File stability issues [SOLVED]

A forum for Classic Amiga specific support issues.

File stability issues [SOLVED]

Postby nyteschayde » Tue Jan 24, 2012 5:58 am

Okay, so, first of all I want to thank Hyperion and all the folks who worked on OS 4.1 for Classic machines. It really is a beauty to behold and I am happy to have it. So much so that I may buy a more modern Amiga to use with this year (AmigaOne 500 or some other similarly capable machine). That being said I am currently running it with

Blizzard PPC 603e/233 68060/50 w/192MB RAM
2MB Chip
Elbox Mediator 1200SX
ATI Radeon 9250
Catweasel MkIV
6-Port SATA II 150 Sil3114 SATA Controller
Spider USB Card (doesn't currently work; please make it work!!)
Subway USB
Indivision AGA
Lyra (or Keyrah; can't remember)
Cocolino PS/2 Mouse Adapter (scroll wheel doesn't work...)

Coming soon
RTL8029AS PCI Network Card
FastATA 1200 MkIV

Blizzard PPC Settings
RAM Speed 70ns
MapRom Off
All Cache settings off as well

Mediator Jumpers
Wait: On
Memory: 8MB
Config: Off
Wake Up: Off
(On means shorted by the jumper, Off means not shorted by the jumper)

Drive Layout
2.5 -> 3.5 IDE Converter (bought locally; not Amiga specific) plugged into motherboard slot
Primary Master -> IDE 30GB PATA Hard Drive 7200RPM
Primary Slave -> 52x IDE PATA CD-RW Drive

120GB 7200RPM SATA II Hard Drive

Boot Config
Currently I have a partition less than 2GB on the PATA drive, DH0:, (at the beginning of the disk) wherein I have a OS4.1 Update 2 Install performed. I then rebooted from the AmigaOS 4.1 CD and installed to another partition less than 2GB on the SATA drive (at the beginning of the disk; just to be sure) called SA0:. On DH0: I uncommented out BootDevice in the KickLayout-A1200 file and created a text file with the word SA0 in a file called Sys:Kickstart/BootDevice. This successfully causes the OS to load from the SATA drive after the kernel has loaded.

I then updated to Update 3 and Update 4. Both times, after each update but before rebooting I copied the Kickstart directory over to DH0:Kickstart and modified the KickLayout-A1200 file to use the BootDevice file I specified. This seems to have applied the correct kickstart updates while still booting from the SATA drive after the kernel has been made resident.

Both drives, DH0: and SA0:, use SFS/00. The MaxTransfer of DH0: (on the internal IDE port) is set to a very low value of 0x0000FE00. I've also tried several other values. The MaxTransfer on the SATA drive is some very large number that was supplied by default. The block size is set to 512 on DH0: and 1024 on SA0:

There seems to be some overall corruption of files greater than a certain (as yet to be determined size). The varying in size is what made me originally think it might be the MaxTransfer but adjusting it doesn't seem to make a difference and I get the same problem with the SATA drive as I do with the internal IDE which doesn't make sense to me.

I am completely unable to install the dev tools. I can subsequently run md5sum on the SDK_53.20.lha file (without even extracting it) and get different values on each run. The total size is somewhere greater than 80MB which leaves very little room (or not enough) if I try to extract it to RAM: Whenever I try to run the file I always have checksum errors (and not always at the same place which lends to the read/write issues).

Smaller files consistently give me a correct md5sum value call after call. Somewhere between 10MB and the much larger 80MB things go wrong. I really really really want to figure out this issue, but I am completely out of ideas. Does anybody else have any suggestions on how I can narrow this down and solve the problem?

Update 1
Just to be sure it wasn't the PowerPC overheating that caused the issue, in addition to the little fan that blows through to the back side (I have a Phase5 model with the black fan less heat sink), I attached a FAN6X1TX3 fan to the black heat sink on the bottom of the BlizzardPPC. I used velcro on the top of the heat sink to the non-moving part of the fan and it barely but nicely fits between the heat sink and the back of the tower.

Update 2
Possibly solved the problem simply by adding the fan. If anybody else is having strange random errors, try adding additional cooling to your PPC. This fix cost me around 9 USD. :) I managed to get the SDK installed for the first time. md5sum run on the SDK file repeatedly (2x since doing it takes a while) reported the correct MD5 value. I will update this later with a truly solved once I've run the system a bit.

Update 3
Considered solved for now. Adding the extra cooling to the PowerPC chip has really cut down on instability issues. Anybody seeing small issues like this should really consider adding a fan. The one I linked to above was thin enough that I was able to add it on the heat sink with some simple velcro and have it work wonders.
Posts: 2
Joined: Mon Jan 23, 2012 11:08 am

Re: File stability issues [SOLVED]

Postby JurassicC » Tue Jan 24, 2012 2:05 pm

I have an upgraded Blizzard at 330Mhz and OS4.1 only became stable when I added an additional heatsink and fan to the phase5 passive cooler.
Others with uprated cyberstorms and blizzards have experienced this too.

As for the spider card.
It would require DMA transfers to be added to the classic kernel.
Even then the spider has a modified flashrom with Elbox's Vendor ID in it, so it may not work with sirion (the OS4 usb stack).
I'd be interested to read about your experiences with the catweasel MK IV in the classic.

User avatar
Beta Tester
Beta Tester
Posts: 454
Joined: Wed Apr 20, 2011 1:21 pm
Location: UK

Return to Platform: Classic

Who is online

Users browsing this forum: Google [Bot] and 6 guests