2021-08-13 12:39:19

by Chris Down

Subject: Re: [PATCH v1] mm: Add configuration to control whether vmpressure notifier is enabled

[email protected] writes:
>From: wangyong <[email protected]>
>Inspired by PSI features, vmpressure inotifier function should also be
>configured to decide whether it is used, because it is an independent
>feature which notifies the user of memory pressure.
>
>Since the vmpressure interface is used in kernel common code, for
>users who do not use the vmpressure function, there will be
>additional overhead.

Could you please demonstrate this additional overhead with profiles or
demonstrations of other real world effects? Thanks.

2021-08-14 08:03:28

by Yong Wang

Subject: Re: [PATCH v1] mm: Add configuration to control whether vmpressure notifier is enabled

Chris Down <[email protected]> 于2021年8月13日周五 下午8:36写道：
>
> [email protected] writes:
> >From: wangyong <[email protected]>
> >Inspired by PSI features, vmpressure inotifier function should also be
> >configured to decide whether it is used, because it is an independent
> >feature which notifies the user of memory pressure.
> >
> >Since the vmpressure interface is used in kernel common code, for
> >users who do not use the vmpressure function, there will be
> >additional overhead.
>
> Could you please demonstrate this additional overhead with profiles or
> demonstrations of other real world effects? Thanks.

Thanks for your reply. In terms of code, it will reduce the operation
of some code.

To demonstrate this additional overhead, It is planned to use lmbench and
Christoph Lamenter’s pagefault tool (https://lkml.org/lkml/2006/8/29/294)
for comparative testing.

Thanks.