On Tue, 2 Dec 2003 10:38:45 -0500 (EST) Ed Phillips <[EMAIL PROTECTED]> wrote:
> Also, from another glance at the source, it appears that clamd will go
> into a "buzz-loop" waiting for a free thread slot if there are already
> MaxThreads threads active. Not the best solution for the "busiest"
> scenario... having the clamd parent process use as much CPU as
> possible when every thread is already busy... but technically, it
Fixed, thanks.
> works, I guess... although I don't see any mutex being used during
> inspection of ths[i].active near line 546 in clamd/server.c - multiple
> threads should use mutexes when accessing the same memory.
There's no need for a mutex there because ths[i].active changes are
atomic.
Best regards,
Tomasz Kojm
--
oo ..... [EMAIL PROTECTED] www.ClamAV.net
(\/)\......... http://www.clamav.net/gpg/tkojm.gpg
\..........._ 0DCA5A08407D5288279DB43454822DC8985A444B
//\ /\ Tue Dec 2 23:15:30 CET 2003
pgp00000.pgp
Description: PGP signature
