|
From: Kailash M. <kai...@ms...> - 2005-06-27 21:42:31
|
Hi, I am a first time user of Valgrind and I am trying to memcheck one of my daemons (some legacy code, some my own) on RHEL3 x86 (linux kernel 2.4). Setting up valgrind and running it worked without a cinch. Thanks for some pretty simple documentation! :-) The daemon relies heavily on NPTL, but nothing out of the ordinary. Just worker threads that are created, stack size is set to 256kb and they communicate over sockets and die very quickly. Both Unix and Internet sockets are used depending on the need. When I look at the output from valgrind, I see some messages regarding uninitialized variables being used in vfprintf() calls etc., After a while, logging stops and all I see is this message. It fills up the log. ==26502== Warning: bad signal number 65 in sigaction() ==26578== Warning: bad signal number 65 in sigaction() ==26674== Warning: bad signal number 65 in sigaction() ==26747== Warning: bad signal number 65 in sigaction() What is the meaning / implication? I know signal 65 is an invalid signo. Where is it coming from? Once valgrind starts spitting this message out, I don't see anything else although the daemon seems to be running happily. Thanks for any advice. Kailash |