Category Archives: Developer Info

Breaking the Memory Barrier

Overview

AmigaOS is a 32 bit OS. There is little we can change about it. The size of an address pointer is intrinsically entangled into the API, and getting rid of this legacy is, for the most part, a matter of replacing all of the API with a new one. Every time a programmer writes something like “sizeof(struct Message)”, the 32 bit nature is fused into his code.

This has some repercussions that cannot be easily ignored. It means that our address space is inherently limited to 32 bits (meaning 4 gigabytes). In reality this space is even smaller than that. PCI space, the kernel, memory buffers, and other memory areas take up a large chunk of the already limited address space, leaving roughly 2 gigabytes for the applications running on the machine – 2 gigs which also are shared between all of the programs running.

Physical Vs. Virtual

A physical address of a memory block is implicitly defined by its position within the memory chips, and the order in which the modules are inserted into the mainboard’s memory slots. They start at zero and go up to a specific maximum.

A virtual address, on the other hand, is what the CPU and hence the application program sees. They might be the same, but as a general rule, they are different. Virtual addresses are given on the fly, but there is a rule that every memory cell must have a unique virtual address, because all references to that cell are stored as the virtual address the application sees.

Modern systems like the X1000 or upcoming models can take more than 4 gigabytes of memory, but so far, the extra memory will never be used. Even in a 4 gigabyte system, there is memory that will never be touched because there is just no free address; and unfortunately, every byte needs to have its own virtual address, and no two bytes can have the same.

Unless…

Extended Memory Object

Extended memory objects (ExtMem) are a means to access memory beyond the 2 gigabyte barrier by applications that are written to make use of them. In a nutshell, an extended memory object is a chunk of physical memory that exists in a “nirvana” state somewhere in the memory of the computer without a virtual address of its own. The memory cannot be accessed by anyone or anything in this state. In order to access it, an application must map part of the object into its own virtual address space. This mapping does make a part of the memory represented by the ExtMem object accessible in a memory window in the application’s own address space.

There is no limit to the number of mappings an application can do. If needed, it can have several mappings active at a time, and add or delete mappings as required. The only restriction is that mappings must not overlap (either in virtual address space or in the memory object itself). Each mappings opens up a view into a part of the memory object, and, depending on how the mapping was performed, the application can read and/or write to the memory as if it were normal memory.

fig2A mapping is defined by the virtual address in application memory (which can be chosen by the application, or picked at random by the OS), the length of the map’s window, and the offset it maps to in the ExtMem object.

There are some caveats though. Most notably, the ExtMem object itself doesn’t have an address. In that sense it should be treated more like a file than a memory block. If an application wants to have permanent references to memory in the ExtMem object, it needs to store them by offset, just like it would with a file. The first offset is zero, so to address the 1000th byte in the memory block, the application needs to reference it by the offset of 1000. Obviously, this offset must be calculated against the base of the mapping’s offset; just like in a file, reading a part of the file into a buffer makes the first byte read the offset zero in the buffer.

As an example, consider the following situation. We want to access byte 3000 of the ExtMem object. We created a mapping that has length 4000 and starts at offset 2000. The resulting address for our byte would be the base address of the mapping plus 1000, since the offset of the beginning is already at 2000.

Downsides of the ExtMem system

If you think now that this all sounds suspiciously like bank switching, then you are right. The method has been used way back in the Home computer age, and even earlier. The Sinclair ZX Spectrum 128K was equipped with twice as much memory as the Z80 CPU could address; the upper 16k of the machine could be swapped between different chunks of the rest of the memory. Similarly, the Commodore 64 used bank switching to address a larger memory than its 6502 CPU could handle. It was the only possibility at the time to add more memory.

This method we employ now is basically the same (with a bit more added comfort).

Obviously, the method is a compromise. A “real” 64 bit system would be better, and much more transparent to use. However, as I already outlined in the beginning, there is a lot of work involved to make AmigaOS 64 bit compatible, and with the method of ExtMem objects, breaking the barrier is possible now as opposed to years down the road.

Who can benefit from ExtMem objects?

Well, every application that, in some way or the other, has to cope with large amounts of data. Even if the dataset is only potentially large (like, for example, a text editor), using an ExtMem object has its advantages. The text editor (or word processor), by its nature, only presents a small subset of the text it is editing to the user. Likewise, a movie editor would only need to have access to a few frames in order to show thumbnails of the movie on a timeline, or display a single frame that the user is working with.

