[PATCH 05/10] Core checkpoint/restart support code

[Matt Helsley]

On Mon, Apr 04, 2011 at 12:42:22PM -0700, Andrew Morton wrote:
> On Mon, 4 Apr 2011 13:51:20 -0500 "Serge E. Hallyn" <serge.hallyn at ubuntu.com> wrote:
> 
> > Quoting Andrew Morton (akpm at linux-foundation.org):
> > > On Mon, 4 Apr 2011 11:27:53 -0500 "Serge E. Hallyn" <serge at hallyn.com> wrote:
> > > 
> > > > Andrew (Cc:d), did you see this thread go by, and it did it look
> > > > in any way more palatable to you?  Have you had any thoughts on
> > > > checkpoint/restart in the last few months?  Or did that horse quietly
> > > > die over winter?
> > > 
> > > argh, it was the victim of LIFO.
> > > 
> > > All I can say at this stage is that I'll be interested next time it
> > > comes past, sorry.
> > 
> > Thanks, that's good to know.
> > 
> > As you know, we started with a minimal patchset, then grew it over time
> > to answer the "but how will you (xyz) without uglifying the kernel".
> > Would you recommend we go back to keeping a separate minimal patchset,
> > or that we develop on the current, pretty feature-full version?  I'm not
> > convinced believe there will be bandwidth to keep two trees and do both
> > justice.
> 
> The minimal patchset is too minimal for Oren's use and the maximal
> patchset seems to have run aground on general kernel sentiment.  So I
> guess you either take the minimal patchset and make it less minimal or
> take the maximal patchset and make it less maximal, ending up with the
> same thing.  How's that for hand-waving useless obviousnesses :)
> 
> One obvious approach is to merge the minimal patchset then, over time,
> sneak more stuff into it so we end up with the maximal patchset which
> people didn't like.  Don't do that :)

Yes, merging this minimal patch set early is obviously premature.

It seems clear from your statement above that "the maximal patchset seems to
have run aground on  general kernel sentiment" -- pushing that set isn't
going to make any progress. So I think we're left with modifying the new
minimal patch set.

However I think we need some review before we continue modifying it. We
had a minimal patch set which evolved into the current maximal set. It
never really got the reviews outside our little group that it needed.
Now we're back with a new minimal patch set. You're asking us to do the same
thing and expect different results -- stack more patches on top and expect to
get it reviewed. OK, but what reason do we have to believe this time will be
any different?

Cheers,
	-Matt Helsley