bug#18032: Crash with etags

bug#18032: Crash with etags

[Paul Eggert]

[This bug report is taken from:
http://lists.gnu.org/archive/html/emacs-devel/2014-07/msg00209.html
I'm filing it here to give it a proper bug number and thread.
Dr. Werner Fink <werner@suse.de> writes:]

Hi,

after installing the linux kernel sources, I see

      cd /usr/src/linux
      ulimit -c unlimited
      make TAGS
        GEN     TAGS
      xargs: etags: terminated by signal 11
      gdb /usr/bin/etags core
      [...]
      Reading symbols from /usr/bin/etags...Reading symbols from 
/usr/lib/debug/usr/bin/etags.debug...done.
      done.
      [New LWP 14768]
      Core was generated by `etags -a 
--regex=^\(ENTRY\|_GLOBAL\)(\([^)]*\)).* \2 / --regex=^SYSCALL_DEFINE['.
      Program terminated with signal SIGSEGV, Segmentation fault.
      #0  __memcpy_sse2_unaligned () at 
../sysdeps/x86_64/multiarch/memcpy-sse2-unaligned.S:37
      37      ../sysdeps/x86_64/multiarch/memcpy-sse2-unaligned.S: No 
such file or directory.
      (gdb) up
      #1  0x000000000040376c in memcpy (__len=90, __src=<optimized out>, 
__dest=0x18ea090) at /usr/include/bits/string3.h:51
      51        return __builtin___memcpy_chk (__dest, __src, __len, 
__bos0 (__dest));
      (gdb) up
      #2  substitute (in=0x17048f0 "ENTRY(lguest_entry)", out=0x1707301 
"2", regs=0x1705dc8, regs=0x1705dc8)
          at /usr/src/debug/emacs-24.3/lib-src/etags.c:5792
      5792            memcpy (t, in + regs->start[dig], diglen);
      (gdb) list
      5787      for (t = result; *out != '\0'; out++)
      5788        if (*out == '\\' && ISDIGIT (*++out))
      5789          {
      5790            dig = *out - '0';
      5791            diglen = regs->end[dig] - regs->start[dig];
      5792            memcpy (t, in + regs->start[dig], diglen);
      5793            t += diglen;
      5794          }
      5795        else
      5796          *t++ = *out;
      (gdb) print *out - '0'
      $1 = 2
      (gdb) print regs->end[dig] - regs->start[dig]
      value has been optimized out
      (gdb) print regs->end[dig]
      value has been optimized out
      (gdb) print regs
      $2 = (struct re_registers *) 0x1705dc8
      (gdb) print *regs
      $3 = {num_regs = 4, start = 0x18d1d40, end = 0x18d4fc0}
      (gdb) print *regs->end
      $4 = 21474836499
      (gdb) print regs->end
      $5 = (regoff_t *) 0x18d4fc0
      (gdb) print *regs->end
      $6 = 21474836499
      (gdb) print *regs->start
      $7 = 0
      (gdb) print *regs->end[2]
      Cannot access memory at address 0x70000000000005f

I've also an other bug report about etags and kernel source tree, see
the cloned bug at http://rudin.suse.de:8888/show_bug.cgi?id=887483

Werner

-- "Having a smoking section in a restaurant is like having a peeing 
section in a swimming pool." -- Edward Burr

bug#18032: Crash with etags

[Lars Ingebrigtsen]

Paul Eggert <eggert@cs.ucla.edu> writes:

> [This bug report is taken from:
> http://lists.gnu.org/archive/html/emacs-devel/2014-07/msg00209.html
> I'm filing it here to give it a proper bug number and thread.
> Dr. Werner Fink <werner@suse.de> writes:]

[...]

>      Reading symbols from /usr/bin/etags...Reading symbols from
>      /usr/lib/debug/usr/bin/etags.debug...done.
>      done.
>      [New LWP 14768]
>      Core was generated by `etags -a
>      --regex=^\(ENTRY\|_GLOBAL\)(\([^)]*\)).* \2 /
>     --regex=^SYSCALL_DEFINE['.
>      Program terminated with signal SIGSEGV, Segmentation fault.
>      #0  __memcpy_sse2_unaligned () at
>       ../sysdeps/x86_64/multiarch/memcpy-sse2-unaligned.S:37
>      37      ../sysdeps/x86_64/multiarch/memcpy-sse2-unaligned.S: No
>      such file or directory.
>      (gdb) up
>      #1  0x000000000040376c in memcpy (__len=90, __src=<optimized
>       out>, __dest=0x18ea090) at /usr/include/bits/string3.h:51
>      51        return __builtin___memcpy_chk (__dest, __src, __len,
>      __bos0 (__dest));

This was five years ago, and unfortunately this wasn't handled at the
time.  But it's unlikely that we'll be making more progress here at this
point, so I'm closing this bug report.  Please reopen if you're still
seeing these crashes in modern versions of Emacs.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no