf_tables_fill_setelem_info(skb, ctx, 0, portid, event, flags, set, elem); if (err < 0) { @@ -6935,7 +6953,7 @@ static int nf_tables_delobj(struct sk_buff *skb, const struct nfnl_info *info, void nft_obj_notify(struct net *net, const struct nft_table *table, struct nft_object *obj, u32 portid, u32 seq, int event, - int family, int report, gfp_t gfp) + u16 flags, int family, int report, gfp_t gfp) { struct nftables_pernet *nft_net = nft_pernet(net); struct sk_buff *skb; @@ -6960,8 +6978,9 @@ void nft_obj_notify(struct net *net, const struct nft_table *table, if (skb == NULL) goto err; - err = nf_tables_fill_obj_info(skb, net, portid, seq, event, 0, family, - table, obj, false); + err = nf_tables_fill_obj_info(skb, net, portid, seq, event, + flags & (NLM_F_CREATE | NLM_F_EXCL), + family, table, obj, false); if (err < 0) { kfree_skb(skb); goto err; @@ -6978,7 +6997,7 @@ static void nf_tables_obj_notify(const struct nft_ctx *ctx, struct nft_object *obj, int event) { nft_obj_notify(ctx->net, ctx->table, obj, ctx->portid, ctx->seq, event, - ctx->family, ctx->report, GFP_KERNEL); + ctx->flags, ctx->family, ctx->report, GFP_KERNEL); } /* @@ -7759,6 +7778,7 @@ static void nf_tables_flowtable_notify(struct nft_ctx *ctx, { struct nftables_pernet *nft_net = nft_pernet(ctx->net); struct sk_buff *skb; + u16 flags = 0; int err; if (!ctx->report && @@ -7769,8 +7789,11 @@ static void nf_tables_flowtable_notify(struct nft_ctx *ctx, if (skb == NULL) goto err; + if (ctx->flags & (NLM_F_CREATE | NLM_F_EXCL)) + flags |= ctx->flags & (NLM_F_CREATE | NLM_F_EXCL); + err = nf_tables_fill_flowtable_info(skb, ctx->net, ctx->portid, - ctx->seq, event, 0, + ctx->seq, event, flags, ctx->family, flowtable, hook_list); if (err < 0) { kfree_skb(skb); @@ -8648,7 +8671,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb) nft_setelem_activate(net, te->set, &te->elem); nf_tables_setelem_notify(&trans->ctx, te->set, &te->elem, - NFT_MSG_NEWSETELEM, 0); + NFT_MSG_NEWSETELEM); nft_trans_destroy(trans); break; case NFT_MSG_DELSETELEM: @@ -8656,7 +8679,7 @@ static int nf_tables_commit(struct net *net, struct sk_buff *skb) nf_tables_setelem_notify(&trans->ctx, te->set, &te->elem, - NFT_MSG_DELSETELEM, 0); + NFT_MSG_DELSETELEM); nft_setelem_remove(net, te->set, &te->elem); if (!nft_setelem_is_catchall(te->set, &te->elem)) { atomic_dec(&te->set->nelems); diff --git a/net/netfilter/nft_quota.c b/net/netfilter/nft_quota.c index 0363f533a42b..c4d1389f7185 100644 --- a/net/netfilter/nft_quota.c +++ b/net/netfilter/nft_quota.c @@ -60,7 +60,7 @@ static void nft_quota_obj_eval(struct nft_object *obj, if (overquota && !test_and_set_bit(NFT_QUOTA_DEPLETED_BIT, &priv->flags)) nft_obj_notify(nft_net(pkt), obj->key.table, obj, 0, 0, - NFT_MSG_NEWOBJ, nft_pf(pkt), 0, GFP_ATOMIC); + NFT_MSG_NEWOBJ, 0, nft_pf(pkt), 0, GFP_ATOMIC); } static int nft_quota_do_init(const struct nlattr * const tb[], -- cgit v1.2.3 From b44d52a50bc6f191f0ae03f65de8401f3ef039b3 Mon Sep 17 00:00:00 2001 From: Andrew Lunn Date: Sun, 3 Oct 2021 17:50:53 +0200 Subject: dsa: tag_dsa: Fix mask for trunked packets A packet received on a trunk will have bit 2 set in Forward DSA tagged frame. Bit 1 can be either 0 or 1 and is otherwise undefined and bit 0 indicates the frame CFI. Masking with 7 thus results in frames as being identified as being from a trunk when in fact they are not. Fix the mask to just look at bit 2. Fixes: 5b60dadb71db ("net: dsa: tag_dsa: Support reception of packets from LAG devices") Signed-off-by: Andrew Lunn Reviewed-by: Vladimir Oltean Signed-off-by: David S. Miller --- net/dsa/tag_dsa.