CF / SD and large drives FAQ

[dJOS]

Originally Posted by Arnie

I have a dual booting Flash drive, first partition Fat32 and second partition FFS (RDB).

I used Aros to create it. You could try Aros to change the partition to Fat.

No need to install it, just use a live CD.

Thanks Arnie, I'll try that

[thomas]

Originally Posted by fgh

HDToolbox from wb2.x/3.0/3.1 only supports CHS addressing and therefore max 7.87GB (8GB cards are ok, they are sligthly smaller than this)
HDToolbox from os3.5/3.9 supports LBA and larger drives

This is not true. HDToolbox and all similar programs only access scsi.device and similar drivers. Thus they don't know anything about CHS, LBA or other hardware-dependent protocols.

All versions of HDToolbox support all size of harddrives up to 2 TB. Versions below 3.5 have a problem printing numbers bigger than 4GB. But this is only a GUI problem. If you ignore the sizes printed in the GUI and just enter the desired values into the LowCyl and HighCyl fields, everything goes well.

Similarly HDToolbox below 3.5 has a problem if the number of cylinders on the drive exeeds a certain value. In this case the partition diagram goes all crazy. Again you can just ignore it and enter values for LowCyl and HighCyl manually.

CHS vs. LBA is a problem of scsi.device and thus affects all kinds of partitioning programs, even HDInstTools.

[fgh]

Thanks Thomas, I've rewritten the part about HDToolbox now. (It's under Maximum drive capacity -> Further information now).

I'm glad you had a look at the FAQ, many parts of it come from EAB posts by you and Toni.
Please let me know if you spot more errors or inaccuracies!

Just to make sure, i couldn't find these information for pfs3-aio,
all settings for pfs3_5.3 and pfs3_aio are the same:
Mask = 0xffffffc
MaxTransfer = 0x1fe00
DosType = 0x00fffffe
and only file required is pfs3_aio-handler in L:

Sorry for noobish question, but there is no information on how to install it in readme file,
searching web gives no better results, only info i've found is regarding pfs3_5.3.

[fgh]

Maxtransfer doesn´t matter with AiO. 0x1fe00 is fine, but you may also leave it at the default. (it makes no difference)
Mask deals with DMA access and is therefore not relevant with scsi.device (A600/1200/4000 IDE), but only for DMA capable controllers.
Dostype should be 0x50445303 (PDS3) or 0x50465303 (PFS3) for all versions of PFS.
Old PFS versions use either Direct SCSI or TD64 depending on the dostype.
PFS AiO supports TD64, NSD and DirectSCSI with both dostypes, but will move directSCSI from last to first priority with PDS3 dostype.

Having the file in L: is only required for setup, as the file system is stored in the RDB of the HDD.

Thanks fgh, all gone smoothly since i have put in proper dostype.

Now i'm fighting with my fast ata, dunno why sometimes ata3prefs
shows pio4, pio5 and another time pio4NF for the very same card in
the very same slot of my CF-IDE adapter (got two slot version).
Furthermore there is no option to change Split - No Split, if
preferences are accessed just after booting with LMB.

Apoligise for messing in this great thread, if someone can point me
to right thread/post i'l be grateful.

[JonSick]

Just to ask; do these cards work OK for those CF/IDE adaptors? KS 3.1, at some point an accelerator too.

http://www.amazon.co.uk/gp/product/B...&s=electronics

[spajdr]

Originally Posted by JonSick

Just to ask; do these cards work OK for those CF/IDE adaptors? KS 3.1, at some point an accelerator too.

http://www.amazon.co.uk/gp/product/B...&s=electronics

These will work just fine, they are best for CF and also for PCMCIA CF adapters.

[JonSick]

Sold! Cheers for the prompt response.

[Toni Wilen]

I think this weird issue should be added to faq.

You partition drive normally, after reboot, some or all FFS partitions don't appear on WB and C:Info lists all missing partitions with "no disk in drive" state, hdtoolbox still shows them correctly.

It is caused by DOS/FFS bug, if first 4 bytes of first block of the partition contains bytes $FFFFFFFF, FFS gets confused and thinks there is no disk.. (I guess it confuses dos disk type $FFFFFFFF = -1 = ID_NO_DISK_PRESENT with real dos type)

This bug can trigger even if drive is new, for example some DOM's or solid state memory cards may contain random looking non-zero data from factory. (Test patterns?) Fix is to wipe the drive completely.

OS3.9BB2 works (reports correct "not a dos disk" state), I haven't tested OS3.5-OS3.9BB1. 3.1 and probably older versions have the bug.

[thomas]

Does the "no disk in drive" also affect the Format program? I.e. if DH2 says "no disk", can you still "format drive dh2: name xy" or does Format reject the command?

[Toni Wilen]

Originally Posted by thomas

Does the "no disk in drive" also affect the Format program? I.e. if DH2 says "no disk", can you still "format drive dh2: name xy" or does Format reject the command?

"Format Failure: no disk in drive"

[Mnemesi]

I'm probably going to be jumping back into the fray on this in the near future. If you recall, I gave a try at installing a 4.32 gig drive in my unit and started to have serious issues with write failures. Although I appreciated the help that was given, that issue resolved itself catastrophically by the Amiga ending up in a drawer for the last half year or so when original unit's 340 meg drive's pins came loose from all the swapping. I recently rearranged my room so I have workspace again, so I'm gonna have another go at it. This time, I'm wondering if anyone would like to offer a more hands-on approach to helping me, possibly via Skype so I can give live updates. I live in EST, so it might be best to organize a time for this since nearly everyone still alive in the Amiga scene is about 6 hours in the future from me. I have WinUAE, some 2.5 drives, and - here's the fun part - a SX-32 Pro based CD32 unit, so here's your chance to see a fairly rare one. Any takers?

