Re: Will stack alignment be broken in the NetBSD-7 release?

[David Laight <david%l8s.co.uk@localhost>]

On Tue, Dec 23, 2014 at 05:41:41PM +0200, Andreas Gustafsson wrote:
> David Laight wrote:
> > Probably an asm function being used to call the destructors.
> > Can you get a stack traceback from within the destructor?
> 
> #0  0x00007f7ff6c3c240 in write () from /usr/lib/libc.so.12
> #1  0x00007f7ff68009de in Test2::~Test2 (this=<optimized out>, 
>     __in_chrg=<optimized out>)
>     at /tmp/bracket/build/2014.12.17.06.54.48-amd64-debug/src/tests/lib/csu/dso/h_initfini3_dso.cxx:30
> #2  0x00007f7ff6cf40be in __cxa_finalize (dso=0x7f7ff6a00e00 <__dso_handle>)
>     at /tmp/bracket/build/2014.12.17.06.54.48-amd64-debug/src/lib/libc/stdlib/atexit.c:198
> #3  0x00007f7ff6800888 in ?? () from /usr/tests/lib/csu/h_initfini3_dso.so
> #4  0x00007f7ff7ff94c0 in ?? ()
> #5  0x00007f7ff68009f9 in _fini () from /usr/tests/lib/csu/h_initfini3_dso.so
> #6  0x0000000000000000 in ?? ()

Hmmm...
I'd guess it is wtong before __cxa_finalise is called.
gdb should be able to display %rsp for each frame.
I'd also guess that traceback is wrong at the bottom, there might
not be the required debug info in an asm stub somewhere.
It might need the stack hand-decoding to find the source code for each
call.
But I'm not in a position to do that at the moment.

	David

-- 
David Laight: david%l8s.co.uk@localhost