ipv6: Add rate limit mask for ICMPv6 messages

To make ICMPv6 closer to ICMPv4, add ratemask parameter. Since the ICMP
message types use larger numeric values, a simple bitmask doesn't fit.
I use large bitmap. The input and output are the in form of list of
ranges. Set the default to rate limit all error messages but Packet Too
Big. For Packet Too Big, use ratemask instead of hard-coded.

There are functions where icmpv6_xrlim_allow() and icmpv6_global_allow()
aren't called. This patch only adds them to icmpv6_echo_reply().

Rate limiting error messages is mandated by RFC 4443 but RFC 4890 says
that it is also acceptable to rate limit informational messages. Thus,
I removed the current hard-coded behavior of icmpv6_mask_allow() that
doesn't rate limit informational messages.

v2: Add dummy function proc_do_large_bitmap() if CONFIG_PROC_SYSCTL
    isn't defined, expand the description in ip-sysctl.txt and remove
    unnecessary conditional before kfree().
v3: Inline the bitmap instead of dynamically allocated. Still is a
    pointer to it is needed because of the way proc_do_large_bitmap work.

Signed-off-by: Stephen Suryaputra <ssuryaextr@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Documentation/networking/ip-sysctl.txt

@@ -1913,11 +1913,26 @@ enhanced_dad - BOOLEAN
 
 icmp/*:
 ratelimit - INTEGER
-	Limit the maximal rates for sending ICMPv6 packets.
+	Limit the maximal rates for sending ICMPv6 messages.
 	0 to disable any limiting,
 	otherwise the minimal space between responses in milliseconds.
 	Default: 1000
 
+ratemask - list of comma separated ranges
+	For ICMPv6 message types matching the ranges in the ratemask, limit
+	the sending of the message according to ratelimit parameter.
+
+	The format used for both input and output is a comma separated
+	list of ranges (e.g. "0-127,129" for ICMPv6 message type 0 to 127 and
+	129). Writing to the file will clear all previous ranges of ICMPv6
+	message types and update the current list with the input.
+
+	Refer to: https://www.iana.org/assignments/icmpv6-parameters/icmpv6-parameters.xhtml
+	for numerical values of ICMPv6 message types, e.g. echo request is 128
+	and echo reply is 129.
+
+	Default: 0-1,3-127 (rate limit ICMPv6 errors except Packet Too Big)
+
 echo_ignore_all - BOOLEAN
 	If set non-zero, then the kernel will ignore all ICMP ECHO
 	requests sent to it over the IPv6 protocol.

include/net/netns/ipv6.h

@@ -8,6 +8,7 @@
 #ifndef __NETNS_IPV6_H__
 #define __NETNS_IPV6_H__
 #include <net/dst_ops.h>
+#include <uapi/linux/icmpv6.h>
 
 struct ctl_table_header;
 
@@ -35,6 +36,8 @@ struct netns_sysctl_ipv6 {
 	int icmpv6_echo_ignore_all;
 	int icmpv6_echo_ignore_multicast;
 	int icmpv6_echo_ignore_anycast;
+	DECLARE_BITMAP(icmpv6_ratemask, ICMPV6_MSG_MAX + 1);
+	unsigned long *icmpv6_ratemask_ptr;
 	int anycast_src_echo_reply;
 	int ip_nonlocal_bind;
 	int fwmark_reflect;

include/uapi/linux/icmpv6.h

@@ -90,6 +90,8 @@ struct icmp6hdr {
 #define ICMPV6_TIME_EXCEED		3
 #define ICMPV6_PARAMPROB		4
 
+#define ICMPV6_ERRMSG_MAX       127
+
 #define ICMPV6_INFOMSG_MASK		0x80
 
 #define ICMPV6_ECHO_REQUEST		128
@@ -110,6 +112,8 @@ struct icmp6hdr {
 
 #define ICMPV6_MRDISC_ADV		151
 
+#define ICMPV6_MSG_MAX          255
+
 /*
  *	Codes for Destination Unreachable
  */

kernel/sysctl.c

@@ -3326,6 +3326,11 @@ int proc_doulongvec_ms_jiffies_minmax(struct ctl_table *table, int write,
     return -ENOSYS;
 }
 
+int proc_do_large_bitmap(struct ctl_table *table, int write,
+			 void __user *buffer, size_t *lenp, loff_t *ppos)
+{
+	return -ENOSYS;
+}
 
