Bug #9666

Segmentation fault while printing out C level backtrace information, when $0 is set

Added by ReiOdaira (Rei Odaira) about 6 years ago. Updated over 5 years ago.

Target version:
ruby -v:
ruby 2.2.0dev (2014-03-24) [powerpc64-linux]


This might be related to [Bug #9654], but when $0 is set, a segmentation fault happens while printing out C-level backtrace. Due to this issue, TestRubyOptions#test_segv_setproctitle fails in my environment (ppc64 linux).

ruby -e '$0="foo.rb"; Process.kill :SEGV, $$'
-e:1: [BUG] Segmentation fault at 0x001f80
ruby 2.2.0dev (2014-03-24) [powerpc64-linux]

-- Control frame information -----------------------------------------------
c:0003 p:---- s:0009 e:000008 CFUNC  :kill
c:0002 p:0021 s:0004 E:0006f4 EVAL   -e:1 [FINISH]
c:0001 p:0000 s:0002 E:000cd4 TOP    [FINISH]

-- Ruby level backtrace information ----------------------------------------
-e:1:in `<main>'
-e:1:in `kill'

-- C level backtrace information -------------------------------------------
foo.rb [0x20774264]
foo.rb [0x207e250c]
foo.rb(rb_bug+0xc4) [0x207e2844]
foo.rb [0x206e64f0]
(__kernel_sigtramp_rt32+0x0) [0x100360]
foo.rb [0x20782ee8]
foo.rb(rb_f_kill+0x98) [0x206e74f8]
foo.rb [0x2075294c]
foo.rb [0x2075c5ac]
foo.rb [0x20771c60]
foo.rb [0x207699a8]
foo.rb [0x2076d8c8]
foo.rb(rb_iseq_eval_main+0x2f8) [0x2076def8]
foo.rb [0x20611884]
foo.rb(ruby_run_node+0xa4) [0x206136c4]
foo.rb [0x2060f77c]
/lib/ fault

Here is the stack trace at the second segmentation fault.

(gdb) bt
#0  0x2030a994 in strlen () from /lib/
#1  0x2085ce70 in kvprintf (fmt=0x208f0c45 "+0x%lx) [0x%lx] %s/%s:%d\n")
    at addr2line.c:1014
#2  kprintf (fmt=0x208f0c45 "+0x%lx) [0x%lx] %s/%s:%d\n") at addr2line.c:776
#3  0x2085e8d8 in rb_dump_backtrace_with_lines (num_traces=18, 
    traces=0x2096790c, syms=0x20c27190) at addr2line.c:678
#4  0x2084428c in rb_print_backtrace () at vm_dump.c:690
#5  rb_vm_bugreport () at vm_dump.c:825
#6  0x208b250c in report_bug (file=<value optimized out>, 
    line=<value optimized out>, fmt=0x208e88dc "Segmentation fault at %p", 
    args=0x209d0034) at error.c:312
#7  0x208b2844 in rb_bug (fmt=0x208e88dc "Segmentation fault at %p")
    at error.c:339
#8  0x207b64f0 in sigsegv (sig=<value optimized out>, info=0x209d00c0, 
    ctx=<value optimized out>) at signal.c:704
#9  <signal handler called>
#10 0x202b674c in kill () from /lib/
#11 0x20852ef4 in ruby_kill (pid=<value optimized out>, 
    sig=<value optimized out>) at thread.c:5185
<<<<< snip >>>>>

Again, line->sname points to some out-of-range address.

Updated by naruse (Yui NARUSE) about 6 years ago

  • Status changed from Open to Closed

It is because glibc's dladdr doesn't return absolute path.
r45394 fixes this with /proc/self/exe.
(this insists it cannot get symbols where ruby cannot fetch the absolute path of the executable like AIX)

Updated by usa (Usaku NAKAMURA) over 5 years ago

  • Backport changed from 2.0.0: UNKNOWN, 2.1: UNKNOWN to 2.0.0: DONTNEED, 2.1: DONTNEED

Also available in: Atom PDF