PFS3 Wrong Index Block ID

[ozymandias]

I got my whdload files from a torrent. I don't know where to find on the Web. They could be corrupted but they works fine on SFS partition.

[Solo761]

I got it from here

http://eab.abime.net/showthread.php?t=68704

I had problems with previous version, 1.3, not the current one 2.1. Although, when I had these problems I had them with CF card, not HDF file. Maybe that's also a factor.

[ozymandias]

Any news? I checked all files but seems equal.

[Toni Wilen]

I am waiting for reply from PFS3 author, I am not sure how this part of filesystem is supposed to work.

I noticed that error goes away if I disable updating of parent directory's modification time (when file is deleted inside directory) but it is only a workaround, not a fix.

[Toni Wilen]

Bug found (Thanks to M.Pelt), fixed version attached. I couldn't duplicate the error anymore.

As usual: this is test version, don't use if you don't have backups.

[ozymandias]

Great! It works! When you release it, add the fixed PFS3Doctor you made in this thread.

[Toni Wilen]

Updated version attached (more optimal fix) and also includes custom version. This version will be officially released soonish.

Change log since previous official version:

- ROM resident priority increased. GVP tekscsi driver (that normally can't load filesystems from RDB due to a driver bug) can now use PFS3 if PFS3 is in ROM.
- Fixed very old PFS3 bug that caused Wrong Index Block ID -error in some rare situations.

[Retrofan]

Thanks a lot

[crabfists]

Sorry to necro this thread but I am having related issues.

I am trying to get a SanDisk 8Gb CF card working with PFS3aio in WinUAE but I am running into problems at the point I copy files to the card. The first attempt at partitioning I got "Wrong index block id" after copying a few gigs of files. I reformatted and reinstalled PFS3aio and now the latest error is "Wrong ablock id" (sic). This was after copying a few hundreds of megabytes of whdload files to the card.

I have one small boot partition (400mb) and a 7.6gig partition.
I am using version 2.3 of PFS3aio with a stock Wb3.1 installed on the boot partition. I am using the v3.0 kickstart ROM.

I am using the HDToolbox from OS3.9 with the following details:

Identifier 0x50465303
Mask 0x7ffffffe
Maxtransfer 0x1fe00
Blocksize 512
Buffers 100

I guess the next step is to try the same setup using a Winuae hdf to rule out it being a hardware issue.

I tried to go past the 4gig barrier a few years ago with similar results. I don't want to give up again!

Can anybody see anything obvious I am doing wrong?

Thanks in advance.

Edit: it is possible there is something wrong with my card or card reader. I tried making a hdf image from the card and WinUAE came back with "Hard drive image file creation failed. Error 5:0(<unknown>)".

[Jope]

Try setting the dostype to PDS\03 on the filesystem and partition and see if that helps. PFS3AIO's autodetect is usually quite robust though.

[thomas]

Add more buffers. PFS manual says it needs at least 150 buffers and up to 500. So 200 should be a good value for the boot partition and 500 for the large partition.

[Toni Wilen]

Lack of buffers "only" causes "Out of buffers" requester, it should not cause filesystem stability problems.

[Toni Wilen]

btw, PFS3 switches to different internal block layout if partition size is 5G or larger.

Do you get same error if you quick format and do the file copy again?
Do you get same error if partition size is only something like 4.8G? (Larger than 4G but smaller than 5G)

Just wondering if there is some rare edge case in PFS3 "large" mode.

[crabfists]

Thanks for the replies.

@Jope When you say "Try setting the dostype to PDS\03" do you mean use the identifier 0x50445303 instead of 0x50465303 but still use the same pfs3_aio-handler file? I don't really understand the difference between PFS3 and PDS3. Does changing the identifier change how the file system behaves?

@Thomas Thanks I will try increasing the buffers on the next attempt.

@Toni I have done a quick format and I am trying the copy again. Will let you know how it goes. Will also try the 4.8 Gig partition test.

By the way, if I use WinUAE to create a HDF how come the identifiers are 0x50445300 for PDS3 and 0x50465300 for PFS3?

[Jope]

Originally Posted by crabfists

When you say "Try setting the dostype to PDS\03" do you mean use the identifier 0x50445303 instead of 0x50465303 but still use the same pfs3_aio-handler file?

Yes.

I don't really understand the difference between PFS3 and PDS3. Does changing the identifier change how the file system behaves?

Yes, it changes the order in which PFS3 aio tries the methods to access the drive. PDS\3 bumps directscsi to the top of the list.

[crabfists]

Thanks for the advice. Really helpful.

PDS\3 bumps directscsi to the top of the list.

Do I need any other libs installed to get directscsi then? Or is it included in PFS3aio? I am just using vanilla Wb3.1 (for that retro feeling). Feel free to tell me to RTFM.

[fgh]

Crabfists: No other libs required, It's included.

Toni: When I had this fault quick formatting did not help.

[crabfists]

@Toni, I did a quick format and copied the files again and I got no errors. I don't really understand why it worked this time.

I don't fully trust it until I can work out why it works sometimes and not others. I thought I would eliminate hardware issues by starting again with a 16gb HDF in RDB mode. I used the PDS/3 identifier this time : 0x50445303. I created these partitions:

- 500mb boot partition with Wb3.1 on it
- 15gb partition for other files

I used PDS3 for both partitions. Not sure if I should have used PFS3 for the 500mb one?

I copied 6gb of whdload games to the 15gb partition without any errors. So far so good.

But now I just got a "PFS-III ALERT: wrong ablock id" when opening certain folders on the Wb3.1 partition (Tools, Locale, L etc). I rebooted and it went away but I really don't know what I am doing wrong.

[supaduper]

Run PFS Doctor that should correct it

[crabfists]

@supaduper I tried PFSDoctor and it found no errors. I have not seen the wrong ablock id errors since rebooting. I don't understand why it seems to have fixed itself.