[Raw Msg Headers][Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: smtpserver 2.99.43b still dying with SEGV

> I hate to say it, but the problem of smtpserver eventually dying (once
> every few days under heavy load) is still here.  This time, stack seems
> be corrupt.  Again, it *might* be because of my modifications dealing
> with on-the-fly charset translation, but might not.


	I am sending to you my developement tree smtpserver version.
	It SHOULD log to the  smtp-server log a line about  "accept()"
	and some error message if the accept() gets some unknown
	error status and dies due to that.

	I have experienced similar problems under heavy loads at
	Solaris 2.3, and Solaris 2.4.  No similar reports so far
	with 2.5(.1).

> =====
> crosser@ariel:/src/zmailer/zmailer-2.99.43/$ gdb smtpserver/smtpserver ~/core
> [...]
> GDB 4.12 (sparc-sun-solaris2.3), 
> Copyright 1994 Free Software Foundation, Inc...
> Core was generated by `/usr/zmailer/bin/smtpserver -asve -8X -l
> /var/log/zmailer/smtpserver'.
> Program terminated with signal 11, Segmentation fault.
> procfs (find_procinfo):  Couldn't locate pid 0
> #0  0xef6cef9c in s_buffer ()
> (gdb) bt
> #0  0xef6cef9c in s_buffer ()
> (gdb) 

	These addresses are -- "interesting".  Propably your
	GDB reports garbage, as s_buffer's address is around
	0x0003f000, and is propably the last named variable
	in the program.  Definitely it is not near the end of
	the address space.  That crash address is most likely
	in libc.    Hmm.. 2.5.1 has everything in user-space
	below 0xe0000000, perhaps that is in kernel space ?

> =====
> Any ideas?
> Eugene

	/Matti Aarnio <mea@nic.funet.fi> <mea@utu.fi>