From: "zhangyi (F)" <yi.zhang@xxxxxxxxxx> [ Upstream commit c044f3d8360d2ecf831ba2cc9f08cf9fb2c699fb ] If we free a metadata buffer which has been failed to async write out in the background, the jbd2 checkpoint procedure will not detect this failure in jbd2_log_do_checkpoint(), so it may lead to filesystem inconsistency after cleanup journal tail. This patch abort the journal if free a buffer has write_io_error flag to prevent potential further inconsistency. Signed-off-by: zhangyi (F) <yi.zhang@xxxxxxxxxx> Link: https://lore.kernel.org/r/20200620025427.1756360-5-yi.zhang@xxxxxxxxxx Signed-off-by: Theodore Ts'o <tytso@xxxxxxx> Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx> --- fs/jbd2/transaction.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c index 1478512ecab3e..cfbf5474bccab 100644 --- a/fs/jbd2/transaction.c +++ b/fs/jbd2/transaction.c @@ -1990,6 +1990,7 @@ int jbd2_journal_try_to_free_buffers(journal_t *journal, { struct buffer_head *head; struct buffer_head *bh; + bool has_write_io_error = false; int ret = 0; J_ASSERT(PageLocked(page)); @@ -2014,11 +2015,26 @@ int jbd2_journal_try_to_free_buffers(journal_t *journal, jbd_unlock_bh_state(bh); if (buffer_jbd(bh)) goto busy; + + /* + * If we free a metadata buffer which has been failed to + * write out, the jbd2 checkpoint procedure will not detect + * this failure and may lead to filesystem inconsistency + * after cleanup journal tail. + */ + if (buffer_write_io_error(bh)) { + pr_err("JBD2: Error while async write back metadata bh %llu.", + (unsigned long long)bh->b_blocknr); + has_write_io_error = true; + } } while ((bh = bh->b_this_page) != head); ret = try_to_free_buffers(page); busy: + if (has_write_io_error) + jbd2_journal_abort(journal, -EIO); + return ret; } -- 2.25.1
