Re: Tree-SSA self checking infrastructure

[Andrew MacLeod <amacleod at redhat dot com>]

On Wed, 2003-11-19 at 16:29, law@redhat.com wrote:
> In message <20031119205044.GL11681@kam.mff.cuni.cz>, Jan Hubicka writes:
>  >> In message <20031119203450.GS16923@atrey.karlin.mff.cuni.cz>, Jan Hubicka w
>  >rite
>  >> s:
>  >>  >> In message <20031119190649.GQ16923@atrey.karlin.mff.cuni.cz>, Jan Hubic
>  >ka w
>  >>  >rite
>  >>  >> s:
>  >>  >>  >This brings me into questions about my tail call updating code.
>  >>  >>  >Perhaps I need to re-do SSA form on call cobbered variables after
>  >>  >>  >removing the call?
>  >>  >> Are you changing the CFG?  Are you changing the dominator tree?  Those
>  >>  >I replace tail call by edge back to the beggining of function and I
>  >>  >possibly split edge fron entry point.  I also take care to construct PHI
>  >>  >nodes for all referenced arguments by hand.
>  >>  >This does change CFG, but does not change dominance relationships.
>  >>  >I am however removing call and perhaps we are maintaining something that
>  >>  >relies on the presence of call...
>  >> Since you're not changing the dominance relationships, I think you are
>  >> safe.  Presumably you create a PHI node for every variable that is set
>  >> inside the newly created loop which is also live at the head of the loop.
>  >> Right?
>  >Hmm, well, does the case of uninitialized variables count?  Can I detect
>  >these somehow?
> Yes, I think they do.  This has always been a little fuzzy to me, but
> I believe folks have argued that you want a PHI if there is a join point
> where the uninitialized version meets with some initialized version.
> 
> jeffa
> 

An SSA version which is uninitialized (apparently there can be only one)
for any given variable can be found via:

          d = default_def (var);



So if the default definition of a variable is an empty stmt, that means
it has no DEF.

so...


 if (EMPTY_STMT_P (SSA_NAME_DEF_STMT (ssa_name))
    {  
       /* This means there is no definition of ssa_var */

       if (default_def (SSA_NAME_VAR (ssa_name)) == ssa_name)
          {
	    /* This is the uninitialized ssa version for the variable.*/
          }
       else
          {
 	    /* This is an error condition. This is not the default
	       definintion, nor is it defined.  */
          }
    }


I thinki its owrks that way. That shjow we check in live_on_entry in
tree-ssa-live.c

If you dont like it, talk to Diego, thats the way he implelemted it
after I pestered him for a long time in order to be able to detect these
:-)

Andrew