Archives for Compiler Core

Leu Icaros64

Posted by ALB42 on 3. Oktober 2019No Comments

Paolo Besser works on a 64bit Version of Icaros (see also his website about it). I only try my AROS64 bit FreePascal stuff with my old linux hosted AROS 64 version, and there it worked fine. But the ABI of x64 AROS is not fixed yet, so still changes can happen. And there are some changes, Leu for example crashes on the Icaros64 alpha version, Paolo sent me. It crashes very early in the startup code. I started to debug that stuff but it’s very hard to debug at this point, that early in the startup. But my initial guess was more or less right, because the AROS devs tried to implement SMP they changed a lot of structures for additional locks and so on, so the Offsets of fields are moved. Usually an easy job just compare the the new C includes with my pascal counter parts. but it makes it much harder if defines and alignment comes to play, here also. The define was easy, there is a __AROSSYSTEM_SMP__ define which seems to be always enabled, even AROS is not compiled for SMP use (seems still too unstable) which includes and additional spinlock_t to MsgPort and Sempahore, so far so well. But it seems this spinlock_t is huge, by the definition I would say an Integer and a Pointer, but the size needed in the structure to cover up the missing offset shift is much more… something around 256 bytes. The structure is 128 aligned, this could be a key for that.

Usually that is not hard to find out, just make a small C program and test the sizes and offsets of the related structs, but… who have guessed it … on Icaros64 that is not that easy, because the gcc installed does not work, it just crashes. And btw. collect-aros, which we need fore linking in FreePascal also crashes, so for sure no compiling directly on the system.

But just by try and fail/crash and comparing the results of process and task structure to the output of Scout I got a somehow working version, it’s a very dirty hack, but for now it works, Leu works, FPC works (until collect-aros crashes when try to link, but without linking, everything is fine), FP-IDE is working (with the same problem like FPC of course).

Leu and FP-IDE on Icaros64

There is some more work to really make that structures correct, but for that I will wait until the gcc is working again. The described way to compile stuff for 64bit in C on the icaros webpage using metamake is hilarious if you just want to compile a hello world it’s way overblown. I tried it, but it does not work, also compiling AROS64 from source does not work, I guess you need some pesky parameter when calling configure which, of course, are not described.

Taming my dragon

Posted by ALB42 on 2. März 20192 Comments

I bought a Draco, if you do not know what it is, its a Amiga compatible Clone used for Video editing. It does not have the traditional Amiga Chipset (ECS/AGA) therefore nearly all games will not work. But Workbench based programs and RTG compatible programs will work. For me only interested in FreePascal and stuff its the perfect RTG Amiga and it’s much cheaper than any other Amiga with RTG, Soundcard and a 060 card + 128 MB RAM.

I replaced the (very noisy) Harddisk with a SCSI2SD V6 and removed one of the fans (which was basically for cooling the HD) and now the Draco is nearly silent and very fast, the SCSI2SD V6 reaches the 10 MB/s easily with the Draco SCSI. The Draco also supplies some standard Zorro II slots, I tried a X-Surf 100 to get network access to it (mainly to bring my FreePascal stuff there) but sadly it does not work, even I found some reports that it should work. Maybe the Zorro in my Draco is somehow broken, I cleaned already all the connectors and checked voltages of the power supply but no luck. To get at least slow network I connected an old Raspi via Serial cable to the Draco and used PPP like in the good old days. It reaches the maximum 115.200 baud but even that feels very very slow, at least it works. Maybe later someone can find out whats wrong with the Zorro.

FreePascal works nicely. But some other programs, which should be working (because only use standard API no hardware access) does not for example Delitracker with AHI-Noteplayer seems not really working nicely. Needs a little bit more investigation.

Datatypes for all

Posted by ALB42 on 14. Mai 20185 Comments

I created the datatypes unit for AmigaOS4 and AROS and rewrote the ones for Amiga68k and MorphOS. OF course the aim is again to make them more or less compatible to each other. Of course also wrote a little test program and compiled for all systems which works nice.

DataTypes test for Amiga68k, MorphOS, AmigaOS4 and AROS

Now I have to learn how to actually use Datatypes in a program, especially seems sometimes the datatypes seems to scale down to 8 bit (256 colors) even the screen has 24 bit which I do not understand and if there is a way to let the datatype scale the image to the right size. I use the BitmapScale from graphics.library but this is awfully slow (At least on AROS, even my straigh forward nearest neighbor routine is faster than that).

