Problem 1
---------

        Draw  a supergraph  of the  following program  (i.e. a  graph in
        which we include  call and return edges  to show interprocedural
        control flow) and observe the  dumps generated by Liveness based
        pointer analysis pass.

        1) What  is  the pointer  information  after  the call  made  to
           function  f and  why?  Do you  find  any pointer  information
           missing? Why?

        2) Why does the pointer information  `x->p' not reach beyond the
           second  call to  function f()  in main?  Explain in  terms of
           paths in the supergraph.

        3) Why does  the pointer information `x->q'  reach beyond second
           call to function f() in main but is not propagated beyond the
           first  call  to  f()?  Explain  in  terms  of  paths  in  the
           supergraph.

Program
------- 

        int *x, *y, p, q;
        
        void f()
        {
         if(p)
           p++; 
        }
        
        int main()
        {
          x = &p;
          f();
          printf("%d",x);
          x = &q;
          f();
          y = &q;
          printf("%d%d%d%d",x,y,p,q);
        
        }
  
Procedure 
---------

        1) Compile   the    test   program   with    following   command
           $(INSTALL)/bin/gcc -O2 -flipta -fdump-ipa-lipta test.c

        2) Observe the *.lipta dump generated
        
        3) `livein' and `liveout' represent  the liveness information at
           a  basic block.  `mayin' and  `mayout' represent  the pointer
           information of the basic  block. Callstring represent calling
           context reaching `f'.