Re: x15

[Olaf Buddenhagen]

Hi,

On Mon, Dec 15, 2014 at 02:43:20PM -0500, Kete wrote:
> On 2014-12-12 18:54, Olaf Buddenhagen wrote:
> >On Mon, Dec 08, 2014 at 06:10:26PM -0500, Kete wrote:

> >>Smartphone users would benefit from this x15
> >
> >Why smartphones in particular?...

> * much less code taking up crucial space on the device

Microkernel architectures actually need more code for the same
functionality than monolithical ones.

> * Things that should be in user mode are highly vulnerable in the
> Android kernel space.

Agreed -- but that's not different from desktops.

> * Android Linux upgrades take a long time to download.

The actual changes for a given fix are probably not bigger than they
would be on a microkernel system. You would need an upgrade method based
on binary diffs though to download only the actual changes. Plus, if the
newer kernel is built with a different compiler, you would still get
many gratuitous changes...

So this is indeed a point where mobile devices are somewhat more
affected -- though I for my part wouldn't consider this a major issue...

> * Hardware vendors do not run their drivers through the Linux
> community, so current drivers do not get updated for subsequent
> kernels.

I don't see how a microkernel architecture is supposed to help with
that? The issues of binary drivers and lack of ABI compatibility for
Linux modules seems entirely unrelated with the kernel architecture to
me...

> * Vulnerabilities are much easier to spot and patch in 100,000 lines
> of code than 14,000,000.

Again, that's right, but not different from other form factors.

All in all, while I certainly agree that microkernel architectures have
a bunch of advantages, I still don't see why this would be any more
important for smartphones than for desktops...

-antrik-