Skip to contentSkip to conversation

nftables: add support for wildcard interfaces

PATCH In review Florian Westphal 2 participants 5 messages Apr 29, 2022 – May 2, 2022 netfilter

lore
Share & tools
lore

Florian Westphal

author maintainer
#0 lore raw

cover nftables: add support for wildcard interfaces

First patch is needed so kernel doesn't think end is before start, second patch allows to dump "x .foo*" correctly, without this nft tries to represent the start/end "name" with a range. This doesn't work well because end range uses \ff padding.

Patch 3 adds tests.

Florian Westphal (3):
netlink: swap byteorder for host-endian concat data
segtree: add pretty-print support for wildcard strings in concatenated
sets
sets_with_ifnames: add test case for concatenated range

 src/netlink.c                                 |  4 +
 src/segtree.c                                 | 38 +++++++-
 .../sets/dumps/sets_with_ifnames.nft          | 25 ++++-
 tests/shell/testcases/sets/sets_with_ifnames  | 93 ++++++++++++++-----
 4 files changed, 134 insertions(+), 26 deletions(-)
Signature · 1 line
2.35.1
author maintainer
#1 lore raw

PATCH 1/3 netlink: swap byteorder for host-endian concat data

All data must be passed in network byte order, else matching won't work respectively kernel will reject the interval because it thinks that start is after end

This is needed to allow use of 'ppp*' in interval sets with concatenations.

---
 src/netlink.c | 4 ++++
 1 file changed, 4 insertions(+)
src/netlink.c +4 -0
index 240c937e3ac5..89d864ed046a 100644
@@ -259,6 +259,10 @@ static int netlink_gen_concat_data_expr(int end, const struct expr *i,
 			mpz_t v;
 
 			mpz_init_bitmask(v, i->len - i->prefix_len);
+
+			if (i->byteorder == BYTEORDER_HOST_ENDIAN)
+				mpz_switch_byteorder(v, i->len / BITS_PER_BYTE);
+
 			mpz_add(v, i->prefix->value, v);
 			count = netlink_export_pad(data, v, i);
 			mpz_clear(v);
Signature · 1 line
2.35.1
author maintainer
#2 lore raw

PATCH 2/3 segtree: add pretty-print support for wildcard strings in concatenated sets

In reply to Florian Westphal

For concat ranges, something like 'ppp*' is translated as a range from 'ppp\0\0\0...' to 'ppp\ff\ff\ff...'.

In order to display this properly, check for presence of string base type and convert to symbolic expression, with appended '*' character.

---
 src/segtree.c | 38 ++++++++++++++++++++++++++++++++++++--
 1 file changed, 36 insertions(+), 2 deletions(-)
src/segtree.c +36 -2
index 4602becc10e6..f9cac373a5f0 100644
@@ -345,6 +345,8 @@ void concat_range_aggregate(struct expr *set)
 		list_for_each_entry_safe(r1, r1_next,
 					 &expr_value(start)->expressions,
 					 list) {
+			bool string_type = false;
+
 			mpz_init(range);
 			mpz_init(p);
 
@@ -356,16 +358,48 @@ void concat_range_aggregate(struct expr *set)
 				goto next;
 			}
 
+			if (expr_basetype(r1)->type == TYPE_STRING &&
+			    expr_basetype(r2)->type == TYPE_STRING) {
+				string_type = true;
+				mpz_switch_byteorder(r1->value, r1->len / BITS_PER_BYTE);
+				mpz_switch_byteorder(r2->value, r2->len / BITS_PER_BYTE);
+			}
+
 			mpz_sub(range, r2->value, r1->value);
 			mpz_sub_ui(range, range, 1);
 			mpz_and(p, r1->value, range);
 
