RISC OS 5 on RISC PC

Hi Everyone,

I’m having some problems getting RISC OS 5 to soft-load on my Kinetic RISC PC (RO4 in ROM, 128mb RAM on kinetic, 32mb on motherboard, 2mb VRAM)

I’ve successfully flashed a 32bit clean ROM onto the NIC and don’t have any other podules installed.
I downloaded and installed a clean !Boot file from the ROOL site and installed the soft load tool in boot.choices

When the machine boots I get the usual RISCOS 4.03 158MB screen and then Error: File ‘System:Modules.Clib’ not found (Error number &D6) and it leaves me with a flashing * promt.

What am I doing wrong?

Regards,

Lee

Did you do a power-on-Delete reset? It sounds like there are some modules unplugged, the meaning of the unplug CMOS RAM settings differs between RO4 and RO5. The unplug lottery has decided to unplug the SharedCLibrary, which means it’s Game Over very soon after.

You also need to download the “system resources” archive (PlingSystem.zip). That contains all the modules that are needed to make the RISC OS 5 hard disc image work properly on pre-RISC OS 5 machines. Without that, I don’t think the boot sequence will get far enough for the softload to kick in.

Did you do a power-on-Delete reset? It sounds like there are some modules unplugged, the meaning of the unplug CMOS RAM settings differs between RO4 and RO5.

The reset is listed in the supplied instructions with the ROM version but I don’t note it in the web site softload instructions

The unplug lottery has decided to unplug the SharedCLibrary, which means it’s Game Over very soon after.

I thought the NIC reflash put a CLib copy in part of the NIC ROM so it is there very early in the boot.

*modules should theoretically show the modules loaded.

My bad, I hadn’t installed the !System update.
On boot up it now asks me if I’d like to softload the new OS. If I press Y I get…

Sofload version 1.19
Platform class: RiscPC
Traditional memory map: Yes
ROM Type: Hard
and then several lines of logaddr, physaddr and checksums etc. at which point the whole computer freezes completely with the blue egg-timer pointer in the middle of the screen not even allowing a ctrl break.

Lee

I’ve tried doing a delete power on and I get the same behaviour

Any ideas anyone?

If it crashes while it’s still displaying the softload tool then that suggests it could be the softload tool which is crashing, rather than the ROM. Try doing the following:

• Boot into the desktop (i.e. decline the softload)
• Hit F12 and enter the following:

spool $.log
run <softload$dir>.softload -v -m 12m -d <softload$dir>.riscos .NONE.
spool

Then post the contents of the $.log file to the forums. The ‘-d’ option should make the softload tool do a dummy run, i.e. it’ll go through all the motions but then bail out just before it would normally start the new ROM. If the tool crashes then you’ll probably have to manually make a note of whatever was printed to the screen, because it probably won’t have all been written out to the log file. This includes making a note of all the numbers – they might be useful!

Hi Jeffrey,

I’ve done that and got a log file but how do I post the output here? It looks like RAW data.

Lee

If I change its type to Text – this is what I get. although the RAW data part at the end is snipped as its pages and pages long

*run <fsoftload$dir>.sorftload -v -m 12m -d <sofltloafd$dir>.riscos .none. spool

Softload version 1.19

Platform class: RiscPC

Traditional memory map: Yes

ROM type: Hard

Hard ROM:

Soft ROM:

