SFS on base 68000

Shadowfire · 14 October 2009, 21:30

I have tested SFS on a dual boot (68000/68030) 3.1 system, and it guru's when booting in 68000 mode.

Has anyone attempted to compile SFS? I'm thinking about setting up a C environment and having a go at it, to try and get a 68000 version. The SFS package uploaded to Aminet has two object files, cachedio.o and deviceio.o, but no associated source files (I presume that these were written in assembler and the author didn't include the source code), which would be a problem if the incompatibility is within these files.

Photon · 14 October 2009, 21:49

SFS is made for 020+ only.

Shadowfire · 15 October 2009, 01:18

Yeah, I was wondering how difficult it would be to modify it for 68000. Going to be very difficult without the source code for the assembler routines, though.

Photon · 15 October 2009, 01:23

Well, it either takes releasing the sources openly, or someone resourcing the binaries and spending the time to make them work, then replace the non-68000 instructions with 68000 instructions.

If the sources are available, it's probably a matter of changing a setting and recompiling them. If the sources were available, that'd probably been done by now?

Dunno why it was originally 68020+ only, but surely it must have something to do with an enormous speed gain.

Sorry. I really must stop nibbling those spotty sarcasm-mushrooms!!!!1one

Cammy · 16 October 2009, 00:03

You might be better off looking for PFS3 or AFS to use on your 68000 partition. Both of these file systems are superior to FFS, and in some ways can perform better than SFS anyway.

alexh · 16 October 2009, 09:39

SFS is an open source project. Hence why there are ports to MorphOS, AmigaOS4 even AROS.

Because it has targeted native x86, PPC and M68k CPU's I doubt very much if there is much raw assembler in the code. Meaning it may be possible to create binaries which do not need 020+ cpus.

Photon · 18 October 2009, 20:22

Yeah, right now it's targeted at a subset of M68k CPUs.

But yes, it would be cool if they could recompile it so all Amiga owners can enjoy it!

alexh · 18 October 2009, 21:27

No-one is going to do it for you. Either you get to grips with Amiga development, C compilers etc. or you go without.

And as I've said in another thread the updates over the last 10 years have not been released. Version 1.84 (i.e. 1.084) is the last for which the source is available. The latest being 1.279

Shadowfire · 20 October 2009, 14:07

Actually, it looks like they have a CVS repository up with the sources. If I get the inclination, I'll pull them & see how much work it is to recompile.

alexh · 20 October 2009, 14:16

All that is, is the source included with SFS 1.084 archive. There have been no commits since the original files.

The sourceforge area was set up by Michal Schulz of AROS not either of the SFS authors.

Photon · 20 October 2009, 19:32

Quote:

Originally Posted by alexh

No-one is going to do it for you. Either you get to grips with Amiga development, C compilers etc. or you go without.

And as I've said in another thread the updates over the last 10 years have not been released. Version 1.84 (i.e. 1.084) is the last for which the source is available. The latest being 1.279

If you knew me (as your last two smug comments suggest you don't), I'd sooner resource all binaries and remove the incompatible instructions than install a c development system on my Amigas. I'm an assembler coder and defender of OCS and 68000 when (it seems) most Amigans think the minimum Amiga to care about is an A1200 with turbo card, and all the rest can be thrown away. There are too many slow and big c programs for Amiga made through history already anyway, so what I do I do in asm for OCS/68000 AND UP.

Surely, though, you see I'm not going to do all that work if someone that already have have a c devsys set up and know what they're doing can simply get the latest released source and set a few flags and recompile it.

Developing for a subset of M68k has been a common issue on Amiga, and I'm sure I'll see it for future software. It's a complete showstopper and criticizing it is always correct, regardless of whether the development progress/circumstances are known or not.

alexh · 20 October 2009, 19:47

Quote:

Originally Posted by Photon

I'd sooner resource all binaries and remove the incompatible instructions than install a c development system on my Amigas.

Heh.

Quote:

Originally Posted by Photon

I'm not going to do all that work if someone that already have have a c devsys set up and know what they're doing can simply get the latest released source and set a few flags and recompile it.

When you first posted I asked BuZz of Exotica to take a look at it. He is a good developer and has ported C-code programs from OS4 to 3.x in the past. He was a bit concious that the only source was the 1.084 and if anyone would actually want it.

Quote:

Originally Posted by Photon

Developing for a subset of M68k has been a common issue on Amiga, and I'm sure I'll see it for future software. It's a complete showstopper and criticizing it is always correct, regardless of whether the development progress/circumstances are known or not.

Depends if the application is impractical for a 68000. Obviously in the case of SFS, a file system, it is not.. but I would question the rationale of say a 3D graphics card driver or a video media player.

Photon · 20 October 2009, 20:47

Well, those would be impractical for any Amiga less than 68060@80 too... unless you want to watch crap :P

3d gfx cards, video player, browser with scripting languages et al... the barrier that divides modern computers from even the fastest least old Motorola Amiga. For any other type of app, I'd say all of them are suited for 68000. Hell, even Real3D was usable on my A500! (Render times were long though, hehe

)

And yeah, I can chat with Buzz what he thinks on irc tonight.

Shadowfire · 23 October 2009, 03:21

I pulled a few of the .c files from the html repository browser.
There is next to no information about the compile environment. A quick check with StormC quickly brought up typedef errors.
Additionally, there are pragmas that are obviously assembler MULU64 directives. So, yes, there is a small amount of embedded assembler in the C files. Which is why it chokes on a 68000.
If I can make some time this weekend I'll see if I can un-optimize the code and get it to compile for a 68000.

