Re: __tls_get_addr problem with recent current

On Thu, Aug 21, 2008 at 05:58:44PM +0300, Kostik Belousov wrote:
> On Thu, Aug 21, 2008 at 04:28:13PM +0200, Bernd Walter wrote:
> > On Thu, Aug 21, 2008 at 04:41:26PM +0300, Kostik Belousov wrote:
> > > On Thu, Aug 14, 2008 at 10:29:49PM +0200, Bernd Walter wrote:
> > > > After updating the kernel and before installworld processes randomly
> > > > crash on startup.
> > > > Userland is ~2 months old amd64 current.
> > > > I wonder if this is because of the OS mix or a general problem right
> > > > now.
> > > > It seems that process only crash if I'm logged in via ssh/telnet but
> > > > not on console, so I might be able to installworld from console.
> > > > 
> > > > One example:
> > > > [59]cicely14# gdb /bin/cat cat.core 
> > > > GNU gdb 6.1.1 [FreeBSD]
> > > > Copyright 2004 Free Software Foundation, Inc.
> > > > GDB is free software, covered by the GNU General Public License, and you are
> > > > welcome to change it and/or distribute copies of it under certain conditions.
> > > > Type "show copying" to see the conditions.
> > > > There is absolutely no warranty for GDB.  Type "show warranty" for details.
> > > > This GDB was configured as "amd64-marcel-freebsd"...(no debugging symbols found)...
> > > > Core was generated by `cat'.
> > > > Program terminated with signal 11, Segmentation fault.
> > > > Reading symbols from /lib/libc.so.7...(no debugging symbols found)...done.
> > > > Loaded symbols for /lib/libc.so.7
> > > > Reading symbols from /libexec/ld-elf.so.1...(no debugging symbols found)...done.
> > > > Loaded symbols for /libexec/ld-elf.so.1
> > > > #0  0x0000000800505523 in __tls_get_addr () from /libexec/ld-elf.so.1
> > > > (gdb) bt
> > > > #0  0x0000000800505523 in __tls_get_addr () from /libexec/ld-elf.so.1
> > > > #1  0x00000008006b34c1 in free () from /lib/libc.so.7
> > > > #2  0x00000008006b4a8e in malloc () from /lib/libc.so.7
> > > > #3  0x00000008006a9600 in regfree () from /lib/libc.so.7
> > > > #4  0x00000008006a918e in regfree () from /lib/libc.so.7
> > > > #5  0x00000008006a6500 in __detect_path_locale () from /lib/libc.so.7
> > > > #6  0x0000000000401b6a in ?? ()
> > > > #7  0x0000000000401035 in ?? ()
> > > > #8  0x000000080052f000 in ?? ()
> > > 
> > > I do not see a problem with today kernel and world (on amd64).
> > > Does the problem persist even with both new kernel and world ?
> > 
> > Yes.
> > What puzzles me is that it only exists if initially log in as a normal
> > user, but not if I directly log in as root.
> > 
> > > If yes, you probably would need to rebuild rtld and libc with debug
> > > symbols and get the backtrace with full debug information.
> > 
> > Ok - I will retry with recent code.
> > If it still exists I'll send you the informations you asked for.
> 
> Does this user have anything `interesting' in profile ? Does the effect
> perstist if you log in as newly created user with clean environment and
> startup scripts ?

Not that I know about, but to be sure I'd also tested with a new user
and empty homedir as well and got the same problems.
login.conf is unmodified.
The only difference is that I use tcsh as shell, while root uses csh.
It is the same physical shell, but who knows how different it behaves.

-- 
B.Walter <bernd_at_bwct.de> http://www.bwct.de
Modbus/TCP Ethernet I/O Baugruppen, ARM basierte FreeBSD Rechner uvm.