Re: [linux-pm] swsusp3: push image reading/writing into userspace

From: Pavel Machek
Date: Thu Sep 15 2005 - 06:17:24 EST


Hi!

> > > You could reclaim it. There are 10 times as many hits (239,000) for
> > > suspend2, and I've never wanted it to be called swsusp2 anyway :). As
> > > for suspend4, at the moment, I'm not planning on ever progressing beyond
> > > 2.x.
> >
> > Sorry, have to ask...
> >
> > "not planning on progressing" == version number stays "2" no matter
> > what changes, or "not planning on progressing" == not plan to use
> > swsusp3/uswsusp infrastructure?
>
> I'm not planning on ever progressing beyond 2.x because I'm seeing the
> code I have now as pretty much feature complete. There are a few small
> areas that I'd like to improve (reinstating module support being one -
> it was a mistake to remove it), but I don't see the need for a complete
> redesign. That said, I was careful to say 'at the moment'. I'm not
> denying for a second that things might change.

Ok.

> Regarding the 'swsusp3/uswsusp infrastructure': as I see it at the
> moment (feel free to correct me), your new revision is only moving as
> much code as you can to userspace (plus changes that are made necessary
> by that). Beyond maybe reducing the kernel size a little, I don't see
> any advantage to that - it just makes things more complicated and
> requires the user to set up more in the way of an initrd or initramfs in
> order to suspend and resume. You end up with more code to maintain to
> get the same functionality (same amount of kernel code or slightly less
> plus extra userland code to do all the reading and writing). I'm

Agreed. OTOH interface code is quite small (~100 lines).

> If the main impetus is seeking to reduce kernel code size, why not just
> provide the option of building your code as a module for those who are
> concerned about that statistic?

It is source code size that I'm concerned about. And flexibility; with
writer in userspace, I (or anyone else :-) can do all kinds of
graphical progress bars, esc-to-cancel, ... that would be too ugly to
do in kernel.
Pavel
--
if you have sharp zaurus hardware you don't need... you know my address
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/