Re: [Lhms-devel] Re: [RFC/PATCH] free_area[] bitmap elimination [3/3]

[Hiroyuki KAMEZAWA <kamezawa.hiroyu@jp.fujitsu.com>]

Dave Hansen wrote:

> On Tue, 2004-08-24 at 05:41, Hiroyuki KAMEZAWA wrote:
> 
>>+static inline int page_is_buddy(struct page *page, int order)
>>+{
>>+       if (page_count(page) == 0 &&
>>+           PagePrivate(page) &&
>>+           !PageReserved(page) &&
>>+            page_order(page) == order) {
>>+               /* check, check... see free_pages_check() */
>>+               if (page_mapped(page) ||
>>+                   page->mapping != NULL ||
>>+                   (page->flags & (
>>+                           1 << PG_lru |
>>+                           1 << PG_locked      |
>>+                           1 << PG_active      |
>>+                           1 << PG_reclaim     |
>>+                           1 << PG_slab        |
>>+                           1 << PG_swapcache |
>>+                           1 << PG_writeback )))
>>+                       bad_page(__FUNCTION__, page);
>>+               return 1;
>>+       }
>>+       return 0;
>>+}
> 
> 
> Please share some code with the free_pages_check() that you stole this
> from.  It's nasty enough to have one copy of it around. :)
Hmm... this part is different from free_pages_check() even if I stoled it from.
Becasuse PG_private bit check is not done here. Sharing some code with
frees_page_check() would make free_pages_check() complex to read.

And this is only a bug checking code and bad_page( __FUNCTION__ , page) is useful
to test this buddy system.

>>+#ifdef CONFIG_VIRTUAL_MEM_MAP  
>>+                       /* This check is necessary when
>>+                          1. there may be holes in zone.
>>+                          2. a hole is not aligned in this order.
>>+                          currently, VIRTUAL_MEM_MAP case, is only case.
>>+                          Is there better call than pfn_valid ?
>>+                       */
>>+                       if (!pfn_valid(zone->zone_start_pfn + (page_idx ^ (1 << order))))
>>+                               break;
>>+#endif         
> 
> 
> This should be hidden in a header somewhere.  We don't want to have to
> see ia64-specific ifdefs in generic code.  
> 
Hmm, I understand what you say. I'll consider better another way.
But why #ifdef is inserted here is that this is RFC and
I want to make it clear this is IA64 specific.

Thank you for your all comments.

-- Kame

-- 
--the clue is these footmarks leading to the door.--
KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>

--
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:"aart@kvack.org"> aart@kvack.org </a>