ip: Extend filter links/addresses
Subsystems (2)
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
@@ -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
@@ -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
@@ -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
@@ -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
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?
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.
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 thertnl_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?