From cc01572e2fb080e279ca625f239aca61f435ebf3 Mon Sep 17 00:00:00 2001 From: Yossi Kuperman Date: Wed, 17 Jan 2018 15:52:41 +0200 Subject: [PATCH 01/19] xfrm: Add SA to hardware at the end of xfrm_state_construct() Current code configures the hardware with a new SA before the state has been fully initialized. During this time interval, an incoming ESP packet can cause a crash due to a NULL dereference. More specifically, xfrm_input() considers the packet as valid, and yet, anti-replay mechanism is not initialized. Move hardware configuration to the end of xfrm_state_construct(), and mark the state as valid once the SA is fully initialized. Fixes: d77e38e612a0 ("xfrm: Add an IPsec hardware offloading API") Signed-off-by: Aviad Yehezkel Signed-off-by: Aviv Heller Signed-off-by: Yossi Kuperman Signed-off-by: Steffen Klassert --- net/xfrm/xfrm_state.c | 10 +++++++--- net/xfrm/xfrm_user.c | 18 +++++++++++------- 2 files changed, 18 insertions(+), 10 deletions(-) diff --git a/net/xfrm/xfrm_state.c b/net/xfrm/xfrm_state.c index 429957412633..2d486492acdb 100644 --- a/net/xfrm/xfrm_state.c +++ b/net/xfrm/xfrm_state.c @@ -2272,8 +2272,6 @@ int __xfrm_init_state(struct xfrm_state *x, bool init_replay, bool offload) goto error; } - x->km.state = XFRM_STATE_VALID; - error: return err; } @@ -2282,7 +2280,13 @@ EXPORT_SYMBOL(__xfrm_init_state); int xfrm_init_state(struct xfrm_state *x) { - return __xfrm_init_state(x, true, false); + int err; + + err = __xfrm_init_state(x, true, false); + if (!err) + x->km.state = XFRM_STATE_VALID; + + return err; } EXPORT_SYMBOL(xfrm_init_state); diff --git a/net/xfrm/xfrm_user.c b/net/xfrm/xfrm_user.c index bdb48e5dba04..7f52b8eb177d 100644 --- a/net/xfrm/xfrm_user.c +++ b/net/xfrm/xfrm_user.c @@ -598,13 +598,6 @@ static struct xfrm_state *xfrm_state_construct(struct net *net, goto error; } - if (attrs[XFRMA_OFFLOAD_DEV]) { - err = xfrm_dev_state_add(net, x, - nla_data(attrs[XFRMA_OFFLOAD_DEV])); - if (err) - goto error; - } - if ((err = xfrm_alloc_replay_state_esn(&x->replay_esn, &x->preplay_esn, attrs[XFRMA_REPLAY_ESN_VAL]))) goto error; @@ -620,6 +613,14 @@ static struct xfrm_state *xfrm_state_construct(struct net *net, /* override default values from above */ xfrm_update_ae_params(x, attrs, 0); + /* configure the hardware if offload is requested */ + if (attrs[XFRMA_OFFLOAD_DEV]) { + err = xfrm_dev_state_add(net, x, + nla_data(attrs[XFRMA_OFFLOAD_DEV])); + if (err) + goto error; + } + return x; error: @@ -662,6 +663,9 @@ static int xfrm_add_sa(struct sk_buff *skb, struct nlmsghdr *nlh, goto out; } + if (x->km.state == XFRM_STATE_VOID) + x->km.state = XFRM_STATE_VALID; + c.seq = nlh->nlmsg_seq; c.portid = nlh->nlmsg_pid; c.event = nlh->nlmsg_type; From aa5dd6fa6f5d4bdc82a67e952bba8ad2e98d77e2 Mon Sep 17 00:00:00 2001 From: Aviad Yehezkel Date: Thu, 18 Jan 2018 15:41:51 +0200 Subject: [PATCH 02/19] xfrm: fix error flow in case of add state fails If add state fails in case of device offload, netdev refcount will be negative since gc task is attempting to dev_free this state. This is fixed by putting NULL in state dev field. Signed-off-by: Aviad Yehezkel Signed-off-by: Boris Pismeny Signed-off-by: Steffen Klassert --- net/xfrm/xfrm_device.c | 1 + 1 file changed, 1 insertion(+) diff --git a/net/xfrm/xfrm_device.c b/net/xfrm/xfrm_device.c index 30e5746085b8..ac9477189d1c 100644 --- a/net/xfrm/xfrm_device.c +++ b/net/xfrm/xfrm_device.c @@ -102,6 +102,7 @@ int xfrm_dev_state_add(struct net *net, struct xfrm_state *x, err = dev->xfrmdev_ops->xdo_dev_state_add(x); if (err) { + xso->dev = NULL; dev_put(dev); return err; } From 5efec5c655dd31944af440ff2b0a93facc4a7762 Mon Sep 17 00:00:00 2001 From: Yossi Kuperman Date: Tue, 23 Jan 2018 00:16:21 +0200 Subject: [PATCH 03/19] xfrm: Fix eth_hdr(skb)->h_proto to reflect inner IP version IPSec tunnel mode supports encapsulation of IPv4 over IPv6 and vice-versa. The outer IP header is stripped and the inner IP inherits the original Ethernet header. Tcpdump fails