net: tunnels: annotate lockless accesses to dev->needed_headroom

IP tunnels can apparently update dev->needed_headroom
in their xmit path.

This patch takes care of three tunnels xmit, and also the
core LL_RESERVED_SPACE() and LL_RESERVED_SPACE_EXTRA()
helpers.

More changes might be needed for completeness.

BUG: KCSAN: data-race in ip_tunnel_xmit / ip_tunnel_xmit

read to 0xffff88815b9da0ec of 2 bytes by task 888 on cpu 1:
ip_tunnel_xmit+0x1270/0x1730 net/ipv4/ip_tunnel.c:803
__gre_xmit net/ipv4/ip_gre.c:469 [inline]
ipgre_xmit+0x516/0x570 net/ipv4/ip_gre.c:661
__netdev_start_xmit include/linux/netdevice.h:4881 [inline]
netdev_start_xmit include/linux/netdevice.h:4895 [inline]
xmit_one net/core/dev.c:3580 [inline]
dev_hard_start_xmit+0x127/0x400 net/core/dev.c:3596
__dev_queue_xmit+0x1007/0x1eb0 net/core/dev.c:4246
dev_queue_xmit include/linux/netdevice.h:3051 [inline]
neigh_direct_output+0x17/0x20 net/core/neighbour.c:1623
neigh_output include/net/neighbour.h:546 [inline]
ip_finish_output2+0x740/0x840 net/ipv4/ip_output.c:228
ip_finish_output+0xf4/0x240 net/ipv4/ip_output.c:316
NF_HOOK_COND include/linux/netfilter.h:291 [inline]
ip_output+0xe5/0x1b0 net/ipv4/ip_output.c:430
dst_output include/net/dst.h:444 [inline]
ip_local_out+0x64/0x80 net/ipv4/ip_output.c:126
iptunnel_xmit+0x34a/0x4b0 net/ipv4/ip_tunnel_core.c:82
ip_tunnel_xmit+0x1451/0x1730 net/ipv4/ip_tunnel.c:813
__gre_xmit net/ipv4/ip_gre.c:469 [inline]
ipgre_xmit+0x516/0x570 net/ipv4/ip_gre.c:661
__netdev_start_xmit include/linux/netdevice.h:4881 [inline]
netdev_start_xmit include/linux/netdevice.h:4895 [inline]
xmit_one net/core/dev.c:3580 [inline]
dev_hard_start_xmit+0x127/0x400 net/core/dev.c:3596
__dev_queue_xmit+0x1007/0x1eb0 net/core/dev.c:4246
dev_queue_xmit include/linux/netdevice.h:3051 [inline]
neigh_direct_output+0x17/0x20 net/core/neighbour.c:1623
neigh_output include/net/neighbour.h:546 [inline]
ip_finish_output2+0x740/0x840 net/ipv4/ip_output.c:228
ip_finish_output+0xf4/0x240 net/ipv4/ip_output.c:316
NF_HOOK_COND include/linux/netfilter.h:291 [inline]
ip_output+0xe5/0x1b0 net/ipv4/ip_output.c:430
dst_output include/net/dst.h:444 [inline]
ip_local_out+0x64/0x80 net/ipv4/ip_output.c:126
iptunnel_xmit+0x34a/0x4b0 net/ipv4/ip_tunnel_core.c:82
ip_tunnel_xmit+0x1451/0x1730 net/ipv4/ip_tunnel.c:813
__gre_xmit net/ipv4/ip_gre.c:469 [inline]
ipgre_xmit+0x516/0x570 net/ipv4/ip_gre.c:661
__netdev_start_xmit include/linux/netdevice.h:4881 [inline]
netdev_start_xmit include/linux/netdevice.h:4895 [inline]
xmit_one net/core/dev.c:3580 [inline]
dev_hard_start_xmit+0x127/0x400 net/core/dev.c:3596
__dev_queue_xmit+0x1007/0x1eb0 net/core/dev.c:4246
dev_queue_xmit include/linux/netdevice.h:3051 [inline]
neigh_direct_output+0x17/0x20 net/core/neighbour.c:1623
neigh_output include/net/neighbour.h:546 [inline]
ip_finish_output2+0x740/0x840 net/ipv4/ip_output.c:228
ip_finish_output+0xf4/0x240 net/ipv4/ip_output.c:316
NF_HOOK_COND include/linux/netfilter.h:291 [inline]
ip_output+0xe5/0x1b0 net/ipv4/ip_output.c:430
dst_output include/net/dst.h:444 [inline]
ip_local_out+0x64/0x80 net/ipv4/ip_output.c:126
iptunnel_xmit+0x34a/0x4b0 net/ipv4/ip_tunnel_core.c:82
ip_tunnel_xmit+0x1451/0x1730 net/ipv4/ip_tunnel.c:813
__gre_xmit net/ipv4/ip_gre.c:469 [inline]
ipgre_xmit+0x516/0x570 net/ipv4/ip_gre.c:661
__netdev_start_xmit include/linux/netdevice.h:4881 [inline]
netdev_start_xmit include/linux/netdevice.h:4895 [inline]
xmit_one net/core/dev.c:3580 [inline]
dev_hard_start_xmit+0x127/0x400 net/core/dev.c:3596
__dev_queue_xmit+0x1007/0x1eb0 net/core/dev.c:4246
dev_queue_xmit include/linux/netdevice.h:3051 [inline]
neigh_direct_output+0x17/0x20 net/core/neighbour.c:1623
neigh_output include/net/neighbour.h:546 [inline]
ip_finish_output2+0x740/0x840 net/ipv4/ip_output.c:228
ip_finish_output+0xf4/0x240 net/ipv4/ip_output.c:316
NF_HOOK_COND include/linux/netfilter.h:291 [inline]
ip_output+0xe5/0x1b0 net/ipv4/ip_output.c:430
dst_output include/net/dst.h:444 [inline]
ip_local_out+0x64/0x80 net/ipv4/ip_output.c:126
iptunnel_xmit+0x34a/0x4b0 net/ipv4/ip_tunnel_core.c:82
ip_tunnel_xmit+0x1451/0x1730 net/ipv4/ip_tunnel.c:813
__gre_xmit net/ipv4/ip_gre.c:469 [inline]
ipgre_xmit+0x516/0x570 net/ipv4/ip_gre.c:661
__netdev_start_xmit include/linux/netdevice.h:4881 [inline]
netdev_start_xmit include/linux/netdevice.h:4895 [inline]
xmit_one net/core/dev.c:3580 [inline]
dev_hard_start_xmit+0x127/0x400 net/core/dev.c:3596
__dev_queue_xmit+0x1007/0x1eb0 net/core/dev.c:4246
dev_queue_xmit include/linux/netdevice.h:3051 [inline]
neigh_direct_output+0x17/0x20 net/core/neighbour.c:1623
neigh_output include/net/neighbour.h:546 [inline]
ip_finish_output2+0x740/0x840 net/ipv4/ip_output.c:228
ip_finish_output+0xf4/0x240 net/ipv4/ip_output.c:316
NF_HOOK_COND include/linux/netfilter.h:291 [inline]
ip_output+0xe5/0x1b0 net/ipv4/ip_output.c:430
dst_output include/net/dst.h:444 [inline]
ip_local_out+0x64/0x80 net/ipv4/ip_output.c:126
iptunnel_xmit+0x34a/0x4b0 net/ipv4/ip_tunnel_core.c:82
ip_tunnel_xmit+0x1451/0x1730 net/ipv4/ip_tunnel.c:813
__gre_xmit net/ipv4/ip_gre.c:469 [inline]
ipgre_xmit+0x516/0x570 net/ipv4/ip_gre.c:661
__netdev_start_xmit include/linux/netdevice.h:4881 [inline]
netdev_start_xmit include/linux/netdevice.h:4895 [inline]
xmit_one net/core/dev.c:3580 [inline]
dev_hard_start_xmit+0x127/0x400 net/core/dev.c:3596
__dev_queue_xmit+0x1007/0x1eb0 net/core/dev.c:4246
dev_queue_xmit include/linux/netdevice.h:3051 [inline]
neigh_direct_output+0x17/0x20 net/core/neighbour.c:1623
neigh_output include/net/neighbour.h:546 [inline]
ip_finish_output2+0x740/0x840 net/ipv4/ip_output.c:228
ip_finish_output+0xf4/0x240 net/ipv4/ip_output.c:316
NF_HOOK_COND include/linux/netfilter.h:291 [inline]
ip_output+0xe5/0x1b0 net/ipv4/ip_output.c:430
dst_output include/net/dst.h:444 [inline]
ip_local_out+0x64/0x80 net/ipv4/ip_output.c:126
iptunnel_xmit+0x34a/0x4b0 net/ipv4/ip_tunnel_core.c:82
ip_tunnel_xmit+0x1451/0x1730 net/ipv4/ip_tunnel.c:813
__gre_xmit net/ipv4/ip_gre.c:469 [inline]
ipgre_xmit+0x516/0x570 net/ipv4/ip_gre.c:661
__netdev_start_xmit include/linux/netdevice.h:4881 [inline]
netdev_start_xmit include/linux/netdevice.h:4895 [inline]
xmit_one net/core/dev.c:3580 [inline]
dev_hard_start_xmit+0x127/0x400 net/core/dev.c:3596
__dev_queue_xmit+0x1007/0x1eb0 net/core/dev.c:4246

