On Wed, 2006-12-13 at 12:40 -0500, Trond Myklebust wrote:
> On Wed, 2006-12-13 at 16:01 +0100, Peter Zijlstra wrote:
> > @@ -386,12 +380,13 @@ int invalidate_inode_pages2_range(struct
> > PAGE_CACHE_SIZE, 0);
> > }
> > }
> > - was_dirty = test_clear_page_dirty(page);
> > + was_dirty = PageDirty(page);
> > if (!invalidate_complete_page2(mapping, page)) {
> > if (was_dirty)
> > set_page_dirty(page);
> ^^^^^^^^^^^^^^^^^^^^^
>
> Why would you still need this?
Hmm yeah, that doesn't make sense. What kind of twist did I get my brain
into.
This should do.
Hmm, does this agree with the DIO path? I just noticed
generic_file_direct_IO() also uses invalidate_inode_pages2_range().
I think it does, its looks like all they want is to ensure nothing
remains in the pagecache after a write, which itself ensures nothing is
dirty to begin with.
---
Delay clearing the dirty page state till after we've invalidated the
page in invalidate_complete_page2(). This gives try_to_release_pages() a
chance to flush dirty data.
Signed-off-by: Peter Zijlstra <[EMAIL PROTECTED]>
---
fs/nfs/file.c | 2 --
mm/truncate.c | 14 ++------------
2 files changed, 2 insertions(+), 14 deletions(-)
Index: linux-2.6-git/fs/nfs/file.c
===================================================================
--- linux-2.6-git.orig/fs/nfs/file.c 2006-12-13 19:41:09.000000000 +0100
+++ linux-2.6-git/fs/nfs/file.c 2006-12-13 19:42:18.000000000 +0100
@@ -320,8 +320,6 @@ static int nfs_release_page(struct page
*/
if (!(gfp & __GFP_FS))
return 0;
- /* Hack... Force nfs_wb_page() to write out the page */
- set_page_dirty(page);
return !nfs_wb_page(page_file_mapping(page)->host, page);
}
Index: linux-2.6-git/mm/truncate.c
===================================================================
--- linux-2.6-git.orig/mm/truncate.c 2006-12-13 19:41:09.000000000 +0100
+++ linux-2.6-git/mm/truncate.c 2006-12-13 19:42:56.000000000 +0100
@@ -306,19 +306,14 @@ invalidate_complete_page2(struct address
if (PagePrivate(page) && !try_to_release_page(page, GFP_KERNEL))
return 0;
+ test_clear_page_dirty(page);
write_lock_irq(&mapping->tree_lock);
- if (PageDirty(page))
- goto failed;
-
BUG_ON(PagePrivate(page));
__remove_from_page_cache(page);
write_unlock_irq(&mapping->tree_lock);
ClearPageUptodate(page);
page_cache_release(page); /* pagecache ref */
return 1;
-failed:
- write_unlock_irq(&mapping->tree_lock);
- return 0;
}
/**
@@ -350,7 +345,6 @@ int invalidate_inode_pages2_range(struct
for (i = 0; !ret && i < pagevec_count(&pvec); i++) {
struct page *page = pvec.pages[i];
pgoff_t page_index;
- int was_dirty;
lock_page(page);
if (page->mapping != mapping) {
@@ -386,12 +380,8 @@ int invalidate_inode_pages2_range(struct
PAGE_CACHE_SIZE, 0);
}
}
- was_dirty = test_clear_page_dirty(page);
- if (!invalidate_complete_page2(mapping, page)) {
- if (was_dirty)
- set_page_dirty(page);
+ if (!invalidate_complete_page2(mapping, page))
ret = -EIO;
- }
unlock_page(page);
}
pagevec_release(&pvec);
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