to properly decode the inner packet in case that h_proto is different than the inner IP version. Fix h_proto to reflect the inner IP version. Signed-off-by: Yossi Kuperman Signed-off-by: Steffen Klassert --- net/ipv4/xfrm4_mode_tunnel.c | 1 + net/ipv6/xfrm6_mode_tunnel.c | 1 + 2 files changed, 2 insertions(+) diff --git a/net/ipv4/xfrm4_mode_tunnel.c b/net/ipv4/xfrm4_mode_tunnel.c index e6265e2c274e..20ca486b3cad 100644 --- a/net/ipv4/xfrm4_mode_tunnel.c +++ b/net/ipv4/xfrm4_mode_tunnel.c @@ -92,6 +92,7 @@ static int xfrm4_mode_tunnel_input(struct xfrm_state *x, struct sk_buff *skb) skb_reset_network_header(skb); skb_mac_header_rebuild(skb); + eth_hdr(skb)->h_proto = skb->protocol; err = 0; diff --git a/net/ipv6/xfrm6_mode_tunnel.c b/net/ipv6/xfrm6_mode_tunnel.c index 02556e356f87..dc93002ff9d1 100644 --- a/net/ipv6/xfrm6_mode_tunnel.c +++ b/net/ipv6/xfrm6_mode_tunnel.c @@ -92,6 +92,7 @@ static int xfrm6_mode_tunnel_input(struct xfrm_state *x, struct sk_buff *skb) skb_reset_network_header(skb); skb_mac_header_rebuild(skb); + eth_hdr(skb)->h_proto = skb->protocol; err = 0; From 545d8ae7affff7fb4f8bfd327c7c7790056535c4 Mon Sep 17 00:00:00 2001 From: "Gustavo A. R. Silva" Date: Mon, 22 Jan 2018 16:34:09 -0600 Subject: [PATCH 04/19] xfrm: fix boolean assignment in xfrm_get_type_offload Assign true or false to boolean variables instead of an integer value. This issue was detected with the help of Coccinelle. Fixes: ffdb5211da1c ("xfrm: Auto-load xfrm offload modules") Signed-off-by: Gustavo A. R. Silva Signed-off-by: Steffen Klassert --- net/xfrm/xfrm_state.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/xfrm/xfrm_state.c b/net/xfrm/xfrm_state.c index 2d486492acdb..a3785f538018 100644 --- a/net/xfrm/xfrm_state.c +++ b/net/xfrm/xfrm_state.c @@ -317,7 +317,7 @@ xfrm_get_type_offload(u8 proto, unsigned short family, bool try_load) if (!type && try_load) { request_module("xfrm-offload-%d-%d", family, proto); - try_load = 0; + try_load = false; goto retry; } From 02612bb05e51df8489db5e94d0cf8d1c81f87b0c Mon Sep 17 00:00:00 2001 From: Guillaume Nault Date: Mon, 22 Jan 2018 18:06:37 +0100 Subject: [PATCH 05/19] pppoe: take ->needed_headroom of lower device into account on xmit In pppoe_sendmsg(), reserving dev->hard_header_len bytes of headroom was probably fine before the introduction of ->needed_headroom in commit f5184d267c1a ("net: Allow netdevices to specify needed head/tailroom"). But now, virtual devices typically advertise the size of their overhead in dev->needed_headroom, so we must also take it into account in skb_reserve(). Allocation size of skb is also updated to take dev->needed_tailroom into account and replace the arbitrary 32 bytes with the real size of a PPPoE header. This issue was discovered by syzbot, who connected a pppoe socket to a gre device which had dev->header_ops->create == ipgre_header and dev->hard_header_len == 0. Therefore, PPPoE didn't reserve any headroom, and dev_hard_header() crashed when ipgre_header() tried to prepend its header to skb->data. skbuff: skb_under_panic: text:000000001d390b3a len:31 put:24 head:00000000d8ed776f data:000000008150e823 tail:0x7 end:0xc0 dev:gre0 ------------[ cut here ]------------ kernel BUG at net/core/skbuff.c:104! invalid opcode: 0000 [#1] SMP KASAN Dumping ftrace buffer: (ftrace buffer empty) Modules linked in: CPU: 1 PID: 3670 Comm: syzkaller801466 Not tainted 4.15.0-rc7-next-20180115+ #97 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 RIP: 0010:skb_panic+0x162/0x1f0 net/core/skbuff.c:100 RSP: 0018:ffff8801d9bd7840 EFLAGS: 00010282 RAX: 0000000000000083 RBX: ffff8801d4f083c0 RCX: 0000000000000000 RDX: 0000000000000083 RSI: 1ffff1003b37ae92 RDI: ffffed003b37aefc RBP: ffff8801d9bd78a8 R08: 1ffff1003b37ae8a R09: 0000000000000000 R10: 0000000000000001 R11: 0000000000000000 R12: ffffffff86200de0 R13: ffffffff84a981ad R14: 0000000000000018 R15: ffff8801d2d34180 FS: 00000000019c4880(0000) GS:ffff8801db300000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00000000208bc000 