Another example is RAM disk. Plans are currently underway to update the RAM disk to make use of the ExtMem object interface, allowing out-of-the-box usage of those normally unassigned memory blocks without draining the valuable main memory. Since (depending on programmer setting) memory blocks can even be allocated “on-demand” instead of ahead of time, this will make RAM disk have an even lower footprint, on top of making it possible to store larger amounts of data than ever before in it.

It needs to be said that the ExtMem system doesn’t require memory beyond the 4 gigabyte bounds. It can work with normal memory as well, even though that is not its purpose.

So, as you can see, a good number of applications have a natural tendency to only access a very small subset of their memory at a given time. All of these are good candidate for using ExtMem objects to break the memory barrier.

FUSE and NTFS for AmigaOS

280px-FUSE_structure.svgFUSE is short for Filesystem in Userspace. FUSE was created to enable non-privileged users to run file systems outside of the kernel which is a big deal for Unix-like operating systems. In AmigaOS, everything runs in userspace so FUSE is not nearly as important for Amiga users. What makes FUSE valuable is all the file system implementations which use FUSE such as NTFS, ext2, ZFS, etc.

The Amiga Operating System implementation of FUSE has been realized via a project called Filesysbox by Leif Salomonsson. A special thanks goes out to Leif for allowing his hard work to be utilized.

Amiga programmer extraordinaire Fredrik Wikström was then commissioned to port Filesysbox over to AmigaOS. Fredrik took the original code and updated it to AmigaOS 4.1 standards. This work included utilizing advanced DOS features such as object notification and the new file system API which seeks to completely avoid the esoteric DOS packet interface. Colin Wenzel is the main man behind the advanced DOS features.

Master_500pxIn order to test whether Filesysbox was working properly we needed a file system to go with it. NTFS-3G by Tuxera was chosen for this purpose. Fredrik also ported a full suite of tools to go along with NTFS itself.

Both Filesysbox and NTFS-3G are contributions being offered to registered AmigaOS users via AmiUpdate. The software licenses require that the source code be made available so registered users can download the matching source code from Hyperion’s web site in the downloads section.

blog_devIt is hoped that 3rd party developers will become interested in porting more file systems in the near future whether they are via the FUSE API or the new DOS file system API. The upcoming SDK will include everything you need. In the mean time, please feel free to utilize the provided source code and the AmigaOS support forum for assistance.

Finally, a big thanks needs to go out to the AmigaOS beta testing team for risking their hard drive partitions while testing NTFS-3G and Filesysbox. It is demanding and potentially destructive work that should not be taken for granted.

AmiWest 2013 AmigaOS Team

AmiWest2013-AmigaOS-Team1200AmiWest 2013 is now over and it was a heck of a lot of fun. We managed to grab a team photo this time.

From left to right in the front row sitting down we have Ken Wilde and Lyle Hazelwood. Next row back we have Bill Borsari, Flip LaFramboise, Trevor Dickinson and Tony Wyatt. In the back row we have Matthew Leaman, Val Marti, Paul Sadlik, Steven Solie and Alex Carmona.

A special thanks to Mike Brantley for taking the photo. Mike also took a lot more photos at the AmiWest 2013 show.

I think the big smiles on those faces says it all really.

I also took some time to update the AmiWest 2013 crowd on the current status of the Amiga Operating System which I will now summarize here:

  1. The netbook project initially announced at AmiWest 2011 has been cancelled. Below is an excerpt from an email to me from Ben Hermans dated October 10, 2013 on this matter.

    Despite best efforts by Hyperion and A-EON we were unable to get acceptable and stable conditions and terms from the Chinese supplier (price point, paying terms, required upfront, etc.)…

    The project was therefore cancelled in favour of a more future proof solution.

  2. A-EON Technology announces the Cyrus Plus Beta Test Programme. Follow the link for all the details and how to apply.
  3. Gallium3D Update
    • Software rendering completed
    • Working on the WinSys part of the implementation
    • Challenges encountered along the way:
      • Must be re-entrant, thread safe and multicore capable
      • Must run on a bare minimum system
      • Efficient
      • Possible to load non-Mesa and non-Gallium drivers
  4. X-Kernel Update
    • Task scheduler rewritten in C
    • Removed reliance on data structures (e.g. ExecBase task lists and ThisTask pointer)
    • Moving scheduler to run on auxiliary cores
    • All cores schedule tasks independently
    • Load balancing between cores
  5. AmigaOS 4.2
    • Depends on Gallium3D release
    • May or may not depend on multicore support
  6. AmigaOS 4.1 Update 7
    • Will be needed for Cyrus Plus product release
    • Consolidates all previous updates
  7. AmigaOS 4.1 Update 6
  8. FUSE and NTFS-3G
  9. New development team members since AmiWest 2012

We also had a chance to have a team meeting on Friday night in one of the hotel rooms while at the show. We had the usual Airing of Grievances and a lengthy discussion on where we are going and how to get there. At one point we were interrupted by an outsider who likened the gathering to a secret Masons meeting.

