Re: Tapset difficulties w/ pointers

["James Dickens" <jamesd dot wi at gmail dot com>]

On 4/28/06, Stone, Joshua I <joshua.i.stone@intel.com> wrote:
> Hi all,
>
> In working on the 'process' tapset, I've encountered a few difficulties
> that I thought I would share, so we can search the collective mind for
> solutions. I've split this into two emails to separate the related
> parts...
>
> For the process.exec probe, I would like to make the pointer to the new
> task_struct available.  This is fine, and then I can make
> access-functions for the user to extract data from this, like so:
>
>   function task_pid:long (task:long) %{ /* pure */
>       if (unlikely(THIS->task == 0))
>           THIS->__retvalue = 0;
>       else {
>           struct task_struct *t = (struct task_struct
> *)(long)THIS->task;
>           THIS->__retvalue = t->tgid;
>       }
>   %}
>
> The obvious problem here is safety.  I can check that the parameter is
> not zero, but what do I do if the user calls task_pid(0xDEADBEEF)?
> Oops...  Or what happens if the user stashes the pointer away in a
> global, and tries to read it later?  Oops...
simple answer, anytime you use a pointer systemtap's framework must be
prepared to fail, you have just scratched the surface of the way it
can fail, its not pretty but you have to catch all the error traps
while running your code and then do your best to back out any changes
your made and exit and leave the system in a runable state, if all
this can be done in a macro, that would be awesome else you will need
pointer access functions.

James Dickens
uadmin.blogspot.com



> One solution is to "hide" task_pid(), perhaps by renaming it to
> __task_pid(), and then have the tapset provide variables for pid, tid,
> etc. and rely on the compiler to elide those that are unused.  Then the
> pointer never needs to be given to the user.  This is still not really
> safe though, because that __task_pid() is only discouraged, not
> protected.  We probably could play a game with naming that only allows
> calling some functions from within the same file, but these functions
> would need to be duplicated multiple tapsets needs them.
>
> A more general solution would be if we had some sort of "handle" type.
> This might be an abstract data type that the script language is unable
> to modify.  It can have metadata to identify the type, and perhaps also
> to limit the valid lifetime of the pointer.  I haven't thought through
> all of the implementation details, but I think that a handle type would
> not be too far a stretch.
>
> I welcome any feedback on this idea, and of course other ideas are
> welcome as well.
>
>
> Josh