CR3: 00000001d9111001 CR4: 00000000001606e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: skb_under_panic net/core/skbuff.c:114 [inline] skb_push+0xce/0xf0 net/core/skbuff.c:1714 ipgre_header+0x6d/0x4e0 net/ipv4/ip_gre.c:879 dev_hard_header include/linux/netdevice.h:2723 [inline] pppoe_sendmsg+0x58e/0x8b0 drivers/net/ppp/pppoe.c:890 sock_sendmsg_nosec net/socket.c:630 [inline] sock_sendmsg+0xca/0x110 net/socket.c:640 sock_write_iter+0x31a/0x5d0 net/socket.c:909 call_write_iter include/linux/fs.h:1775 [inline] do_iter_readv_writev+0x525/0x7f0 fs/read_write.c:653 do_iter_write+0x154/0x540 fs/read_write.c:932 vfs_writev+0x18a/0x340 fs/read_write.c:977 do_writev+0xfc/0x2a0 fs/read_write.c:1012 SYSC_writev fs/read_write.c:1085 [inline] SyS_writev+0x27/0x30 fs/read_write.c:1082 entry_SYSCALL_64_fastpath+0x29/0xa0 Admittedly PPPoE shouldn't be allowed to run on non Ethernet-like interfaces, but reserving space for ->needed_headroom is a more fundamental issue that needs to be addressed first. Same problem exists for __pppoe_xmit(), which also needs to take dev->needed_headroom into account in skb_cow_head(). Fixes: f5184d267c1a ("net: Allow netdevices to specify needed head/tailroom") Reported-by: syzbot+ed0838d0fa4c4f2b528e20286e6dc63effc7c14d@syzkaller.appspotmail.com Signed-off-by: Guillaume Nault Reviewed-by: Xin Long Signed-off-by: David S. Miller --- drivers/net/ppp/pppoe.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) diff --git a/drivers/net/ppp/pppoe.c b/drivers/net/ppp/pppoe.c index 4e1da1645b15..5aa59f41bf8c 100644 --- a/drivers/net/ppp/pppoe.c +++ b/drivers/net/ppp/pppoe.c @@ -842,6 +842,7 @@ static int pppoe_sendmsg(struct socket *sock, struct msghdr *m, struct pppoe_hdr *ph; struct net_device *dev; char *start; + int hlen; lock_sock(sk); if (sock_flag(sk, SOCK_DEAD) || !(sk->sk_state & PPPOX_CONNECTED)) { @@ -860,16 +861,16 @@ static int pppoe_sendmsg(struct socket *sock, struct msghdr *m, if (total_len > (dev->mtu + dev->hard_header_len)) goto end; - - skb = sock_wmalloc(sk, total_len + dev->hard_header_len + 32, - 0, GFP_KERNEL); + hlen = LL_RESERVED_SPACE(dev); + skb = sock_wmalloc(sk, hlen + sizeof(*ph) + total_len + + dev->needed_tailroom, 0, GFP_KERNEL); if (!skb) { error = -ENOMEM; goto end; } /* Reserve space for headers. */ - skb_reserve(skb, dev->hard_header_len); + skb_reserve(skb, hlen); skb_reset_network_header(skb); skb->dev = dev; @@ -930,7 +931,7 @@ static int __pppoe_xmit(struct sock *sk, struct sk_buff *skb) /* Copy the data if there is no space for the header or if it's * read-only. */ - if (skb_cow_head(skb, sizeof(*ph) + dev->hard_header_len)) + if (skb_cow_head(skb, LL_RESERVED_SPACE(dev) + sizeof(*ph))) goto abort; __skb_push(skb, sizeof(*ph)); From e9191ffb65d8e159680ce0ad2224e1acbde6985c Mon Sep 17 00:00:00 2001 From: Ben Hutchings Date: Mon, 22 Jan 2018 20:06:42 +0000 Subject: [PATCH 06/19] ipv6: Fix getsockopt() for sockets with default IPV6_AUTOFLOWLABEL Commit 513674b5a2c9 ("net: reevalulate autoflowlabel setting after sysctl setting") removed the initialisation of ipv6_pinfo::autoflowlabel and added a second flag to indicate whether this field or the net namespace default should be used. The getsockopt() handling for this case was not updated, so it currently returns 0 for all sockets for which IPV6_AUTOFLOWLABEL is not explicitly enabled. Fix it to return the effective value, whether that has been set at the socket or net namespace level. Fixes: 513674b5a2c9 ("net: reevalulate autoflowlabel setting after sysctl ...") Signed-off-by: Ben Hutchings Signed-off-by: David S. Miller --- include/net/ipv6.h | 1 + net/ipv6/ip6_output.c | 2 +- net/ipv6/ipv6_sockglue.c | 2 +- 3 files changed, 3 insertions(+), 2 deletions(-) diff --git a/include/net/ipv6.h b/include/net/ipv6.h index f73797e2fa60..221238254eb7 100644 --- a/include/net/ipv6.h +++ b/include/net/ipv6.h @@ -331,6 +331,7 @@ int ipv6_flowlabel_opt_get(struct sock *sk, struct in6_flowlabel_req *freq, int flags); int ip6_flowlabel_init(void); void ip6_flowlabel_cleanup(void); +bool ip6_autoflowlabel(struct net *net, const struct ipv6_pinfo *np); static inline void fl6_sock_release(struct ip6_flowlabel *fl) { diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c index 4f7d8de56611..3763dc01e374 100644 --- a/net/ipv6/ip6_output.c +++ b/net/ipv6/ip6_output.c @@ -166,7 +166,7 @@ int ip6_output(struct net *net, struct sock *sk, struct sk_buff *skb) !