28 October 2009, 13:36

Quote:

Originally Posted by Shadowfire

If I can make some time this weekend I'll see if I can un-optimize the code and get it to compile for a 68000.

I hope you can compile it for 68000 !!!

14 October 2009, 21:30	#1
Shadowfire Registered User Join Date: Aug 2001 Location: Connecticut USA Posts: 617	SFS on base 68000 I have tested SFS on a dual boot (68000/68030) 3.1 system, and it guru's when booting in 68000 mode. Has anyone attempted to compile SFS? I'm thinking about setting up a C environment and having a go at it, to try and get a 68000 version. The SFS package uploaded to Aminet has two object files, cachedio.o and deviceio.o, but no associated source files (I presume that these were written in assembler and the author didn't include the source code), which would be a problem if the incompatibility is within these files.

16 October 2009, 09:39	#6
alexh Thalion Webshrine Join Date: Jan 2004 Location: Oxford Posts: 14,337	SFS is an open source project. Hence why there are ports to MorphOS, AmigaOS4 even AROS. Because it has targeted native x86, PPC and M68k CPU's I doubt very much if there is much raw assembler in the code. Meaning it may be possible to create binaries which do not need 020+ cpus. Last edited by alexh; 16 October 2009 at 09:45.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Sierra, all your base are belong to us...	Fragger	Nostalgia & memories	15	16 June 2010 10:31
Base Jumpers	Graham Humphrey	request.Old Rare Games	22	16 July 2009 20:51
Base Jumpers CD32	sut	support.WinUAE	1	13 March 2006 22:15
Game Base 64 V2	plasmatron	Retrogaming General Discussion	14	03 September 2004 23:02
Base Jumpers CD³²	MethodGit	request.Old Rare Games	4	22 January 2002 05:32

14 October 2009, 21:49	#2
Photon Moderator Join Date: Nov 2004 Location: Eksjö / Sweden Posts: 5,602	SFS is made for 020+ only.

15 October 2009, 01:18	#3
Shadowfire Registered User Join Date: Aug 2001 Location: Connecticut USA Posts: 617	Yeah, I was wondering how difficult it would be to modify it for 68000. Going to be very difficult without the source code for the assembler routines, though.

15 October 2009, 01:23	#4
Photon Moderator Join Date: Nov 2004 Location: Eksjö / Sweden Posts: 5,602	Well, it either takes releasing the sources openly, or someone resourcing the binaries and spending the time to make them work, then replace the non-68000 instructions with 68000 instructions. If the sources are available, it's probably a matter of changing a setting and recompiling them. If the sources were available, that'd probably been done by now? Dunno why it was originally 68020+ only, but surely it must have something to do with an enormous speed gain. Sorry. I really must stop nibbling those spotty sarcasm-mushrooms!!!!1one

16 October 2009, 00:03	#5
Cammy Registered User Join Date: Aug 2007 Location: Tasmania, Australia Age: 39 Posts: 1,189	You might be better off looking for PFS3 or AFS to use on your 68000 partition. Both of these file systems are superior to FFS, and in some ways can perform better than SFS anyway.

18 October 2009, 20:22	#7
Photon Moderator Join Date: Nov 2004 Location: Eksjö / Sweden Posts: 5,602	Yeah, right now it's targeted at a subset of M68k CPUs. But yes, it would be cool if they could recompile it so all Amiga owners can enjoy it!

18 October 2009, 21:27	#8
alexh Thalion Webshrine Join Date: Jan 2004 Location: Oxford Posts: 14,337	No-one is going to do it for you. Either you get to grips with Amiga development, C compilers etc. or you go without. And as I've said in another thread the updates over the last 10 years have not been released. Version 1.84 (i.e. 1.084) is the last for which the source is available. The latest being 1.279

20 October 2009, 14:07	#9
Shadowfire Registered User Join Date: Aug 2001 Location: Connecticut USA Posts: 617	Actually, it looks like they have a CVS repository up with the sources. If I get the inclination, I'll pull them & see how much work it is to recompile.

20 October 2009, 14:16	#10
alexh Thalion Webshrine Join Date: Jan 2004 Location: Oxford Posts: 14,337	All that is, is the source included with SFS 1.084 archive. There have been no commits since the original files. The sourceforge area was set up by Michal Schulz of AROS not either of the SFS authors.

20 October 2009, 20:47	#13
Photon Moderator Join Date: Nov 2004 Location: Eksjö / Sweden Posts: 5,602	Well, those would be impractical for any Amiga less than 68060@80 too... unless you want to watch crap :P 3d gfx cards, video player, browser with scripting languages et al... the barrier that divides modern computers from even the fastest least old Motorola Amiga. For any other type of app, I'd say all of them are suited for 68000. Hell, even Real3D was usable on my A500! (Render times were long though, hehe ) And yeah, I can chat with Buzz what he thinks on irc tonight.

23 October 2009, 03:21	#14
Shadowfire Registered User Join Date: Aug 2001 Location: Connecticut USA Posts: 617	I pulled a few of the .c files from the html repository browser. There is next to no information about the compile environment. A quick check with StormC quickly brought up typedef errors. Additionally, there are pragmas that are obviously assembler MULU64 directives. So, yes, there is a small amount of embedded assembler in the C files. Which is why it chokes on a 68000. If I can make some time this weekend I'll see if I can un-optimize the code and get it to compile for a 68000.

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)