This is the mail archive of the mailing list for the elfutils project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [patch v7 3/5] x86* unwinder: libdwfl/

On Mon, 28 Oct 2013 11:13:38 +0100, Mark Wielaard wrote:
> On Sun, 2013-10-27 at 15:29 +0100, Jan Kratochvil wrote:
> > Valgrind would always warn if the application did not set it.  Even if the
> > appplication never used it.  It would warn on lines:
> >   if (! process->callbacks->set_initial_registers (thread,
> >                                                    thread->callbacks_arg))
> >         process->callbacks->thread_detach (thread, thread->callbacks_arg);
> > 
> > as CALLBACKS_ARG is passed there by value, not by reference.
> BTW. If you configure --with-valgrind then make check runs all tests
> under valgrind.


It fails on:
backtrace: backtrace.c:445: selfdump: Assertion `plt != ((void *)0)' failed.

The testcase looks a bit at some ELF internals...  Maybe the testcase could
use some asm volatile () instead of adapting to what compiler produced.

> I don't think valgrind memcheck would warn in the above case for passing
> around "undefined values". It only warns when those undefined bits are
> actually used in a condition or end up being passed as syscall arguments
> it knows should have defined values.

OK, you are right, I did not expect that.  It can safely pass undefined values
in registers.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]