The Neo FreeRunner smart phone

The Neo FreeRunner is a touch screen smart phone, which operating system is based on Linux. The phone is intended to be used by software developers familiar with Linux as the general consumer, but in time as development progress, this audience should broaden to all phone consumers. Software developers have taken the Neo FreeRunner to their heart because of the total freedom they have to use it, and to design software for it. In time the broader consumer will appreciate the high quality product, the performance of the phone, and the wide range of free software packages expected to emerge in time. This will allow users, to make the maximum use of the hardware by tailoring it to fit their needs.

Is it usable?

At present time the Neo FreeRunner is fully functional. It is used by many as a day-to-day phone. There are many phones with fully functional GPS, and there is good software (Locations, TangoGPS, Navit (for use in cars) and many more) to use, especially with OpenStreetMap. With the current shipping, the battery life is only a little over one day, and it has an alarm clock, media player, internet browser, game console, email reader and contacts manager: all in different version, which can be downloaded and installed (all open source and free). The new version called GTA07 is available.

How much does it cost?

The Neo FreeRunner is sold through many channels (this includes www.openmoko.com) and the official price has been set to ca. €290 , but the price may vary slightly from country to country.

Hardware specification of the Neo FreeRunner GTA02

• Display- Topply o2.8, 480 x 640 pixels, VGA, 200 NIT minimum, resistance type touch
• User Interface Navigation- Touch screen on LCD, 2 control “buttons”, 1 Power button, 1 Aux for 911 emergency call
• Built-in 802.11b/g Radio (Atheros chipset AR6001 Flash version)
• Built-in Bluetooth 2.0 + EDR (CSR and support PCM audio , BC4 firmware version)
• Built-in 2D/3D graphics acceleration chip (S-Media 3362)
• 2 built-in Tri-Axis sensors (ST accelerometer LIS302DL)
• Built-in GPS Radio – -130 dBm with internal antenna, -157 dBm tracking on chipset specification, TTFF under 40 seconds with -130 dBm signal strength, and tracking (u-Blox)
• Antenna – Specialized antenna for best in hand hold GPS, GPRS and Wi-Fi/Bluetooth performance are required, -105dBm on receiving, Tx 30dbm+2 on GSM
• External Antennae – MMCX GPS connector
• GPRS Radio –GSM/GPRS radio. A Pre-PTCRB certified module will be preferred
• Linux – Linux kernel 2.6.24 or later Openmoko kernel
• USB - Client and Host-mode switchable (to be used for software downloading), provide host 5V power

The dicussion with back ground

Linux kernel 2.6 (Monolithic kernel)

The Linux kernel is developed by Linus Torvalds, but maintained by thousands of individuals. It is a monolithic operating systems structure, which is constructed in a layer fashion. The system is constantly enhanced to make it fit the needs of the marked (in embedded systems, PC systems and servers). It is enhanced so that it often changes its structure. System functions like the process and memory management, the process and thread scheduling, drivers and I/O is implemented in kernel space - where I/O communication are provided by modules, which can be inserted/removed in runtime – they are all built against the kernel. Now when the all that functionality is built against the kernel, this also means that if the kernel changes, so does the set of modules. To add or change features provided by the hardware, all the “layers” above the changed one also has to be changed, this is of course a worst case scenario, but it still if the concept change too much in the kernel, things such as modules ( I/O communication and parts of drivers) needs more than just a recompilation, it needs a complete code modification. The Linux 2.6 kernel does support soft real-time performance. This is done through a configuration of the kernel, to make the kernel fully preemptable, this is done with a new configuration option for the kernel, that changes the behavior of the kernel by allowing processes to be preempted high priority work is available to be done (if the standard 2.6 kernel is used, then when a userspace process makes a system call, a high priority process have to wait until the call is complete before getting access to the CPU).

