Position independent code and global destructor order (devel/ice)

Michael Gmelin freebsd at grem.de
Fri Dec 6 02:32:34 UTC 2013



On Thu, 5 Dec 2013 18:20:58 +0100
Michael Gmelin <freebsd at grem.de> wrote:

> On Thu, 5 Dec 2013 19:07:28 +0200
> Konstantin Belousov <kostikbel at gmail.com> wrote:
> 
> > If you provided link to the original discussion, it would take much
> > less energy and time to refresh the memory.
> 
> Forgot to include it, sorry. Thanks for you response and the patch.
> 
> > 
> > I looked over this, and I think that r211706 can be refined, by only
> > calling atexit hooks for 'wrong' dso when the call to
> > __cxa_finalize() is from the real dlclose(), as opposed to the
> > process exit.  For me, with the world where the patch is applied,
> > your 'major tom' test passes.
> 
> Makes sense.
> 
> > 
> > Of course, when dso1 registers __cxa_atexit hook located in dso2,
> > and dso2 is unloaded before dso1, the hooks are called in the wrong
> > order still, but the only alternatives there are either crash or do
> > not call such hooks at all.
> 
> Sounds like a good solution. PHP (and similar) crashes are avoided
> while clean code isn't affected. I will try the patch in my specific
> use case to make sure it works there as well.
> 

I tested your patch on 10-BETA2 and 10-BETA3. I built our complete
production environment from scratch, ran a huge number of unit and
integration tests. Everything works as expected, including Ice's
"Schwarz Counter" implementation
(http://en.wikibooks.org/wiki/More_C%2B%2B_Idioms/Nifty_Counter).

So as far as I can tell this fixes the problem, while still
accomplishing the goals of r211706.

Thanks for work,
Michael

-- 
Michael Gmelin


More information about the freebsd-ports mailing list