 			/* Check if we are forced, or if it's anyway preferable,
-			 * to express the range as two points instead of a
-			 * netmask.
+			 * to express the range as a wildcard string, or two points
+			 * instead of a netmask.
 			 */
 			prefix_len = range_mask_len(r1->value, r2->value,
 						    r1->len);
+			if (string_type) {
+				mpz_switch_byteorder(r1->value, r1->len / BITS_PER_BYTE);
+				mpz_switch_byteorder(r2->value, r2->len / BITS_PER_BYTE);
+			}
+
+			if (prefix_len >= 0 &&
+			    (prefix_len % BITS_PER_BYTE) == 0 &&
+			    string_type) {
+				unsigned int str_len = prefix_len / BITS_PER_BYTE;
+				char data[str_len + 2];
+
+				mpz_export_data(data, r1->value, BYTEORDER_HOST_ENDIAN, str_len);
+				data[str_len] = '*';
+
+				tmp = constant_expr_alloc(&r1->location, r1->dtype,
+							  BYTEORDER_HOST_ENDIAN,
+							  (str_len + 1) * BITS_PER_BYTE, data);
+				tmp->len = r2->len;
+				list_replace(&r2->list, &tmp->list);
+				r2_next = tmp->list.next;
+				expr_free(r2);
+				free_r1 = 1;
+				goto next;
+			}
+
 			if (prefix_len < 0 ||
 			    !(r1->dtype->flags & DTYPE_F_PREFIX)) {
 				tmp = range_expr_alloc(&r1->location, r1,
Signature · 1 line
2.35.1
author maintainer
#3 lore raw

PATCH 3/3 sets_with_ifnames: add test case for concatenated range

In reply to Florian Westphal

Refactor existing test case for simple interace name ranges (without concatenations) to also cover "addr . ifname".

---
 .../sets/dumps/sets_with_ifnames.nft          | 25 ++++-
 tests/shell/testcases/sets/sets_with_ifnames  | 93 ++++++++++++++-----
 2 files changed, 94 insertions(+), 24 deletions(-)
tests/shell/testcases/sets/dumps/sets_with_ifnames.nft +24 -1
index 12c1aa960a66..6b073ae2d090 100644
@@ -14,6 +14,21 @@ table inet testifsets {
 			     "ppp0" }
 	}
 
+	set concat {
+		type ipv4_addr . ifname
+		elements = { 10.1.2.2 . "abcdef0",
+			     10.1.2.2 . "abcdef1" }
+	}
+
+	set concat_wild {
+		type ipv4_addr . ifname
+		flags interval
+		elements = { 10.1.2.2 . "abcdef*",
+			     10.1.2.1 . "bar",
+			     1.1.2.0/24 . "abcdef0",
+			     12.2.2.0/24 . "abcdef*" }
+	}
+
 	chain v4icmp {
 		iifname @simple counter packets 0 bytes 0
 		iifname @simple_wild counter packets 0 bytes 0
@@ -21,8 +36,16 @@ table inet testifsets {
 		iifname { "abcdef*", "eth0" } counter packets 0 bytes 0
 	}
 
+	chain v4icmpc {
+		ip saddr . iifname @concat counter packets 0 bytes 0
+		ip saddr . iifname @concat_wild counter packets 0 bytes 0
+		ip saddr . iifname { 10.1.2.2 . "abcdef0" } counter packets 0 bytes 0
+		ip saddr . iifname { 10.1.2.2 . "abcdef*" } counter packets 0 bytes 0
+	}
+
 	chain input {
 		type filter hook input priority filter; policy accept;
-		ip protocol icmp goto v4icmp
+		ip protocol icmp jump v4icmp
+		ip protocol icmp goto v4icmpc
 	}
 }
tests/shell/testcases/sets/sets_with_ifnames +70 -23

124 lines not on this page · raw message

Signature · 1 line
2.35.1
#4 lore raw
In reply to Florian Westphal
Quoted 6 lines · Florian Westphal

First patch is needed so kernel doesn't think end is before start, second patch allows to dump "x .foo*" correctly, without this nft tries to represent the start/end "name" with a range. This doesn't work well because end range uses \ff padding.

Patch 3 adds tests.

LGTM, thanks