oclHashcat-plus fails to stop cleanly
#1
oclHashcat-plus 0.15 runs fine most of the time.

When a GPU temp limit or any other thing causes Hashcat to abort, CPU usage of the oclHashcat-plus process uses nearly all available CPU

kill -s SIGINT does nothing
SIGTERM causes kernel messages to start appearing:
kernel:[345325.053414] BUG: soft lockup - CPU#3 stuck for 23s!
kernel:[345496.877879] BUG: soft lockup - CPU#1 stuck for 24s!
etc.
Eventually oclHashcat-plus goes zombie according to top but still consumes >100% of CPU and the machine has to be rebooted to get back to a responsive state.

Environment:
oclHashcat-plus64 v0.15
Suse 12.3
Catalyst 13.4
3xGigabyte HD7970
AMD FX6300

Usually, I'm going through WPA hashes when this happens, haven't seen it any other time yet...
Most recently I was typing in the wrong terminal and killed it. (I was typing 'ping' Tongue ) and it said GPU limit reached even though temps were fine.
#2
You should probably report this in Trac. This sounds a bit like a bug.
#3
(11-12-2013, 06:39 PM)unix-ninja Wrote: You should probably report this in Trac. This sounds a bit like a bug.

done.
To add to this, now GPU#1 isn't working... guess I need to reseat the cards.
#4
it sounds like an asic hang to me.