On freshly installed gentoo system I've compiled some program that I was using for months on FC3. Both systems have guile 1.6.4. Most important difference is that gentoo was compiled with -fstack-protector CFLAGS option put into /etc/make.conf. I've tried to emerge guile without this option and compile my program with -fno-stack-protector option but it didn't change anything: my program aborts on function scm_init_guile(); Reproducible: Always Steps to Reproduce: 1. create some file, say libguile.c: #include <libguile.h> #include <stdio.h> int main() { printf("Initializing Scheme... "); fflush(stdout); scm_init_guile(); printf("OK\n"); return 0; } 2. compile it: cc -Wall -o libguile libguile.c `guile-config compile` `guile-config link` 3. run: ./libguile Actual Results: Initializing Scheme... Aborted Expected Results: Initializing Scheme... OK $ uname -a Linux bajka 2.4.28-hardened-r3 #2 SMP Thu Jan 27 11:22:23 CET 2005 i686 Pentium Pro GenuineIntel GNU/Linux $ cat /etc/make.conf CFLAGS="-O2 -mcpu=i686 -fomit-frame-pointer -fstack-protector" CHOST="i386-pc-linux-gnu" CXXFLAGS="${CFLAGS}" MAKEOPTS="-j1" $ cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 1 model name : Pentium Pro stepping : 9 cpu MHz : 199.439 cache size : 256 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov bogomips : 398.13 processor : 1 vendor_id : GenuineIntel cpu family : 6 model : 1 model name : Pentium Pro stepping : 9 cpu MHz : 199.439 cache size : 256 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov bogomips : 398.13
I've just checked the dmesgs and found: grsec: signal 6 sent to /home/newchief/libguile[libguile:14194] uid/euid:1000/1000 gid/egid:407/407, parent /bin/bash[bash:19261] uid/euid:1000/1000 gid/egid:407/407 grsec: attempted resource overstep by requesting 4096 for RLIMIT_CORE against limit 0 by /home/newchief/libguile[libguile:14194] uid/euid:1000/1000 gid/egid:407/407, parent /bin/bash[bash:19261] uid/euid:1000/1000 gid/egid:407/407 That's the reason for the nasty behavior.
Please don't file bugs under the Gentoo Developer Relations bug product unless you need to log a human resources issue. File bugs under the Gentoo Linux product.
could you do a strace to verify/invalidate that rlimit is the cause of the signal, and that it doesn't come from a signal-handler inside libguile.
need more info before we can verify this bug.
During recent kernel upgrade I've chosen less intrusive options of grsecurity and now I'm able to use libguile depending programs, so I think, case is closed. I was suspecting stack-protector to be the reason of the failure, but it wasn't. The only problem was the grsecurity kernel options concerning executables.