Installing FreePascal …

Posted by ALB42 on 8. April 20182 Comments

Working on a Installer script to installing FreePascal for Amiga using the official way. In the past I often used Installer scripts but never wrote one. It looks very close to LISP (brackets everywhere 😉 ) luckily the scripts are just saved as text. You can peek into other scripts, how to make some of the more difficult parts. For the basic structure I found a nice program in Aminet InstallerGen, it provides a nice MUI Gui to create the install steps, which also helps to understand the language.
The first version is finished now. You can select the installation type, minimal, typical, full and customized. On customized you can select the packages to install. (see image) Of course it also adds the needed entries for assign and path to the user-startup. The plan is to create some more such installers to install more packages, afterwards, like LCL, MUIClass or the FPU-enabled units.
For MorphOS and OS4 this should be easy to create from it. On AROS it will be a little bit different, there is no need to alter the user-startup, you only need to register it as a package.

Vampire V2.7 with FPU

Posted by ALB42 on 2. März 20188 Comments

The new Vampire firmware is released V2.7 which contains a Hardware FPU in the FPGA (some seldom 68881/68882 commands are still emulated, like the 68060/68040 also do) but nevertheless, thats very nice for my MUIMapparium (you remember the problem?). Of course I flashed directly the new version, first the bad news, it’s VERY unstable for me, it’s said it needs some soldering because there are some errors on the early Vampire cards which make the power supply to the FPGA bad… something like this and I’m affected with that … so it’s a little bit annoying to work with it because there are drawing errors on the screen and it crashes often. I reduced the screen resolution and ended all background program which made it much better. But nevertheless to really try it out I have to wait until someone fix my card. I can’t do that myself, most scary thing in the world a coder with a screwdriver let alone a soldering equipment :-P.
But this was not the topic of this post. I tried MUIMapparium FPU version on my new Vampire 2.7… good news it starts does not crash, bad news the map stays empty. The same Executable worked well with FEmu (I checked especially before I flashed the new one) on the old Gold 2 and still work in UAE. But the FPU calculation seems to work because the mouse pointer movement shows reasonable coordinates. I was a little bit surprised because even the GUI in the map window was gone. I checked the code ah yes there is a tiny floating point calculation, fine let’s see whats that. An my guess was right it is the floating point calculation, the Button size is calculated by the Font Size * 1.2 to have a little bit more space around it. After adding some debug output it seems that the floating point calculation works well but the rounding always return zero, so I wrote a little test program to test the rounding here the outputs of the testprogram in my setups:

Vampire 2.7 Amiga 1200/030/68882 UAE 68060 emul
a := 5 = 5
a * 1.0 = 5.000000000E+00
Round(a * 1.0) = 0
b := a * 1.3 =  6.499999523E+00
Round(b) = 0
Ceil(b) = 7
Floor(b) = 6
Floor(b) = 6
press enter
end
a := 5 = 5
a * 1.0 = 5.000000000E+00
Round(a * 1.0) = 5
b := a * 1.3 =  6.500000000E+00
Round(b) = 6
Ceil(b) = 7
Floor(b) = 6
Floor(b) = 6
press enter
end
a := 5 = 5
a * 1.0 = 5.000000000E+00
Round(a * 1.0) = 5
b := a * 1.3 =  6.500000000E+00
Round(b) = 6
Ceil(b) = 7
Floor(b) = 6
Floor(b) = 6
press enter
end

Na? who spots the difference. Funny that only Round() have this problem but ceil, trunc, floor not. This also explains why MUIMapparium shows no maps at all, if all is rounded to 0. Ok I have to wait until they fix that… yeah I could replace all Round(a) by Floor(a+0.5) but why should I do that, here is clearly something broken in the FPGA.
 
You want to try on your own computer – Exe for m68k with FPU: TestFPU and the source: TestFPU.pas

Updated vlink and vasm for fpc

Posted by ALB42 on 12. Februar 20182 Comments

The creation of vlink and vasm for Amiga is always a little bit difficult, hence I seldom do that. But Charlie told me, that the creator of vasm and vlink also supply executables for most of the platforms. Even the rather uncommon 68k asm with standard syntax (instead of Motorola syntax) are there. Long text, short meaning, the vlink and vasm for Amiga 68k, Amiga OS4 and MorphOS are updated in the nightly Release packages. And with this it will be much easier to keep them up to date.