(IP6CB(skb)->flags & IP6SKB_REROUTED)); } -static bool ip6_autoflowlabel(struct net *net, const struct ipv6_pinfo *np) +bool ip6_autoflowlabel(struct net *net, const struct ipv6_pinfo *np) { if (!np->autoflowlabel_set) return ip6_default_np_autolabel(net); diff --git a/net/ipv6/ipv6_sockglue.c b/net/ipv6/ipv6_sockglue.c index 2d4680e0376f..e8ffb5b5d84e 100644 --- a/net/ipv6/ipv6_sockglue.c +++ b/net/ipv6/ipv6_sockglue.c @@ -1336,7 +1336,7 @@ static int do_ipv6_getsockopt(struct sock *sk, int level, int optname, break; case IPV6_AUTOFLOWLABEL: - val = np->autoflowlabel; + val = ip6_autoflowlabel(sock_net(sk), np); break; case IPV6_RECVFRAGSIZE: From 848b159835ddef99cc4193083f7e786c3992f580 Mon Sep 17 00:00:00 2001 From: Neil Horman Date: Mon, 22 Jan 2018 16:06:37 -0500 Subject: [PATCH 07/19] vmxnet3: repair memory leak with the introduction of commit b0eb57cb97e7837ebb746404c2c58c6f536f23fa, it appears that rq->buf_info is improperly handled. While it is heap allocated when an rx queue is setup, and freed when torn down, an old line of code in vmxnet3_rq_destroy was not properly removed, leading to rq->buf_info[0] being set to NULL prior to its being freed, causing a memory leak, which eventually exhausts the system on repeated create/destroy operations (for example, when the mtu of a vmxnet3 interface is changed frequently. Fix is pretty straight forward, just move the NULL set to after the free. Tested by myself with successful results Applies to net, and should likely be queued for stable, please Signed-off-by: Neil Horman Reported-By: boyang@redhat.com CC: boyang@redhat.com CC: Shrikrishna Khare CC: "VMware, Inc." CC: David S. Miller Acked-by: Shrikrishna Khare Signed-off-by: David S. Miller --- drivers/net/vmxnet3/vmxnet3_drv.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/vmxnet3/vmxnet3_drv.c b/drivers/net/vmxnet3/vmxnet3_drv.c index d1c7029ded7c..cf95290b160c 100644 --- a/drivers/net/vmxnet3/vmxnet3_drv.c +++ b/drivers/net/vmxnet3/vmxnet3_drv.c @@ -1616,7 +1616,6 @@ static void vmxnet3_rq_destroy(struct vmxnet3_rx_queue *rq, rq->rx_ring[i].basePA); rq->rx_ring[i].base = NULL; } - rq->buf_info[i] = NULL; } if (rq->data_ring.base) { @@ -1638,6 +1637,7 @@ static void vmxnet3_rq_destroy(struct vmxnet3_rx_queue *rq, (rq->rx_ring[0].size + rq->rx_ring[1].size); dma_free_coherent(&adapter->pdev->dev, sz, rq->buf_info[0], rq->buf_info_pa); + rq->buf_info[0] = rq->buf_info[1] = NULL; } } From 1ecdaea02ca6bfacf2ecda500dc1af51e9780c42 Mon Sep 17 00:00:00 2001 From: Yuval Mintz Date: Wed, 24 Jan 2018 10:02:09 +0100 Subject: [PATCH 08/19] mlxsw: spectrum_router: Don't log an error on missing neighbor Driver periodically samples all neighbors configured in device in order to update the kernel regarding their state. When finding an entry configured in HW that doesn't show in neigh_lookup() driver logs an error message. This introduces a race when removing multiple neighbors - it's possible that a given entry would still be configured in HW as its removal is still being processed but is already removed from the kernel's neighbor tables. Simply remove the error message and gracefully accept such events. Fixes: c723c735fa6b ("mlxsw: spectrum_router: Periodically update the kernel's neigh table") Fixes: 60f040ca11b9 ("mlxsw: spectrum_router: Periodically dump active IPv6 neighbours") Signed-off-by: Yuval Mintz Reviewed-by: Ido Schimmel Signed-off-by: Jiri Pirko Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c | 10 ++-------- 1 file changed, 2 insertions(+), 8 