[strongSwan-dev] exception in malloc, mainly during list operations

[SM K]

Hi Gerd and Martin,

Thank you for your responses. The box does not have faulty memory, It has
crashes on multiple systems, while other applications havent. I believe it
is caused by an invalid free somewhere, but I cannot point where at. I have
run with some tools that could have detected such problems, but I did not
run into such a problem. I might have to put in more instrumentation to
detect invalid memory frees.

I wonder if this is caused by a race condition somewhere. I was thinking
for reducing the number of threads strongswan uses, at least until we can
find a reason for the crash. What is the minimum number of threads needed
for strongswan to run? I currently have defined 16 threads, and we run on a
16 core box.

-sk

On Fri, Mar 27, 2015 at 5:53 AM, Gerd v. Egidy <lists at egidy.de> wrote:

> > Hm, not sure what actually we could do wrong in this call. It is a plain
> > malloc() with 24 bytes. Either we have inconsistent state by some
> > invalid malloc/free use, or it could even be an internal malloc issue.
>
> another issue could be faulty memory. I have seen segfaults or other fatal
> signals sent to some processes in this case.
>
> Does your system use memory with ECC?
>
> Do other programs also segfault on this machine?
>
> Please run memtest ( http://www.memtest.org/ ) for a few hours on your
> system
> and see if it finds any memory errors.
>
> Kind regards,
>
> Gerd
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.strongswan.org/pipermail/dev/attachments/20150327/6f87e0f3/attachment.html>