This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [GOOGLE] recalculate dominance before update_ssa in AutoFDO pass


ok.

David

On Wed, Oct 15, 2014 at 4:00 PM, Dehao Chen <dehao@google.com> wrote:
> It's destroyed by gimple_value_profile_transformations.
>
> Patch updated to move the update_ssa early.
>
> Index: gcc/auto-profile.c
> ===================================================================
> --- gcc/auto-profile.c (revision 216278)
> +++ gcc/auto-profile.c (working copy)
> @@ -1674,7 +1674,14 @@ afdo_annotate_cfg (const stmt_set &promoted_stmts)
>        counts_to_freqs ();
>      }
>    if (flag_value_profile_transformations)
> -    gimple_value_profile_transformations ();
> +    {
> +      gimple_value_profile_transformations ();
> +      free_dominance_info (CDI_DOMINATORS);
> +      free_dominance_info (CDI_POST_DOMINATORS);
> +      calculate_dominance_info (CDI_POST_DOMINATORS);
> +      calculate_dominance_info (CDI_DOMINATORS);
> +      update_ssa (TODO_update_ssa);
> +    }
>  }
>
>  /* Wrapper function to invoke early inliner.  */
> @@ -1756,7 +1763,6 @@ auto_profile (void)
>        early_inline ();
>        autofdo::afdo_annotate_cfg (promoted_stmts);
>        compute_function_frequency ();
> -      update_ssa (TODO_update_ssa);
>
>        /* Local pure-const may imply need to fixup the cfg.  */
>        if (execute_fixup_cfg () & TODO_cleanup_cfg)
>
> Dehao
>
> On Wed, Oct 15, 2014 at 10:50 AM, Xinliang David Li <davidxl@google.com> wrote:
>> Is it destroyed by value profile transformations? Can you move the
>> dominance recomputing code closer to where it gets invalidated?
>>
>> David
>>
>> On Wed, Oct 15, 2014 at 10:37 AM, Dehao Chen <dehao@google.com> wrote:
>>> This patch recalculates dominance info before update_ssa call in
>>> AutoFDO. This fixes bug when dominance info is out-of-date and causes
>>> segfaults during update_ssa.
>>>
>>> Bootstrapped and regression test on-going.
>>>
>>> OK for google-4_9 branch?
>>>
>>> Thanks,
>>> Dehao
>>>
>>> Index: gcc/auto-profile.c
>>> ===================================================================
>>> --- gcc/auto-profile.c (revision 216278)
>>> +++ gcc/auto-profile.c (working copy)
>>> @@ -1756,6 +1756,11 @@ auto_profile (void)
>>>        early_inline ();
>>>        autofdo::afdo_annotate_cfg (promoted_stmts);
>>>        compute_function_frequency ();
>>> +
>>> +      free_dominance_info (CDI_DOMINATORS);
>>> +      free_dominance_info (CDI_POST_DOMINATORS);
>>> +      calculate_dominance_info (CDI_POST_DOMINATORS);
>>> +      calculate_dominance_info (CDI_DOMINATORS);
>>>        update_ssa (TODO_update_ssa);
>>>
>>>        /* Local pure-const may imply need to fixup the cfg.  */


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]