On Sun, 02 Aug 2015 16:07:06 +0200,
Al Viro wrote:

On Wed, Jul 29, 2015 at 07:45:11AM -0700, Dmitry Torokhov wrote:

quoted

quoted

This seems coming from evdev_flush().  As there is no fd leak, it's no
big problem per se.  But, now the question is whether returning such
an error code is correct behavior at all.  At least, it doesn't seem
defined in POSIX:
  http://pubs.opengroup.org/onlinepubs/009695399/functions/close.html

Hmm, if I checked the right version of the code close_nointr_nofail()
expects only 0 as the return code so even if we change the kernel to
use more conforming -EIO instead of -ENODEV systemd will still die...

The question is whether we really need to propagate return value from
f_op->flush() up to userspace in filp_close(). Why don't we ask Al?

That's the whole damn point of having ->flush().  And yes, we do need that -
things like NFS (not to mention tapes, etc.) do rely on that.

Whether it makes sense to do this kind of "do something that might have
a failure to report on each close()" for evdev is up to driver, obviously.

So, the behavior of VFS layer is as designed.  Then I suppose the fix
should be rather in evdev.c.  Dmitry, could you paper over it?


thanks,

Takashi