From 1d3fd8a10bedb09006cfc963bfcf051c3021f626 Mon Sep 17 00:00:00 2001 From: Stephen Suryaputra Date: Sat, 27 Apr 2019 09:14:33 -0400 Subject: vrf: Use orig netdev to count Ip6InNoRoutes and a fresh route lookup when sending dest unreach When there is no route to an IPv6 dest addr, skb_dst(skb) points to loopback dev in the case of that the IP6CB(skb)->iif is enslaved to a vrf. This causes Ip6InNoRoutes to be incremented on the loopback dev. This also causes the lookup to fail on icmpv6_send() and the dest unreachable to not sent and Ip6OutNoRoutes gets incremented on the loopback dev. To reproduce: * Gateway configuration: ip link add dev vrf_258 type vrf table 258 ip link set dev enp0s9 master vrf_258 ip addr add 66:1/64 dev enp0s9 ip -6 route add unreachable default metric 8192 table 258 sysctl -w net.ipv6.conf.all.forwarding=1 sysctl -w net.ipv6.conf.enp0s9.forwarding=1 * Sender configuration: ip addr add 66::2/64 dev enp0s9 ip -6 route add default via 66::1 and ping 67::1 for example from the sender. Fix this by counting on the original netdev and reset the skb dst to force a fresh lookup. v2: Fix typo of destination address in the repro steps. v3: Simplify the loopback check (per David Ahern) and use reverse Christmas tree format (per David Miller). Signed-off-by: Stephen Suryaputra Reviewed-by: David Ahern Tested-by: David Ahern Reviewed-by: David Ahern Signed-off-by: David S. Miller --- net/ipv6/route.c | 23 +++++++++++++++++------ 1 file changed, 17 insertions(+), 6 deletions(-) (limited to 'net/ipv6/route.c') diff --git a/net/ipv6/route.c b/net/ipv6/route.c index 7178e32eb15d..b4899f0de0d0 100644 --- a/net/ipv6/route.c +++ b/net/ipv6/route.c @@ -3668,23 +3668,34 @@ int ipv6_route_ioctl(struct net *net, unsigned int cmd, void __user *arg) static int ip6_pkt_drop(struct sk_buff *skb, u8 code, int ipstats_mib_noroutes) { - int type; struct dst_entry *dst = skb_dst(skb); + struct net *net = dev_net(dst->dev); + struct inet6_dev *idev; + int type; + + if (netif_is_l3_master(skb->dev) && + dst->dev == net->loopback_dev) + idev = __in6_dev_get_safely(dev_get_by_index_rcu(net, IP6CB(skb)->iif)); + else + idev = ip6_dst_idev(dst); + switch (ipstats_mib_noroutes) { case IPSTATS_MIB_INNOROUTES: type = ipv6_addr_type(&ipv6_hdr(skb)->daddr); if (type == IPV6_ADDR_ANY) { - IP6_INC_STATS(dev_net(dst->dev), - __in6_dev_get_safely(skb->dev), - IPSTATS_MIB_INADDRERRORS); + IP6_INC_STATS(net, idev, IPSTATS_MIB_INADDRERRORS); break; } /* FALLTHROUGH */ case IPSTATS_MIB_OUTNOROUTES: - IP6_INC_STATS(dev_net(dst->dev), ip6_dst_idev(dst), - ipstats_mib_noroutes); + IP6_INC_STATS(net, idev, ipstats_mib_noroutes); break; } + + /* Start over by dropping the dst for l3mdev case */ + if (netif_is_l3_master(skb->dev)) + skb_dst_drop(skb); + icmpv6_send(skb, ICMPV6_DEST_UNREACH, code, 0); kfree_skb(skb); return 0; -- cgit v1.2.3 From 0e2338749192ce0e52e7174c5352f627632f478a Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Sun, 28 Apr 2019 12:22:25 -0700 Subject: ipv6: fix races in ip6_dst_destroy() We had many syzbot reports that seem to be caused by