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/dsa/tag_dsa.c b/net/dsa/tag_dsa.c index 77d0ce89ab77..e5127b7d1c6a 100644 --- a/net/dsa/tag_dsa.c +++ b/net/dsa/tag_dsa.c @@ -210,7 +210,7 @@ static struct sk_buff *dsa_rcv_ll(struct sk_buff *skb, struct net_device *dev, cmd = dsa_header[0] >> 6; switch (cmd) { case DSA_CMD_FORWARD: - trunk = !!(dsa_header[1] & 7); + trunk = !!(dsa_header[1] & 4); break; case DSA_CMD_TO_CPU: -- cgit v1.2.3 From dbe0b88064494b7bb6a9b2aa7e085b14a3112d44 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Mon, 4 Oct 2021 18:05:07 -0700 Subject: net: bridge: use nla_total_size_64bit() in br_get_linkxstats_size() bridge_fill_linkxstats() is using nla_reserve_64bit(). We must use nla_total_size_64bit() instead of nla_total_size() for corresponding data structure. Fixes: 1080ab95e3c7 ("net: bridge: add support for IGMP/MLD stats and export them via netlink") Signed-off-by: Eric Dumazet Cc: Nikolay Aleksandrov Cc: Vivien Didelot Acked-by: Nikolay Aleksandrov Signed-off-by: David S. Miller --- net/bridge/br_netlink.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/bridge/br_netlink.c b/net/bridge/br_netlink.c index 6c58fc14d2cb..29b8f6373fb9 100644 --- a/net/bridge/br_netlink.c +++ b/net/bridge/br_netlink.c @@ -1666,7 +1666,7 @@ static size_t br_get_linkxstats_size(const struct net_device *dev, int attr) } return numvls * nla_total_size(sizeof(struct bridge_vlan_xstats)) + - nla_total_size(sizeof(struct br_mcast_stats)) + + nla_total_size_64bit(sizeof(struct br_mcast_stats)) + nla_total_size(0); } -- cgit v1.2.3 From 0854a0513321cf70bea5fa483ebcaa983cc7c62e Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Mon, 4 Oct 2021 18:05:08 -0700 Subject: net: bridge: fix under estimation in br_get_linkxstats_size() Commit de1799667b00 ("net: bridge: add STP xstats") added an additional nla_reserve_64bit() in br_fill_linkxstats(), but forgot to update br_get_linkxstats_size() accordingly. This can trigger the following in rtnl_stats_get() WARN_ON(err == -EMSGSIZE); Fixes: de1799667b00 ("net: bridge: add STP xstats") Signed-off-by: Eric Dumazet Cc: Vivien Didelot Cc: Nikolay Aleksandrov Acked-by: Nikolay Aleksandrov Signed-off-by: David S. Miller --- net/bridge/br_netlink.c | 1 + 1 file changed, 1 insertion(+) (limited to 'net') diff --git a/net/bridge/br_netlink.c b/net/bridge/br_netlink.c index 29b8f6373fb9..5c6c4305ed23 100644 --- a/net/bridge/br_netlink.c +++ b/net/bridge/br_netlink.c @@ -1667,6 +1667,7 @@ static size_t br_get_linkxstats_size(const struct net_device *dev, int attr) return numvls * nla_total_size(sizeof(struct bridge_vlan_xstats)) + nla_total_size_64bit(sizeof(struct br_mcast_stats)) + + (p ? nla_total_size_64bit(sizeof(p->stp_xstats)) : 0) + nla_total_size(0); } -- cgit v1.2.3 From a56d447f196fa9973c568f54c0d76d5391c3b0c0 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Mon, 4 Oct 2021 12:55:22 -0700 Subject: net/sched: sch_taprio: properly cancel timer from taprio_destroy() There is a comment in qdisc_create() about us not calling ops->reset() in some cases. err_out4: /* * Any broken qdiscs that would require a ops->reset() here? * The qdisc was never in action so it shouldn't