Linux 4.16 Gets Three New Driver Subsystems Plus VirtualBox Guest Driver

[Happy Heyoka]

SIOX? Shouldn't the kernel be trimmed down and not have bloatware added? Why does a single company have to have their own peculiar software added to the kernel?

here's a link to a PDF about the protocol:

http://www.sioxsolutions.se/download/pdf/sioxsys4_uk.pdf

it looks like a fairly standard industrial serial bus protocol and I've seen a dozen just like it. It has some neat features but nothing that jumps out immediately (maybe the fact that the devices are bus powered - and the bus is running at around 24 volts).
At a maximum speed of around 19.2kbits it's not exactly in need of kernel implementation but possibly there's some kind of serial driver wrinkle that is needed to implement it easily.

Maybe it's just easy fallout work from the SERDEV stuff? https://www.phoronix.com/scan.php?pa...ial-Bus-SERDEV

[starshipeleven]

SIOX? Shouldn't the kernel be trimmed down and not have bloatware added?

It's a modular subsystem controlled by a compile option. Most distros won't enable that so it won't be compiled at all and won't bloat anything.
The same is for CANbus or many other features that make sense only in some specific embedded products.

Why does a single company have to have their own peculiar software added to the kernel?

It's an open standard for automation control and communication.
Eckelmann is a company that builds industrial automation products https://can-newsletter.org/engineeri...000_eckelmann/ . It's not their own internal stuff so your comparison with "toilet control software" is invalid.

Having it mainlined helps them spread its use as others can use that instead of coming up with their own local protocol.

[DrYak]

Oh wait, there's a 7-11 in Paramus that has their own software that unlocks the toilets. Maybe we should add that to the next kernel.

Note that if that software is written by a big company that is in charge of assuring it's maintenance for the next decade, and has the willingness and budget to maintain the linux kernel drivers and keep them stable and uptodate with the kernel development :
then why wouldn't it by acceptable to upstream the drivers if they are stable ?

- Distribution which are "Fedora Toilet Edition Pro" aren't going to enable the module anyway. (So it's definitely not going to overload your Raspberry Pi. Unless you are actually in the toilet automation business. In that case you would be benefiniting from a stable solid driver with corporate backing)

- There are resource to maintain it, so when the big "Kernel 4.19 massive overhaul of all toilet protocols" happens in the future, the company that is in charge of the module will maintain it and make sure that it's not causing instabilities in crashes.

As a counter example, see IPX and the decision to remove it from mainstream as it is just sitting there rotting, with nobody paying attention to it.

[treba]

I may be wrong but I think these are the *guest* drivers, to run Linux inside a VirtualBox VM. That does nothing for your problem. To run VirtualBox on a Linux host, you still need VB's third party kernel drivers and so you still can't use secureboot (unless you sign them yourself).

Ah alright. Thanks for pointing that out