On 25.03.22 21:09, Andrew Morton wrote:
> On Thu, 24 Mar 2022 10:43:33 +0000 cgel.zte@xxxxxxxxx wrote:
>
>> From: Yang Yang <yang.yang29@xxxxxxxxxx>
>>
>> Users may use ksm by calling madvise(, , MADV_MERGEABLE) when they want
>> to save memory, it's a tradeoff by suffering delay on ksm cow. Users can
>> get to know how much memory ksm saved by reading
>> /sys/kernel/mm/ksm/pages_sharing, but they don't know what's the costs
>> of ksm cow, and this is important of some delay sensitive tasks.
>>
>> So add ksm cow events to help users evaluate whether or how to use ksm.
>
> It's unclear (to me) how anyone will actually use this, how they will
> interpret the output.
>
> Some tutorial words added to Documentation/vm/ksm.rst would be helpful.
> While in there, please check for any other /proc/vmstat fields which
> we forgot to document.
>
>> --- a/include/linux/vm_event_item.h
>> +++ b/include/linux/vm_event_item.h
>> @@ -131,6 +131,7 @@ enum vm_event_item { PGPGIN, PGPGOUT, PSWPIN, PSWPOUT,
>> SWAP_RA_HIT,
>> #ifdef CONFIG_KSM
>> KSM_SWPIN_COPY,
>> + COW_KSM,
>
> I agree that this name looks unpleasingly backwards. Do we have an
> expectation that we actually will be adding more COW_* fields?
As raised previously (also when proposing this), I'd like to have
COW_ANON, COW_ZERO, COW_OTHER. Ideally, we'd have added all via a single
patch for them. They would at least be of value to me.
--
Thanks,
David / dhildenb
