I really don't understand why this makes sense to do. What is so special and secret about the baseband that makes it worth the effort of running a hypervisor on such an underpowered phone?

Even the cheapest baseband chip will add a significant cost to a phone. If they can get rid of the physical baseband processor they get a major cost savings.

As for why user a hypervisor vs a user mode or kernel driver: Modularity.

Some company can specialize in producing soft-basebands (software eats the world!) and ship a simple image blob that you can drop into your virtualization host, set it to use 5% of CPU time and you're away to the races.

Might be able to bring up a baseband on a new platform inside a day.

(Not that a shanzhai manufacturer making a $34 phone cares in the slightest about those things)

In that case, and with a CPU and peripherals designed with keeping this concept in mind, the overhead will be almost non-existant compared to, for example, running the baseband functionality as a "regular kernel module"; with the added advantage of clearly separating both worlds, and isolating the realtime-requirements of the baseband from the Linux kernel.

For some similar concept, have a loot at "RTlinux", where the Linux-Kernel runs as the lowest-priority task in a very thin realtime-OS to achieve guaranteed reaction times for, e.g. motor control applications.

http://en.wikipedia.org/wiki/RTLinux

Would that imply hand-assembly in a small factory rather than large batch production?

I gather from the OA that this is basically a reference design using SoC components. Just wondering if there will ever be a Web site where I could spec a made-to-order phone in the same way that I can spec a cheap laptop? So I could have (say) a high quality sound recorder chip with 3.5mm input socket for an external microphone added to an otherwise standard phone? Or a really good camera chip?

I dislike buying laptops because I can't find anywhere that does them custom for a reasonable price.

Instead of a battery, I often use a constant voltage constant current 3.7V SMPS regulator board to switch on my old phones when I want to experiment with them. Connect the positive lead, ground the other two leads, and most of them (i.e., 7 out of 8) turn on just fine.

Here are prices on some Spreadtrum parts:

http://www.europe-asia-electronics.com/index.php?op=stocksea....

The SR1019A is $0.564, SC2330B is $0.876, and SC6825C is $3.83. I found the VC5282 on a different site for 2.30RMB ($0.37 USD), and prices for KMN9X000RM seem to vary a lot but vary between $0.14 (possibly a scam) and $15.

It's hard to estimate the BOM cost because of all the other pieces' unknown specs, but I don't think it's more than ~$20.

When the phone is assembled you can trigger it by holding ‘Volume Down’ while the phone is off, and simultaneously plugging it into USB. I haven’t tried ‘ResearchDownload’ yet, unfortunately it sounds like it can only write images not read them – taking a backup involves flashing a working CWM for the phone model first.

That's the same method used on Mediatek smartphones; they have the bootloader in ROM too, and this is also what is used to load initial firmware on blank devices at the factory. (It's also how to recover from a "hard brick".) The ResearchDownload utility looks like a clone of Mediatek's SpFlashTool, and if the bootloader mode works in a similar fashion, dumping the full eMMC should also be possible like it is in MTK. All the bootloader does is provide commands to write data into memory and execute it.

I'd be very surprised if they were actually virtualising the baseband and applications processors on the same CPU, since the former needs hard realtime constraints. The MTK SoCs use a separate Cortex-R4 core for each modem (two in the case of dual-SIM) to do this. They could be using vLM to emulate a "virtual" SoC, allowing them to change the underlying hardware while still running the same Linux kernel. The possibility that they're virtualising Mediatek's Linux kernel on their hardware is not so far-fetched...

Although, the article didn't say it had an odd number of cores, so I guess not (unless it's really a tri- or penta-core SoC)

Maybe they don't allow direct linking.

Likely the device have a number of smaller partitions (A huawei phone i have handy have no less that 19 different ones) that for various reasons are not mounted.

We don't really want to go the route of using unlicensed phones like this one or hacks with a rpi and dozens of shields, so it'd be cool if there's anything available...

Long story short. Mozilla are making dirt cheap phones, Buy their motherboard direct from china that goes into the phone and put a headless build of firefoxOS on it. All the hardware interfacing has been done for you.

I've no idea how this project is going. Still hard to get those phones in the UK so I've not bothered to look more than the talk and the site.

http://www.seeedstudio.com/wiki/LinkIt_ONE

Not on a single chip , but it uses components that are aimed at smartwatches , so optimized for size.

It is very outdated from both a hardware and software perspective.

You can walk into any walmart or dollar store in the united states and buy a far more modern phone for $30

I suspect there were better choices in that market they bought it.

Let's talk about it: If you have something to say, add it here:

http://glitchclub.com/?topic=34