Issues with PFS3AIO

[PoulpSquad]

Top of the day to all of you,

I'm having a bit of trouble with an HDF partitioned and formatted with PFS3AIO 3.1. It is attached to uaehf.device (using WinUAE 4.9.1 x64).

The HDF is 4095 MB big, with partition DH0: of 55 MB at the beginning, and DH1: taking up the rest of the disk. Mask=0x7FFFFFFF, Max transfer=0x7FFFFFFF, buffers=512 for both partitions.

Both partitions are PFS3/00 ID 0x50465300, fast-formatted in Workbench (I'm using Amiga Forever 9).

My setting has been working great so far, but a couple of hours ago I got a pretty bad surprise when I decided it would be a wise idea to check the partitions with PFSDoctor (and not diskvalid as I wrote origninally) from PFS3_53.lha.

PFSDoctor reports no errors on DH0:, but finds thousands of errors on DH1: and chokes after a couple of seconds, then a guru meditation happens.

Trying psfdoctor found in pfsdoctor_experimental.zip reports no errors, but I'm not sure if it would report any in an understandable way by human beings.

Did I somehow screw the HDF without noticing all these days and thousands of files got somehow corrupted?

PS: I have attached the logs.

[PeterK]

I'm not using any filesystems under WinUAE, but I prefer to access the Windows directories directly instead. So I can't tell you what might go wrong in your PFS3AIO setup, but I noticed that the latest WinUAE versions 4.9.x are not always working as reliable as WinUAE 4.4.0. It seems that there are still some new issues in 4.9.x, especially with the Jit.

I would suggest to try out older WinUAE versions like 4.4.0, before assuming a bug in PFS3AIO or these tools.
https://www.winuae.net/download-old-versions/

[PoulpSquad]

Good call, I'll take a look at it, thank you for the advice!

What I'd like to know is how pfsdoctor from pfsdoctor_experimental.zip reports errors...

I don't think it fixes them silently with the check parameter?

*Edit*

Same happens with WinUAE 4.4.0.

I'm more and more convinced the problem isn't with PFS3AIO but with the old version of PFSDoctor included in the 5.3 archive.

PFSDoctor from Toni's experimental archive remains stubbornly non-descript, generating the same outputs with check or repair parameters. I guess that's a good thing...

[Toni Wilen]

It probably is stack overflow and/or long file names/very long paths that cause the crash (very long file names/paths was very rare back in the day)

[PoulpSquad]

Indeed, with 5 million stacks old PFSDoctor manages to complete a volume check, but still manages to generate a 2.6 MB log file with almost 4000 errors...

New PFSDoctor reports nothing out of the ordinary.

I'll simply ditch all these tools from older versions of PFS and only use PFSDoctor experimental for the time being.

Thank your for your help!