Modems and external dialers.

[Liam Proven]

On Mon, 10 Jun 2019 at 15:45, Peter Corlett via cctalk
<cctalk at classiccmp.org> wrote:
>
> So long as said companies don't just make yet another Android device based on a
> cheap-and-nasty Mediatek SOC which requires proprietary Android-only drivers to
> work well, and then make misleading claims about Linux support.

Don't all phone chipsets require Android drivers?

And on that point, so does the RasPi.

> The Gemini's keyboard was very much a take-my-money-now feature when I saw it,
> but since it was being crowdfunded on Indiegogo, the platform for stuff too
> dodgy for Kickstarter, I decided to exercise caution and wait to see what, if
> anything, would be delivered. When they finally admitted it had a Mediatek
> chipset, I lost all interest. Been there, done that, never again.

You pays your money, etc. I'm quite happy with mine. I don't use it as
a phone but for taking notes at conferences and events, for instance,
it's _superb_.

> Planet are right now crowdfunding their new "Cosmo Communicator". They have
> apparently learned nothing as it also has a Mediatek chipset, and yet they
> continue to disingenuously claim Linux support. I shall be giving this one a
> wide berth too.

On their sales volumes, I think they have to go with whatever is cheap
and customisable on the Chinese market.

One of the sad things about the ARM market is that there is no
industry standard, no baseline to aim for. There isn't even standard
firmware. Lots of devices don't have firmware at all, so every Linux
port is a bare-metal thing, starting with hardware initialisation. A
year or 2 after it goes off the market, it's junk, as nothing will
support it any more.

ARM64 is trying to impose a requirement for UEFI, I believe, but [a]
the legions of cheap kit makers don't care and just ignore it, and [b]
UEFI is horrible.

> Third time lucky, eh? Maybe they should start talking to the Raspberry Pi
> people who actually know a thing or two about getting Linux working well on
> mobile chipsets.

Via big binary BLOBs, yeah, and a weird bootloader that means that the
GPU initialises the system and (I hear) retains some degree of control
over interrupts, making it more or less impossible to run a proper
hypervisor on the things.

-- 
Liam Proven - Profile: https://about.me/liamproven
Email: lproven at cix.co.uk - Google Mail/Hangouts/Plus: lproven at gmail.com
Twitter/Facebook/Flickr: lproven - Skype/LinkedIn: liamproven
UK: +44 7939-087884 - ČR (+ WhatsApp/Telegram/Signal): +420 702 829 053