| Message ID | 1267534262-13196-1-git-send-email-dmonakhov@openvz.org |
|---|---|
| State | Not Applicable, archived |
| Headers | show |
On Tue 02-03-10 15:51:02, Dmitry Monakhov wrote: > - There is theoretical possibility to perform writepage on > RO superblock. Add explicit check for what case. > - Page must being locked before writepage. Thanks. Merged into my tree. Honza > > Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org> > --- > fs/ext3/inode.c | 7 +++++++ > 1 files changed, 7 insertions(+), 0 deletions(-) > > diff --git a/fs/ext3/inode.c b/fs/ext3/inode.c > index 8ac4acf..14d40a4 100644 > --- a/fs/ext3/inode.c > +++ b/fs/ext3/inode.c > @@ -1528,6 +1528,7 @@ static int ext3_ordered_writepage(struct page *page, > int err; > > J_ASSERT(PageLocked(page)); > + WARN_ON_ONCE(IS_RDONLY(inode)); > > /* > * We give up here if we're reentered, because it might be for a > @@ -1600,6 +1601,9 @@ static int ext3_writeback_writepage(struct page *page, > int ret = 0; > int err; > > + J_ASSERT(PageLocked(page)); > + WARN_ON_ONCE(IS_RDONLY(inode)); > + > if (ext3_journal_current_handle()) > goto out_fail; > > @@ -1642,6 +1646,9 @@ static int ext3_journalled_writepage(struct page *page, > int ret = 0; > int err; > > + J_ASSERT(PageLocked(page)); > + WARN_ON_ONCE(IS_RDONLY(inode)); > + > if (ext3_journal_current_handle()) > goto no_write; > > -- > 1.6.6 >
diff --git a/fs/ext3/inode.c b/fs/ext3/inode.c index 8ac4acf..14d40a4 100644 --- a/fs/ext3/inode.c +++ b/fs/ext3/inode.c @@ -1528,6 +1528,7 @@ static int ext3_ordered_writepage(struct page *page, int err; J_ASSERT(PageLocked(page)); + WARN_ON_ONCE(IS_RDONLY(inode)); /* * We give up here if we're reentered, because it might be for a @@ -1600,6 +1601,9 @@ static int ext3_writeback_writepage(struct page *page, int ret = 0; int err; + J_ASSERT(PageLocked(page)); + WARN_ON_ONCE(IS_RDONLY(inode)); + if (ext3_journal_current_handle()) goto out_fail; @@ -1642,6 +1646,9 @@ static int ext3_journalled_writepage(struct page *page, int ret = 0; int err; + J_ASSERT(PageLocked(page)); + WARN_ON_ONCE(IS_RDONLY(inode)); + if (ext3_journal_current_handle()) goto no_write;
- There is theoretical possibility to perform writepage on RO superblock. Add explicit check for what case. - Page must being locked before writepage. Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org> --- fs/ext3/inode.c | 7 +++++++ 1 files changed, 7 insertions(+), 0 deletions(-)