êÎcÌŷò©H@‡+TzÑ¡\LÇó×á‰6!×ü/=¹€ô)ú⇓EƒÉN±Ë²©Üíb⇐ÎÏú_¬jÑ*,ñ@+„qmõãG†Òå¤ÙæöŶ&\×ŷGè—¸´®Ì²g4o¥Ïe£êTŒõESnÄgïaÉ4
¢#öƹlÅÍs‡‰½ºUt;I†!✘†å8”åôœ$‰†¨²¹kt$dã®æyð‡ü}ªµŶÁŴ3Êj,4⇐¥4fm•ÊH9˸妢—©`’ÉÊ`ç¬Ò·L(ñkkÖF ²ª"`Øu&O+€#<¨–éo5Àê‹Œ⇒⇒øOÌuŷ-¦IòÈ2dŃ”ðÜÿ‰»ÕÄù礣…ø@*lEÚi<¯’¡Îks°`z⇓3ÔáLLe£ÏÉSÖ¸H²ñV4R^µ‡e’"mÐ ô’÷/ŵõå·öâ]FyÉhèrPS0½ÇgƇ(|j¹‡`ûP臽©ÇüWPãȾ5GüÄPj@_BNGÒ=k2¤⇓ì(2ÆHká™­Có@8fit›Ûh¥üÚôï­¤*±l‰ÊeÑk^“Tt✘ٯЗ`É−HOt.nã1t }éÈg
ÌóiÀ⇒Z
Ń>¨’mß•ä_a»®þh«zT^´R/ M…ä⇒âD)/ïJC6Vzî
íŶ+o¦mYNóTV°ñzVüfp¶£(hë#xòÂx.[ òœfl·GºÈSõùŒ”¢−?ø#—A¹H&øƒùqÝE­óC”ñ÷U‹OÊq{¢xÒwùÞÔ€œ9MøóÞ=qɤï…e³5æÂX/¡£-Ð ¬f{½\³‰¼7™yŒu([ò†dæ<=»C׿êQíän›ãýÅøy,½XL ŒÒ
ÃÞ−ƒØT‰X)ÜqÕ\f4

Hmm, that suggests it crashed, or didn’t close the log properly for some reason. Maybe copy the text by hand? (It should just be whatever the softload tool prints out to the command line when you run it)

This is what I get….

Softload version 1.19
Platform class: RiscPC
Traditional memory map: Yes
ROM type: Hard
Hard ROM:
logaddr: 80000000
physaddr: 00000000
size: 00400000
checksum: 0ff0324a
Soft ROM:
logaddr: 03800000
physaddr: 00000000
size: 00400000
checksum: Off0324a
soft-loading ROM ‘<Softload$Dir>.riscos’…
size: 00400000
checksum: e845082b
Npages=40448 FreeSize=153288k
physmem @ 000001c0
chunk addr chunk size
02000000 00200000
20400000 03c00000
30000000 04000000
10000000 01000000
18000000 01000000
foundpage=39424 foundphysaddr=18c00000
_

and there it hangs

Lee

I’ve tried a few things but still can’t get it to work…

So far I’ve tried removing one stick of RAM on the mainboard (leaving just 16MB in Slot 0 and 128MB on the Kinetic), removing the 586 PC Card, several delete power ons, *unplugged the CallASWI and ShardCLibrary that are flashed on the NIC (ROOL flash update) and it still hangs as above…

Maybe I should stick to RO6 :-(

Lee

No need to go crazy and tear your machine apart! I think you might be the first person to have tried using the softload tool on RISC OS 4.03, so a couple of teething problems aren’t too unexpected (the development of Kinetic support for RISC OS 5 was done using an older Kinetic card that didn’t have the RISC OS 4 flash ROMs on board)

After checking the code it looks like there are a few places where it could be crashing:

• Creating the dynamic area to load the ROM image into (possible)
• Loading the ROM into the dynamic area (unlikely)
• Patching the ROM (unlikely)

If those steps had completed OK you would have seen a “dummy run complete” message (or it would have started the new ROM if the -d option wasn’t present).

When I get home I’ll make a new build of the softload tool with some extra debugging code, to see if we can narrow down the problem further.

A couple of extra angles here:
You don’t mention anywhere which ROM you’re attempting to softload, whether it’s the stable 5.20 , or bleeding edge 5.21? It’s possible something’s got broken since the stable release.

The softload tool itself certainly works on a normal StrongARM and 4.02, it doesn’t look to have changed for 12+ months and I understand 4.03 is almost the same as 4.02. One option therefore is to switch the Kinetic for a normal StrongARM temporarily (either via the jumper link on the Kinetic, or an actual Acorn card).

The IOMD HAL does check to see if it’s already loaded from RAM (so as not to mess up the read only RAM controller settings).