Unify ASL

Posted by ALB42 on 21. August 2017No Comments

I checked the ASL.library units of MorphOS and Amiga 68k against the official C includes of the SDKs. Especially the TFileRequester structure was always a little bit trouble because the old Amiga asl unit still used the old field names rf_* but from V38 of the Library this fields are all names fr_* and some other tiny dame differences (e.g. Dir vs. Drawer). In AROS and AmigaOS4 I only added the newer names because here I do not have any “old” code. This resulted in a big inconsistency between the platforms and need ifdef’s in the final programs. To prevent a direct breakage of the existing sources (e.g. LCL, MUIMapparium) Amiga and MorphOS have both field names in the structure (as case). The aim will be to remove that ifdef’s from the sources.

Amiga Linux the next day.

Posted by ALB42 on 3. Juli 2017No Comments

Of course the freepascal compilation did not work, still needs some link library and stuff. (Hint for later at least libc6-dev and libgcc-5-dev is needed) I restarted first a plain compilation and hmm it still runs 😉 the tests will need to wait a little bit. Besides that of course I played more with the system, of course its even slower that way because the compilation is running in parallel. Installed some packages I need on there, telnetd for example, sshd just slows the poor 68030 to much down. For file transfer I want to join it the normal NFS network I have already. But when installing nfs-common I was really shocked it has python as recommend, really? Python for NFS? On my normal computer I do not care it’s usually there already but on Amiga, space (and more important installation time) is short 😉 so need to with “–no-install-recommends” But my personal opinion is when a simple network file protocol from the early years of Unices pull the most hyped script language monster (15.5 MB additional crap!) then there is clearly something wrong in your package system.

IPv6 on an Amiga

As you can see the compilation still runs 😉 (the Assembler as runs currently on other TTY) and even IPV6 works! An Amiga doing IPV6, welcome to 21. century (no Amiga OS is currently able to do that). The system runs rock stable (uptime already 28 hours) even I stress it with parallel installation and other changes and let the compilation run. Very nice and stable just a little bit slow.

ALB goes crazy episode 9842: Linux on 68030 Amiga

Posted by ALB42 on 2. Juli 20174 Comments

To help ChainQ with the FreePascal for m68k compiler bug finding I installed a qemu-m68k on my server together with a chroot environment to let the FreePascal testsuite run. An ChainQ really found a lot of bugs. I also improved the chroot environment. The number of fails falled down from 160 to 34 which is a good value more than i386 (around 20) but better than PowerPc (around 40) 😉 already a nice achievement. But I thought it would be even cooler to start the test suite directly natively in a m68k Linux environment. (also qemu-m68k maybe has some bugs :-P) The only real hardware I have available for that is my Amiga 1200 with 68030, luckily it has a MMU so Linux will be possible. (The Amiga 600 with Vampire does have no MMU therefore Linux is not possible to use) Informations about this topic are very hard to get and most of the links are dead today. So in the end I compiled the kernel (3.16.7) on my own and created a initramfs by hand. For initial root I used the chroots debian environment I use already for the fpc tests. But systemd made me much headache, it was much easier to write a own little init script. Finally the systems boot and even the PCMCIA card works, I can SSH to that server.

Linux on a Amiga 1200 with 68030

The next is installing FreePascal and try to run the compilation and finally the test of course. There are some unexpected problems here. Usually the latest FPC from trunk requires the latest stable FreePascal version, that means at the moment 3.0.0 or 3.0.2 but both can’t compile the trunk because some features missing in 3.0. I tried to merge them from the trunk but failed to many things missing and not easy to identify which else should be merged. So I will use the 3.1.1 as starting compiler (and ignore the warning that this is not supported 😛 yeah I’m a real outlaw) sadly it also have some problems I have to check later. For first try I just start the testsuite with the compiler cross compiled on my other Linux (x64) computer. The testrun started today at 22:15 will it run completely and if yes how long will that need? a week? 🙂 We will see.

Compiling/Testing FPC on m68k Linux

Playing with locale

Posted by ALB42 on 22. Mai 2017No Comments

Still playing with locale settings on Amiga systems. FreePascal offer a possibility to care about international settings like decimalseparator, date formats and currency symbol. So I looked a little bit into and the system is comparable, still much work, so I’m not sure if really put to every start of a program, or maybe it would be better to put it to a separate unit and you have to include and invoke it by hand. We will discuss that.

Testprogram with Germany, USA and UK Locale Setting

Starting the same test program 3 times and changed the Locale Country setting in between. (For me rather funny that the currency format for positive and negative Values are so different not so nice if you have them in a table) But so far it works very nicely.