be necessary. */ As taprio sets a timer before actually receiving a packet, we need to cancel it from ops->destroy, just in case ops->reset has not been called. syzbot reported: ODEBUG: free active (active state 0) object type: hrtimer hint: advance_sched+0x0/0x9a0 arch/x86/include/asm/atomic64_64.h:22 WARNING: CPU: 0 PID: 8441 at lib/debugobjects.c:505 debug_print_object+0x16e/0x250 lib/debugobjects.c:505 Modules linked in: CPU: 0 PID: 8441 Comm: syz-executor813 Not tainted 5.14.0-rc6-syzkaller #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 RIP: 0010:debug_print_object+0x16e/0x250 lib/debugobjects.c:505 Code: ff df 48 89 fa 48 c1 ea 03 80 3c 02 00 0f 85 af 00 00 00 48 8b 14 dd e0 d3 e3 89 4c 89 ee 48 c7 c7 e0 c7 e3 89 e8 5b 86 11 05 <0f> 0b 83 05 85 03 92 09 01 48 83 c4 18 5b 5d 41 5c 41 5d 41 5e c3 RSP: 0018:ffffc9000130f330 EFLAGS: 00010282 RAX: 0000000000000000 RBX: 0000000000000003 RCX: 0000000000000000 RDX: ffff88802baeb880 RSI: ffffffff815d87b5 RDI: fffff52000261e58 RBP: 0000000000000001 R08: 0000000000000000 R09: 0000000000000000 R10: ffffffff815d25ee R11: 0000000000000000 R12: ffffffff898dd020 R13: ffffffff89e3ce20 R14: ffffffff81653630 R15: dffffc0000000000 FS: 0000000000f0d300(0000) GS:ffff8880b9d00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007ffb64b3e000 CR3: 0000000036557000 CR4: 00000000001506e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: __debug_check_no_obj_freed lib/debugobjects.c:987 [inline] debug_check_no_obj_freed+0x301/0x420 lib/debugobjects.c:1018 slab_free_hook mm/slub.c:1603 [inline] slab_free_freelist_hook+0x171/0x240 mm/slub.c:1653 slab_free mm/slub.c:3213 [inline] kfree+0xe4/0x540 mm/slub.c:4267 qdisc_create+0xbcf/0x1320 net/sched/sch_api.c:1299 tc_modify_qdisc+0x4c8/0x1a60 net/sched/sch_api.c:1663 rtnetlink_rcv_msg+0x413/0xb80 net/core/rtnetlink.c:5571 netlink_rcv_skb+0x153/0x420 net/netlink/af_netlink.c:2504 netlink_unicast_kernel net/netlink/af_netlink.c:1314 [inline] netlink_unicast+0x533/0x7d0 net/netlink/af_netlink.c:1340 netlink_sendmsg+0x86d/0xdb0 net/netlink/af_netlink.c:1929 sock_sendmsg_nosec net/socket.c:704 [inline] sock_sendmsg+0xcf/0x120 net/socket.c:724 ____sys_sendmsg+0x6e8/0x810 net/socket.c:2403 ___sys_sendmsg+0xf3/0x170 net/socket.c:2457 __sys_sendmsg+0xe5/0x1b0 net/socket.c:2486 do_syscall_x64 arch/x86/entry/common.c:50 [inline] do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80 Fixes: 44d4775ca518 ("net/sched: sch_taprio: reset child qdiscs before freeing them") Signed-off-by: Eric Dumazet Cc: Davide Caratti Reported-by: syzbot Acked-by: Vinicius Costa Gomes Acked-by: Davide Caratti Signed-off-by: David S. Miller --- net/sched/sch_taprio.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'net') diff --git a/net/sched/sch_taprio.c b/net/sched/sch_taprio.c index 1ab2fc933a21..b9fd18d98646 100644 --- a/net/sched/sch_taprio.c +++ b/net/sched/sch_taprio.c @@ -1641,6 +1641,10 @@ static void taprio_destroy(struct Qdisc *sch) list_del(&q->taprio_list); spin_unlock(&taprio_list_lock); + /* Note that taprio_reset() might not be called if an error + * happens in qdisc_create(), after taprio_init() has been called. + */ + hrtimer_cancel(&q->advance_timer); taprio_disable_offload(dev, q, NULL); -- cgit v1.2.3 From 7707a4d01a648e4c655101a469c956cb11273655 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Mon, 4 Oct 2021 14:24:15 -0700 Subject: