netdev - [PATCH iproute2-next 0/6] ip: Use netlink to walk through network device list

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <1517577055-23788-1-git-send-email-serhe.popovych@gmail.com>
Date:   Fri,  2 Feb 2018 15:10:49 +0200
From:   Serhey Popovych <serhe.popovych@...il.com>
To:     netdev@...r.kernel.org
Subject: [PATCH iproute2-next 0/6] ip: Use netlink to walk through network device list

In this seris I replace /proc/net/dev and /sys/class/net usage for walk
through network device list in iptunnel/ip6tunnel and iptuntap with
netlink dump.

Following changed since RFC was sent:

  1) Treat @struct rtnl_link_stats and @struct rtnl_link_stats64 as
     array with __u32 and __u64 elements respectively in
     copy_rtnl_link_stats64() as suggested by Stephen Hemminger.

  2) Remove @name and @size parameters from @struct tnl_print_nlmsg_info
     since we can get them easily from other data.

Testing.
========

Following script is used to ensure I didn't broke things too much:

\#!/bin/bash

iproute2_dir="$1"
iface='gre1'

pushd "$iproute2_dir" &>/dev/null

for i in new old; do
	DIR="/tmp/$i"
	mkdir -p "$DIR"

	ln -snf ip.$i ip/ip

	for o in '' -s -d; do
		ip/ip $o tunnel show           >"$DIR/ip${o}-tunnel-show"
		ip/ip -4 $o tunnel show        >"$DIR/ip-4${o}-tunnel-show"
		ip/ip -6 $o tunnel show        >"$DIR/ip-6${o}-tunnel-show"
		ip/ip $o tunnel show dev "$iface" \
			>"$DIR/ip${o}-tunnel-show-$iface"
		ip/ip $o tuntap show           >"$DIR/ip${o}-tuntap-show"
	done
done
rm -f ip/ip

diff -urN /tmp/{old,new} |sed -n -Ee'/^(-{3}|\+{3})[[:space:]]+/!p'
rc=$?

popd &>/dev/null
exit $rc

Results:
========

...
fopen /sys/class/net/ipip1/tun_flags: No such file or directory
fopen /sys/class/net/ipip2/tun_flags: No such file or directory
fopen /sys/class/net/gre10/tun_flags: No such file or directory
...
diff -urN /tmp/old/ip-d-tuntap-show /tmp/new/ip-d-tuntap-show
@@ -1,4 +1,4 @@
-tun1: tap user 1004 group 27
-	Attached to processes:
 tun0: tun user 1000 group 27
 	Attached to processes:
+tun1: tap user 1004 group 27
+	Attached to processes:
diff -urN /tmp/old/ip-s-tuntap-show /tmp/new/ip-s-tuntap-show
@@ -1,2 +1,2 @@
-tun1: tap user 1004 group 27
 tun0: tun user 1000 group 27
+tun1: tap user 1004 group 27
diff -urN /tmp/old/ip-tuntap-show /tmp/new/ip-tuntap-show
@@ -1,2 +1,2 @@
-tun1: tap user 1004 group 27
 tun0: tun user 1000 group 27
+tun1: tap user 1004 group 27

So basically only print order for ip tuntap get changes. Rest is intact.

Thanks,
Serhii

Serhey Popovych (6):
  ipaddress: Unify print_link_stats() and print_link_stats64()
  ip: Introduce get_rtnl_link_stats_rta() to get link statistics
  tunnel: Split statistic getting and printing
  iptunnel/ip6tunnel: Code cleanups
  iptunnel/ip6tunnel: Use netlink to walk through tunnels list
  tuntap: Use netlink to walk through tuntap list

 include/utils.h |    3 +
 ip/ip6tunnel.c  |  115 +++++++++++----------------------
 ip/ipaddress.c  |  189 ++++---------------------------------------------------
 ip/iptunnel.c   |   93 +++++++++------------------
 ip/iptuntap.c   |  121 ++++++++++++++++++++++++++---------
 ip/tunnel.c     |  114 ++++++++++++++++++++++++++-------
 ip/tunnel.h     |   17 ++++-
 lib/utils.c     |   45 +++++++++++++
 8 files changed, 324 insertions(+), 373 deletions(-)

-- 
1.7.10.4