rsync Corrupted MAC on input.

[drew einhorn]

On Tue, Jul 14, 2009 at 3:37 PM, NoOp <glgxg at sbcglobal.net> wrote:
>
> > I do not think it is flow control.  TCP should take care that.  Others
> > have seen the same error.  The MAC is SSH's Message Authentication Code.
> > Here is a good explanation of what is happening.
> >
> > http://blogs.sun.com/janp/entry/ssh_messages_code_bad_packet
> >
>
> And a few places to look:
> <https://launchpad.net/+search?field.text="Corrupted+MAC"&field.actions.search=Search>

Whoa!!  There's currently 122 hits for this search.  Nearly all of
them refer to Bug #60764
Is there an easy way to hide all the Bug #60764 references and take a
close look at
what's left.

I am running rsync from the command line on jaunty on a Lenovo s10e netbook.

It opens a ssh connection to another jaunty box on an old Dell workstation.

running:

      ethtool -K eth0 tx off

on the s10e netbook solves the problem,

> dmesg | grep eth0
> [    4.115706] eth0: Tigon3 [partno(BCM95906) rev c002 PHY(5906)] (PCI Express) 10/100Base-TX Ethernet 00:23:8b:64:3d:c3
> [    4.115720] eth0: RXcsums[1] LinkChgREG[0] MIirq[0] ASF[0] WireSpeed[0] TSOcap[0]
> [    4.115729] eth0: dma_rwctrl[76180000] dma_mask[64-bit]
> [   33.858072] ADDRCONF(NETDEV_UP): eth0: link is not ready
> [   35.433322] tg3: eth0: Link is up at 100 Mbps, half duplex.
> [   35.433335] tg3: eth0: Flow control is off for TX and off for RX.
> [   35.434010] ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
> [   45.452075] eth0: no IPv6 routers present

Only half duplex, flow control is off.
Hmm, maybe it is a flow control problem.

sudo ethtool -k eth0

Offload parameters for eth0:
Cannot get device flags: Operation not supported
rx-checksumming: on
tx-checksumming: off
scatter-gather: off
tcp segmentation offload: off
udp fragmentation offload: off
generic segmentation offload: on
large receive offload: off

On the old Dell box we have we have a lot more stuff offloaded
and a 100 Mbs Full Duplex link with Flow Control RX/TX


sudo ethtool -k eth4

Offload parameters for eth4:
Cannot get device flags: Operation not supported
rx-checksumming: on
tx-checksumming: on
scatter-gather: on
tcp segmentation offload: on
udp fragmentation offload: off
generic segmentation offload: on
large receive offload: off

and

dmesg | grep eth4

[    3.651527] e1000: eth4: e1000_probe: Intel(R) PRO/1000 Network Connection
[   21.154221] ADDRCONF(NETDEV_UP): eth4: link is not ready
[   21.156424] e1000: eth4: e1000_watchdog: NIC Link is Up 100 Mbps
Full Duplex, Flow Control: RX/TX
[   21.178066] ADDRCONF(NETDEV_CHANGE): eth4: link becomes ready
[   31.740018] eth4: no IPv6 routers present

Bug #60764 Is probably the place to report this but it might be one of
the other bugs.

Hmm.  The thing  between the netbook and old Dell is a Netgear DS108
Dual speed Hub
not a switch.  I'm not sure I understand exactly what a dual speed hub is.
But I'll bet it is causing the problem.

So maybe it's not a bug after all.

But it should go into the bug comments.

There's got to be a better way of diagnosing this problem.
Or at least and an automated tool that diagnoses it.

--
Drew Einhorn