[PATCH] Updates for page migration
This adds some additional comments in order to help others figure out how
exactly the code works. And fix a variable name.
Also swap_page does need to ignore all reference bits when unmapping a
page. Otherwise we may have to repeatedly unmap a frequently touched page.
So change the try_to_unmap parameter to 1.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 5a61080..5db32fd 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -632,7 +632,7 @@
struct address_space *mapping = page_mapping(page);
if (page_mapped(page) && mapping)
- if (try_to_unmap(page, 0) != SWAP_SUCCESS)
+ if (try_to_unmap(page, 1) != SWAP_SUCCESS)
goto unlock_retry;
if (PageDirty(page)) {
@@ -839,7 +839,7 @@
* pages are swapped out.
*
* The function returns after 10 attempts or if no pages
- * are movable anymore because t has become empty
+ * are movable anymore because to has become empty
* or no retryable pages exist anymore.
*
* Return: Number of pages not migrated when "to" ran empty.
@@ -928,12 +928,21 @@
goto unlock_both;
if (mapping->a_ops->migratepage) {
+ /*
+ * Most pages have a mapping and most filesystems
+ * should provide a migration function. Anonymous
+ * pages are part of swap space which also has its
+ * own migration function. This is the most common
+ * path for page migration.
+ */
rc = mapping->a_ops->migratepage(newpage, page);
goto unlock_both;
}
/*
- * Trigger writeout if page is dirty
+ * Default handling if a filesystem does not provide
+ * a migration function. We can only migrate clean
+ * pages so try to write out any dirty pages first.
*/
if (PageDirty(page)) {
switch (pageout(page, mapping)) {
@@ -949,9 +958,10 @@
; /* try to migrate the page below */
}
}
+
/*
- * If we have no buffer or can release the buffer
- * then do a simple migration.
+ * Buffers are managed in a filesystem specific way.
+ * We must have no buffers or drop them.
*/
if (!page_has_buffers(page) ||
try_to_release_page(page, GFP_KERNEL)) {
@@ -966,6 +976,11 @@
* swap them out.
*/
if (pass > 4) {
+ /*
+ * Persistently unable to drop buffers..... As a
+ * measure of last resort we fall back to
+ * swap_page().
+ */
unlock_page(newpage);
newpage = NULL;
rc = swap_page(page);