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c index 6c0391c13fe0..7042c855a5d6 100644 --- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c +++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_router.c @@ -1942,11 +1942,8 @@ static void mlxsw_sp_router_neigh_ent_ipv4_process(struct mlxsw_sp *mlxsw_sp, dipn = htonl(dip); dev = mlxsw_sp->router->rifs[rif]->dev; n = neigh_lookup(&arp_tbl, &dipn, dev); - if (!n) { - netdev_err(dev, "Failed to find matching neighbour for IP=%pI4h\n", - &dip); + if (!n) return; - } netdev_dbg(dev, "Updating neighbour with IP=%pI4h\n", &dip); neigh_event_send(n, NULL); @@ -1973,11 +1970,8 @@ static void mlxsw_sp_router_neigh_ent_ipv6_process(struct mlxsw_sp *mlxsw_sp, dev = mlxsw_sp->router->rifs[rif]->dev; n = neigh_lookup(&nd_tbl, &dip, dev); - if (!n) { - netdev_err(dev, "Failed to find matching neighbour for IP=%pI6c\n", - &dip); + if (!n) return; - } netdev_dbg(dev, "Updating neighbour with IP=%pI6c\n", &dip); neigh_event_send(n, NULL); From 560a66075d694e6ec24c60967b4d93d97cbb33d1 Mon Sep 17 00:00:00 2001 From: Wolfgang Bumiller Date: Thu, 18 Jan 2018 11:32:35 +0100 Subject: [PATCH 09/19] net: sched: em_nbyte: don't add the data offset twice 'ptr' is shifted by the offset and then validated, the memcmp should not add it a second time. Signed-off-by: Wolfgang Bumiller Signed-off-by: David S. Miller --- net/sched/em_nbyte.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/sched/em_nbyte.c b/net/sched/em_nbyte.c index df3110d69585..07c10bac06a0 100644 --- a/net/sched/em_nbyte.c +++ b/net/sched/em_nbyte.c @@ -51,7 +51,7 @@ static int em_nbyte_match(struct sk_buff *skb, struct tcf_ematch *em, if (!tcf_valid_offset(skb, ptr, nbyte->hdr.len)) return 0; - return !memcmp(ptr + nbyte->hdr.off, nbyte->pattern, nbyte->hdr.len); + return !memcmp(ptr, nbyte->pattern, nbyte->hdr.len); } static struct tcf_ematch_ops em_nbyte_ops = { From d3303a65a00c94372ddab831570647488e6c06e2 Mon Sep 17 00:00:00 2001 From: Wolfgang Bumiller Date: Thu, 18 Jan 2018 11:32:36 +0100 Subject: [PATCH 10/19] net: sched: fix TCF_LAYER_LINK case in tcf_get_base_ptr TCF_LAYER_LINK and TCF_LAYER_NETWORK returned the same pointer as skb->data points to the network header. Use skb_mac_header instead. Signed-off-by: Wolfgang Bumiller Signed-off-by: David S. Miller --- include/net/pkt_cls.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/include/net/pkt_cls.h b/include/net/pkt_cls.h index 8e08b6da72f3..753ac9361154 100644 --- a/include/net/pkt_cls.h +++ b/include/net/pkt_cls.h @@ -522,7 +522,7 @@ static inline unsigned char * tcf_get_base_ptr(struct sk_buff *skb, int layer) { switch (layer) { case TCF_LAYER_LINK: - return skb->data; + return skb_mac_header(skb); case TCF_LAYER_NETWORK: return skb_network_header(skb); case TCF_LAYER_TRANSPORT: From 581e7226a5d43f629eb6399a121f85f6a15f81be Mon Sep 17 00:00:00 2001 From: Tom Herbert Date: Wed, 24 Jan 2018 12:35:40 -0800 Subject: [PATCH 11/19] kcm: Only allow TCP sockets to be attached to a KCM mux TCP sockets for IPv4 and IPv6 that are not listeners or in closed stated are allowed to be attached to a KCM mux. Fixes: ab7ac4eb9832 ("kcm: Kernel Connection Multiplexor module") Reported-by: syzbot+8865eaff7f9acd593945@syzkaller.appspotmail.com Signed-off-by: Tom Herbert Reviewed-by: Eric Dumazet Signed-off-by: David S. Miller --- net/kcm/kcmsock.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/net/kcm/kcmsock.c b/net/kcm/kcmsock.c index d4e98f20fc2a..7632797fb68e 100644 --- a/net/kcm/kcmsock.c +++ b/net/kcm/kcmsock.c @@ -1387,8 +1387,13 @@ static int kcm_attach(struct socket *sock, struct socket *csock, if (!csk) return -EINVAL; - /* We must prevent loops or risk deadlock ! */ - if (csk->sk_family == PF_KCM) + /* Only allow TCP sockets to be attached for now */ + if ((csk->sk_family != AF_INET && csk->sk_family != AF_INET6) || + csk->sk_protocol != IPPROTO_TCP) + return -EOPNOTSUPP; + + /* Don't allow listeners or closed sockets */ + if (csk->sk_state == TCP_LISTEN || csk->sk_state == TCP_CLOSE) return -EOPNOTSUPP; psock = kmem_cache_zalloc(kcm_psockp, GFP_KERNEL); From e5571240236c5652f3e079b1d5866716a7ad819c Mon Sep 17 00:00:00 2001 From: Tom Herbert Date: Wed, 24 Jan 2018 12:35:41 -0800 Subject: [PATCH 12/19] kcm: Check if sk_user_data already set in kcm_attach This is needed to prevent sk_user_data being overwritten. The check is done under the callback lock. This should prevent a socket from being attached twice to a KCM mux. It also prevents a socket from being attached for other use cases of sk_user_data as long as the other cases set sk_user_data under the lock. Followup work is needed to unify all the use cases of sk_user_data to use the same locking. Reported-by: syzbot+114b15f2be420a8886c3@syzkaller.appspotmail.com Fixes: ab7ac4eb9832 ("kcm: Kernel Connection Multiplexor module") Signed-off-by: Tom Herbert Reviewed-by: Eric Dumazet Signed-off-by: David S. Miller --- net/kcm/kcmsock.c | 16 ++++++++++++++-- 1 file changed, 14 insertions(+), 2 deletions(-) diff --git a/net/kcm/kcmsock.c b/net/kcm/kcmsock.c index 7632797fb68e..4a8d407f8902 100644 --- a/net/kcm/kcmsock.c +++ b/net/kcm/kcmsock.c @@ -1410,9 +1410,18 @@ static int kcm_attach(struct socket *sock, struct socket *csock, return err; } - sock_hold(csk); - write_lock_bh(&csk->sk_callback_lock); + + /* Check if sk_user_data is aready by KCM or someone else. + * Must be done under lock to prevent race conditions. + */ + if (csk->sk_user_data) { + write_unlock_bh(&csk->sk_callback_lock); + strp_done(&psock->strp); + kmem_cache_free(kcm_psockp, psock); + return -EALREADY; + } + psock->save_data_ready = csk->sk_data_ready; psock->save_write_space = csk->sk_write_space; psock->save_state_change = csk->sk_state_change; @@ -1420,8 +1429,11 @@ static int kcm_attach(struct socket *sock, struct socket *csock, csk->sk_data_ready = psock_data_ready; csk->sk_write_space = psock_write_space; csk->sk_state_change = psock_state_change; + write_unlock_bh(&csk->sk_callback_lock); + sock_hold(csk); + /* Finished initialization, now add the psock to the MUX. */ spin_lock_bh(&mux->lock); head = &mux->psocks; From 4de49474b18936d62797d1dd451c6c4db1a7b119 Mon Sep 17 00:00:00 2001 From: Michal Kalderon Date: Tue, 23 Jan 2018 11:33:46 +0200 Subject: [PATCH 13/19] qed: Remove reserveration of dpi for kernel Double reservation for kernel dedicated dpi was performed. Once in the core module and once in qedr. Remove the reservation from core. Signed-off-by: Michal Kalderon Signed-off-by: Ariel Elior Signed-off-by: David S. Miller --- drivers/net/ethernet/qlogic/qed/qed_rdma.c | 3 --- 1 file changed, 3 deletions(-) diff --git a/drivers/net/ethernet/qlogic/qed/qed_rdma.c b/drivers/net/ethernet/qlogic/qed/qed_rdma.c index c8c4b3940564..9d6e2d43d4de 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_rdma.c +++ b/drivers/net/ethernet/qlogic/qed/qed_rdma.c @@ -615,9 +615,6 @@ static int qed_rdma_reserve_lkey(struct qed_hwfn *p_hwfn) { struct qed_rdma_device *dev = p_hwfn->p_rdma_info->dev; - /* The first DPI is reserved for the Kernel */ - __set_bit(0, p_hwfn->p_rdma_info->dpi_map.bitmap); - /* Tid 0 will be used as the key for "reserved MR". * The driver should allocate memory for it so it can be loaded but no * ramrod should be passed on it. From 1fe280a056dff50774bd59c3e61187cf8c0ccf10 Mon Sep 17 00:00:00 2001 From: Michal Kalderon Date: Tue, 23 Jan 2018 11:33:47 +0200 Subject: [PATCH 14/19] qed: Free reserved MR tid A tid was allocated for reserved MR during initialization but not freed. This lead to an annoying output message during rdma unload flow. Signed-off-by: Michal Kalderon Signed-off-by: Ariel Elior Signed-off-by: David S. Miller --- drivers/net/ethernet/qlogic/qed/qed_rdma.c | 28 +++++++++++++--------- 1 file changed, 17 insertions(+), 11 deletions(-) diff --git a/drivers/net/ethernet/qlogic/qed/qed_rdma.c b/drivers/net/ethernet/qlogic/qed/qed_rdma.c index 9d6e2d43d4de..b7abb8205d3a 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_rdma.c +++ b/drivers/net/ethernet/qlogic/qed/qed_rdma.c @@ -358,10 +358,27 @@ static void qed_rdma_resc_free(struct qed_hwfn *p_hwfn) kfree(p_rdma_info); } +static void qed_rdma_free_tid(void *rdma_cxt, u32 itid) +{ + struct qed_hwfn *p_hwfn = (struct qed_hwfn *)rdma_cxt; + + DP_VERBOSE(p_hwfn, QED_MSG_RDMA, "itid = %08x\n", itid); + + spin_lock_bh(&p_hwfn->p_rdma_info->lock); + qed_bmap_release_id(p_hwfn, &p_hwfn->p_rdma_info->tid_map, itid); + spin_unlock_bh(&p_hwfn->p_rdma_info->lock); +} + +static void qed_rdma_free_reserved_lkey(struct qed_hwfn *p_hwfn) +{ + qed_rdma_free_tid(p_hwfn, p_hwfn->p_rdma_info->dev->reserved_lkey); +} + static void qed_rdma_free(struct qed_hwfn *p_hwfn) { DP_VERBOSE(p_hwfn, QED_MSG_RDMA, "Freeing RDMA\n"); + qed_rdma_free_reserved_lkey(p_hwfn); qed_rdma_resc_free(p_hwfn); } @@ -794,17 +811,6 @@ static struct qed_rdma_device *qed_rdma_query_device(void *rdma_cxt) return p_hwfn->p_rdma_info->dev; } -static void qed_rdma_free_tid(void *rdma_cxt, u32 itid) -{ - struct qed_hwfn *p_hwfn = (struct qed_hwfn *)rdma_cxt; - - DP_VERBOSE(p_hwfn, QED_MSG_RDMA, "itid = %08x\n", itid); - - spin_lock_bh(&p_hwfn->p_rdma_info->lock); - qed_bmap_release_id(p_hwfn, &p_hwfn->p_rdma_info->tid_map, itid); - spin_unlock_bh(&p_hwfn->p_rdma_info->lock); -} - static void qed_rdma_cnq_prod_update(void *rdma_cxt, u8 qz_offset, u16 prod) { struct qed_hwfn *p_hwfn; From b7051cb8dadd69f85da5989017af2bb35b418950 Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Tue, 23 Jan 2018 00:08:40 -0800 Subject: [PATCH 15/19] i40e: flower: check if TC offload is enabled on a netdev Since TC block changes drivers are required to check if the TC hw offload flag is set on the interface themselves. Fixes: 2f4b411a3d67 ("i40e: Enable cloud filters via tc-flower") Fixes: 44ae12a768b7 ("net: sched: move the can_offload check from binding phase to rule insertion phase") Signed-off-by: Jakub Kicinski Reviewed-by: Simon Horman Acked-by: Jiri Pirko Acked-by: Amritha Nambiar Acked-by: Jeff Kirsher Signed-off-by: David S. Miller --- drivers/net/ethernet/intel/i40e/i40e_main.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c index 42dcaefc4c19..af792112a2d3 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_main.c +++ b/drivers/net/ethernet/intel/i40e/i40e_main.c @@ -7505,6 +7505,8 @@ static int i40e_setup_tc_cls_flower(struct i40e_netdev_priv *np, { struct i40e_vsi *vsi = np->vsi; + if (!tc_can_offload(vsi->netdev)) + return -EOPNOTSUPP; if (cls_flower->common.chain_index) return -EOPNOTSUPP; From e9cb4239134c860e5f92c75bf5321bd377bb505b Mon Sep 17 00:00:00 2001 From: Jason Wang Date: Tue, 23 Jan 2018 17:27:25 +0800 Subject: [PATCH 16/19] vhost: use mutex_lock_nested() in vhost_dev_lock_vqs() We used to call mutex_lock() in vhost_dev_lock_vqs() which tries to hold mutexes of all virtqueues. This may confuse lockdep to report a possible deadlock because of trying to hold locks belong to same class. Switch to use mutex_lock_nested() to avoid false positive. Fixes: 6b1e6cc7855b0 ("vhost: new device IOTLB API") Reported-by: syzbot+dbb7c1161485e61b0241@syzkaller.appspotmail.com Signed-off-by: Jason Wang Acked-by: Michael S. Tsirkin Signed-off-by: David S. Miller --- drivers/vhost/vhost.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/vhost/vhost.c b/drivers/vhost/vhost.c index 33ac2b186b85..549771a0cd8b 100644 --- a/drivers/vhost/vhost.c +++ b/drivers/vhost/vhost.c @@ -904,7 +904,7 @@ static void vhost_dev_lock_vqs(struct vhost_dev *d) { int i = 0; for (i = 0; i < d->nvqs; ++i) - mutex_lock(&d->vqs[i]->mutex); + mutex_lock_nested(&d->vqs[i]->mutex, i); } static void vhost_dev_unlock_vqs(struct vhost_dev *d) From 6f3180afbb22106d96a1320e175562f36a4d3506 Mon Sep 17 00:00:00 2001 From: Jason Wang Date: Tue, 23 Jan 2018 17:27:26 +0800 Subject: [PATCH 17/19] vhost: do not try to access device IOTLB when not initialized The code will try to access dev->iotlb when processing VHOST_IOTLB_INVALIDATE even if it was not initialized which may lead to NULL pointer dereference. Fixes this by check dev->iotlb before. Fixes: 6b1e6cc7855b0 ("vhost: new device IOTLB API") Signed-off-by: Jason Wang Acked-by: Michael S. Tsirkin Signed-off-by: David S. Miller --- drivers/vhost/vhost.