Re: [PATCH 0/2] ARM: Allow either FLATMEM or SPARSEMEM on the multiplatform build

[Russell King - ARM Linux admin <linux@armlinux.org.uk>]

On Tue, Jun 02, 2020 at 02:37:45PM +0200, Geert Uytterhoeven wrote:
> Hi Russell,
> 
> On Tue, Jun 2, 2020 at 2:22 PM Russell King - ARM Linux admin
> <linux@armlinux.org.uk> wrote:
> > On Tue, Jun 02, 2020 at 02:18:48PM +0200, Geert Uytterhoeven wrote:
> > > On Thu, May 21, 2020 at 4:50 PM Russell King - ARM Linux admin
> > > <linux@armlinux.org.uk> wrote:
> > > > On Thu, May 21, 2020 at 05:07:45PM +0300, Mike Rapoport wrote:
> > > > > Ah, when either of these patforms will become a part of the
> > > > > multiplatform build, the only option for multiplatform build will be
> > > > > sparsemem.
> > > > > So it would be nice if somebody could check the cost of using sparsemem
> > > > > vs flatmem, espessially on low end machines.
> > > >
> > > > Do you think they will become part of multiplatform?
> > > >
> > > > If they're low-end machines, then adding:
> > > >
> > > > (a) the additional memory overhead of a multiplatform kernel
> > > > (b) the additional runtime overhead of the complexities of multiplatform
> > > >     kernels
> > > >
> > > > is surely an odd thing to do, especially when few really care about
> > > > these platforms becoming part of a multiplatform kernel, except those
> > > > who like the idea of multiplat.
> > >
> > > The fallacy of "multi-platform", again?
> > >
> > > Isn't it a requirement for any new ARM v7-A platforms to be part of
> > > ARCH_MULTI_V7, even if it's a low-end machine?
> >
> > What does ARMv7 have to do with this thread, that is discussing the
> > older ARMv4 platforms? I find your comments above irrelevent to
> > this discussion, and seems to be worded to provoke a reaction.
> 
> Sorry, I used ARMv7 as an example, as it's rare for any new ARMv4
> platforms to show up.  But the recently introduced sam9x60 (ARMv5) is
> part of ARCH_MULTI_V5, and I expect it will be used with less than the
> 256 MiB provided on the dev board.
> 
> Nevertheless, there's a movement to convert everything to ARCH_MULTI_V*
> where possible. So it matters for all variants.

First, please understand that I have nothing against multiplatform.

What I do have problems with is the "lets move everything to
multiplatform" without fully reasoning it out, especially when it
comes to older platforms that have limited attraction for the whole
motivation behind multiplatform, which is the convenience of
distributions.

multiplatform is the current fad-de-jour, and everyone is in a "you
MUST convert EVERYTHING to it, because it's the best thing since
sliced bread".

multiplatform comes with it a load of forced-enabled options that
are force-enabled.  Should RiscPC be forced to have the common clock
support built into the kernel, when it doesn't' have _any_ controllable
clocks, when it's likely that the machines have limited memory
available?  Yes, I have recently booted the RiscPC with a 5.x kernel,
and no, it didn't need very much fixing, the code is quite *stable*
(oh, I blaspheme, we can't have stable code in Linux, we must change
it every five minutes.)

I think, maybe, you should leave these decisions to people who actually
have the platforms, know them inside out, and can assess whether
conversion really makes sense or not, rather than blindly joining the
zombie march repeating "multiplatform is great, everything must be
multiplatform".

What is patently obvious is that in Linux, there is no tolerance of
someone who has a differing view, and they end up receiving reaction
provoking comments designed to inflame the situation.  Difference of
opinion must be suppressed at all costs!

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTC for 0.8m (est. 1762m) line in suburbia: sync at 13.1Mbps down 424kbps up

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel