On 07/01/2011 05:21 PM, Michael Scheidell wrote: > > > On 6/30/11 7:35 AM, Török Edwin wrote: >> The patch above does 2 things: >> - if pthread_cond_timedwait returns any error (other than ETIMEDOUT) it >> logs it, and breaks the loop >> - make sure pthread_cond_timedwait's timeout parameter is valid until >> pthread_cond_timedwait wakes up >> > so far, so good. > still running.. with the side effect of using a LOT less threads than 0.97 or > 0.97.1. also seems to use a lot less ram (doesn't start to use un non > recoverable ram). I assume due to less threads > active. > > by now, I would have peeked at 14 threads (13 idle, one running stats) when I > looked at it with clamdtop.
It probably peaked at 14 threads before because all 14 threads became deadlocked (and inflooping when pthread_cond_timedwait returned error). > > now, I have 4 (3 idle and one running stats) Sounds good. I will disable the JIT bytecode engine in 0.97.1 with the next daily.cvd update (0.97, 0.96.x and master will not be affected). 0.97.1 should then fall back to interpreter, so that others who don't have the patch applied don't get their clamds hung. Best regards, --Edwin _______________________________________________ Help us build a comprehensive ClamAV guide: visit http://wiki.clamav.net http://www.clamav.net/support/ml
