Re: syncing the disks when entering sleep

["Rafael J. Wysocki" <rjw@sisk.pl>]

On Thursday 28 January 2010, Pavel Machek wrote:
> > On Wednesday 27 January 2010, Pavel Machek wrote:
> > > On Wed 2010-01-27 01:51:03, Rafael J. Wysocki wrote:
> > > > On Tuesday 26 January 2010, Pavel Machek wrote:
> > ...
> > > > > 
> > > > > AFAICT no new interface is needed. Just do SNAPSHOT_FREEZE, then
> > > > > _S2RAM then _UNFREEZE.
> > > > 
> > > > That's not quite straightforward and I wouldn't seriously suggest that to
> > > > anyone.
> > > 
> > > Three ioctls, seems pretty much ok.
> > 
> > They are for something else.  Yes, you could use them for this purpose,
> > incidentally.  No, you shouldn't do that.  OK?
> 
> Why not? It seems less ugly then the other solution.

Well, not to me at least.

> > Please note that merely _opening_ /dev/snapshot has side effects, so that's
> > not just "three ioctls".
> 
> I search Doc*/power/userland-swsusp.txt, and could not find mentions
> of sideeffects. Perhaps that should be fixed?

Please feel free to make the docs more complete, then.

BTW, if you look at snapshot_open(), you'll immediately see what I mean.

Moreover, the SNAPSHOT_S2RAM ioctl won't execute the suspend notifiers
as appropriate, so that simply doesn't fly.

Rafael