Skip to contentSkip to conversation

ip: Extend filter links/addresses

PATCH In review Anton Danilov 2 participants 4 messages Jan 8, 2022 – Jan 12, 2022

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 ip: Extend filter links/addresses

This patch improves the filtering of links/addresses with the next features:
1. Additional types: ether, loopback, ppp
2. Exclude of specific interface type with 'exclude_type' option

Examples:
ip link show type ether
ip address show exclude_type ppp

---
 ip/ip_common.h           |  1 +
 ip/ipaddress.c           | 52 ++++++++++++++++++++++++++++++++++++----
 ip/iplink.c              |  2 +-
 man/man8/ip-address.8.in | 16 ++++++++++++-
 man/man8/ip-link.8.in    | 13 +++++++---
 5 files changed, 75 insertions(+), 9 deletions(-)
ip/ip_common.h +1 -0
index ea04c8ff..38377be4 100644
@@ -26,6 +26,7 @@ struct link_filter {
 	int master;
 	char *kind;
 	char *slave_kind;
+	char *exclude_kind;
 	int target_nsid;
 };
 
ip/ipaddress.c +48 -4

82 lines not on this page · raw message

ip/iplink.c +1 -1
index a3ea775d..e0d49cab 100644
@@ -120,7 +120,7 @@ void iplink_usage(void)
 		"		[ gso_max_size BYTES ] | [ gso_max_segs PACKETS ]\n"
 		"\n"
 		"	ip link show [ DEVICE | group GROUP ] [up] [master DEV] [vrf NAME] [type TYPE]\n"
-		"		[nomaster]\n"
+		"		[exclude_type TYPE] [nomaster]\n"
 		"\n"
 		"	ip link xstats type TYPE [ ARGS ]\n"
 		"\n"
man/man8/ip-address.8.in +15 -1
index 65f67e06..21de3d77 100644
@@ -45,6 +45,8 @@ ip-address \- protocol address management
 .IR PATTERN " ] [ "
 .B  master
 .IR DEVICE " ] [ "
+.B  exclude_type
+.IR TYPE " ] [ "
 .B  type
 .IR TYPE " ] [ "
 .B vrf
@@ -138,7 +140,10 @@ ip-address \- protocol address management
 .BR ipvlan " |"
 .BR lowpan " |"
 .BR geneve " |"
-.BR macsec " ]"
+.BR macsec " |"
+.BR ether " |"
+.BR loopback " |"
+.BR ppp " ]"
 
 .SH "DESCRIPTION"
 The
@@ -337,6 +342,10 @@ interface list by comparing it with the relevant attribute in case the kernel
 didn't filter already. Therefore any string is accepted, but may lead to empty
 output.
 
+.TP
+.BI exclude_type " TYPE"
+don't list linterfaces of the given type.
+
 .TP
 .B up
 only list running interfaces.
@@ -441,6 +450,11 @@ Same as above except that only addresses assigned to active network interfaces
 are shown.
 .RE
 .PP
+ip address show type ether
+.RS 4
+Shows IPv4 and IPv6 addresses assigned to all physical ethernetl interfaces
+.RE
+.PP
 ip address show dev eth0
 .RS 4
 Shows IPv4 and IPv6 addresses assigned to network interface eth0.
man/man8/ip-link.8.in +10 -3
index 19a0c9ca..60cbb5c7 100644
@@ -176,10 +176,12 @@ ip-link \- network device configuration
 .BR up " ] ["
 .B master
 .IR DEVICE " ] ["
-.B type
-.IR ETYPE " ] ["
 .B vrf
 .IR NAME " ] ["
+.B type
+.IR ETYPE " ] ["
+.B exclude_type
+.IR ETYPE " ] ["
 .BR nomaster " ]"
 
 .ti -8
@@ -237,7 +239,7 @@ ip-link \- network device configuration
 
 .ti -8
 .IR ETYPE " := [ " TYPE " |"
-.BR bridge_slave " | " bond_slave " ]"
+.BR ether " | " loopback " | " ppp " | " bridge_slave " | " bond_slave " ]"
 
 .ti -8
 .IR VFVLAN-LIST " := [ "  VFVLAN-LIST " ] " VFVLAN
@@ -2630,6 +2632,11 @@ ip link show type vlan
 Shows the vlan devices.
 .RE
 .PP
+ip link show exclude_type ppp
+.RS 4
+List the network interfaces except PPP devices.
+.RE
+.PP
 ip link show master br0
 .RS 4
 Shows devices enslaved by br0
Signature · 1 line
2.20.1
#1 lore raw
Quoted 14 lines · Anton Danilov
ip/ipaddress.c +10 -0
@@ -227,6 +227,28 @@ static int match_link_kind(struct rtattr **tb, const char *kind, bool slave)
 	return strcmp(parse_link_kind(tb[IFLA_LINKINFO], slave), kind);
 }
 
