SysInfo 4.3 ready, finally

[npagonis73]

Originally Posted by TjLaZer

When I try SysInfo 4.1 to 4.3 on an A1200/Vampire 2 with Classic WB 3.9 the machine crashes and locks up. When I run it from early startup, it runs fine. Any ideas why?

As Hercules said I also think it is a stack problem. Either install stackattack2 or increase stack size on the sysinfo icon.

[TjLaZer]

Originally Posted by Hercules

Could be stack problem? Try running it with a bigger stack and see if it helps.

Tried 8,000 and 20,000 stack and didn't help. Has to be something with the ClassicWB. No idea what it could be.

[TuKo]

Hi Geijer,

There is a long lasting but that I would love to see to be phased out : under some circumstances, the SysInfo drive benchmark can fail and will output a bogus result as 6'553'600.

Is that any chance for you to fix that and output at least an error message instead of that misleading result ?

Thanks in advance

Best Regards

[Geijer]

Do you know how to trigger that result? Sounds like overflow in calculations. Maybe some JIT magic in some emulation that screws the calculations?

Ahh, sorry, missed that you referred to the disk benchmark. I'm currently reorking that part to support much higher drive speeds...

[Aladin]

https://d0.se/

Last version v4.4 (11/11/2020)
https://sysinfo.d0.se/#Download

Last beta v4.5 beta 12 (28/01/2021)
https://sysinfo.d0.se/#Beta

[trixster]

Still wondering if we can get Super DMAC identification code back into sysinfo?

[Geijer]

Originally Posted by trixster

Still wondering if we can get Super DMAC identification code back into sysinfo?

It is on the todo list but unfortunantely not at the top, yet...

[lioneer]

Its possible now to use sysinfo with something other than classil PAL? When i had rtg card, sysinfo was unusable (or with DraCo)

[Geijer]

Originally Posted by lioneer

Its possible now to use sysinfo with something other than classil PAL? When i had rtg card, sysinfo was unusable (or with DraCo)

No, not yet. SysInfo uses PAL/NTSC to produce comparable numbers. We are working on finding a solution but there is a few things that has higher priority right now.

[paraj]

It guru's (0x8000000B) on my phase 5 blizzard 1260/50 after showing the main screen (with and without running startup-sequence). Same for v. 4.4 and v. 4.5b12 (also tried some older versions). It works if I disable it (by pressing 2 during boot), but otherwise let e.g. IDEfix run and the reboot).

TBH I haven't done much "modern" (68000+) amiga dev, but if anyone can recommend a decent debugging setup for this I can probably provide better info.

[daxb]

You may have a look at the mmulib archive: http://aminet.net/package/util/libs/MMULib

[paraj]

Originally Posted by daxb

You may have a look at the mmulib archive: http://aminet.net/package/util/libs/MMULib

Just to be clear: If I install it and run MuForce I'm supposed to see something useful, right? Because it still crashes in the same way. I know this isn't a support thread for mmulib, just making sure I'm not missing something obvious (Also tried also running MuGuardianAngel and just plain enforcer with the same result).

[Geijer]

Originally Posted by paraj

Just to be clear: If I install it and run MuForce I'm supposed to see something useful, right? Because it still crashes in the same way. I know this isn't a support thread for mmulib, just making sure I'm not missing something obvious (Also tried also running MuGuardianAngel and just plain enforcer with the same result).

daxb is correct, the common solution to Guru 8000000B on '060 CPUs is to use 68060.library. Please try booting from the disk version (dms) from http://sysinfo.d0.se/#Download instead that has a general 68060.library.

[daxb]

I was not correct because I was only referring to the debugging setup and not the bug itself.

However, I would recommend to read the docs carefully. Usually debug output is sent to serial and can be captured via cable connected to another computer which runs a software listening to the serial port. Then there is sashimi which can redirect the output to a console window. I usually used this. Next thing you would like to use is SegTracker.

[paraj]

Thanks guys, I've resolved it, and can now spend the day enjoying my glorious 28 MFlops!

The disk version worked, but showed a yellow fullscreen warning saying "68040 installation alert - do not use 68040.library on a 68060" during startup (also shows up when I tried to fiddle with CPU settings, but only in 4.4). This was the clue I needed.

I already had mmulib installed (with 68060.library), but it doesn't (or atleast didn't for me) update 68040.library so I was using version 37.30 of that. Copying in the new version from mmulib allowed me to run both v.4.4 and 4.5b12 from WB.

Noticed a few things apart from that: v4.5b12 doesn't show the CPU type (it's correctly displayed as 68060 in v4.4) and both versions shows the FPU type as 68060+68882, don't know if that's intended? (I found a screenshot online of v4.0 that just shows FPU as 68060 for someone else). In the "Drives" screen my PFS partitions show up as either "Not a DOS disk"/"FFS International" respectively.


In v4.5b12 if I press "SPEED" and then try to scroll the list libraries list it hangs (I have to hold down ctrl+amiga+amiga to force a reboot), this doesn't happen in 4.4. (NM, I guess this is a known bug)

[TuKo]

Here is attached a small 4GB HDF booting directly to SysInfo, no geometry or partition fancy, plain FFS.

On WinUAE, it will crash on Drive Benchmark (test without JIC, 020, 040, etc.).

[TuKo]

Further testing revealed that there is a div by zero in the drive benchmark located here :



With a quick and dirty patch, that crash can be avoided :



Attached all files for that issue and how to quickly fix it.


Best would be of course to fix sources.