write to 0xffff88815b9da0ec of 2 bytes by task 2379 on cpu 0:
ip_tunnel_xmit+0x1294/0x1730 net/ipv4/ip_tunnel.c:804
__gre_xmit net/ipv4/ip_gre.c:469 [inline]
ipgre_xmit+0x516/0x570 net/ipv4/ip_gre.c:661
__netdev_start_xmit include/linux/netdevice.h:4881 [inline]
netdev_start_xmit include/linux/netdevice.h:4895 [inline]
xmit_one net/core/dev.c:3580 [inline]
dev_hard_start_xmit+0x127/0x400 net/core/dev.c:3596
__dev_queue_xmit+0x1007/0x1eb0 net/core/dev.c:4246
dev_queue_xmit include/linux/netdevice.h:3051 [inline]
neigh_direct_output+0x17/0x20 net/core/neighbour.c:1623
neigh_output include/net/neighbour.h:546 [inline]
ip6_finish_output2+0x9bc/0xc50 net/ipv6/ip6_output.c:134
__ip6_finish_output net/ipv6/ip6_output.c:195 [inline]
ip6_finish_output+0x39a/0x4e0 net/ipv6/ip6_output.c:206
NF_HOOK_COND include/linux/netfilter.h:291 [inline]
ip6_output+0xeb/0x220 net/ipv6/ip6_output.c:227
dst_output include/net/dst.h:444 [inline]
NF_HOOK include/linux/netfilter.h:302 [inline]
mld_sendpack+0x438/0x6a0 net/ipv6/mcast.c:1820
mld_send_cr net/ipv6/mcast.c:2121 [inline]
mld_ifc_work+0x519/0x7b0 net/ipv6/mcast.c:2653
process_one_work+0x3e6/0x750 kernel/workqueue.c:2390
worker_thread+0x5f2/0xa10 kernel/workqueue.c:2537
kthread+0x1ac/0x1e0 kernel/kthread.c:376
ret_from_fork+0x1f/0x30 arch/x86/entry/entry_64.S:308