use-after-free of struct fib6_info. ip6_dst_destroy(), fib6_drop_pcpu_from() and rt6_remove_exception() are writers vs rt->from, and use non consistent synchronization among themselves. Switching to xchg() will solve the issues with no possible lockdep issues. BUG: KASAN: user-memory-access in atomic_dec_and_test include/asm-generic/atomic-instrumented.h:747 [inline] BUG: KASAN: user-memory-access in fib6_info_release include/net/ip6_fib.h:294 [inline] BUG: KASAN: user-memory-access in fib6_info_release include/net/ip6_fib.h:292 [inline] BUG: KASAN: user-memory-access in fib6_drop_pcpu_from net/ipv6/ip6_fib.c:927 [inline] BUG: KASAN: user-memory-access in fib6_purge_rt+0x4f6/0x670 net/ipv6/ip6_fib.c:960 Write of size 4 at addr 0000000000ffffb4 by task syz-executor.1/7649 CPU: 0 PID: 7649 Comm: syz-executor.1 Not tainted 5.1.0-rc6+ #183 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 Call Trace: __dump_stack lib/dump_stack.c:77 [inline] dump_stack+0x172/0x1f0 lib/dump_stack.c:113 kasan_report.cold+0x5/0x40 mm/kasan/report.c:321 check_memory_region_inline mm/kasan/generic.c:185 [inline] check_memory_region+0x123/0x190 mm/kasan/generic.c:191 kasan_check_write+0x14/0x20 mm/kasan/common.c:108 atomic_dec_and_test include/asm-generic/atomic-instrumented.h:747 [inline] fib6_info_release include/net/ip6_fib.h:294 [inline] fib6_info_release include/net/ip6_fib.h:292 [inline] fib6_drop_pcpu_from net/ipv6/ip6_fib.c:927 [inline] fib6_purge_rt+0x4f6/0x670 net/ipv6/ip6_fib.c:960 fib6_del_route net/ipv6/ip6_fib.c:1813 [inline] fib6_del+0xac2/0x10a0 net/ipv6/ip6_fib.c:1844 fib6_clean_node+0x3a8/0x590 net/ipv6/ip6_fib.c:2006 fib6_walk_continue+0x495/0x900 net/ipv6/ip6_fib.c:1928 fib6_walk+0x9d/0x100 net/ipv6/ip6_fib.c:1976 fib6_clean_tree+0xe0/0x120 net/ipv6/ip6_fib.c:2055 __fib6_clean_all+0x118/0x2a0 net/ipv6/ip6_fib.c:2071 fib6_clean_all+0x2b/0x40 net/ipv6/ip6_fib.c:2082 rt6_sync_down_dev+0x134/0x150 net/ipv6/route.c:4057 rt6_disable_ip+0x27/0x5f0 net/ipv6/route.c:4062 addrconf_ifdown+0xa2/0x1220 net/ipv6/addrconf.c:3705 addrconf_notify+0x19a/0x2260 net/ipv6/addrconf.c:3630 notifier_call_chain+0xc7/0x240 kernel/notifier.c:93 __raw_notifier_call_chain kernel/notifier.c:394 [inline] raw_notifier_call_chain+0x2e/0x40 kernel/notifier.c:401 call_netdevice_notifiers_info+0x3f/0x90 net/core/dev.c:1753 call_netdevice_notifiers_extack net/core/dev.c:1765 [inline] call_netdevice_notifiers net/core/dev.c:1779 [inline] dev_close_many+0x33f/0x6f0 net/core/dev.c:1522 rollback_registered_many+0x43b/0xfd0 net/core/dev.c:8177 rollback_registered+0x109/0x1d0 net/core/dev.c:8242 unregister_netdevice_queue net/core/dev.c:9289 [inline] unregister_netdevice_queue+0x1ee/0x2c0 net/core/dev.c:9282 unregister_netdevice include/linux/netdevice.h:2658 [inline] __tun_detach+0xd5b/0x1000 drivers/net/tun.c:727 tun_detach drivers/net/tun.c:744 [inline] tun_chr_close+0xe0/0x180 drivers/net/tun.c:3443 __fput+0x2e5/0x8d0 fs/file_table.c:278 ____fput+0x16/0x20 fs/file_table.c:309 task_work_run+0x14a/0x1c0 kernel/task_work.c:113 exit_task_work include/linux/task_work.h:22 [inline] do_exit+0x90a/0x2fa0 kernel/exit.c:876 do_group_exit+0x135/0x370 kernel/exit.c:980 __do_sys_exit_group kernel/exit.c:991 [inline] __se_sys_exit_group kernel/exit.c:989 [inline] __x64_sys_exit_group+0x44/0x50 kernel/exit.c:989 do_syscall_64+0x103/0x610 arch/x86/entry/common.c:290 entry_SYSCALL_64_after_hwframe+0x49/0xbe RIP: 0033:0x458da9 Code: ad b8 fb ff c3 66 2e 0f 1f 84 00 00 00 00 00 66 90 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 0f 83 7b b8 fb ff c3 66 2e 0f 1f 84 00 00 00 00 RSP: 002b:00007ffeafc2a6a8 EFLAGS: 00000246 ORIG_RAX: 00000000000000e7 RAX: ffffffffffffffda RBX: 000000000000001c RCX: 0000000000458da9 RDX: 0000000000412a80 RSI: 0000000000a54ef0 RDI: 0000000000000043 RBP: 00000000004be552 R08: 000000000000000c R09: 000000000004c0d1 R10: 0000000002341940 R11: 0000000000000246 R12: 00000000ffffffff R13: 00007ffeafc2a7f0 R14: 000000000004c065 R15: 00007ffeafc2a800 Fixes: a68886a69180 ("net/ipv6: Make from in rt6_info rcu protected") Signed-off-by: Eric Dumazet Reported-by: syzbot Cc: David Ahern Reviewed-by: David Ahern Acked-by: Martin KaFai Lau Acked-by: Wei Wang Signed-off-by: David S. Miller --- net/ipv6/ip6_fib.c | 4 +--- net/ipv6/route.c | 9 ++------- 2 files changed, 3 insertions(+), 10 deletions(-) (limited to 'net/ipv6/route.c') diff --git a/net/ipv6/ip6_fib.c b/net/ipv6/ip6_fib.c index 6613d8dbb0e5..91247a6fc67f 100644 --- a/net/ipv6/ip6_fib.c +++ b/net/ipv6/ip6_fib.c @@ -921,9 +921,7 @@ static void fib6_drop_pcpu_from(struct fib6_info *f6i, if (pcpu_rt) { struct fib6_info *from; - from = rcu_dereference_protected(pcpu_rt->from, - lockdep_is_held(&table->tb6_lock)); - rcu_assign_pointer(pcpu_rt->from, NULL); + from = xchg((__force struct fib6_info **)&pcpu_rt->from, NULL); fib6_info_release(from); } } diff --git a/net/ipv6/route.c b/net/ipv6/route.c index b4899f0de0d0..2cc166bc978d 100644 --- a/net/ipv6/route.c +++ b/net/ipv6/route.c @@ -379,11 +379,8 @@ static void ip6_dst_destroy(struct dst_entry *dst) in6_dev_put(idev); } - rcu_read_lock(); - from = rcu_dereference(rt->from); - rcu_assign_pointer(rt->from, NULL); + from = xchg((__force struct fib6_info **)&rt->from, NULL); fib6_info_release(from); - rcu_read_unlock(); } static void ip6_dst_ifdown(struct dst_entry *dst, struct net_device *dev, @@ -1288,9 +1285,7 @@ static void rt6_remove_exception(struct rt6_exception_bucket *bucket, /* purge completely the exception to allow releasing the held resources: * some [sk] cache may keep the dst around for unlimited time */ - from = rcu_dereference_protected(rt6_ex->rt6i->from, - lockdep_is_held(&rt6_exception_lock)); - rcu_assign_pointer(rt6_ex->rt6i->from, NULL); + from = xchg((__force struct fib6_info **)&rt6_ex->rt6i->from, NULL); fib6_info_release(from); dst_dev_put(&rt6_ex->rt6i->dst); -- cgit v1.2.3 From 886b7a50100a50f1cbd08a6f8ec5884dfbe082dc Mon Sep 17 00:00:00 2001 From: Martin KaFai Lau Date: Tue, 30 Apr 2019 10:45:12 -0700 Subject: ipv6: A few fixes on dereferencing rt->from It is a followup after the fix in commit 9c69a1320515 ("route: Avoid crash from dereferencing NULL rt->from") rt6_do_redirect(): 1. NULL checking is needed on rt->from because a parallel fib6_info delete could happen that sets rt->from to NULL. (e.g. rt6_remove_exception() and fib6_drop_pcpu_from()). 