 #endif /* CONFIG_PROC_SYSCTL */
 
@@ -3366,3 +3371,4 @@ EXPORT_SYMBOL(proc_dointvec_ms_jiffies);
 EXPORT_SYMBOL(proc_dostring);
 EXPORT_SYMBOL(proc_doulongvec_minmax);
 EXPORT_SYMBOL(proc_doulongvec_ms_jiffies_minmax);
+EXPORT_SYMBOL(proc_do_large_bitmap);

net/ipv6/af_inet6.c

@@ -850,6 +850,15 @@ static int __net_init inet6_net_init(struct net *net)
 	net->ipv6.sysctl.icmpv6_echo_ignore_all = 0;
 	net->ipv6.sysctl.icmpv6_echo_ignore_multicast = 0;
 	net->ipv6.sysctl.icmpv6_echo_ignore_anycast = 0;
+
+	/* By default, rate limit error messages.
+	 * Except for pmtu discovery, it would break it.
+	 * proc_do_large_bitmap needs pointer to the bitmap.
+	 */
+	bitmap_set(net->ipv6.sysctl.icmpv6_ratemask, 0, ICMPV6_ERRMSG_MAX + 1);
+	bitmap_clear(net->ipv6.sysctl.icmpv6_ratemask, ICMPV6_PKT_TOOBIG, 1);
+	net->ipv6.sysctl.icmpv6_ratemask_ptr = net->ipv6.sysctl.icmpv6_ratemask;
+
 	net->ipv6.sysctl.flowlabel_consistency = 1;
 	net->ipv6.sysctl.auto_flowlabels = IP6_DEFAULT_AUTO_FLOW_LABELS;
 	net->ipv6.sysctl.idgen_retries = 3;

net/ipv6/icmp.c

@@ -168,22 +168,21 @@ static bool is_ineligible(const struct sk_buff *skb)
 	return false;
 }
 
-static bool icmpv6_mask_allow(int type)
+static bool icmpv6_mask_allow(struct net *net, int type)
 {
-	/* Informational messages are not limited. */
-	if (type & ICMPV6_INFOMSG_MASK)
+	if (type > ICMPV6_MSG_MAX)
 		return true;
 
-	/* Do not limit pmtu discovery, it would break it. */
-	if (type == ICMPV6_PKT_TOOBIG)
+	/* Limit if icmp type is set in ratemask. */
+	if (!test_bit(type, net->ipv6.sysctl.icmpv6_ratemask))
 		return true;
 
 	return false;
 }
 
-static bool icmpv6_global_allow(int type)
+static bool icmpv6_global_allow(struct net *net, int type)
 {
-	if (icmpv6_mask_allow(type))
+	if (icmpv6_mask_allow(net, type))
 		return true;
 
 	if (icmp_global_allow())
@@ -202,7 +201,7 @@ static bool icmpv6_xrlim_allow(struct sock *sk, u8 type,
 	struct dst_entry *dst;
 	bool res = false;
 
-	if (icmpv6_mask_allow(type))
+	if (icmpv6_mask_allow(net, type))
 		return true;
 
 	/*
@@ -511,7 +510,7 @@ static void icmp6_send(struct sk_buff *skb, u8 type, u8 code, __u32 info,
 	local_bh_disable();
 
 	/* Check global sysctl_icmp_msgs_per_sec ratelimit */
-	if (!(skb->dev->flags&IFF_LOOPBACK) && !icmpv6_global_allow(type))
+	if (!(skb->dev->flags & IFF_LOOPBACK) && !icmpv6_global_allow(net, type))
 		goto out_bh_enable;
 
 	mip6_addr_swap(skb);
@@ -731,6 +730,11 @@ static void icmpv6_echo_reply(struct sk_buff *skb)
 	if (IS_ERR(dst))
 		goto out;
 
+	/* Check the ratelimit */
+	if ((!(skb->dev->flags & IFF_LOOPBACK) && !icmpv6_global_allow(net, ICMPV6_ECHO_REPLY)) ||
+	    !icmpv6_xrlim_allow(sk, ICMPV6_ECHO_REPLY, &fl6))
+		goto out_dst_release;
+
 	idev = __in6_dev_get(skb->dev);
 
 	msg.skb = skb;
@@ -751,6 +755,7 @@ static void icmpv6_echo_reply(struct sk_buff *skb)
 		icmpv6_push_pending_frames(sk, &fl6, &tmp_hdr,
 					   skb->len + sizeof(struct icmp6hdr));
 	}
+out_dst_release:
 	dst_release(dst);
 out:
 	icmpv6_xmit_unlock(sk);
@@ -1137,6 +1142,13 @@ static struct ctl_table ipv6_icmp_table_template[] = {
 		.mode		= 0644,
 		.proc_handler = proc_dointvec,
 	},
+	{
+		.procname	= "ratemask",
+		.data		= &init_net.ipv6.sysctl.icmpv6_ratemask_ptr,
+		.maxlen		= ICMPV6_MSG_MAX + 1,
+		.mode		= 0644,
+		.proc_handler = proc_do_large_bitmap,
+	},
 	{ },
 };
 
@@ -1153,6 +1165,7 @@ struct ctl_table * __net_init ipv6_icmp_sysctl_init(struct net *net)
 		table[1].data = &net->ipv6.sysctl.icmpv6_echo_ignore_all;
 		table[2].data = &net->ipv6.sysctl.icmpv6_echo_ignore_multicast;
 		table[3].data = &net->ipv6.sysctl.icmpv6_echo_ignore_anycast;
+		table[4].data = &net->ipv6.sysctl.icmpv6_ratemask_ptr;
 	}
 	return table;
 }