I was just about to post how I've been running 2.6.1 with my BFL singles for 2 days straight without problem, and then I caught these errors. I restarted cgminer and it's working again for now.
This is almost certainly caused by Kano's reduction of the timeout from 30 seconds to 0.1 second... I skimmed over this one and assumed it was 1 second (which is at least reasonable). Bet this is what was causing the other problems people reported too... thanks for the tip
I'll fix it for the next BFGMiner.
If my change is the cause of all the problems then you only need to do a single one line change of driver_bitforce.c line 41
The value '1' - change it
So if that's the cause of it all as you are implying then you will have a fix ready in a few minutes ...
Anyway,
at least try to understand the old code
On windows it was 30 seconds
On linux is was either 0.1 seconds or never (HANG if BFL didn't reply)
(yeah a bug)
If you entered 'bitforce:' on the front in linux it wouldn't set a timeout (HANG on no reply)
If you didn't put 'bitforce:' on the front in linux, Icarus would set a 0.1s timeout
(bitforce still didn't set a timeout of course but the Icarus one affected the bitforce code)
I changed the code coz whenever it got a comms error, it would hang trying to flush the buffer coz I use 'bitforce:'
(my BFL device is crappy and thus I get to see what happens with a comms error sometimes)