Jeffrey RISC OS 4.03 is on the main board in 2 ROMs as usual not on the kinetic card. I believe that RISC OS 4.04 was the version flashed to some kinetics onboard flash memory (my kinetic hasn’t got the flash chips on it)

Sprow ive tried both the stable and the development versions and get the same results for both. I have got a vanilla strongarm card so ill give that a go, but I would obviously prefer to get the kinetic card running.

Over here is a new build of the softload tool. I’ve added a bit more debug output, some extra error checks, and a couple of options to enable some potential fixes (assuming it’s the dynamic area bit which is going wrong). Try extracting it to the root of your hard disc and then running it from the command line, e.g.

$.softload -v -m 12m -d <softload$dir>.riscos .NONE.

Make a note of anything new that’s printed out after the ‘foundpage=…’ bit. Assuming this new version still crashes, there are also the two options you can use to enable the potential fixes. These are -fix1 and -fix2. E.g.

$.softload -v -m 12m -d -fix1 -fix2 <softload$dir>.riscos .NONE.

Note that it’s important to include the fix options before the ROM filename, otherwise they’ll be ignored!

If those fixes work, i.e. it gets as far as ‘dummy run complete’, then can you also try again without one of the two fix options to see if we can work out which one was necessary? E.g. fix1 disables the cache for some of the code, which isn’t ideal. You should also be able to try without the -d to see if it actually reboots properly!

If this new version does work then you should be able to copy it into !SoftLoad, and modify Loader to add the relevant fix option(s).

Hi Jeffrey,

Thanks for your continued help but I’m afraid it’s still a no go. It hangs at exactly the same place with or without the fixes. Out of interest I unplugged the kinetic and installed a 233Mhz strongARM and using the standard softload tool this allowed me to boot into RISC OS 5, but only after unplugging callASWI and sharedCLibrary from the NIC, if i leave them plugged in I get an internal error.

Lee

Any more thoughts on this?
I would like to get the softload running and do some testing before I splash out on a new set of ROMs.

Lee

I still can’t get RISC OS 5 to softload on my kinetic :-(
Please help… I’ve had enough of RISC OS 6’s bugs!

Lee

Please help… I’ve had enough of RISC OS 6’s bugs!

bugs? Apart from the zero development status I thought it was a stable release.

There’s plenty of little niggles with it… its like a poor beta. Two bugs that immediately spring to mind are the configure options for button colours don’t ‘stick’, neither does the shareFS on/off button. To me it feels much less polished than RISC OS 4.39 did or RISC OS 5 on the Iyonix. I have all kinds of weird stability issues with RISC OS 6 too. YMMV

There’s plenty of little niggles with it… its like a poor beta. Two bugs that immediately spring to mind are the configure options for button colours don’t ‘stick’, neither does the shareFS on/off button. To me it feels much less polished than RISC OS 4.39 did or RISC OS 5 on the Iyonix.

I thought 6 was supposed to be the fully tested stable release of the 4.39 stream, I will take you word for the state of play (see below)

I have all kinds of weird stability issues with RISC OS 6 too. YMMV

My mileage is nil – when they were touting Select and I had money I had an Iyonix and spending on 32 bit related stuff seemed more sensible than spending on an OS that wouldn’t run on my machine.

I’ve used RISC OS 6 on my main machine 5-8 hours a day every working day since it came out. I’ve had stability issues may be three or four times a year! I’ve had very very few stability issues with RISC OS 5 but my usage is more than an order of magnitude smaller. I expect that the features not in RISC OS 5 that I find very useful will arrive, but unfortunately progress is not quick and don’t get me started on CTRL K…
I now run a RO 6 RiscPC and a PandaRO on same desk and use the PandaRO when speed is important. Having two machines, keyboard and mice, even if they were the same OS is very useful I don’t know why I didn’t do it before.

Lee, in case you (or anyone else) comes back to this topic in future, please see my post in the thread at https://www.riscosopen.org/forum/forums/5/topics/2872 where I describe a recipe for booting (more or less) directly into RO5 on a RO4 machine.

It is not a simple “dual-boot” (I have two !Boot directories) so far but I suspect some streamlining will be possible – for instance, much of what is in !Boot.Resources is 26/32 neutral so should not need to be duplicated in full (I could set respective Boot$Resources variables to point to a common directory, or use a symlink utility)