Re: [patch PR39806] Potential fix for spurious code generation

[Olga Golovanevsky <OLGA at il dot ibm dot com>]

As for now none of us can reproduce this bug any more.
What should I do: commit or drop?

Regtested on ppc-suse-linux64.

Thanks,
Olga


Olga Golovanevsky/Haifa/IBM wrote on 26/11/2009 17:29:28:

> This patch is a potential fix for spurious code generation due to
> struct reorg (as suggested by Andrew).
> As I cannot reproduce it on ppc, please check it on your platform.
>
> Thanks,
> Olga
>
> Index: ipa-struct-reorg.c
> ===================================================================
> --- ipa-struct-reorg.c (revision 154644)
> +++ ipa-struct-reorg.c (working copy)
> @@ -569,7 +569,7 @@ static new_var
>  is_in_new_vars_htab (tree decl, htab_t new_vars_htab)
>  {
>    return (new_var) htab_find_with_hash (new_vars_htab, decl,
> -     htab_hash_pointer (decl));
> +     DECL_UID (decl));
>  }
>
>  /* Given original variable ORIG_VAR, this function returns
> @@ -1962,7 +1962,7 @@ add_to_new_vars_htab (new_var new_node,
>    void **slot;
>
>    slot = htab_find_slot_with_hash (new_vars_htab, new_node->orig_var,
> -       htab_hash_pointer (new_node->orig_var),
> +       DECL_UID (new_node->orig_var),
>         INSERT);
>    *slot = new_node;
>  }
> @@ -2254,15 +2254,19 @@ create_new_var (tree var_decl, htab_t ne
>  static hashval_t
>  new_var_hash (const void *x)
>  {
> -  return htab_hash_pointer (((const_new_var)x)->orig_var);
> +  return DECL_UID (((const_new_var)x)->orig_var);
>  }
>
> -/* This function returns nonzero if orig_var of new_var X is equal to Y.
*/
> +/* This function returns nonzero if orig_var of new_var X
> +   and tree Y have equal UIDs.  */
>
>  static int
>  new_var_eq (const void *x, const void *y)
>  {
> -  return ((const_new_var)x)->orig_var == (const_tree)y;
> +  if (DECL_P ((const_tree)y))
> +    return DECL_UID (((const_new_var)x)->orig_var) == DECL_UID
> ((const_tree)y);
> +  else
> +    return 0;
>  }
>
>  /* This function check whether a structure type represented by STR