If you would like to come to AmiWest 2014 and join in the fun then keep October 24, 25 and 26 open. There is also a programming seminar planned for October 23 and 24. See you there!

An SDK Update (finally)

blog_devIt has been a long time coming but we finally got around to releasing an updated Software Development Kit (SDK) for the Amiga Operating System. You can download it from Hyperion’s server.

This SDK includes all the usual includes and autodocs you need to use all the latest released AmigaOS features. The AmigaOS Documentation Wiki contains all the higher level information you need and will continue to be updated to help explain everything. The wiki also has a new Frequently Asked Questions section where we will post the most common problems and solutions.

This SDK is also a tad incomplete because I ran out of time to prepare it before AmiWest 2013. Therefore, there will be another SDK update or two sometime after the AmiWest show which will include even more.

We will also try harder to provide an updated SDK much more regularly from now on. Thanks to AmiUpdate we now have a way to deliver all sorts of SDK updates as needed with minimal effort.

Support for the SDK is available from the official AmigaOS support forum. You may also want to give OS4Coding a try if you get stuck on something.

Have fun!

New AmigaOS Core Developers

blog_devWe are pleased to announce that we have joined the AmigaOS development team. We hope that we are able to contribute some good things to AmigaOS using our skills in coding (Frank) and graphics (Thomas).

The follow is a short description of us and our products.

EntwicklerX:
We are currently working full time, self-employed, developing our games on multiple platforms (AmigaOS, Android, iOS, Windows Phone, Xbox Indies) as EntwicklerX. For about 10 years we have worked together on various software projects in our spare time. The focus in the last 5-6 years was to earn some money with the small bonus that it has allowed both of us to live from it (not enough for a Ferrari but enough to have fun at work). A strict separation between design and programming logic helped us build our products in an effective way with our limited time. While Frank takes care of programming, Thomas builds the graphics. Our subsystem is so sophisticated that Thomas can create and test Layouts without the help of Frank and Frank can use this directly in his code. This saves a lot of time and nerves. Before this, Frank had to compile every time if Thomas wanted to move a graphic a few pixels. ;)

AmiBoing Games

Our Amiga platform AmiBoing is used to bring an online connection to our games for high scores and achievements and it is also used for distributing our games. With more than 200 Users we can say this is the leading online gaming community on the Amiga platform. ;)

avatarImago_small

Thomas:
I care about what the user sees, how users interact with a game, paint graphics, create the levels and take care of our website. I have invested a lot of time into understanding how to create themes for AmigaOS and am still learning. The seamless design of an operating system is very important to me and I hope to play a part in this within AmigaOS.

avatarGoos_small

Frank:
I try to bring Thomas’ graphics to life and love to get the maximum out of AmigaOS using any available techniques to optimize (e.g. compositing). In our projects, I take care of all the programming. Together with Thomas, I am working on the game play and new game ideas. I can do my part in helping the Amiga in all areas of coding and to help current developers.

We are a Team:
Even with the split of responsibilities we will usually work as a team because ideas and their implementation always occur together. We look forward to working with the existing developers and also contributing to any interesting discussions. We will do our best but please note that we are only human and have a finite amount of time to work on everything. Give us time to understand how things work within the developer team. ;)

Best regards and let us say thank you for adding us to the list of AmigaOS core developers,
Thomas “imagodespira” Claus and Frank “Goos McGuile” Menzel.

Some Links:
EntwicklerX: www.entwickler-x.de
AmiBoing: www.amiboing.de
Themes: www.amiboing.de/themes.php

M.A.C.E

ThemeScreenshot

Yet Another AmigaOS Core Developer

Please welcome Fredrik Wikström as our newest AmigaOS core developer.

Fredrik has been busy on the Amiga scene for many years and has produced a lot of very useful software. More information about Fredrik’s work can be found on his web site.

Given Fredrik’s exemplary track record, I think we can all expect some very good things from him.

If you have used Fredrik’s software in the past, please also consider a donation to show your appreciation. If bounties are more your thing, he has been working on a few at Open Amiga.