netlink: annotate data races around nlk->bound While existing code is correct, KCSAN is reporting a data-race in netlink_insert / netlink_sendmsg [1] It is correct to read nlk->bound without a lock, as netlink_autobind() will acquire all needed locks. [1] BUG: KCSAN: data-race in netlink_insert / netlink_sendmsg write to 0xffff8881031c8b30 of 1 bytes by task 18752 on cpu 0: netlink_insert+0x5cc/0x7f0 net/netlink/af_netlink.c:597 netlink_autobind+0xa9/0x150 net/netlink/af_netlink.c:842 netlink_sendmsg+0x479/0x7c0 net/netlink/af_netlink.c:1892 sock_sendmsg_nosec net/socket.c:703 [inline] sock_sendmsg net/socket.c:723 [inline] ____sys_sendmsg+0x360/0x4d0 net/socket.c:2392 ___sys_sendmsg net/socket.c:2446 [inline] __sys_sendmsg+0x1ed/0x270 net/socket.c:2475 __do_sys_sendmsg net/socket.c:2484 [inline] __se_sys_sendmsg net/socket.c:2482 [inline] __x64_sys_sendmsg+0x42/0x50 net/socket.c:2482 do_syscall_x64 arch/x86/entry/common.c:50 [inline] do_syscall_64+0x3d/0x90 arch/x86/entry/common.c:80 entry_SYSCALL_64_after_hwframe+0x44/0xae read to 0xffff8881031c8b30 of 1 bytes by task 18751 on cpu 1: netlink_sendmsg+0x270/0x7c0 net/netlink/af_netlink.c:1891 sock_sendmsg_nosec net/socket.c:703 [inline] sock_sendmsg net/socket.c:723 [inline] __sys_sendto+0x2a8/0x370 net/socket.c:2019 __do_sys_sendto net/socket.c:2031 [inline] __se_sys_sendto net/socket.c:2027 [inline] __x64_sys_sendto+0x74/0x90 net/socket.c:2027 do_syscall_x64 arch/x86/entry/common.c:50 [inline] do_syscall_64+0x3d/0x90 arch/x86/entry/common.c:80 entry_SYSCALL_64_after_hwframe+0x44/0xae value changed: 0x00 -> 0x01 Reported by Kernel Concurrency Sanitizer on: CPU: 1 PID: 18751 Comm: syz-executor.0 Not tainted 5.14.0-rc1-syzkaller #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 Fixes: da314c9923fe ("netlink: Replace rhash_portid with bound") Signed-off-by: Eric Dumazet Reported-by: syzbot Signed-off-by: David S. Miller --- net/netlink/af_netlink.c | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/netlink/af_netlink.c b/net/netlink/af_netlink.c index 24b7cf447bc5..ada47e59647a 100644 --- a/net/netlink/af_netlink.c +++ b/net/netlink/af_netlink.c @@ -594,7 +594,10 @@ static int netlink_insert(struct sock *sk, u32 portid) /* We need to ensure that the socket is hashed and visible. */ smp_wmb(); - nlk_sk(sk)->bound = portid; + /* Paired with lockless reads from netlink_bind(), + * netlink_connect() and netlink_sendmsg(). + */ + WRITE_ONCE(nlk_sk(sk)->bound, portid); err: release_sock(sk); @@ -1012,7 +1015,8 @@ static int netlink_bind(struct socket *sock, struct sockaddr *addr, if (nlk->ngroups < BITS_PER_LONG) groups &= (1UL << nlk->ngroups) - 1; - bound = nlk->bound; + /* Paired with WRITE_ONCE() in netlink_insert() */ + bound = READ_ONCE(nlk->bound); if (bound) { /* Ensure nlk->portid is up-to-date. */ smp_rmb(); @@ -1098,8 +1102,9 @@ static int netlink_connect(struct socket *sock, struct sockaddr *addr, /* No need for barriers here as we return to user-space without * using any of the bound attributes. + * Paired with WRITE_ONCE() in netlink_insert(). */ - if (!nlk->bound) + if (!READ_ONCE(nlk->bound)) err = netlink_autobind(sock); if (err == 0) { @@ -1888,7 +1893,8 @@ static int netlink_sendmsg(struct socket *sock, struct msghdr *msg, size_t len) dst_group = nlk->dst_group; } - if (!nlk->bound) { + /* Paired with WRITE_ONCE() in netlink_insert() */ + if (!READ_ONCE(nlk->bound)) { err = netlink_autobind(sock); if (err) goto out; -- cgit v1.2.3 From d0c6416bd7091647f6041599f396bfa19ae30368 Mon Sep 17 00:00:00 2001 From: Jiang Wang Date: Mon, 4 Oct 2021 23:25:28 +0000 Subject: unix: Fix an issue in unix_shutdown causing the other end read/write failures Commit 94531cfcbe79 ("af_unix: Add unix_stream_proto for sockmap") sets unix domain socket peer state to TCP_CLOSE in unix_shutdown. This could happen when the local end is shutdown but the other end is not. Then, the other end will get read or write failures which is not expected. Fix the issue by setting the local state to shutdown. Fixes: 94531cfcbe79 ("af_unix: Add unix_stream_proto for sockmap") Reported-by: Casey Schaufler Suggested-by: Cong Wang Signed-off-by: Jiang Wang Signed-off-by: Daniel Borkmann Tested-by: Casey Schaufler Reviewed-by: Casey Schaufler Acked-by: Song Liu Link: https://lore.kernel.org/bpf/20211004232530.2377085-1-jiang.wang@bytedance.com --- net/unix/af_unix.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c index f505b89bda6a..915afcae6a12 100644 --- a/net/unix/af_unix.c +++ b/net/unix/af_unix.c @@ -2860,6 +2860,9 @@ static int unix_shutdown(struct socket *sock, int mode) unix_state_lock(sk); sk->sk_shutdown |= mode; + if ((sk->sk_type == SOCK_STREAM || sk->sk_type == SOCK_SEQPACKET) && + mode == SHUTDOWN_MASK) + sk->sk_state = TCP_CLOSE; other = unix_peer(sk); if (other) sock_hold(other); @@ -2882,12 +2885,10 @@ static int unix_shutdown(struct socket *sock, int mode) other->sk_shutdown |= peer_mode; unix_state_unlock(other); other->sk_state_change(other); - if (peer_mode == SHUTDOWN_MASK) { + if (peer_mode == SHUTDOWN_MASK) sk_wake_async(other, SOCK_WAKE_WAITD, POLL_HUP); - other->sk_state = TCP_CLOSE; - } else if (peer_mode & RCV_SHUTDOWN) { + else if (peer_mode & RCV_SHUTDOWN) sk_wake_async(other, SOCK_WAKE_WAITD, POLL_IN); - } } if (other) sock_put(other); -- cgit v1.2.3 From d34367991933d28bd7331f67a759be9a8c474014 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Tue, 5 Oct 2021 14:04:17 -0700 Subject: rtnetlink: fix if_nlmsg_stats_size() under estimation rtnl_fill_statsinfo() is filling skb with one mandatory if_stats_msg structure. nlmsg_put(skb, pid, seq, type, sizeof(struct if_stats_msg), flags); But if_nlmsg_stats_size() never considered the needed storage. This bug did not show up because alloc_skb(X) allocates skb with extra tailroom, because of added alignments. This could very well be changed in the future to have deterministic behavior. Fixes: 10c9ead9f3c6 ("rtnetlink: add new RTM_GETSTATS message to dump link stats") Signed-off-by: Eric Dumazet Cc: Roopa Prabhu Acked-by: Roopa Prabhu Signed-off-by: David S. Miller --- net/core/rtnetlink.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net') diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c index 972c8cb303a5..8ccce85562a1 100644 --- a/net/core/rtnetlink.c +++ b/net/core/rtnetlink.c @@ -5262,7 +5262,7 @@ nla_put_failure: static size_t if_nlmsg_stats_size(const struct net_device *dev, u32 filter_mask) { - size_t size = 0; + size_t size = NLMSG_ALIGN(sizeof(struct if_stats_msg)); if (stats_attr_valid(filter_mask, IFLA_STATS_LINK_64, 0)) size += nla_total_size_64bit(sizeof(struct rtnl_link_stats64)); -- cgit v1.2.3 From 8d6c414cd2fb74aa6812e9bfec6178f8246c4f3a Mon Sep 17 00:00:00 2001 From: Mike Manning Date: Tue, 5 Oct 2021 