netdev - Re: [PATCH] gianfar: Add gfar_change

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: <d8e83e50ed440ef46a7746ed7abdb9fffc4d7491.camel@infinera.com>
Date:   Fri, 7 Dec 2018 17:01:11 +0000
From:   Joakim Tjernlund <Joakim.Tjernlund@...inera.com>
To:     "andrew@...n.ch" <andrew@...n.ch>
CC:     "claudiu.manoil@....com" <claudiu.manoil@....com>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>
Subject: Re: [PATCH] gianfar: Add gfar_change_carrier()

On Fri, 2018-12-07 at 15:15 +0100, Andrew Lunn wrote:
> 
> 
> > Been a bit busy today but now I have played with dormant using ip link and got some odd results:
> > # > ifconfig eth0
> > eth0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
> >         inet 172.20.0.246  netmask 255.255.0.0  broadcast 172.20.255.255
> >         inet6 fe80::ad9c:b230:1da8:1821  prefixlen 64  scopeid 0x20<link>
> >         ether 8c:16:45:89:cf:c6  txqueuelen 1000  (Ethernet)
> >         RX packets 1848903  bytes 736764445 (702.6 MiB)
> >         RX errors 0  dropped 0  overruns 0  frame 0
> >         TX packets 627462  bytes 222453345 (212.1 MiB)
> >         TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0
> >         device interrupt 16  memory 0xdc200000-dc220000
> > # > ip link set mode dormant dev eth0
> >  ping sunet.se
> > PING sunet.se (192.36.171.231) 56(84) bytes of data.
> > 64 bytes from webc.sunet.se (192.36.171.231): icmp_seq=1 ttl=54 time=2.22 ms
> > 64 bytes from webc.sunet.se (192.36.171.231): icmp_seq=2 ttl=54 time=2.17 ms
> > 
> > # > ifconfig eth0
> > eth0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
> >         inet 172.20.0.246  netmask 255.255.0.0  broadcast 172.20.255.255
> >         inet6 fe80::ad9c:b230:1da8:1821  prefixlen 64  scopeid 0x20<link>
> >         ether 8c:16:45:89:cf:c6  txqueuelen 1000  (Ethernet)
> >         RX packets 1905479  bytes 753549572 (718.6 MiB)
> >         RX errors 0  dropped 0  overruns 0  frame 0
> >         TX packets 648266  bytes 224421617 (214.0 MiB)
> >         TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0
> >         device interrupt 16  memory 0xdc200000-dc220000
> > still RUNNING ..
> > 
> > #> ip link show eth0
> > 5: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP mode DORMANT group default qlen
> > 
> > state is still UP ?
> > 
> > # > ip link set state dormant dev eth0
> > # > ip link show eth0
> > 5: eth0: <NO-CARRIER,BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state DORMANT mode DORMANT group default qlen 1000
> > # > ifconfig eth0
> > eth0: flags=4099<UP,BROADCAST,MULTICAST>  mtu 1500
> > ...
> > 
> > Now both state and not RUNNING :)
> > but ...
> >  # ping sunet.se
> > PING sunet.se (192.36.171.231) 56(84) bytes of data.
> > 64 bytes from webc.sunet.se (192.36.171.231): icmp_seq=1 ttl=54 time=2.43 ms
> > 64 bytes from webc.sunet.se (192.36.171.231): icmp_seq=2 ttl=54 time=2.31 ms
> > 
> > I can still ping though. Is this how it is supposed to work ? No sure how state and mode relate to each other either.
> 
> I don't actually know. I know some things are supposed to work while
> dormant. You should be able to perform 802.1x negotiation, etc.
> 
> You might find the people on the wireless list know more. I think it
> is used by wpa_supplicant and hostapd during device authentication.


I am not sure why .ndo_change_carrier() is a no go in real drivers. Consider PHY less (aka Fixed PHYs)
devices. Here it makes sense to be able to control carrier from /sys I think.

Would you be happier if .ndo_change_carrier() only acted on Fixed PHYs?
I could also rework ndo_change_carrier to only use netif_carrier_on/off like team and dummy do.

 Jocke