Steven Solie
AmigaOS Development Team Lead

New AmigaOS Core Developers

Please welcome our newest AmigaOS core developers: Alex Carmona and Lyle Hazelwood.

Alex has been a dedicated beta tester and programmer for many years. He has decided to take the leap to OS development. Now he can fix all those bugs he has been filing…

Lyle took the plunge and purchased an AmigaOne X-1000 system. He has been a very active developer ever since. More information about Lyle can be found at his web site.

Their names have already been added to the credits page on the AmigaOS web site. You can also find both of them actively helping people on the AmigaOS support forum.

Memory Page Locking Confusion

With the introduction of the AmigaOS Documentation Wiki and the uploading of the latest Amiga ROM Kernel Reference Manuals we have been given the opportunity to bring the Amiga Operating System’s documentation up to date. Although not complete, several dedicated volunteers have been slowly improving on that documentation.

One of the most confusing changes to the API has proven to be Exec’s new LockMem() and UnlockMem() function calls. These functions are use to lock and unlock memory pages. Normally, a programmer should not have to worry about memory pages nor worry about what the swapper is doing. Such details should be left to the operating system. However, in special circumstances (e.g. device drivers) it may be necessary to use LockMem() and UnlockMem().

What really confused programmers is the fact that we advocated the use of UnlockMem() even when the programmer did not explicitly call LockMem(). The issue was primarily centered on MEMF_SHARED memory which, by default, is automatically locked. That in itself isn’t really a problem. Add in the fact MEMF_ANY memory is translated to MEMF_SHARED memory and you have the potential for a lot of locked memory pages and thus ineffective use of system memory.

AmigaOS programmers are always trying to be helpful. Pretty much everyone jumped on the bandwagon and declared this API change as a travesty of enormous consequence. Why didn’t the FreeVec() function unlock the memory for them? Why must UnlockMem() be provided the size of the memory to unlock? They were very close to rioting.

After much discussion and fact gathering, the AmigaOS development team have officially changed our minds. The new advice is to never call UnlockMem() unless you explicitly called LockMem() in the first place. The wiki has already been updated to reflect this.

AmigaOS programmers are also rather attentive. “What happens to those implicitly locked memory pages when I don’t call UnlockMem()?” The truth of the matter is that some pages will actually be leaked. That is, some of those memory pages will not be moveable and thus not swappable. This does not mean the memory is not available to the system. This means the underlying memory pages cannot be swapped.

Now before resuming the riots, please take a moment and think about this. The system has been running fine for years now without programmers calling UnlockMem() on implicitly locked memory. We only recently explained this on the new documentation wiki. Something deeper is indeed going on here.

Without going into all the internal details, the system will automatically unlock memory that has been freed even if you forgot to call UnlockMem() on that memory. This is not true for every memory allocation but it is true for the majority of the memory in the system. This is why everything has been running fine for years now.

So what is going to happen to those lonely locked pages that can’t be automatically unlocked? The plan is to modify Exec’s discrimination between MEMF_SHARED and MEMF_PRIVATE memory. They will be moving to their own page caches. This effectively eliminates the problem entirely without a single line of code changed in your applications.

We are very sorry for the confusion. In the end, we figured it all out and I think this new advice (only call UnlockMem() if you called LockMem()) is very easy for programmers to follow.

Steven Solie
AmigaOS Development Team Lead

Xena – Signs of Life

The first tools for programming Xena (no, not the Warrior Princess) are now ready for distribution.

The XMOS XS1-L2 chip built into each AmigaOne X1000 with a unique local bus connected directly to the PA6T.

These tools provide the ability to load programs directly into the the Xena chip as well as monitor most of the internal registers as the code is executed.

Unlike other XTAG devices, no adapter board, USB port, or JTAG cables are needed. We can access all of this directly from the AmigaOS command line.

In addition to the new tools, a new AmigaOS resource named xena.resource is included to arbitrate access to the new hardware features.

I have also included a simple .xe file that will flash two LEDs on the AmigaOne X1000 board, as a “proof of programming” demonstration.

These tools were ported from code provided by Segher Boessenkool. Please be sure to thank him if you see him around the XCore Exhange Forums.

The tools are available now and can be downloaded directly from Hyperion’s web site by all registered X1000 users. The new tools and documentation will be added to a future release of the AmigaOS SDK as well.

Information on XMOS chips, compilers and other tools are available at their web site.
Excellent community resources may be found at the XCore Exchange.

I hope these tools provide a beginning into the exploration of these Xciting new features.

For more information about me please visit my web site.

LyleHaze