[RFC][PATCH 3/3] optimize follow_pages()

[Peter Zijlstra <a.p.zijlstra@chello.nl>]

Ofcourse I got the return paths tangled :-(

---

From: Peter Zijlstra <a.p.zijlstra@chello.nl>

Christoph Lameter suggested I pull set_page_dirty() out from under the 
pte lock.

I reviewed the current calls and found the one in follow_page() a candidate
for the same treatment.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>

---

 include/linux/mm.h |    1 +
 mm/memory.c        |   17 ++++++++++++++---
 2 files changed, 15 insertions(+), 3 deletions(-)

Index: linux-2.6/include/linux/mm.h
===================================================================
--- linux-2.6.orig/include/linux/mm.h	2006-05-10 08:11:18.000000000 +0200
+++ linux-2.6/include/linux/mm.h	2006-05-10 08:12:47.000000000 +0200
@@ -1015,6 +1015,7 @@ struct page *follow_page(struct vm_area_
 #define FOLL_TOUCH	0x02	/* mark page accessed */
 #define FOLL_GET	0x04	/* do get_page on page */
 #define FOLL_ANON	0x08	/* give ZERO_PAGE if no pgtable */
+#define FOLL_DIRTY	0x10	/* the page was dirtied */
 
 #ifdef CONFIG_PROC_FS
 void vm_stat_account(struct mm_struct *, unsigned long, struct file *, long);
Index: linux-2.6/mm/memory.c
===================================================================
--- linux-2.6.orig/mm/memory.c	2006-05-10 08:12:40.000000000 +0200
+++ linux-2.6/mm/memory.c	2006-05-10 08:16:33.000000000 +0200
@@ -962,18 +962,28 @@ struct page *follow_page(struct vm_area_
 	if (unlikely(!page))
 		goto unlock;
 
-	if (flags & FOLL_GET)
+	if (flags & (FOLL_GET | FOLL_TOUCH))
 		get_page(page);
 	if (flags & FOLL_TOUCH) {
 		if ((flags & FOLL_WRITE) &&
 		    !pte_dirty(pte) && !PageDirty(page))
+			flags |= FOLL_DIRTY;
+	}
+
+	pte_unmap_unlock(ptep, ptl);
+
+	if (flags & FOLL_TOUCH) {
+		if (flags & FOLL_DIRTY)
 			set_page_dirty(page);
 		mark_page_accessed(page);
 	}
+	if (!(flags & FOLL_GET))
+		put_page(page);
+	goto out;
+
 unlock:
 	pte_unmap_unlock(ptep, ptl);
-out:
-	return page;
+	goto out;
 
 no_page_table:
 	/*
@@ -986,6 +996,7 @@ no_page_table:
 			get_page(page);
 		BUG_ON(flags & FOLL_WRITE);
 	}
+out:
 	return page;
 }
 


--
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>