On 5/7/19 4:44 PM, Richard Biener wrote:
> On May 7, 2019 4:33:08 PM GMT+02:00, "Martin Liška" <mli...@suse.cz> wrote:
>> On 5/7/19 2:56 PM, Richard Biener wrote:
>>> But that can use the existing get_hot_bb_threshold since we never
>> want
>>> to dump -1 in case min_count was never initialized.
>>
>> Yes. But the function will call:
>>
>> get_hot_bb_threshold ()
>> {
>> if (min_count == -1)
>> {
>> gcov_type t = profile_info->sum_max / PARAM_VALUE
>> (HOT_BB_COUNT_FRACTION);
>> set_hot_bb_threshold (t);
>> ...
>>
>> which will cause a segfault in non-PGO run. Note that:
>> static gcov_type min_count = -1;
>>
>> is a non-exported variable so that's why I simply added the getter.
>>
>> Hope it's fine as is?
>
> Oh, I see. Hmm, so we should get away with no min_coubt when all counter
> kinds are non-pgo?
Yes, in that situation we don't want to print:
/* --param=gimple-fe-computed-hot-bb-threshold=xyz */
Martin
>
> Richard.
>
>>
>> Martin
>
