Re: [PATCH] swsusp: separate swap-writing/reading code

[Nigel Cunningham]

Hi Pavel.

On Monday 27 March 2006 20:26, Pavel Machek wrote:
> On Ät 23-03-06 17:48:58, Mark Lord wrote:
> > Rafael J. Wysocki wrote:
> > >I agree it probably may be improved.  Still it seems to be good enough.
> > >Further,
> > >it's more efficient than the previous solution, so I consider it as an
> > >improvement.
> > >Also this code has been tested for quite some time in -mm and appears to
> > >behave properly, at least we haven't got any bug reports related to it
> > > so far.
> >
> > I find the in-kernel swsusp to be quite slow, and it seems to use
> > an awful lot of memory for book-keeping.  So count that as encouragement
> > to improve the performance when you can.
>
> Extents will provide 0.01% speedup at most, and with increase of code
> complexity. Not a nice tradeoff if you ask me.

My point wasn't speed, but efficient use of memory. A bitmap is certainly 
better than storing n*sizeof(swap_entry_t) byes. Extents would be better 
again, but perhaps not as big after the bitmap switch.

> If you want faster suspend, that should be easy. You'll need *current*
> 2.6.16-git , and userland tools from suspend.sf.net . There's HOWTO
> that explains how to set it up. We can even do LZF these days...
>
> > >Currently I'm not working on any better solution.  If you can provide
> > > any patches to implement one, please submit them, but I think they'll
> > > have to be
> > >tested for as long as this code, in -mm.
> >
> > It would be *really nice* if you guys could stop being so underhandedly
> > nasty in every single reply to anything from Nigel.
> >
> > He really is trying to help, you know.
>
> Actually Rafael was *very* nice at him, I'd say. Pointing for tiny
> inefficiencies, without patch attached is not really helpful.

Yes, I didn't think Rafael was harsh in that email, but that's just my 
opinion. Regarding "without patch attached", could you please remember that 
comment next time you comment on my patches?

> I have repeatedly pointed him on ways how he can *really* help. There
> are ways to do suspend2 in userspace these days, but Nigel refuses to
> use them.

You know that I disagree that doing suspend in userspace is the right 
approach, and you know that current uswsusp can't do everything Suspend2 does 
without further substantial modification. Please stop painting me as the bad 
guy because I won't roll over and play dead for you. Please also stop 
encouraging people to use uswsusp when you have also warned that it might eat 
their partitions.

Regards,

Nigel

Attachment: pgp00000.pgp
Description: PGP signature