14:03:42 +0100 Subject: net: prefer socket bound to interface when not in VRF The commit 6da5b0f027a8 ("net: ensure unbound datagram socket to be chosen when not in a VRF") modified compute_score() so that a device match is always made, not just in the case of an l3mdev skb, then increments the score also for unbound sockets. This ensures that sockets bound to an l3mdev are never selected when not in a VRF. But as unbound and bound sockets are now scored equally, this results in the last opened socket being selected if there are matches in the default VRF for an unbound socket and a socket bound to a dev that is not an l3mdev. However, handling prior to this commit was to always select the bound socket in this case. Reinstate this handling by incrementing the score only for bound sockets. The required isolation due to choosing between an unbound socket and a socket bound to an l3mdev remains in place due to the device match always being made. The same approach is taken for compute_score() for stream sockets. Fixes: 6da5b0f027a8 ("net: ensure unbound datagram socket to be chosen when not in a VRF") Fixes: e78190581aff ("net: ensure unbound stream socket to be chosen when not in a VRF") Signed-off-by: Mike Manning Reviewed-by: David Ahern Link: https://lore.kernel.org/r/cf0a8523-b362-1edf-ee78-eef63cbbb428@gmail.com Signed-off-by: Jakub Kicinski --- net/ipv4/inet_hashtables.c | 4 +++- net/ipv4/udp.c | 3 ++- net/ipv6/inet6_hashtables.c | 2 +- net/ipv6/udp.c | 3 ++- 4 files changed, 8 insertions(+), 4 deletions(-) (limited to 'net') diff --git a/net/ipv4/inet_hashtables.c b/net/ipv4/inet_hashtables.c index 80aeaf9e6e16..bfb522e51346 100644 --- a/net/ipv4/inet_hashtables.c +++ b/net/ipv4/inet_hashtables.c @@ -242,8 +242,10 @@ static inline int compute_score(struct sock *sk, struct net *net, if (!inet_sk_bound_dev_eq(net, sk->sk_bound_dev_if, dif, sdif)) return -1; + score = sk->sk_bound_dev_if ? 2 : 1; - score = sk->sk_family == PF_INET ? 2 : 1; + if (sk->sk_family == PF_INET) + score++; if (READ_ONCE(sk->sk_incoming_cpu) == raw_smp_processor_id()) score++; } diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index 2a7825a5b842..8536b2a7210b 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -390,7 +390,8 @@ static int compute_score(struct sock *sk, struct net *net, dif, sdif); if (!dev_match) return -1; - score += 4; + if (sk->sk_bound_dev_if) + score += 4; if (READ_ONCE(sk->sk_incoming_cpu) == raw_smp_processor_id()) score++; diff --git a/net/ipv6/inet6_hashtables.c b/net/ipv6/inet6_hashtables.c index 55c290d55605..67c9114835c8 100644 --- a/net/ipv6/inet6_hashtables.c +++ b/net/ipv6/inet6_hashtables.c @@ -106,7 +106,7 @@ static inline int compute_score(struct sock *sk, struct net *net, if (!inet_sk_bound_dev_eq(net, sk->sk_bound_dev_if, dif, sdif)) return -1; - score = 1; + score = sk->sk_bound_dev_if ? 2 : 1; if (READ_ONCE(sk->sk_incoming_cpu) == raw_smp_processor_id()) score++; } diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c index e505bb007e9f..8d785232b479 100644 --- a/net/ipv6/udp.c +++ b/net/ipv6/udp.c @@ -133,7 +133,8 @@ static int compute_score(struct sock *sk, struct net *net, dev_match = udp_sk_bound_dev_eq(net, sk->sk_bound_dev_if, dif, sdif); if (!dev_match) return -1; - score++; + if (sk->sk_bound_dev_if) + score++; if (READ_ONCE(sk->sk_incoming_cpu) == raw_smp_processor_id()) score++; -- cgit v1.2.3