Re: [RFC PATCH 1/5] x86/Kconfig: Add Page Cache Accounting entry

[Liu Yuan <namei.unix@xxxxxxxxx>]

On Thu, Mar 3, 2011 at 12:24 AM, Randy Dunlap <rdunlap@xxxxxxxxxxxx> wrote:

On Wed,  2 Mar 2011 16:38:06 +0800 Liu Yuan wrote:

> From: Liu Yuan <tailai.ly@taobao.com>
>
> Signed-off-by: Liu Yuan <tailai.ly@taobao.com>
> ---
>  arch/x86/Kconfig.debug |    9 +++++++++
>  1 files changed, 9 insertions(+), 0 deletions(-)
>
> diff --git a/arch/x86/Kconfig.debug b/arch/x86/Kconfig.debug
> index 615e188..f29e32d 100644
> --- a/arch/x86/Kconfig.debug
> +++ b/arch/x86/Kconfig.debug
> @@ -304,4 +304,13 @@ config DEBUG_STRICT_USER_COPY_CHECKS
>
>         If unsure, or if you run an older (pre 4.4) gcc, say N.
>
> +config PAGE_CACHE_ACCT
> +     bool "Page cache accounting"
> +     ---help---
> +       Enabling this options to account for page cache hit/missed number of
> +       times. This would allow user space applications get better knowledge
> +       of underlying page cache system by reading virtual file. The statitics
> +       per partition are collected.
> +
> +       If unsure, say N.
>  endmenu
> --

rewrite:

         Enable this option to provide for page cache hit/miss counters.
         This allows userspace applications to obtain better knowledge of the
         underlying page cache subsystem by reading a virtual file.
         Statistics are collect per partition.

questions:
       what virtual file?
       what kind of partition?

Hi Randy,

Thanks for your correction.

'virtual file' in this patch context means files in sysfs mounted at /sys.
'partition' indicates that every disk partition (like /dev/sda/{sda1,sda2...} has its own accountings for page cache information.

I am not confident about phrasing. so please correct it if you think it is way better.

Thanks,
Liu Yuan