Re: [RFC][PATCH -mm] PM: Do not sync filesystems from within the freezer

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

On Friday, 6 July 2007 11:31, Miklos Szeredi wrote:
> > > > On Fri, 2007-07-06 at 09:13 +0200, Miklos Szeredi wrote:
> > > > > 
> > > > > Another myth, that has been debunked already.  The problem is: how do
> > > > > you define fuse processes?  There's no theoretical or even practial
> > > > > way to do that.
> > > > 
> > > > It could if they told the kernel via some black magic ...
> > > > 
> > > > But that still suck. The freezer sucks :-)
> > > 
> > > Yeah, and it wouldn't work in practice, since the auxilary tasks might
> > > be part of a library which is not even aware of being used by a "fuse
> > > task".
> > 
> > This is why I think the whole concept of filesystems in userspace is broken. 
> > Trying to shift things that need special privilege and special handling to 
> > userspace is just asking for trouble.
> 
> I'm not claiming fuse doesn't introduce problems.  I've been tackling
> those problems for a number of years now.  Suspend/hibernate is just
> the next thing that needs looking at.
> 
> Just blaming either suspend or fuse for these problems is unproductive
> and stupid.

Well, the "stupid" part is a bit offensive, don't you think?

> The problems _are_ solvable if not always simply. 

Agreed.

> > You can say it's the kernel code's fault, but then you have to
> > explain why it's only fuse (yeah ok, and XFS) that have problems.
> 
> I'm not faulting anythig.  My opinion, is that the freezer is not
> needed for suspend, and removing it will not just solve the problems
> with fuse but several others.  It probably needs a lot of work, but
> hey, that's why we are here.
> 
> The patch posted by Rafael, which lets the freezer skip uninterrupible
> tasks is a step in the right direction.

Thanks.

BTW, I think that this is just less radical than the change proposed by Matthew
and it should address the underlying problems.

Greetings,
Rafael


-- 
"Premature optimization is the root of all evil." - Donald Knuth
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/