file corruption

[crazyc]

OK this is driving me mad. I was experimenting with booting from the TF1260 and ended up trying with what I assume was too long a cable to the cf adapter coming out the back of my checkmate case. It was booting fine but but seemed to have errors when writing.
Somehow this issue is now persisting when I have dropped back to a short cable etc.
Example: I edit user-startup and save it. I then go to reopen it and it wont - if I try via shell it just says file contains binary.
I rolled back to an early image, tested it in winuae and successfully amended user startup etc, saved it, rebooted, it was all fine.
Took it and put it in my miggy and got the error.
So I am thinking I have a hardware issue. BUT I can use the CF card that came with the TF1260 fine, with no such error.
OK I thought, my CF card is toast, so went back and put the image over an older card that is known working but not used with the TF1260.
Still got the issue. Works under WINUAE but not in the amiga.
Its driving me nuts.
Any suggestions?

[nogginthenog]

I'm not familiar with the TF1260 but are you sure it's not a MaxTransfer issue?

[crazyc]

As I am using PFS AIO I thought the max transfer issues were resolved? but in any case this is the same base that I have been using all along and that has worked fine since AoS3.2was released and I did a clean install.
I forked my setup a while ago for my other miggy which has RTG etc and doesnt have this issue. Gonna go back to a pre fork build and see if that also has the issue.
I would swear it was hardware if it were not for the one working card with an old OS setup.

[thomas]

Originally Posted by crazyc

Any suggestions?

Instead of always reverting to another precondition I would rather try to analyse the problem.

Start with a hex dump of the corrupted file. What is the binary it complains about?

Also do some other tests. For example boot without startup-sequence and edit the file from the initial CLI prompt. Does the same happen?

The MaxTransfer issue needs a buffer larger than 128K. How big is the file you edit?

I am not familiar with TF accelerators. Does it allow firmware updates? Is there an update available?

What if you connect the CF card to the internal IDE bus rather than the TF?

To find the difference between the CF which came with the TF and the current one, please run this program once with each card: http://thomas-rapp.homepage.t-online.../hddreport.lha
Attach both reports here.

[crazyc]

Thanks Thomas, I am in the process of gathering the info, but for various reasons my amiga time is a little sporadic. I did just check something following one of your points.
Using only the dodgy disk:
Booting using 3.2 failsafe startup and copying and editing the file via Shell works fine.
However with the same startup doing it via dopus causes errors.
Booting normally causes errors in both methods.
Anyway, am going to copy over the hd test program and get results back later.
Cheers
C.

[thomas]

Originally Posted by crazyc

However with the same startup doing it via dopus causes errors.

Does the TF sit properly and tight? Is the RAM on the TF known good?

This sounds very much like a RAM issue. The more RAM is used, the more likely the error occurs.

Which OS is installed on the CF card that came with the TF?

[crazyc]

I wonder if your last point may be key, and I want to test that in a minute. I forgot one thing that changed around the right time to be suspicious.
To use the tf1260 ide you need to have the ehide driver loaded either via loadmodule off the motherboard ide or by being burned into a rom.
Although almost everything does seem to work fine I did switch to custom roms around that point. These were 3.2 roms purely with the ehide driver loaded as well.
The working card is not 3.2 based and so did need tweaking to load off my 3.2 roms anyway (workbench and icon library) If the rom is to blame that would explain why the bad cards seem to work OK under winuae which is off a clean 3.2 rom image.
Will swap out the custom roms in a minute after running a hddreport on the bad card.

[crazyc]

No its not that. Replacing the roms with original 3.2 roms made no difference.

[crazyc]

Well, were it not for the one unaffected scalos based card I would assume I have messed up the hardware when changing cf adapters/cables etc, but then the error should be consistent.
But if its software how come it doesn't affect my other RTG a1200 which is off the same base.
Does the lack of a RTC make a difference, the TF1260 doesnt have one whereas my previous accelerators did.

Anyway, here are the HDD reports.

[thomas]

Not sure if this is the reason, but the "corruption" report shows an old version of PFS3AIO and a partition larger than 6GB. PFS3 had some issues with such large partitions which might have been fixed in later versions.

And the DH0 partition has a high MaxTransfer value. I don't know if ehide.device has an issue with MaxTransfer values higher than the ATA limit.

Besides that I don't see any issues in the reports.

[crazyc]

Thanks Thomas.

I have just put the card that suffers corruption into my towered amiga and it does not suffer corruption there, (Blizzard SCSI kit) so maybe the ehide is sensitive to the Max Transfer Value.
Am I right in thinking I can simply correct this in HD toolbox, to 0x1FE00 without reformating stuff, but changing the file system to the latest PFS3AIO from Aminet will require a reformat?

[thomas]

Originally Posted by crazyc

Am I right in thinking I can simply correct this in HD toolbox, to 0x1FE00 without reformating stuff, but changing the file system to the latest PFS3AIO from Aminet will require a reformat?

You can update both without reformat. There is an "update file system" button in the Add/Update section of HDToolbox. This just replaces the old code by the new version.

[crazyc]

Thanks Thomas for all your help with this.
Preliminary testing (had to leave to pick up the wife) suggests this has resolved the issue!