Re: [tree-ssa] Merge results as of 2003-05-06

[law at redhat dot com]

In message <1052408596.13871.2068.camel@p4>, Andrew MacLeod writes:
 >> So what I do is set my used bit during the walk over the tree structures
 >> that occurs in create_ssa_var_map.    At first I only marked variables
 >> which appeared as a use or def.  But I quickly ran into variables which
 >> appeared only in vuse/vdefs such as operands to asms, so I have to walk
 >> and mark the virtual operands as well.
 >> 
 >
 >The out_of_ssa_tag will be set on anything that went through the
 >partitioning process. That amounts to anything which is a USE, a DEF, or
 >occurs in a PHI node somewhere. I guess its possible that there will be
 >variables which are used like this, but not in the partition map. I
 >suppose you can have all ASM VDEF's added to the partition during
 >create_ssa_var_map... then they'd have that bit set for sure.
So I went back and tried to see what objects weren't being marked with
the out_of_ssa_tag that I need marked.  Here's an example:
typedef int DItype __attribute__ ((mode (DI)));
typedef union
{
  DItype ll;
} DWunion;
DItype
__negdi2 (DItype u)
{
  DWunion w;
  return w.ll;
}

Which turns into:

__negdi2 (u)
{
  union DWunion w;

  
  # BLOCK 0 (k.c:10).  PRED: -1.  SUCC: -2.
  
  #   DItype_3 = VDEF <DItype_2>;
  #   VUSE <w_1>;
  return w.ll;
}


The VAR_DECL for "w" is never marked with the out_of_ssa_tag.

Anyway, we need to look at the VUSE operands to ensure that w gets marked
as being used.

I don't know what further work you've got planned, but at least for now
I don't think we can really rely on the out_of_ssa_tag bit to tell us
what variables were used since we care about variables which are effectively
ignored by the out-of-ssa pass.

jeff