Re: ptrace PEEKTEXT IO error?!

[Daniel Jacobowitz <drow@false.org>]

On Wed, Oct 05, 2005 at 07:20:11PM +0100, David Lecomber wrote:
> On Wed, 2005-10-05 at 13:11 -0400, Daniel Jacobowitz wrote:
> > On Wed, Oct 05, 2005 at 06:04:00PM +0100, David Lecomber wrote:
> > > Dear all,
> > > 
> > > I have a multithreaded code which is misbehaving in GDB (verified on
> > > latest CVS).  After threads are created, the target suddenly seems to
> > > become unwriteable, and even unreadable in some bits.
> > 
> > This usually means the thread is not stopped.  Failing that, check with
> > your kernel.
> 
> 'ps' says the processes is stopped - so I guess that makes it a kernel
> woe.
> 
> I've knocked up a ptrace test code, and it also shows the same behaviour
> independently of GDB.   
> 
> It seems ok at reading memory from addresses on the heap and stack --
> but it barfs at reading memory from the text segment.  Is there
> something that the user code  (one of the libraries called by my
> program) could do to prevent access to memory by a ptracing program?

No, this is most likely a kernel bug.  This is mapped code, right, not
some mmaped device file (which is not generally accessible by ptrace)?

-- 
Daniel Jacobowitz
CodeSourcery, LLC