On 2017-12-14 22:35, Stefan Nilsson wrote: > Hey Laurent, > > I have been using s6 for some time now to bring up and monitor services > in a custom embedded system. The system is built using buildroot and > (due to customer requirements) built with glibc 2.26. > [snip] > > When buildroot 2017.11 was released, I was thrilled to see new versions > of s6 in it (s6-rc 0.3.0.0, skalibs 220.127.116.11 and s6 18.104.22.168), but when > booting up I immediately got a SIGFPE in s6-rc-init,... I ran it through > strace and the result can be seen here: http://steffe.net/s6-sigfpe.txt > > It crashes at the exact same place if I re-run it, and it crashes every > time.
Quite interesting that you can consistently reproduce a crash there. The tail of that strace, from the open(.../supervise/control)->ENOENT onwards, exactly matches the crash I've reported at http://skarnet.org/cgi-bin/archive.cgi?1:msp:1033:jfbcmmfpnlnhnahkgkfh but: I haven't been able to find a way to reproduce it, and both the two occurrences I've seen have different symptoms, namely assert()s inside glibc failing [hence the open(/dev/tty) onwards is from glibc's assert handler, and can be ignored]. But that was with glibc 2.24, and I suppose it's possible that some code has been reorganized so that you're hitting a div-by-0 instead. Of course, it's also possible that the crashes are unrelated, but that seems unlikely. Rasmus