Re: Cgraph alias reorg 15/14 (New infrastructure for same body aliases)

["H.J. Lu" <hjl.tools@gmail.com>]

On Sun, Jun 12, 2011 at 8:01 AM, H.J. Lu <hjl.tools@gmail.com> wrote:
> On Sun, Jun 12, 2011 at 7:54 AM, Jan Hubicka <hubicka@ucw.cz> wrote:
>>>
>>> This also pretty much destroyed C++ for ia32:
>>>
>>> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49378
>>> http://gcc.gnu.org/ml/gcc-regression/2011-06/msg00159.html
>>
>> Hi,
>> It seems somewhat amazing that we hit kernel sensitive miscompilation here.
>> The problem most probably is the fact that thunks and functions with thunks can become
>> local. This is correct since thunks are represented as direct calls now, but this
>> makes i386 to use local ABI when calling or compiling them.
>>
>> Does the following patch help? We may also need to look for the presence of thunk
>> callers.
>>
>> Index: ipa.c
>> ===================================================================
>> --- ipa.c       (revision 174958)
>> +++ ipa.c       (working copy)
>> @@ -120,6 +120,7 @@ static bool
>>  cgraph_non_local_node_p_1 (struct cgraph_node *node, void *data ATTRIBUTE_UNUSED)
>>  {
>>    return !(cgraph_only_called_directly_or_aliased_p (node)
>> +           && !ipa_ref_has_aliases_p (&node->ref_list)
>>            && node->analyzed
>>            && !DECL_EXTERNAL (node->decl)
>>            && !node->local.externally_visible
>> @@ -132,7 +133,11 @@ cgraph_non_local_node_p_1 (struct cgraph
>>  static bool
>>  cgraph_local_node_p (struct cgraph_node *node)
>>  {
>> -   return !cgraph_for_node_and_aliases (cgraph_function_or_thunk_node (node, NULL),
>> +   struct cgraph_node *n = cgraph_function_or_thunk_node (node, NULL);
>> +
>> +   if (n->thunk.thunk_p)
>> +     return false;
>> +   return !cgraph_for_node_and_aliases (n,
>>                                        cgraph_non_local_node_p_1, NULL, true);
>>
>>  }
>>
>
> I am testing it now.  Will know the results in 2 hours.
>

Yes, it fixes C++ regressions.

Thanks.

-- 
H.J.