Re: [RFA]: LOC_COMPUTED + abstracted dwarf2 evaluator, again

[Jim Blandy <jimb@redhat.com>]

Daniel Berlin <dberlin@dberlin.org> writes:
> + /* Evaluate a location description, given in THEBLOCK, in the
> +    context of frame FRAME. */
> + static struct value *
> + evaluate_loc_desc (struct symbol *var, struct frame_info *frame,
> + 		   struct dwarf_block *theblock, struct type *type)
> + {
> +   CORE_ADDR result;
> +   struct value * retval;
> +   struct dwarf_expr_baton *baton = xmalloc (sizeof (struct dwarf_expr_baton));
> +   struct dwarf_expr_context *ctx;
> +   
> +   retval = allocate_value(type);
> +   baton->var = var;
> +   baton->frame = frame;
> +   
> +   VALUE_LVAL (retval) = lval_memory;
> +   VALUE_BFD_SECTION (retval) = SYMBOL_BFD_SECTION (var);
> +   
> +   ctx = new_dwarf_expr_context ();
> +   ctx->read_reg_baton = baton;
> +   ctx->read_reg = dwarf_expr_read_reg;
> +   ctx->read_mem_baton = baton;
> +   ctx->read_mem = dwarf_expr_read_mem;
> +   ctx->get_frame_base_baton = baton;
> +   ctx->get_frame_base = dwarf_expr_frame_base;
> +   ctx->error = error;
> +   
> +   dwarf_expr_eval (ctx, theblock->data, theblock->size);
> +   
> +   if (ctx->in_reg)
> +     {
> +       store_typed_address (VALUE_CONTENTS_RAW (retval), 
> + 			   SYMBOL_TYPE (var), dwarf_expr_fetch (ctx, 0));
> +       VALUE_LVAL (retval) = not_lval;
> +     }
> +   else
> +     {
> +       result = dwarf_expr_fetch (ctx, 0);
> +       VALUE_LAZY (retval) = 1;
> +       VALUE_ADDRESS (retval) = result;
> +     }

This looks wrong.  If evaluating the Dwarf location expression yields
an `in_reg' result, shouldn't the value be an lval_register, with its
the address set to the register number?

I think the evaluator is confused in this regard, too; see my upcoming
post about that.