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/drivers/vhost/vhost.c b/drivers/vhost/vhost.c index 549771a0cd8b..5727b186b3ca 100644 --- a/drivers/vhost/vhost.c +++ b/drivers/vhost/vhost.c @@ -1015,6 +1015,10 @@ static int vhost_process_iotlb_msg(struct vhost_dev *dev, vhost_iotlb_notify_vq(dev, msg); break; case VHOST_IOTLB_INVALIDATE: + if (!dev->iotlb) { + ret = -EFAULT; + break; + } vhost_vq_meta_reset(dev); vhost_del_umem_range(dev->iotlb, msg->iova, msg->iova + msg->size - 1); From 45d6e545505fd32edb812f085be7de45b6a5c0af Mon Sep 17 00:00:00 2001 From: Ivan Mikhaylov Date: Wed, 24 Jan 2018 15:53:24 +0300 Subject: [PATCH 18/19] net/ibm/emac: add 8192 rx/tx fifo size emac4syn chips has availability to use 8192 rx/tx fifo buffer sizes, in current state if we set it up in dts 8192 as example, we will get only 2048 which may impact on network speed. Signed-off-by: Ivan Mikhaylov Signed-off-by: David S. Miller --- drivers/net/ethernet/ibm/emac/core.c | 6 ++++++ drivers/net/ethernet/ibm/emac/emac.h | 2 ++ 2 files changed, 8 insertions(+) diff --git a/drivers/net/ethernet/ibm/emac/core.c b/drivers/net/ethernet/ibm/emac/core.c index 7feff2450ed6..241db3199b88 100644 --- a/drivers/net/ethernet/ibm/emac/core.c +++ b/drivers/net/ethernet/ibm/emac/core.c @@ -494,6 +494,9 @@ static u32 __emac_calc_base_mr1(struct emac_instance *dev, int tx_size, int rx_s case 16384: ret |= EMAC_MR1_RFS_16K; break; + case 8192: + ret |= EMAC4_MR1_RFS_8K; + break; case 4096: ret |= EMAC_MR1_RFS_4K; break; @@ -516,6 +519,9 @@ static u32 __emac4_calc_base_mr1(struct emac_instance *dev, int tx_size, int rx_ case 16384: ret |= EMAC4_MR1_TFS_16K; break; + case 8192: + ret |= EMAC4_MR1_TFS_8K; + break; case 4096: ret |= EMAC4_MR1_TFS_4K; break; diff --git a/drivers/net/ethernet/ibm/emac/emac.h b/drivers/net/ethernet/ibm/emac/emac.h index 5afcc27ceebb..d0a0e3b3f283 100644 --- a/drivers/net/ethernet/ibm/emac/emac.h +++ b/drivers/net/ethernet/ibm/emac/emac.h @@ -151,9 +151,11 @@ struct emac_regs { #define EMAC4_MR1_RFS_2K 0x00100000 #define EMAC4_MR1_RFS_4K 0x00180000 +#define EMAC4_MR1_RFS_8K 0x00200000 #define EMAC4_MR1_RFS_16K 0x00280000 #define EMAC4_MR1_TFS_2K 0x00020000 #define EMAC4_MR1_TFS_4K 0x00030000 +#define EMAC4_MR1_TFS_8K 0x00040000 #define EMAC4_MR1_TFS_16K 0x00050000 #define EMAC4_MR1_TR 0x00008000 #define EMAC4_MR1_MWSW_001 0x00001000 From 624ca9c33c8a853a4a589836e310d776620f4ab9 Mon Sep 17 00:00:00 2001 From: Ivan Mikhaylov Date: Wed, 24 Jan 2018 15:53:25 +0300 Subject: [PATCH 19/19] net/ibm/emac: wrong bit is used for STA control register write STA control register has areas of mode and opcodes for opeations. 18 bit is using for mode selection, where 0 is old MIO/MDIO access method and 1 is indirect access mode. 19-20 bits are using for setting up read/write operation(STA opcodes). In current state 'read' is set into old MIO/MDIO mode with 19 bit and write operation is set into 18 bit which is mode selection, not a write operation. To correlate write with read we set it into 20 bit. All those bit operations are MSB 0 based. Signed-off-by: Ivan Mikhaylov Signed-off-by: David S. Miller --- drivers/net/ethernet/ibm/emac/emac.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/ibm/emac/emac.h b/drivers/net/ethernet/ibm/emac/emac.h index d0a0e3b3f283..c26d2631ca30 100644 --- a/drivers/net/ethernet/ibm/emac/emac.h +++ b/drivers/net/ethernet/ibm/emac/emac.h @@ -244,7 +244,7 @@ struct emac_regs { #define EMAC_STACR_PHYE 0x00004000 #define EMAC_STACR_STAC_MASK 0x00003000 #define EMAC_STACR_STAC_READ 0x00001000 -#define EMAC_STACR_STAC_WRITE 0x00002000 +#define EMAC_STACR_STAC_WRITE 0x00000800 #define EMAC_STACR_OPBC_MASK 0x00000C00 #define EMAC_STACR_OPBC_50 0x00000000 #define EMAC_STACR_OPBC_66 0x00000400