A detached loop device can hit an I/O error on fsync() if it was previously associated/attached. This happens because the write cache is enabled in loop_configure() in the attach path, but not disabled anywhere in the detach path. So, disable write cache in the detach path. Before: # losetup /dev/loop7 disk.img # losetup -d /dev/loop7 # strace -e fsync parted -s /dev/loop7 unit s print 2>&1 | grep fsync fsync(3) = -1 EIO (Input/output error) Warning: Error fsyncing/closing /dev/loop7: Input/output error [ 34.696133] blk_update_request: I/O error, dev loop7, sector 0 op 0x1:(WRITE) flags 0x800 phys_seg 0 prio class 0 After: # losetup /dev/loop7 disk.img # losetup -d /dev/loop7 # strace -e fsync parted -s /dev/loop7 unit s print 2>&1 | grep fsync fsync(3) = 0 Signed-off-by: Mauricio Faria de Oliveira <mfo@xxxxxxxxxxxxx> Co-developed-by: Eric Desrochers <eric.desrochers@xxxxxxxxxxxxx> Signed-off-by: Eric Desrochers <eric.desrochers@xxxxxxxxxxxxx> --- drivers/block/loop.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/block/loop.c b/drivers/block/loop.c index a58084c2ed7c..0e23ab151667 100644 --- a/drivers/block/loop.c +++ b/drivers/block/loop.c @@ -1223,6 +1223,9 @@ static int __loop_clr_fd(struct loop_device *lo, bool release) goto out_unlock; } + if (!(lo->lo_flags & LO_FLAGS_READ_ONLY) && filp->f_op->fsync) + blk_queue_write_cache(lo->lo_queue, false, false); + /* freeze request queue during the transition */ blk_mq_freeze_queue(lo->lo_queue); -- 2.25.1
