myotherpcisapdp11
Posts: 12
Joined: Wed Mar 09, 2016 5:46 am

Ethernet oddness - Pi4 B 4GB

Sat Apr 03, 2021 12:16 pm

Hi! Got my lady-fiend a Pi4B (4 gig model) for lockdown, but having a recurring problem with it (and with several router/switches - her own is the current Virgin hub ting..).. My own (identical, but I have aged 3com 100MHz switches on a far-to-big-for-a-normal-house sized lan, have tried one at her house, cascaded from the VM router, with no improvement) Pi4B has the same problem (again on multiple dissimilar switches), but is spurious.

In essence, the ethernet is flapping up and down fairly frequently for a while, then behaving, for up to a few hours, then going off and having another funny turn.. When it's happening, the interweb is pretty much incommunicado.

It does it with the standard current raspbios distro fresh on a card, and with my *slightly customised* version quite equally.

[Start of log sample]
Mar 28 14:30:58 hostname kernel: [609890.507167] bcmgenet fd580000.ethernet eth0: Link is Down
Mar 28 14:31:20 hostname kernel: [609912.347524] bcmgenet fd580000.ethernet eth0: Link is Up - 100Mbps/Full - flow control rx/tx
Mar 28 14:37:56 hostname kernel: [610308.593716] bcmgenet fd580000.ethernet eth0: Link is Down
Mar 28 14:38:17 hostname kernel: [610329.394105] bcmgenet fd580000.ethernet eth0: Link is Up - 100Mbps/Full - flow control rx/tx
Mar 28 14:57:25 hostname kernel: [611477.572371] bcmgenet fd580000.ethernet eth0: Link is Down
Mar 28 14:57:46 hostname kernel: [611498.372777] bcmgenet fd580000.ethernet eth0: Link is Up - 100Mbps/Full - flow control rx/tx
Mar 28 14:58:19 hostname kernel: [611531.653217] bcmgenet fd580000.ethernet eth0: Link is Down
Mar 28 14:58:40 hostname kernel: [611552.453693] bcmgenet fd580000.ethernet eth0: Link is Up - 100Mbps/Full - flow control rx/tx
Mar 28 15:10:08 hostname kernel: [612239.904667] bcmgenet fd580000.ethernet eth0: Link is Down
Mar 28 15:10:28 hostname kernel: [612260.705084] bcmgenet fd580000.ethernet eth0: Link is Up - 100Mbps/Full - flow control rx/tx
Mar 28 15:25:49 hostname kernel: [613181.119691] bcmgenet fd580000.ethernet eth0: Link is Down
Mar 28 15:26:11 hostname kernel: [613202.960147] bcmgenet fd580000.ethernet eth0: Link is Up - 100Mbps/Full - flow control rx/tx
Mar 28 16:01:42 hostname kernel: [615333.952630] bcmgenet fd580000.ethernet eth0: Link is Down
Mar 28 16:02:02 hostname kernel: [615354.753052] bcmgenet fd580000.ethernet eth0: Link is Up - 100Mbps/Full - flow control rx/tx
Mar 28 16:46:10 hostname kernel: [618002.635845] bcmgenet fd580000.ethernet eth0: Link is Down
Mar 28 16:46:32 hostname kernel: [618024.476318] bcmgenet fd580000.ethernet eth0: Link is Up - 100Mbps/Full - flow control rx/tx
Mar 28 17:59:45 hostname kernel: [622417.506213] bcmgenet fd580000.ethernet eth0: Link is Down
Mar 28 18:00:06 hostname kernel: [622438.306647] bcmgenet fd580000.ethernet eth0: Link is Up - 100Mbps/Full - flow control rx/tx
Mar 28 18:52:39 hostname kernel: [625591.632373] bcmgenet fd580000.ethernet eth0: Link is Down
Mar 28 18:53:00 hostname kernel: [625612.432787] bcmgenet fd580000.ethernet eth0: Link is Up - 100Mbps/Full - flow control rx/tx
[End of log sample]

Any ideas as to the cause/fix of this?

epoch1970
Posts: 6371
Joined: Thu May 05, 2016 9:33 am
Location: Paris, France

Re: Ethernet oddness - Pi4 B 4GB

Sat Apr 03, 2021 4:15 pm

Power source for that Pi?
"S'il n'y a pas de solution, c'est qu'il n'y a pas de problème." Les Shadoks, J. Rouxel

swampdog
Posts: 630
Joined: Fri Dec 04, 2015 11:22 am

Re: Ethernet oddness - Pi4 B 4GB

Sat Apr 03, 2021 5:13 pm

It's not explicitly clear. Assuming you mean..

