This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC 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: Is this a bug in libffi testsuite on solaris2.8?


> Date: Thu, 30 Oct 2003 22:13:26 +0100
> From: Andreas Tobler <toa@pop.agri.ch>
> To: Bradley Lucier <lucier@math.purdue.edu>
> 
> Bradley Lucier wrote:
> 
> > My systems administrators told me they are getting the following in in 
> > /var/adm/messages on the sparc-sun-solaris2.8 box I use to bootstrap and 
> > regression-test mainline every night:
> > 
> > Oct 30 04:58:37 banach.math.purdue.edu genunix: [ID 533030 kern.notice] 
> > NOTICE: unwindtest.exe[7957]  attempt to execute code on stack by uid <xxx>
> > 
> > and I happen to be the lucky owner of uid <xxx>.  The unwindtest.cc 
> > tests pass on libffi, but is this a bug in the testsuite, a bogus 
> > warning, or something else?

Recent versions of Solaris -- by default -- do not allow code on the
stack to be executed.  The intent is to make the system more secure as
you can't use a malicious buffer overflow of a stack buffer to modify
the return address (to a stack address), write arbitrary code, and
then execute it when the function returns.

BTW, GDB configured for Solaris used to execute code on the stack when
doing inferior function calls.  Recent GDB's don't.

I would argue that the warning is not bogus.


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