+static int match_if_type_name(unsigned short if_type, const char *type_name)
+{
+
+	char *expected_type_name;
+
+	switch (if_type) {
+	case ARPHRD_ETHER:
+		expected_type_name = "ether";
+		break;
+	case ARPHRD_LOOPBACK:

ip/ipaddress.c static int match_link_kind(struct rtattr **tb, const char *kind, bool slave) · line 249

Quoted 10 lines
ip/ipaddress.c +10 -0
+		expected_type_name = "loopback";
+		break;
+	case ARPHRD_PPP:
+		expected_type_name = "ppp";
+		break;
+	default:
+		expected_type_name = "";
+	}
+
+	return !strcmp(type_name, expected_type_name);

current 'type' filtering is the 'kind' string in the rtnl_link_ops -- bridge, veth, vlan, vrf, etc. You are now wanting to add 'exclude_type' and make it based on hardware type. That is a confusing user api.

What type of interface filtering is motivating this change? e.g., link / address lists but ignoring say vlan or veth devices?

#2 lore raw

Hello, David.

Quoted 3 lines

current 'type' filtering is the 'kind' string in the rtnl_link_ops -- bridge, veth, vlan, vrf, etc. You are now wanting to add 'exclude_type' and make it based on hardware type. That is a confusing user api.

The 'exclude_type' options first checks the 'kind' in the rtnl_link_ops, then the hardware type.

Quoted 2 lines

What type of interface filtering is motivating this change? e.g., link / address lists but ignoring say vlan or veth devices?

Yep, sure. If you use the kvm hypervisors, you can see a lot of veth devices. I think an ability to hide them is a good feature.

Filtering by 'kind' value has not been touched.

Quoted 33 lines · David Ahern
Quoted 14 lines · Anton Danilov
ip/ipaddress.c +10 -0
@@ -227,6 +227,28 @@ static int match_link_kind(struct rtattr **tb, const char *kind, bool slave)
      return strcmp(parse_link_kind(tb[IFLA_LINKINFO], slave), kind);
 }
 
+static int match_if_type_name(unsigned short if_type, const char *type_name)
+{
+
+     char *expected_type_name;
+
+     switch (if_type) {
+     case ARPHRD_ETHER:
+             expected_type_name = "ether";
+             break;
+     case ARPHRD_LOOPBACK:

ip/ipaddress.c static int match_link_kind(struct rtattr **tb, const char *kind, bool slave) · line 249

Quoted 10 lines
ip/ipaddress.c +10 -0
+             expected_type_name = "loopback";
+             break;
+     case ARPHRD_PPP:
+             expected_type_name = "ppp";
+             break;
+     default:
+             expected_type_name = "";
+     }
+
+     return !strcmp(type_name, expected_type_name);

current 'type' filtering is the 'kind' string in the rtnl_link_ops -- bridge, veth, vlan, vrf, etc. You are now wanting to add 'exclude_type' and make it based on hardware type. That is a confusing user api.

What type of interface filtering is motivating this change? e.g., link / address lists but ignoring say vlan or veth devices?

Signature · 1 line
Anton Danilov.
#3 lore raw
Quoted 8 lines · Anton Danilov

Hello, David.

Quoted 3 lines

current 'type' filtering is the 'kind' string in the rtnl_link_ops -- bridge, veth, vlan, vrf, etc. You are now wanting to add 'exclude_type' and make it based on hardware type. That is a confusing user api.

The 'exclude_type' options first checks the 'kind' in the rtnl_link_ops, then the hardware type.

ok; missed that on the first pass.

Update the commit message to say filtering by "hardware type"

Quoted 5 lines
Quoted 5 lines · Anton Danilov
Quoted 4 lines
ip/ipaddress.c +0 -0
@@ -227,6 +227,28 @@ static int match_link_kind(struct rtattr **tb, const char *kind, bool slave)
      return strcmp(parse_link_kind(tb[IFLA_LINKINFO], slave), kind);
 }
 
Quoted 14 lines
Quoted 6 lines
Quoted 6 lines
ip/ipaddress.c +6 -0
+static int match_if_type_name(unsigned short if_type, const char *type_name)
+{
+
+     char *expected_type_name;
+
+     switch (if_type) {
Quoted 8 lines

ip/ipaddress.c static int match_link_kind(struct rtattr **tb, const char *kind, bool slave) · line 243

Quoted 8 lines
ip/ipaddress.c +8 -0
+     case ARPHRD_ETHER:
+             expected_type_name = "ether";
+             break;
+     case ARPHRD_LOOPBACK:
+             expected_type_name = "loopback";
+             break;
+     case ARPHRD_PPP:
+             expected_type_name = "ppp";

ppp devices have kind set, so ARPHRD_PPP should not be needed.

Also, you have supported hardware types in multiple places - this match function and the filter.kind check. Make 1 table with supported types and use that table with helpers or both paths.

Why not allow exclude by "_slave" type? e.g., should all devices but bridge ports?