[RFC][PATCH] nameing reserved pages [1/3]

[RFC][PATCH] nameing reserved pages [1/3]

[KAMEZAWA Hiroyuki]

[-- Attachment #1: Type: text/plain, Size: 43 bytes --]

inline functions for naming pages.
-- Kame

[-- Attachment #2: name_reserved.patch --]
[-- Type: text/plain, Size: 1804 bytes --]


Adding page_type definitions and funcs for naming reserved pages.

Reserved page's information is stored into page->private.

This is a weak naming method and anyone can overwrite it. 

This information is used in /dev/memstate in following patch.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>


---

 linux-2.6.12-rc2-kamezawa/include/linux/mm.h |   31 +++++++++++++++++++++++++++
 1 files changed, 31 insertions(+)

diff -puN include/linux/mm.h~name_reserved include/linux/mm.h
--- linux-2.6.12-rc2/include/linux/mm.h~name_reserved	2005-04-20 09:37:48.000000000 +0900
+++ linux-2.6.12-rc2-kamezawa/include/linux/mm.h	2005-04-20 10:38:01.000000000 +0900
@@ -348,6 +348,37 @@ static inline void put_page(struct page 
 #endif		/* CONFIG_HUGETLB_PAGE */
 
 /*
+ * Type of Pages. This is used in /dev/memstate.
+ * value range is 0-255.
+ */
+enum page_type {
+	Page_Common = 0,
+	Min_Reserved_Types = 1,
+	Rserved_Unknwon = 1,
+	Reserved_At_Boot,
+	Max_Reserved_Types,
+	Page_Invalid = 0xff
+};
+/*
+ * Basically, page->private has no meaning without PG_private.
+ * Here, we use page->private for PG_reserved pages to record type of a page.
+ * Because a page is reserved, anyone will not modify page->private.
+ * When it is freed, page->private will be overwritten by some code.
+ */
+static inline void set_page_reserved(struct page *page, unsigned char type)
+{
+	SetPageReserved(page);
+	page->private = type;
+}
+
+static inline unsigned char reserved_page_type(struct page *page)
+{
+	if (!PageReserved(page))
+		return 0;
+	return (unsigned char)page->private;
+}
+
+/*
  * Multiple processes may "see" the same page. E.g. for untouched
  * mappings of /dev/null, all processes see the same page full of
  * zeroes, and text pages of executables and shared libraries have

_

Re: [RFC][PATCH] nameing reserved pages [1/3]

[Pavel Machek]

Hi!

> inline functions for naming pages.
> -- Kame

> 
> Adding page_type definitions and funcs for naming reserved pages.
> 
> Reserved page's information is stored into page->private.
> 
> This is a weak naming method and anyone can overwrite it. 
> 
> This information is used in /dev/memstate in following patch.
> 
> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
> 
> 
> ---
> 
>  linux-2.6.12-rc2-kamezawa/include/linux/mm.h |   31 +++++++++++++++++++++++++++
>  1 files changed, 31 insertions(+)
> 
> diff -puN include/linux/mm.h~name_reserved include/linux/mm.h
> --- linux-2.6.12-rc2/include/linux/mm.h~name_reserved	2005-04-20 09:37:48.000000000 +0900
> +++ linux-2.6.12-rc2-kamezawa/include/linux/mm.h	2005-04-20 10:38:01.000000000 +0900
> @@ -348,6 +348,37 @@ static inline void put_page(struct page 
>  #endif		/* CONFIG_HUGETLB_PAGE */
>  
>  /*
> + * Type of Pages. This is used in /dev/memstate.
> + * value range is 0-255.
> + */
> +enum page_type {
> +	Page_Common = 0,
> +	Min_Reserved_Types = 1,
> +	Rserved_Unknwon = 1,
         ~
	  missing e?

> +	Reserved_At_Boot,
> +	Max_Reserved_Types,
> +	Page_Invalid = 0xff
> +};

You certainly use unusual naming convention here. Could we get
reserved_at_boot instead? (I.e. all lowercase).


> +/*
> + * Basically, page->private has no meaning without PG_private.
> + * Here, we use page->private for PG_reserved pages to record type of a page.
> + * Because a page is reserved, anyone will not modify page->private.
> + * When it is freed, page->private will be overwritten by some code.
> + */
> +static inline void set_page_reserved(struct page *page, unsigned char type)
> +{

Make it enum page_type type.

								Pavel
-- 
Boycott Kodak -- for their patent abuse against Java.

Re: [RFC][PATCH] nameing reserved pages [1/3]

[KAMEZAWA Hiroyuki]

Hi!
Pavel Machek wrote:
>          ~
> 	  missing e?
Oh, it's typo :(. thanks!

> 
> 
>>+	Reserved_At_Boot,
>>+	Max_Reserved_Types,
>>+	Page_Invalid = 0xff
>>+};
> 
> 
> You certainly use unusual naming convention here. Could we get
> reserved_at_boot instead? (I.e. all lowercase).
> 
Okay.
> 
> 
>>+static inline void set_page_reserved(struct page *page, unsigned char type)
>>+{
> 
> 
> Make it enum page_type type.

I'll do.

Thanks
-- Kame <kamezawa.hiroyu@jp.fujitsu.com>