2. fib6_info_hold() is not enough. Same reason as (1). Meaning, holding dst->__refcnt cannot ensure rt->from is not NULL or rt->from->fib6_ref is not 0. Instead of using fib6_info_hold_safe() which ip6_rt_cache_alloc() is already doing, this patch chooses to extend the rcu section to keep "from" dereference-able after checking for NULL. inet6_rtm_getroute(): 1. NULL checking is also needed on rt->from for a similar reason. Note that inet6_rtm_getroute() is using RTNL_FLAG_DOIT_UNLOCKED. Fixes: a68886a69180 ("net/ipv6: Make from in rt6_info rcu protected") Signed-off-by: Martin KaFai Lau Acked-by: Wei Wang Reviewed-by: David Ahern Reviewed-by: Eric Dumazet Signed-off-by: David S. Miller --- net/ipv6/route.c | 38 ++++++++++++++++++-------------------- 1 file changed, 18 insertions(+), 20 deletions(-) (limited to 'net/ipv6/route.c') diff --git a/net/ipv6/route.c b/net/ipv6/route.c index 2cc166bc978d..0520aca3354b 100644 --- a/net/ipv6/route.c +++ b/net/ipv6/route.c @@ -3392,11 +3392,8 @@ static void rt6_do_redirect(struct dst_entry *dst, struct sock *sk, struct sk_bu rcu_read_lock(); from = rcu_dereference(rt->from); - /* This fib6_info_hold() is safe here because we hold reference to rt - * and rt already holds reference to fib6_info. - */ - fib6_info_hold(from); - rcu_read_unlock(); + if (!from) + goto out; nrt = ip6_rt_cache_alloc(from, &msg->dest, NULL); if (!nrt) @@ -3408,10 +3405,7 @@ static void rt6_do_redirect(struct dst_entry *dst, struct sock *sk, struct sk_bu nrt->rt6i_gateway = *(struct in6_addr *)neigh->primary_key; - /* No need to remove rt from the exception table if rt is - * a cached route because rt6_insert_exception() will - * takes care of it - */ + /* rt6_insert_exception() will take care of duplicated exceptions */ if (rt6_insert_exception(nrt, from)) { dst_release_immediate(&nrt->dst); goto out; @@ -3424,7 +3418,7 @@ static void rt6_do_redirect(struct dst_entry *dst, struct sock *sk, struct sk_bu call_netevent_notifiers(NETEVENT_REDIRECT, &netevent); out: - fib6_info_release(from); + rcu_read_unlock(); neigh_release(neigh); } @@ -5023,16 +5017,20 @@ static int inet6_rtm_getroute(struct sk_buff *in_skb, struct nlmsghdr *nlh, rcu_read_lock(); from = rcu_dereference(rt->from); - - if (fibmatch) - err = rt6_fill_node(net, skb, from, NULL, NULL, NULL, iif, - RTM_NEWROUTE, NETLINK_CB(in_skb).portid, - nlh->nlmsg_seq, 0); - else - err = rt6_fill_node(net, skb, from, dst, &fl6.daddr, - &fl6.saddr, iif, RTM_NEWROUTE, - NETLINK_CB(in_skb).portid, nlh->nlmsg_seq, - 0); + if (from) { + if (fibmatch) + err = rt6_fill_node(net, skb, from, NULL, NULL, NULL, + iif, RTM_NEWROUTE, + NETLINK_CB(in_skb).portid, + nlh->nlmsg_seq, 0); + else + err = rt6_fill_node(net, skb, from, dst, &fl6.daddr, + &fl6.saddr, iif, RTM_NEWROUTE, + NETLINK_CB(in_skb).portid, + nlh->nlmsg_seq, 0); + } else { + err = -ENETUNREACH; + } rcu_read_unlock(); if (err < 0) { -- cgit v1.2.3