f2fs: fix to add refcount once page is tagged PG_private

As Gao Xiang reported in bugzilla:

https://bugzilla.kernel.org/show_bug.cgi?id=202749

f2fs may skip pageout() due to incorrect page reference count.

The problem here is that MM defined the rule [1] very clearly that
once page was set with PG_private flag, we should increment the
refcount in that page, also main flows like pageout(), migrate_page()
will assume there is one additional page reference count if
page_has_private() returns true.

But currently, f2fs won't add/del refcount when changing PG_private
flag. Anyway, f2fs should follow MM's rule to make MM's related flows
running as expected.

[1] https://lore.kernel.org/lkml/2b19b3c4-2bc4-15fa-15cc-27a13e5c7af1@aol.com/

Reported-by: Gao Xiang <gaoxiang25@huawei.com>
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
This commit is contained in:
Chao Yu 2019-03-06 17:30:59 +08:00 committed by Jaegeuk Kim
parent 25720cc05e
commit 240a59156d
6 changed files with 36 additions and 23 deletions

View file

@ -191,8 +191,7 @@ void f2fs_register_inmem_page(struct inode *inode, struct page *page)
f2fs_trace_pid(page);
set_page_private(page, (unsigned long)ATOMIC_WRITTEN_PAGE);
SetPagePrivate(page);
f2fs_set_page_private(page, (unsigned long)ATOMIC_WRITTEN_PAGE);
new = f2fs_kmem_cache_alloc(inmem_entry_slab, GFP_NOFS);
@ -280,8 +279,7 @@ next:
ClearPageUptodate(page);
clear_cold_data(page);
}
set_page_private(page, 0);
ClearPagePrivate(page);
f2fs_clear_page_private(page);
f2fs_put_page(page, 1);
list_del(&cur->list);
@ -370,8 +368,7 @@ void f2fs_drop_inmem_page(struct inode *inode, struct page *page)
kmem_cache_free(inmem_entry_slab, cur);
ClearPageUptodate(page);
set_page_private(page, 0);
ClearPagePrivate(page);
f2fs_clear_page_private(page);
f2fs_put_page(page, 0);
trace_f2fs_commit_inmem_page(page, INMEM_INVALIDATE);