[LTP] [PATCH] syscalls/madvise06: Add madvise WILLNEED performance problem tag

xuyang2018.jy@fujitsu.com xuyang2018.jy@fujitsu.com
Wed Aug 4 03:23:12 CEST 2021


Hi Cyril
> Hi!
>> diff --git a/testcases/kernel/syscalls/madvise/madvise06.c b/testcases/kernel/syscalls/madvise/madvise06.c
>> index 63d8d5452..f2f65ff96 100644
>> --- a/testcases/kernel/syscalls/madvise/madvise06.c
>> +++ b/testcases/kernel/syscalls/madvise/madvise06.c
>> @@ -35,6 +35,12 @@
>>    * faults. Two faults are allowed incase some tasklet or something
>>    * else unexpected, but irrelevant procedure, registers a fault to
>>    * our process.
>> + *
>> + * It also can reproduce the MADV_WILLNEED preformance problem.
>> + * It was introduced since 5.9 kernel with the following commit
>> + *   e6e88712e43b ("mm: optimise madvise WILLNEED")
>> + * and fixed since 5.10-rc5 kernel with the following commit
>> + *   66383800df9c ("mm: fix madvise WILLNEED performance problem").
>>    */
>>
>>   #include<errno.h>
>> @@ -242,6 +248,8 @@ static struct tst_test test = {
>>   	.tags = (const struct tst_tag[]) {
>>   		{"linux-git", "55231e5c898c"},
>>   		{"linux-git", "8de15e920dc8"},
>> +		{"linux-git", "e6e88712e43b"},
>> +		{"linux-git", "66383800df9c"},
>
> We only put patches that fixed the problem into the tags since we use
> these to print failure hints about missing kernel patches.
Oh, yes. Thanks for your review, I merged this patch with removing 
introduced patch tag.

Best Regards
Yang Xu
>
> Other than that it looks fine.
>
>>   		{}
>>   	}
>>   };
>> --
>> 2.23.0
>>
>>
>> --
>> Mailing list info: https://lists.linux.it/listinfo/ltp
>


More information about the ltp mailing list