netdev - Re: [PATCH iproute2-next 8/9] utils: Introduce and use print_name_and

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

Message-ID: <4b5094fc-7cf7-74a2-4d54-1dd9fe60c77b@gmail.com>
Date:   Wed, 7 Feb 2018 09:36:55 +0200
From:   Serhey Popovych <serhe.popovych@...il.com>
To:     David Ahern <dsahern@...il.com>, netdev@...r.kernel.org
Subject: Re: [PATCH iproute2-next 8/9] utils: Introduce and use
 print_name_and_link() to print name@...k

David Ahern wrote:
> On 2/5/18 12:49 PM, Serhey Popovych wrote:
>> There is at least three places implementing same things: two in
>> ipaddress.c print_linkinfo() & print_linkinfo_brief() and one in
>> bridge/link.c.
>>
>> They are diverge from each other very little: bridge/link.c does not
>> support JSON output at the moment and print_linkinfo_brief() does not
>> handle IFLA_LINK_NETNS case.
>>
>> Introduce and use print_name_and_link() routine to handle name@...k
>> output in all possible variations; respect IFLA_LINK_NETNS attribute to
>> handle case when link is in different namespace; use ll_idx_n2a() for
>> interface name instead of "<nil>" to share logic with other code (e.g.
>> ll_name_to_index() and ll_index_to_name()) supporting such template.
>>
>> Signed-off-by: Serhey Popovych <serhe.popovych@...il.com>
>> ---
>>  bridge/link.c   |   13 +++----------
>>  include/utils.h |    4 ++++
>>  ip/ipaddress.c  |   44 ++------------------------------------------
>>  lib/utils.c     |   49 +++++++++++++++++++++++++++++++++++++++++++++++++
>>  4 files changed, 58 insertions(+), 52 deletions(-)
>>
> 
> This patch is causing a diff on my system:
> 
> # ip  -br add sh > /tmp/1
> # ip/ip  -br add sh > /tmp/2
> # diff /tmp/1 /tmp/2
> 8c8
> < veth-out@br3     UP             fe80::18a8:89ff:fee7:55c5/64
> ---
>> veth-out@if7     UP             fe80::18a8:89ff:fee7:55c5/64
> 
> So the current ip resolves ifindex 7 to br3:
> 
> # ip li sh dev br3
> 7: br3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master
> vrf3 state UP mode DEFAULT group default qlen 1000
> 
> where your patch causes if%d to be printed.
> 

That's interesting. I guess output comes from ll_idx_n2a() in this
change when both IFLA_LINK and IFLA_LINK_NETNS is seen.

My guess about this case is following:

  1) veth-out is of "veth" rtnl kind. (ip -d li sh dev veth-out).

  2) according to drivers/net/veth.c veth_get_iflink() and
     veth_get_link_net() IFLA_LINK and IFLA_LINK_NETNS are taken
     from peer device.

  3) seeing @br3 in current ip output looks confusing according to (2)
     as veth do not link to something other than it's peer that is in
     different network namespace.

From (3) I guess @br3 is incorrect value and caused by missing
IFLA_LINK_NETNS handling in old print_linkinfo_brief(): it always
calls ll_index_to_name().

Could you provide some more details about your setup if above guess is
wrong.

Especially following ones:

  1) ip -d li sh dev veth-out  (get the rtnl kind)

  2) ip -d li sh dev br3 (get the rtnl kind)

  3) uname -r or cat /proc/version



Download attachment "signature.asc" of type "application/pgp-signature" (491 bytes)