Yeah, looking forward to see if it can handle 2 or 3 Single SC's.
Me too, just to make sure that I don't spread mis-information though it is NOT lib-usb that has the issue, it seems to be only the serial-usb that has 'issues', and even then I have no serial-usb FPGA devices to test with (like Kano did, and found problems).
And added to that, no one has tested ASIC with the Pi (lib-usb or serial-usb) so it may turn out to be a non issue.
The problem with serial-USB is that each device has it's own kernel driver.
Thus each driver's bugs are able to take hold and cause problems (and cause incompatibilities between them)
I've not bothered to rewrite the Icarus driver coz ... I've not had the time and not really something I care too much about ...
Everything else (non-GPU) on cgminer uses libusb and mining with two different devices (BFL and MMQ) going through the same libusb has not had any major problems (i.e. no problems that were rpi specific)
All the new ASIC drivers we write are (and will be) libusb also.
... I'll have a few ASICs soon ... well ... certainly before the end of the world at least
Well that was interesting.
I had made a small change to my cgminer run script to log on the rpi instead of my main fileserver ...
I had a look at the two logs (for the above crashes) and they showed cgminer still running.
The rpi hadn't died, just the network had.
So the problem may be some interaction between the "Prolific Technology, Inc. PL2303 Serial Port" and the network driver?
I've written a crontab script to check if the network has been down for 5 minutues and then request a cgminer quit and shutdown the rpi
... coz I don't (usually) have a display or keyboard attached, it means it shuts down properly rather than pulling the power out
The cgminer logs show that happening also.
Anyway, the logs show:
cgminer [2013-03-14 09:57:47] Stratum connection to pool 0 interrupted
[2013-03-14 09:58:02] Pool 0
http://au.ozco.in:3333 not responding!
syslogMar 14 09:58:03 raspberrypi ntpd[1908]: Deleting interface #2 eth0, 192.168#123, interface stats: received=99, sent=103, dropped=0, active_time=1204 secs
...
Mar 14 09:58:03 raspberrypi kernel: [ 1267.800680] smsc95xx 1-1.1:1.0: eth0: Failed to write register index 0x00000014
Mar 14 09:58:03 raspberrypi kernel: [ 1267.800711] smsc95xx 1-1.1:1.0: eth0: Failed to write HW_CFG_LRST_ bit in HW_CFG register, ret = -110
...
kern.logMar 14 09:44:36 raspberrypi kernel: [ 460.821299] smsc95xx 1-1.1:1.0: eth0: Failed to read register index 0x00000114
Mar 14 09:56:41 raspberrypi kernel: [ 1185.918029] smsc95xx 1-1.1:1.0: eth0: Failed to read register index 0x00000114
... that 2nd one repeated every ~5 seconds until ...
Mar 14 09:58:03 raspberrypi kernel: [ 1267.800680] smsc95xx 1-1.1:1.0: eth0: Failed to write register index 0x00000014
Mar 14 09:58:03 raspberrypi kernel: [ 1267.800711] smsc95xx 1-1.1:1.0: eth0: Failed to write HW_CFG_LRST_ bit in HW_CFG register, ret = -110
... then more of the above until shutdown
So basically it's the nic/nic driver that's the problem (smsc95xx) and I've seen nic errors in the logs before without mining with my Icarus
