commit e259221763a40403d5bb232209998e8c45804ab8 ("fs: simplify the generic_write_sync prototype") reworked callers of generic_write_sync(), and ended up dropping the error return for the directio path. Prior to that commit, in dio_complete(), an error would be bubbled up the stack, but after that commit, errors passed on to dio_complete were eaten up.
This was reported on the list earlier, and a fix was proposed in https://lore.kernel.org/lkml/20160921141539.GA17898@infradead.org/, but never followed up with. We recently hit this bug in our testing where fencing io errors, which were previously erroring out with EIO, were being returned as success operations after this commit.
The fix proposed on the list earlier was a little short -- it would have still called generic_write_sync() in case `ret` already contained an error. This fix ensures generic_write_sync() is only called when there's no pending error in the write.
CC: stable@vger.kernel.org Reported-by: Ravi Nankani rnankani@amazon.com Signed-off-by: Maximilian Heyne mheyne@amazon.de Signed-off-by: Torsten Mehlan tomeh@amazon.de Signed-off-by: Uwe Dannowski uwed@amazon.de Signed-off-by: Amit Shah aams@amazon.de Signed-off-by: David Woodhouse dwmw@amazon.co.uk --- fs/direct-io.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/fs/direct-io.c b/fs/direct-io.c index 093fb54cd316..199146036093 100644 --- a/fs/direct-io.c +++ b/fs/direct-io.c @@ -325,8 +325,8 @@ static ssize_t dio_complete(struct dio *dio, ssize_t ret, unsigned int flags) */ dio->iocb->ki_pos += transferred;
- if (dio->op == REQ_OP_WRITE) - ret = generic_write_sync(dio->iocb, transferred); + if (ret > 0 && dio->op == REQ_OP_WRITE) + ret = generic_write_sync(dio->iocb, ret); dio->iocb->ki_complete(dio->iocb, ret, 0); }