The Microkernel

The idea of the Microkernel appeared in the late 1980’s, and the concept was to move as much as possible out of the kernel space (and into the processors first-level cache) and in to user space, meaning to reduce the kernel to basic process communication and I/O control and let the other system services live in user space in the form of normal processes also called daemons. There exist daemons for handling memory management, one for process management, one for managing drivers, and so on. Because the daemons do not run in kernel space, a context switches are needed to allow user processes access to the kernel (to run in kernel mode). In the memory management of the microkernel (L4) every process consists of three primitives. A process maps its memory pages to another process if its wants to share these pages. When a process gives the pages to another process, it can no longer access them, and they are under the control of the other process, as long as the first process does not flush the pages. This means that if a process flushes its shared memory pages, it retrieves the control back on those pages. The system now works as follows, the microkernel reserves all memory at startup, to one process called a base process (which live in user space, all processes do), and if another process needs memory, it can directly ask the base process for memory, which saves the trip down through the kernel. Because of the way processes are built in the microkernel (with the three primitives: grant, map and flush), a process can only grant, map or flush memory pages, memory protection still exists and the overhead of context switches is reduced.

What is the ultimate operating system for the Neo FreeRunner?

I believe that the Microkernel structure is the best way to make the mobile phone stable. When looking at the kernel structure of the microkernel, there are several smart phones on the marked based on the microkernel structure, purely based on the stability of the kernel, and because the further development on the kernel has a higher degree of maintainability. If one were to add a new features to a monolithic system (e.g. a new memory management routine), it means recompilation of the whole kernel, often including the whole driver infrastructure. If you have a new memory management routine and want to implement it into a monolithic architecture, modification of other parts of the system could be needed. In case of a microkernel the services are isolated from each other through the message system (Microkernel’s uses message queues build in as FIFO queue). It is enough to re-implement the new memory manager. The processes which formerly used the other manager, do not notice the change. Microkernel’s also show their flexibility in removing features. That way a microkernel can be the base of real-time appliances in single chip systems. On the other hand microkernel’s itself must be highly optimized for the processor they are intended to run on. To try to sum up the monolithic structure, it suffers from the fact that it includes all of the basic services inside kernel space, which makes the kernel lack on extensibility and maintainability, on the sheer size of the kernel growing big, and the fact that bug fixing or the adding of new features mean a complete re-compilation of the kernel which is time and memory consuming. The fact that microkernel has – for a long time been used – in mobile phones should also give an indication of the stability of the kernel. Toshiba W47T is one of many that run on a microkernel, and there are many different microkernel structured operating systems out there for mobile/smart phones, and there are all in use in different phones. One exiting system is the Cosmos microkernel developed by a former developer from Microsoft, and the operating system is .NET based, and open source.

I had the pleasure of attending the DEMO MANIA session on Windows Embedded Standard 2011, also called “Quebec” or plain Windows 7 Embedded.

Due to the fact that it was a DEMO session, the amount information given were minimal and the visual presentation was in focus.

The embedded team has come up with a treat for all of us technical guys. They have prepared Windows 7 for embedded devices, and in such a way that when you boot up for the first time, a menu appears to let you chose which Device drivers you need for the system (the ones it did not find it self). After this, the system let’s you pick and chose your desired configuration, meaning that you now have a chance to deselect IE, or just the parts of IE you don’t want… And off course not only IE, but every little piece of Windows 7 you select and deselect as of your likings (That is WOW!). In Short it is a full Windows 7 with all features, (silverlight etc.) where you deselect features to make it fit your device and wishes.

I got a hold of an early preview version of Windows Standard 2011 and when I got home, I immediately started to install the version onto my Asus Eee 901 just to see the how it worked! And you can really optimize a laptop with this kind of version :-)

But…. It was brought to my attention that it was illegal to install Windows Embedded Standard on a Notebook, and therefore I have now erased it, and cannot show any pictures of my progress with the adventure (I will not even explain about it).

Windows 7 Embedded Standard 2011 as OS for Windows Mobile?

But with all of this power at your hands, and the statements from Microsoft that the future of their mobile phones will run on at least 1GHz processors…. Hmmm wouldn’t be weird if Microsoft didn’t use this as a new operating system (OS) for Their Mobile products too?? Let me answer that! Yes it would.. With this product at their hands, Microsoft is finallly able to have – One Operating System For All platforms – A vision they had for years, and a vision Mr. Steve Balmer talked about on his last visit in Denmark.

