[PATCH] Fix PR ipa/65557

[Uros Bizjak]

On Tue, May 12, 2015 at 4:54 PM, Martin Liška <mliska@suse.cz> wrote:
> On 04/30/2015 11:11 PM, Uros Bizjak wrote:
>> Hello!
>>
>>> 2015-03-29  Martin Liska  <mliska@suse.cz>
>>>
>>>    PR ipa/65557
>>>    * ipa-icf.c (sem_function::equals_wpa): Check if IPA CP
>>>    has already filled up function summary.
>>>    (sem_item_optimizer::update_hash_by_addr_refs): Likewise.
>>>
>>> gcc/testsuite/ChangeLog:
>>>
>>> 2015-03-29  Martin Liska <mliska@suse.cz>
>>>
>>>    * g++.dg/ipa/pr65557.C: New test.
>>
>> --- /dev/null
>> +++ b/gcc/testsuite/g++.dg/ipa/pr65557.C
>> @@ -0,0 +1,19 @@
>> +/* { dg-do compile } */
>> +/* { dg-options "-O2 -fdump-ipa-icf-details"  } */
>> +
>> +struct S0
>> +{
>> +  S0 ()
>> +  {
>> +  }
>> +};
>> +
>> +struct S1
>> +{
>> +  S1 ()
>> +  {
>> +  }
>> +};
>> +
>> +S0 s0;
>> +S1 s1;
>>
>> The testcase doesn't clean its IPA dump. However, there is also no
>> scan dump function, so it is questionable, what the testcase tries to
>> do with the dump. Also, the flags that trigger the bug in the PR are
>> different: "-fdevirtualize -fipa-cp -fipa-icf-functions".
>>
>> Can you please check the testcase?
>>
>> Uros.
>>
>
> Hi.
>
> The test case caused ICE before the patch was applied. So removing -fdump*
> is the right fix for the test.
>
> I hope the patch is obvious. Should I also apply the patch for 5.1.0 branch?

Yes, please. Plenty of efforts were put in the cleaning of various dumps.

Thanks,
Uros.