On Tue, 2020-12-08 at 10:39 +0800, Ian Kent wrote: > On Sun, 2020-12-06 at 06:49 +0800, Ian Kent wrote: > > On Sat, 2020-12-05 at 12:38 -0300, Andreas Hasenack wrote: > > > Hi, > > > > > > On Sat, Dec 5, 2020 at 6:32 AM Ian Kent <raven@xxxxxxxxxx> wrote: > > > > > > > Removing the LDFLAGS option -Wl,-Bsymbolic-functions makes the > > > > problem go away. > > > > > > Related: > > > https://www.spinics.net/lists/autofs/msg01903.html > > > > > > In Ubuntu, -Wl,-Bsymbolic-functions is a system-wide default and > > > we > > > remove it from the flags when building autofs in d/rules: > > > > > > # Taken from #1636781 as it's the same problem here. Also see > > > # > > > https://lists.ubuntu.com/archives/ubuntu-devel/2018-June/040369.html > > > ifeq ($(shell dpkg-vendor --derives-from Ubuntu && echo yes),yes) > > > export DEB_LDFLAGS_MAINT_STRIP = -Wl,-Bsymbolic-functions > > > endif > > > > > > > > > > I'm not sure why this happens, the best I could find is this: > > > > https://stackoverflow.com/questions/7216973/is-there-a-downside-to-using-bsymbolic-functions/20729291 > > > > The unfortunate fact is that automount is a threaded application, > > and > > it does allocate and initialize that configuration variable earlier > > than any accesses from functions in lib/defaults.c, so the > > configuration variable should not be NULL in any thread at some > > later point. > > > > Admittedly that initialization function return should be checked > > but I don't see how that can cause this. The validity of this > > variable shouldn't need to be checked elsewhere in the config. > > handling functions. > > > > It's fine to call this a bug in autofs but what is the bug, IDGT? > > Is it certain the optimisation option isn't broken in some way? > > Thinking about this further it may be due to my use of a static > library that is linked to each object, automount and the shared > lookup, parse and mount shared libraries. > > Still, this problem occurring entirely within in the main() function, > before any shared libraries are used where only the fork() to become > a daemon has occurred when this crash occurs, makes me wonder about > this link option. > > It may be that to fix this I need to change that static library to a > shared library. I'm not sure how much work that would be but I > suspect > it would be quite a bit ... or maybe the use that link option is > broken > and so misguided in itself ... So making the static library a shared library fixes this and, if I haven't made any mistakes that haven't showed up yet, seems fairly straight forward to do. Ian
