[etherlab-users] Problem changing states while bus scanning

[Eugene Bordenkircher]

I've found a case where the Master stops a bus scan mid scan and never
recovers.  I'll try to explain how we got there.

I'm pretty sure this is not specific to the hardware but I'm providing
as much detail as possible.  We have a Beckhoff setup of a EK1100 ->
EL1002 -> EL2002.  The application is configured to return all data
from the Input card and send data to all points in the Output card.

When everything is in the operational state and we are transferring
data we simulate a failure by pulling the Output card from the
coupler.  The master correctly starts a rescan of the bus.

The problem arises at this point if we call ecrt_master_deactivate()
while the bus_scan is still active.  The slaves end up in the half
scanned state as shown by the attachment below, and they never
recover.  The only way to get it out of this state is to unload the
master and reload it.

I've also attached the output of the kernel log with debug = 1.

Any help would be much appreciated.

Eugene
-------------- next part --------------
A non-text attachment was scrubbed...
Name: slaves.log
Type: text/x-log
Size: 98 bytes
Desc: not available
URL: <http://lists.etherlab.org/pipermail/etherlab-users/attachments/20100623/ae6ac1b3/attachment-0006.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: kern.log
Type: text/x-log
Size: 12669 bytes
Desc: not available
URL: <http://lists.etherlab.org/pipermail/etherlab-users/attachments/20100623/ae6ac1b3/attachment-0007.bin>