On Wed, 2015-07-08 at 00:11 +0200, Rafael J. Wysocki wrote:

On Tuesday, July 07, 2015 04:38:26 PM Oliver Neukum wrote:
 

quoted

That is a tough nut. But that's not a reason to make it worse.
I'd say there's no reason not to use a secondary interface to
suspend without syncing or to extend or introduce such an interface
if the API is deficient.

Well, the point here is that the sync we have doesn't prevent all potentially
possible bad things from happening.  It's a partial measure at best in that
respect.

Well, removed hardware doesn't work. That is a very basic limitation.
But can we guarantee that any returned syscall actually wrote to disk?
Yes, but it must be done in kernel space. So doing a sync has a true
benefit.
I don't see why you would want to generally remove it. What is wrong
with an interface allowing a selection there to those who don't care
about additional guarantees?

	Regards
		Oliver