Her pi works on your lan.
Your pi does not work on hers.

An rpi4 connected to my lan..

Code: Select all

foo@pi18:~ $ sudo ethtool eth0
Settings for eth0:
	Supported ports: [ TP MII ]
	Supported link modes:   10baseT/Half 10baseT/Full 
	                        100baseT/Half 100baseT/Full 
	                        1000baseT/Half 1000baseT/Full 
	Supported pause frame use: Symmetric Receive-only
	Supports auto-negotiation: Yes
	Supported FEC modes: Not reported
	Advertised link modes:  10baseT/Half 10baseT/Full 
	                        100baseT/Half 100baseT/Full 
	                        1000baseT/Half 1000baseT/Full 
	Advertised pause frame use: Symmetric Receive-only
	Advertised auto-negotiation: Yes
	Advertised FEC modes: Not reported
	Link partner advertised link modes:  10baseT/Half 10baseT/Full 
	                                     100baseT/Half 100baseT/Full 
	                                     1000baseT/Full 
	Link partner advertised pause frame use: Symmetric
	Link partner advertised auto-negotiation: Yes
	Link partner advertised FEC modes: Not reported
	Speed: 1000Mb/s
	Duplex: Full
	Port: MII
	PHYAD: 1
	Transceiver: internal
	Auto-negotiation: on
	Supports Wake-on: gs
	Wake-on: d
        SecureOn password: 00:00:00:00:00:00
	Current message level: 0x00000007 (7)
			       drv probe link
	Link detected: yes
Which log are you posting?

Code: Select all

foo@pi18:~ $ sudo cat /var/log/kern.log | grep bcmge
Apr  1 08:34:35 pi18 kernel: [    2.360329] bcmgenet fd580000.ethernet: failed to get enet clock
Apr  1 08:34:35 pi18 kernel: [    2.362997] bcmgenet fd580000.ethernet: GENET 5.0 EPHY: 0x0000
Apr  1 08:34:35 pi18 kernel: [    2.365674] bcmgenet fd580000.ethernet: failed to get enet-wol clock
Apr  1 08:34:35 pi18 kernel: [    2.368275] bcmgenet fd580000.ethernet: failed to get enet-eee clock
Apr  1 08:34:35 pi18 kernel: [    2.370850] bcmgenet: Skipping UMAC reset
Apr  1 08:34:35 pi18 kernel: [    2.394123] libphy: bcmgenet MII bus: probed
Apr  1 08:34:35 pi18 kernel: [  118.551653] bcmgenet: Skipping UMAC reset
Apr  1 08:34:35 pi18 kernel: [  118.554331] bcmgenet fd580000.ethernet: configuring instance for external RGMII
Apr  1 08:34:35 pi18 kernel: [  118.554802] bcmgenet fd580000.ethernet eth0: Link is Up - 1Gbps/Full - flow control rx/tx
Apr  1 08:34:52 pi18 kernel: [  137.036155] bcmgenet fd580000.ethernet eth0: Link is Down
Apr  1 08:34:52 pi18 kernel: [  137.051679] bcmgenet: Skipping UMAC reset
Apr  1 08:34:52 pi18 kernel: [  137.053452] bcmgenet fd580000.ethernet: configuring instance for external RGMII
Apr  1 08:34:52 pi18 kernel: [  137.053743] bcmgenet fd580000.ethernet eth0: Link is Up - 1Gbps/Full - flow control rx/tx
Apr  1 08:34:53 pi18 kernel: [  137.296119] bcmgenet fd580000.ethernet eth0: Link is Down
Apr  1 08:34:53 pi18 kernel: [  137.327517] bcmgenet: Skipping UMAC reset
Apr  1 08:34:53 pi18 kernel: [  137.329294] bcmgenet fd580000.ethernet: configuring instance for external RGMII
Apr  1 08:34:53 pi18 kernel: [  137.329580] bcmgenet fd580000.ethernet eth0: Link is Up - 1Gbps/Full - flow control rx/tx
I have my VM hub (VMDG490) in plain cable modem mode(*). I can get a web browser at it on http://192.168.100.1/ at which point I have a "router status" tab. You do not have to log in for that. You can then peruse the logs. It's (ahem) normal for the network log to spit out a few errors.