value changed: 0x0dd4 -> 0x0e14

Reported by Kernel Concurrency Sanitizer on:
CPU: 0 PID: 2379 Comm: kworker/0:0 Not tainted 6.3.0-rc1-syzkaller-00002-g8ca09d5fa354-dirty #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 03/02/2023
Workqueue: mld mld_ifc_work

Fixes: 8eb30be ("ipv6: Create ip6_tnl_xmit")
Reported-by: syzbot <[email protected]>
Signed-off-by: Eric Dumazet <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>

Author: edumazet

include/linux/netdevice.h

@@ -297,9 +297,11 @@ struct hh_cache {
  * relationship HH alignment <= LL alignment.
  */
 #define LL_RESERVED_SPACE(dev) \
-	((((dev)->hard_header_len+(dev)->needed_headroom)&~(HH_DATA_MOD - 1)) + HH_DATA_MOD)
+	((((dev)->hard_header_len + READ_ONCE((dev)->needed_headroom)) \
+	  & ~(HH_DATA_MOD - 1)) + HH_DATA_MOD)
 #define LL_RESERVED_SPACE_EXTRA(dev,extra) \
-	((((dev)->hard_header_len+(dev)->needed_headroom+(extra))&~(HH_DATA_MOD - 1)) + HH_DATA_MOD)
+	((((dev)->hard_header_len + READ_ONCE((dev)->needed_headroom) + (extra)) \
+	  & ~(HH_DATA_MOD - 1)) + HH_DATA_MOD)
 
 struct header_ops {
 	int	(*create) (struct sk_buff *skb, struct net_device *dev,

net/ipv4/ip_tunnel.c

@@ -614,10 +614,10 @@ void ip_md_tunnel_xmit(struct sk_buff *skb, struct net_device *dev,
 	}
 
 	headroom += LL_RESERVED_SPACE(rt->dst.dev) + rt->dst.header_len;
-	if (headroom > dev->needed_headroom)
-		dev->needed_headroom = headroom;
+	if (headroom > READ_ONCE(dev->needed_headroom))
+		WRITE_ONCE(dev->needed_headroom, headroom);
 
-	if (skb_cow_head(skb, dev->needed_headroom)) {
+	if (skb_cow_head(skb, READ_ONCE(dev->needed_headroom))) {
 		ip_rt_put(rt);
 		goto tx_dropped;
 	}
@@ -800,10 +800,10 @@ void ip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev,
 
 	max_headroom = LL_RESERVED_SPACE(rt->dst.dev) + sizeof(struct iphdr)
 			+ rt->dst.header_len + ip_encap_hlen(&tunnel->encap);
-	if (max_headroom > dev->needed_headroom)
-		dev->needed_headroom = max_headroom;
+	if (max_headroom > READ_ONCE(dev->needed_headroom))
+		WRITE_ONCE(dev->needed_headroom, max_headroom);
 
-	if (skb_cow_head(skb, dev->needed_headroom)) {
+	if (skb_cow_head(skb, READ_ONCE(dev->needed_headroom))) {
 		ip_rt_put(rt);
 		DEV_STATS_INC(dev, tx_dropped);
 		kfree_skb(skb);

net/ipv6/ip6_tunnel.c

@@ -1240,8 +1240,8 @@ int ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev, __u8 dsfield,
 	 */
 	max_headroom = LL_RESERVED_SPACE(dst->dev) + sizeof(struct ipv6hdr)
 			+ dst->header_len + t->hlen;
-	if (max_headroom > dev->needed_headroom)
-		dev->needed_headroom = max_headroom;
+	if (max_headroom > READ_ONCE(dev->needed_headroom))
+		WRITE_ONCE(dev->needed_headroom, max_headroom);
 
 	err = ip6_tnl_encap(skb, t, &proto, fl6);
 	if (err)