Raspberry Pi Network Data

[Ian Lepore]

On Tue, 2013-02-26 at 13:17 -0500, Sean Cavanaugh wrote:
> 
> > -----Original Message-----
> > From: owner-freebsd-arm at freebsd.org [mailto:owner-freebsd-
> > arm at freebsd.org] On Behalf Of Brett Wynkoop
> > Sent: Tuesday, February 26, 2013 12:04 PM
> > To: freebsd-arm at freebsd.org
> > Subject: Raspberry Pi Network Data
> > 
> > Greeting-
> > 
> > For a couple of days I have been building software from ports.  Mostly
> these
> > builds are for things I just want to try on the Pi, or Bone, but the
> secondary
> > reason is to put some pre-compiled packages up for those that do not have
> > the patience to build them.
> > 
> > While my Pi has been stable for a couple of weeks I have noted that
> > sometimes it stops talking on the network.  At those times if I get on the
> > console and ifconfig down and back up the interface it starts talking on
> the
> > net just fine again.
> > 
> > Last night I believe I found a link between disk i/o and network non-
> > responsiveness.  During a period of high disk i/o to the USB connected
> flash
> > drive I lost network.  The console had messages about retrys to the disk
> and
> > the console was slow to respond.  It took me for ever to get logged in
> > because the console kept dropping characters while I typed.  I am using
> usb-
> > keyboard and composite video for the console.
> > 
> > When I got logged in I still had trouble typing ifconfig ue0 down ;
> ifconfig ue0
> > up, but once I did everything went back to normal.
> > Keyboard response was fine, disk i/o no longer seemed to be reporting
> > errors and of course the network came back on line.
> > 
> > I went to sleep with zoneminder building.  Now 6 hours later I find the
> > machine in the same state. Since the disk, keyboard, and ethernet are all
> usb
> > devices could we have a bug in the usb sub-system?
> > 
> > As soon as the ifconfig ue0 down happens the console keyboard becomes
> > properly responsive again.  Could we have some sort of interrupt problem
> > going on here?
> > 
> > This is food for thought for you kernel hackers.  If there is anything you
> want
> > me to specifically try or do the next time I have this problem, probably
> in the
> > next 24 hours, please let me know.
> > 
> > Your fellow ARM hacker,
> > 
> > -Brett
> > 
> 
> Keep in mind that the network port, the SD card slot,  and obviously the USB
> ports themselves are all on the same USB bus. That may be part of the issue.
> Definitely agree that it should be able to swap between them easier than
> manually shunting it.

Not the sd card, it has its own dedicated sdhci controller in the SoC.

The only usb thing active on my rpi is the onboard hardware (hub and
network interface) and it has a tendency to occasionally drop off the
bus and return.  Actually, it's not just the ethernet, the whole hub
(onboard hub, not external) disappears and reappears.  This happens
intermittantly, sometimes several times a day.  Twice it has failed to
recover -- the hub never reattached until I rebooted.

smsc0: warning: Failed to read register 0x114
smsc0: warning: MII is busy
ugen0.2: <vendor 0x0424> at usbus0 (disconnected)
uhub1: at uhub0, port 1, addr 2 (disconnected)
ugen0.3: <vendor 0x0424> at usbus0 (disconnected)
smsc0: warning: Failed to read register 0x114
smsc0: warning: MII is busy
smsc0: warning: Failed to read register 0x114
smsc0: warning: MII is busy
smsc0: warning: Failed to read register 0x114
smsc0: warning: MII is busy
smsc0: at uhub1, port 1, addr 3 (disconnected)
smsc0: warning: Failed to read register 0x114
smsc0: warning: MII is busy
smsc0: warning: Failed to read register 0x114
smsc0: warning: MII is busy
smsc0: warning: Failed to read register 0x114
smsc0: warning: MII is busy
smsc0: warning: Failed to read register 0x114
smsc0: warning: MII is busy
smsc0: warning: Failed to read register 0x114
smsc0: warning: MII is busy
ukphy0: detached
miibus0: detached
Feb 26 03:13:05 rpi dhclient[246]: connection closed
Feb 26 03:13:05 rpi dhclient[246]: exiting.
Feb 26 03:13:05 rpi ntpd[519]: sendto(172.22.42.240) (fd=22): No route
to host
ugen0.2: <vendor 0x0424> at usbus0
uhub1: <vendor 0x0424 product 0x9512, class 9/0, rev 2.00/2.00, addr 2>
on usbus0
uhub1: MTT enabled
uhub1: 3 ports with 2 removable, self powered
Feb 26 03:13:06 rpi ntpd[519]: sendto(172.22.42.254) (fd=22): No route
to host
ugen0.3: <vendor 0x0424> at usbus0
smsc0: <vendor 0x0424 product 0xec00, rev 2.00/2.00, addr 3> on usbus0
smsc0: chip 0xec00, rev. 0002
miibus0: <MII bus> on smsc0
ukphy0: <Generic IEEE 802.3u media interface> PHY 1 on miibus0
ukphy0:  none, 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto
ue0: <USB Ethernet> on smsc0
ue0: Ethernet address: b8:27:eb:33:7c:02
smsc0: chip 0xec00, rev. 0002

-- Ian