[LEDE-DEV] XRX200 (TD-W8970v1) USB instability

[David Woodhouse]

I added some USB devices to my TD-W8970 last week. It started crashing
— with nothing on the serial port, just complete death followed
eventually by a (presumably watchdog-triggered) restart.

I replaced the USB hub with a powered hub, which seemed to help so I
assumed the crashes were power-related. Domoticz seemed to be mostly
working fine after that, and I put it into production (and left the
country a few days later). I did see this message occasionally:

 kernel: [387424.281851] usb usb1: clear tt 1 (9083) error -22

... but it seemed mostly harmless. Then last night, during the large
nightly rsync download, it looks like it took the whole machine down:

domoticz: evohome: WARNING unrecognised message structure - possible corrupt message '03069B' (3)
domoticz: evohome: WARNING unrecognised message structure - possible corrupt message '04' (3)
domoticz: evohome: WARNING unrecognised message structure - possible corrupt message '0105DC00FFFFFF' (2)
kernel: [387424.281851] usb usb1: clear tt 1 (9083) error -22
kernel: [387639.895022] usb usb1: clear tt 1 (9083) error -22
hostapd: wlan0: STA 00:d0:2d:74:ad:67 WPA: group key handshake completed (RSN)
odhcpd[21324]: Initial RA router lifetime 1, 0 address(es) available on eth0.2
upsd[25103]: Data for UPS [openups] is stale - check driver
upsd[25103]: UPS [openups] data is no longer stale
hostapd: nl80211: nl80211_recv_beacons->nl_recvmsgs failed: -5
upsd[25103]: Data for UPS [openups] is stale - check driver
upsd[25103]: UPS [openups] data is no longer stale
kernel: [388115.224649] usb usb1: clear tt 1 (9083) error -22
upsd[25103]: Data for UPS [openups] is stale - check driver
upsd[25103]: UPS [openups] data is no longer stale
pppd[3506]: System time change detected.
pppd[3506]: System time change detected.
upsd[25103]: Data for UPS [openups] is stale - check driver
hostapd: nl80211: nl80211_recv_beacons->nl_recvmsgs failed: -5
upsd[25103]: UPS [openups] data is no longer stale
hostapd: wlan0: STA 00:d0:2d:74:ad:67 WPA: group key handshake completed (RSN)
hostapd: nl80211: nl80211_recv_beacons->nl_recvmsgs failed: -5
upsd[25103]: Data for UPS [openups] is stale - check driver
upsd[25103]: UPS [openups] data is no longer stale
pppd[3506]: System time change detected.
pppd[3506]: System time change detected.
upsd[25103]: Data for UPS [openups] is stale - check driver
upsd[25103]: UPS [openups] data is no longer stale
hostapd: nl80211: nl80211_recv_beacons->nl_recvmsgs failed: -5
hostapd: nl80211: nl80211_recv_beacons->nl_recvmsgs failed: -5
pppd[3506]: System time change detected.
upsd[25103]: Data for UPS [openups] is stale - check driver
pppd[3506]: System time change detected.
pppd[3506]: LCP terminated by peer
pppd[3506]: Connect time 3134.3 minutes.
pppd[3506]: Sent 1836195634 bytes, received 1705672340 bytes.
upsd[25103]: UPS [openups] data is no longer stale
netifd: Network device 'pppoe-vdsl' link is down
netifd: Network alias 'pppoe-vdsl' link is down
netifd: Interface 'vdsl_6' has link connectivity loss

The VDSL came up and down a few times after that for the next hour or
so, and then the machine crashed hard until I noticed in the morning
and had it reset.

Should I expect USB to be flaky on this machine, and attempt to run the
data acquisition elsewhere? It's running LEDE 17.01 (4.4.47).
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 4938 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/lede-dev/attachments/20170222/38a40c14/attachment.bin>