Generically, I have(**) two common problems. The first is VM attempt an update and it fails. Typically it sorts itself out after "time". The other is they change my ip address. The former issue is solved by watching the logs, seeing it is still struggling the next morning so hit the power on the hub. The latter almost never happens these days but is a pig when it does: same solution, hit the power on the hub(***).

(*) recommended for any isp but some VM hubs are particularly buggy. I have an rpi "firewall" attached directly to the VM ethernet. My wireless routers are inside my intranet.

(**) not had a problem in months, not since they decided to change my ip address.

(***) leave it turned off for two or three minutes or the hub can get even more confused.

If you can, put the VM hub into CM mode and plug the pi directly into the CM eth port. See if the problem goes away.

myotherpcisapdp11
Posts: 12
Joined: Wed Mar 09, 2016 5:46 am

Re: Ethernet oddness - Pi4 B 4GB

Sun Apr 04, 2021 12:42 am

epoch1970 wrote:
Sat Apr 03, 2021 4:15 pm
Power source for that Pi?
Both of our Pi4s have the official/recommended one, have tried swapping them.. no power problem indicated etc.

myotherpcisapdp11
Posts: 12
Joined: Wed Mar 09, 2016 5:46 am

Re: Ethernet oddness - Pi4 B 4GB

Sun Apr 04, 2021 1:09 am

swampdog wrote:
Sat Apr 03, 2021 5:13 pm
It's not explicitly clear. Assuming you mean..

Her pi works on your lan.
Your pi does not work on hers.
Both play up on all of our lans, (i have multiple switches/routers on my "home" LAN.. It's a fairly odd computer lab kinda place rather than a conventional house..these days (i'm downsizing it) it's mostly 3 old 100baseT Dlink switches, though also 1 3com 8 port 100baseT, a netgear 100baseT switch, which seem to "just work", and a few other old hubs i just used to connect to 10base2 and 10base5 for the older DEC (PDP-11/Vax) and IBM (5150/5160 etc.) stuff. Things just play up a lot worse on her (very simple, kinda typical out of the box) home LAN, with both Pi4s, and her pi4 plays up worse on my LAN than my pi4 does - but both do, on both LANs, it's just a matter of to what degree. My own, on my LAN, is quite tolerable, her's has days it's not actually possible to use the network meaningfully at all.. Right this moment (I'm there, using it) it seems to be almost passable.

In the case of Her pi4, it's connected (via a 2M cat5E cable, i've done switcheroos to eliminate the cable and port from the investigation) directly to the switch part of the "Hub".

That log is an excerpt from /var/log/messages. Large swathes of the messages log on that pi consist of similar stuff, it's a frequent problem.
Other stuff connected (via UTP) to that router/switch work just fine, including Pi2/Pi3, only multiple Pi4s are having this problem.

$ sudo ethtool eth0
Settings for eth0:
Supported ports: [ TP MII ]
Supported link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Half 1000baseT/Full
Supported pause frame use: Symmetric Receive-only
Supports auto-negotiation: Yes
Supported FEC modes: Not reported
Advertised link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Half 1000baseT/Full
Advertised pause frame use: Symmetric Receive-only
Advertised auto-negotiation: Yes
Advertised FEC modes: Not reported
Link partner advertised link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
Link partner advertised pause frame use: Symmetric Receive-only
Link partner advertised auto-negotiation: Yes
Link partner advertised FEC modes: Not reported
Speed: 100Mb/s
Duplex: Full
Port: MII
PHYAD: 1
Transceiver: internal
Auto-negotiation: on
Supports Wake-on: gsf
Wake-on: d
SecureOn password: 00:00:00:00:00:00
Current message level: 0x00000007 (7)
drv probe link
Link detected: yes

It /looks/ like a hardware problem from where I'm sitting, and i would have called it that, had it not occured on both (seperately sourced, months apart, different suppliers) pi4s. :-/ Which now leaves me thinking it's a soft/firmware issue on the Pi4, so i'm looking to see if there is a fix.

It's not related to dhcp, that I checked weeks ago, nor thermal. It just seems to be random, start "misbehaving" for a while (from half hour to a couple of days", then behave itself for a while, and so on.. <sigh> I do get unexplained drops at my place too, but less frequent by far.

It's a VM Hub3.0 btw. I just checked. Not sure of the specs on it, but it seems to "just work" with everything. (except when VM do one of their pesky updates to turn off features (like logging, for example!) to make space for more graphics or so.. :-/ )

CM mode won't wash here. It's a "vanilla" household.. many devices, many children, established applecart. The Pi would have to take the walk of shame instead. :-(

User avatar
rpdom
Posts: 18541
Joined: Sun May 06, 2012 5:17 am
Location: Chelmsford, Essex, UK

Re: Ethernet oddness - Pi4 B 4GB

Sun Apr 04, 2021 5:34 am

I had lots of problems with my VM Superhub 3.0. I gave up, put it in Modem mode and got a TP-Link router to link it to my network. End of problems.
Unreadable squiggle

swampdog
Posts: 630
Joined: Fri Dec 04, 2015 11:22 am

Re: Ethernet oddness - Pi4 B 4GB

Tue Apr 13, 2021 9:29 pm

Sorry it's been a while..

..if you can't do the CM thing you can just burn two standard sdcards and plug the rpi4's back-to-back to see if the error goes away. Assuming it does, plug a 100M switch between them and retest. Repeat with each VM hub.

Alternatively, just use wireless?

swampdog
Posts: 630
Joined: Fri Dec 04, 2015 11:22 am

Re: Ethernet oddness - Pi4 B 4GB

Sat Apr 17, 2021 3:50 am

Fwiw I just caught my superhub going nuts..

Code: Select all

17/04/2021 02:27:06 GMT 17/04/2021 02:27:06 GMT Warning (5) 84020200 Lost MDD Timeout
17/04/2021 02:26:44 GMT 17/04/2021 02:26:44 GMT Warning (5) 84020200 Lost MDD Timeout
17/04/2021 02:26:25 GMT 17/04/2021 02:26:25 GMT Warning (5) 84020200 Lost MDD Timeout
17/04/2021 02:26:06 GMT 17/04/2021 02:26:06 GMT Warning (5) 84020200 Lost MDD Timeout
17/04/2021 02:25:47 GMT 17/04/2021 02:25:47 GMT Warning (5) 84020200 Lost MDD Timeout
17/04/2021 02:25:18 GMT 17/04/2021 02:25:18 GMT Error (4) 68000407 TOD established
..while I was there to see it. Lots of other log entries omitted from the above. Suffice it to say my entire network misbehaved for best part of an hour while the superhub tried unsuccessfully to apply a maintenance update. I run my own DNS/DHCP and even that broke for a while. Also my own firewall/NAT. Interestingly while this was going on the external ethernet (ie: connected to the superhub) kept being toggled between its normal external ip and the superhub internal ip so my NAT was chucking all my external traffic at 192.168.100.1 which is nice.

jj_0
Posts: 128
Joined: Wed Jul 11, 2012 7:07 am

Re: Ethernet oddness - Pi4 B 4GB

Sat Apr 17, 2021 1:34 pm

Have you checked out disabling EEE, see viewtopic.php?f=28&t=305820?

Return to “Troubleshooting”