And here is the more technical information:

Processor Architecture
Support for multiple processor architectures
· x86
· x64

Tools
Improved developer experiences:
· Wizard experience with Image Build Wizard (IBW)
· Advanced experience with Image Configuration Editor (ICE)

Componentization
The right level of granularity to build special-purpose devices
· Hundreds of feature packages based on latest innovations for Windows 7
· Embedded enabling features such as Enhanced Write Filter, File Based Write Filter, Registry Filter, Hibernate Once Resume Many, and Custom Shell to fulfill embedded-specific requirements
· Large number of driver sets for compatibility with growing set of device hardware and peripherals

Application Compatibility
Applications and drivers for Windows 7 can work on Windows Embedded Standard 2011 without difficult, expensive, and time-consuming porting effort

Enterprise Connectivity and Manageability
Support for Active Directory, Domain Join, Group Policies, Network Access Protection, and IPv6 to enable connectivity and manageability with Windows Server, System Center Configuration Manager, System Center Operations Manager, and Windows Server Update Services

User Experiences
Rich, interactive user experiences with Windows Aero and Windows Touch and Gesture. A stable framework Windows Presentation Foundation for building new and innovative experiences.

Video resources

 http://channel9.msdn.com/posts/Charles/Windows-Embedded-Past-Present-and-Future/

Websites

 http://windowsembedded.com

http://msdn.microsoft.com/en-us/windowsembedded/ce/default.aspx

http://msdn.microsoft.com/en-us/windowsembedded/standard/default.aspx

Blogs

Olivier Bloch - http://blogs.msdn.com/obloch/

http://blogs.msdn.com/mikehall

http://blogs.msdn.com/jcoyne

“All my bags are packed, I’m ready to go……”

I’m am very exited to get to Tech ED Berlin 2009 (which is now sold out!). I’ve been trying to wrap my head around what track and sessions I should attend, because there are MANY! to chose from. As a person I believe it’s importent to focus on the fact that this is a unique opportunity to evolve your knowledge. Therefore I have chosen to expand my choice of sessions to not just include areas of immediate focus, but to also include areas of common interest.

I spend a few hours on the track & sessions page, figuring out pros and cons on sessions that unfortunately collided on day and hours. But I think I hit an amazing agenda for the entire week:

The one session that I look really forward to is Windows 7 and Windows Server 2008 R2 Kernel Changes!! I’ve been following the discussion on the internet, and following the videos from Channel 9 on Windows 7.

The official note on the session leaves me with the impression that this is going to be really good:

“This session goes beneath the hood of Windows 7 and Windows Server 2008 R2 to describe and demonstrate the key changes in the kernel. Topics include: scalability improvements such as removal of the global scheduler lock, support for more than 64 logical processors, and user mode scheduling, virtual service accounts, core parking and timer coalescing for power efficiency, trigger-started services, core architecture changes to modularize Windows ("Minwin") and more.”

And the embedded part with demo’s are also something to be looking forward to:

“This demo-packed session presents all sorts of devices, tools, and technologies used in today's and tomorrow's embedded projects. If you want to learn how Windows Embedded can be used to build stunning devices or if you are just a geek, you can't miss this session.”

And yes, I’m a geek!

I’m really looking forward to this mix of Desktop, embedded, web technology and architechture sessions. And not forgetting all the demo presentations that makes this Tech ED a must-see event.

Appart from all of these sessions, I will be on the lookout for the expo areas, where I will get a chance to have a look at what all other developers are doing right now.

I’m also loking forward to meeting all the other Microsoft Student Partners from all over the world. And this will ofcourse introduce them to my network, and me to theirs.

If you find the subjects listet above interesting, then I will attempt to blog on every single one of them, and hopefully provide you with awesome pictures from the whole event. If you find subjects of interests missing in my list, I will suggest that you have a look at some of the other Microsoft Student Partner blogs. We are well represented at the Tech ED event, and maybe they will attend more convenient subjects. You can find links to there blogs on the left column.