Re: [PATCH 04/10] sched/debug: Increase SCHEDSTAT_VERSION to 16

From: Swapnil Sapkal <Swapnil.Sapkal@amd.com>
To: Shrikanth Hegde <sshegde@linux.ibm.com>
Cc: linux-kernel@vger.kernel.org,
	Dietmar Eggemann <dietmar.eggemann@arm.com>,
	Linus Torvalds <torvalds@linux-foundation.org>,
	Peter Zijlstra <peterz@infradead.org>,
	Valentin Schneider <vschneid@redhat.com>,
	Vincent Guittot <vincent.guittot@linaro.org>,
	"Gautham R. Shenoy" <gautham.shenoy@amd.com>,
	Ingo Molnar <mingo@kernel.org>
Subject: Re: [PATCH 04/10] sched/debug: Increase SCHEDSTAT_VERSION to 16
Date: Fri, 15 Mar 2024 19:33:33 +0530	[thread overview]
Message-ID: <fafaf2ec-a997-4802-88a8-1c1c9e3b60f5@amd.com> (raw)
In-Reply-To: <40325480-238a-4839-98e5-018e1890ab78@linux.ibm.com>

On 3/15/2024 2:50 PM, Shrikanth Hegde wrote:
> On 3/15/24 2:36 PM, Swapnil Sapkal wrote:
>> Hello Shrikanth,
>>
>> Thank you for reviewing.
>>
>> On 3/15/2024 10:07 AM, Shrikanth Hegde wrote:
>>>> 19 20
>>>> 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36
>>>> +domain<N> <cpumask> 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
>>>> 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43
>>>> 44 45
>>>>      The first field is a bit mask indicating what cpus this domain
>>>> operates
>>>> over.
>>>>    
>>> IIUC, this is editing the content of Version 15, But changes would
>>> happen on Version 16.
>>> Instead can we add the section for Version 16 and not modify for 15?
>> This file contains the field details of current schedstat version and
>> the short
>> summary of what change across the versions. Maintaining each versions
>> details will
>> increase the file size. Instead I will add the links to previous version's
>> documentation.
>>   
> I hadn't seen that. what you are saying is right. it would bloat up.
>
>> Thoughts on this?
>>
>>>> -The next 24 are a variety of sched_balance_rq() statistics in grouped
>>>> into types
>>>> +The next 33 are a variety of sched_balance_rq() statistics in grouped
>>>> into types
>>>>    of idleness (idle, busy, and newly idle):
> Please change this to
>   of idleness (busy, idle, and newly idle):

Thanks for catching this. I have updated this in v2. v2 is available at:
https://lore.kernel.org/all/20240315135501.1778620-1-swapnil.sapkal@amd.com/

>>>>          1)  # of times in this domain sched_balance_rq() was called
>>>> when the
>>>> +        cpu was busy
>>>> +    2)  # of times in this domain sched_balance_rq() checked but
>>>> found the
>>>> +        load did not require balancing when busy

--
Thanks and Regards,
Swapnil