shrink_mmap ()?

shrink_mmap ()?

[Chirayu Patel]

Hi,

I was going through the source for shrink_mmap.......

The attached code puzzled me..

We are freeing a page with count = 1 (referenced by one process only) but
we are not manipulating any page table entries. Why? Shouldnt we be
manipulating the page table entries or where are the page table entries
getting manipulated?

I know I have missed something terribly obvious over here. Can someone
please help me out. 

Thanks.

-- Chirayu

-----------------------------------------------------------------------
switch (atomic_read(&page->count)) {
	case 1:
		/* is it a swap-cache or page-cache page? */
		if (page->inode) {
			if (test_and_clear_bit(PG_referenced,
                                               &page->flags)) {
				touch_page(page);
				break;
			}
			age_page(page);
			if (page->age)
				break;
			if (PageSwapCache(page)) {
				delete_from_swap_cache(page);
				return 1;
			}
			remove_page_from_hash_queue(page);
			remove_page_from_inode_queue(page);
			__free_page(page);
			return 1;
		}
------------------------------------------------------------------------------

Re: shrink_mmap ()?

[Rik van Riel]

On Thu, 26 Mar 1998, Chirayu Patel wrote:

> I was going through the source for shrink_mmap.......
> 
> We are freeing a page with count = 1 (referenced by one process only) but
> we are not manipulating any page table entries. Why? Shouldnt we be
> manipulating the page table entries or where are the page table entries
> getting manipulated?
> 
> I know I have missed something terribly obvious over here. Can someone
> please help me out. 

When the count is 1, the page cache _is_ the only reference to
the page.

Rik.
+-------------------------------------------+--------------------------+
| Linux: - LinuxHQ MM-patches page          | Scouting       webmaster |
|        - kswapd ask-him & complain-to guy | Vries    cubscout leader |
|     http://www.fys.ruu.nl/~riel/          | <H.H.vanRiel@fys.ruu.nl> |
+-------------------------------------------+--------------------------+