On 2018/08/06 20:56, Tetsuo Handa wrote:

On 2018/08/06 19:09, Jan Kara wrote:

quoted

Looks like some kind of a race where device block size gets changed while
getblk() runs (and creates buffers for underlying page). I don't have time
to nail it down at this moment can have a look into it later unless someone
beats me to it.

It seems that loop device is relevant to this problem.

Speak of loop device, I'm waiting for Jens for three months
http://lkml.kernel.org/r/1527297408-4428-1-git-send-email-penguin-kernel@I-love.SAKURA.ne.jp
but Jens is too busy to come up with an alternative. Since that is a big patch, I wish we can
start testing that patch before Jan starts writing a patch for fixing getblk() race problem.