PAGE_CACHE_SIZE vs. PAGE_SIZE

PAGE_CACHE_SIZE vs. PAGE_SIZE

[Kirill A. Shutemov]

[-- Attachment #1: Type: text/plain, Size: 726 bytes --]

Hi,

PAGE_CACHE_* macros were introduced long time ago in hope to implement
page cache with larger chunks than one page in future.

In fact it was never done.

Some code paths assume PAGE_CACHE_SIZE <= PAGE_SIZE. E.g. we use
zero_user_segments() to clear stale parts of page on cache filling, but
the function is implemented only for individual small page.

It's unlikely that global switch to PAGE_CACHE_SIZE > PAGE_SIZE will never
happen since it will affect to much code at once.

I think support of larger chunks in page cache can be in implemented in
some form of THP with per-fs enabling.

Is it time to get rid of PAGE_CACHE_* macros?
I can prepare patchset if it's okay.

-- 
 Kirill A. Shutemov

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: PAGE_CACHE_SIZE vs. PAGE_SIZE

[Andrew Morton]

On Fri, 18 Jan 2013 17:57:25 +0200
"Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> wrote:

> Hi,
> 
> PAGE_CACHE_* macros were introduced long time ago in hope to implement
> page cache with larger chunks than one page in future.
> 
> In fact it was never done.
> 
> Some code paths assume PAGE_CACHE_SIZE <= PAGE_SIZE. E.g. we use
> zero_user_segments() to clear stale parts of page on cache filling, but
> the function is implemented only for individual small page.
> 
> It's unlikely that global switch to PAGE_CACHE_SIZE > PAGE_SIZE will never
> happen since it will affect to much code at once.
> 
> I think support of larger chunks in page cache can be in implemented in
> some form of THP with per-fs enabling.
> 
> Is it time to get rid of PAGE_CACHE_* macros?
> I can prepare patchset if it's okay.

The distinct PAGE_CACHE_SIZE has never been used for anything, but I do
kinda like it for documentary reasons: PAGE_SIZE is a raw, low-level
thing and PAGE_CACHE_SIZE is the specialized
we're-doing-pagecache-stuff thing.

But I'm sure I could get used to not having it ;)

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: PAGE_CACHE_SIZE vs. PAGE_SIZE

[Michel Lespinasse]

On Fri, Feb 1, 2013 at 6:40 AM, Andrew Morton <akpm@linux-foundation.org> wrote:
> On Fri, 18 Jan 2013 17:57:25 +0200
> "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> wrote:
>
>> Hi,
>>
>> PAGE_CACHE_* macros were introduced long time ago in hope to implement
>> page cache with larger chunks than one page in future.
>>
>> In fact it was never done.
>>
>> Some code paths assume PAGE_CACHE_SIZE <= PAGE_SIZE. E.g. we use
>> zero_user_segments() to clear stale parts of page on cache filling, but
>> the function is implemented only for individual small page.
>>
>> It's unlikely that global switch to PAGE_CACHE_SIZE > PAGE_SIZE will never
>> happen since it will affect to much code at once.
>>
>> I think support of larger chunks in page cache can be in implemented in
>> some form of THP with per-fs enabling.
>>
>> Is it time to get rid of PAGE_CACHE_* macros?
>> I can prepare patchset if it's okay.
>
> The distinct PAGE_CACHE_SIZE has never been used for anything, but I do
> kinda like it for documentary reasons: PAGE_SIZE is a raw, low-level
> thing and PAGE_CACHE_SIZE is the specialized
> we're-doing-pagecache-stuff thing.
>
> But I'm sure I could get used to not having it ;)

Personally I always find such distinctions without a difference - like
page_cache_release vs put_page - rather confusing, especially when
working near the fs/mm boundary (for example in and under
handle_pte_fault())

-- 
Michel "Walken" Lespinasse
A program is never fully debugged until the last user dies.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: PAGE_CACHE_SIZE vs. PAGE_SIZE

[Simon Jeons]

On 01/18/2013 11:57 PM, Kirill A. Shutemov wrote:
> Hi,
>
> PAGE_CACHE_* macros were introduced long time ago in hope to implement
> page cache with larger chunks than one page in future.
>
> In fact it was never done.
>
> Some code paths assume PAGE_CACHE_SIZE <= PAGE_SIZE. E.g. we use
> zero_user_segments() to clear stale parts of page on cache filling, but
> the function is implemented only for individual small page.
>
> It's unlikely that global switch to PAGE_CACHE_SIZE > PAGE_SIZE will never
> happen since it will affect to much code at once.
>
> I think support of larger chunks in page cache can be in implemented in
> some form of THP with per-fs enabling.

IIRC, you try to implement THP support page cache, then PAGE_CACHE_SIZE 
maybe don't need any more.

>
> Is it time to get rid of PAGE_CACHE_* macros?
> I can prepare patchset if it's okay.
>

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>