[Wepl]

Is it possible that PFS AIO has problems with many files in a directory?
If I try a dir command in disk_11 of BeneathASteelSky-CD32 which contains 3632 files I get a PFS requester 'Read attempt outside partition'.
It's a 8G CF-Card in A1200.

[fgh]

Small update: (Finally) added the submission above from Toni Wilen to the FAQ. (Thanks!)

Wepl: I don't know. If you didn't get a reply from Toni, try contacting him directly.

[ShK]

OS3.1-, OS3.9 HDToolBox and HDInstTools Drive Definition gives different values ​​for my 8GB SanDisk:

HDToolBox 40.4:
Cylinders: 15400
Heads: 16
Blocks per Track: 63
Block per Cylinder: 1008

HDToolBox 44.23:
Cylinders: 12128
Heads: 5
Blocks per Track: 256
Block per Cylinder: 1280

HDInstTools 6.8
Cylinders: 41068
Heads: 1
Blocks per Track: 378

http://s30.postimg.org/3p2by9k4h/HDTool_Box31.png
http://s30.postimg.org/wfz5obpy9/HDTool_Box39.png
http://s30.postimg.org/6fqvacc9t/HDInst_Tools.png

All of these can make partitions to the entire ~7.4G region.

Which values ​​should I use?

Cylinders * Heads * Sectors = total number of blocks a.k.a. highest logical block address (LBA)
Cylinders * Heads * Sectors * 512 = size of the disk in bytes
Blocks per Track = Sectors
Tracks per Cylinder = Heads
Blocks per Cylinder = Sectors * Heads
Cylinders * Blocks per Cylinder = total number of blocks on the drive

[mark_k]

You might as well use the one which gives the largest useable amount of sectors (HDToolBox 44.23 here).

Or you can find out the actual card capacity and make up your own fake geometry values to cover the entire drive. You can use a program like SCSIQuery to print the number of blocks.

[ShK]

Ok, so I can make the fake geometry just like Cylinders * Blocks per Cylinder?

SCSIQuery results:

Device capacity :

Capacity in blocks : 15523704.
Block length       : 512.
Length in the middle of the medium in blocks: 0.

[mark_k]

Hmmm, that's interesting...

SCSIQuery reports the number of blocks as 15,523,704. But the HDToolBox 44.23 fake geometry gives 12128×5×256 = 15,523,840 blocks.

Try creating a partition at the end of the drive using HDToolBox 44.23, then format it and copy files there, and compare with the source files. Are there any errors?

HDInstTool's fake geometry does match the 15,523,704 blocks figure. You might as well use that as-is, but you could adjust the geometry.

15,523,704 = 2×2×2×3×3×3×7×10267, so for example an alternative fake geometry could have
Cylinders = 10267×3 = 30801
Heads = 4
BlocksPerTrack = 126

[ShK]

Before before I do that copying test, I installed the memory card in a real Amiga 600.

It gives Capacity in blocks : 15523200 (the memory card has been cleaned with diskpart)

Warning: Required memory type for device not found.
Assuming MEMF_24BITDMA.
SCSIQuery 1.33 (22.11.99) © THOR
Generic sense data : 

Good.

Device inquiry data : 

Device status           : connected.
Device class            : optical memory device.
Removable medium device.
Device type modifier    : 00.
ISO version             : 00.
ECMA version            : 00.
ANSI version            : SCSI-2.
Response data format    : SCSI-2.
Additional data length 1F.
Flags:
	Relative addressing not supported.
	32 bit wide SCSI not supported.
	16 bit wide SCSI supported.
	Synchron data transfer not supported.
	Linked commands not supported.
	Tagged queues not supported.
	Device reacts with hard reset.
Vendor                  : SD      
Product                 : to CF Adapter   
Version                 : 45.0
Date                    : 
Comment                 : 

Extended inquiry data : 

	00 00 00 00 00 00 00 00  ........
	00 00 00 00 00 00 00 00  ........
	00 00 00 00 00 00 00 00  ........
	00 00 00 00 00 00 00     ....... 



Control page information:

Mode Data Length : 5
Medium Type      : 0
Device specific  : 0
Block des. len   : 0



Device capacity : 

Capacity in blocks : 15523200.
Block length       : 512.
Length in the middle of the medium in blocks: 0.


Primary defect list:
Read Defect List Header returned error : 
Check Condition         : Details as below.
LBA not valid.
Error code of last command.
Segment ID              : 00.
SenseFlags              :
Sense Key               : Illegal request.
Additional sense length : 10.
Cmd information bytes   : 00 00 00 00.
Drive specific flags    :
Extended sense code ASC : 0x20.
Extended sense code ASCQ: 0x00.
ASC,ASCQ encoded error is invalid command operation code.
Field replaceable unit code FRUC 0x00.
Additional flags        :
	Error data output.
	Byte error.
	Bit position 0x00.
Byte position           : 0x00 0x00.


Grown defect list:
Read Defect List Header returned error : 
Check Condition         : Details as below.
LBA not valid.
Error code of last command.
Segment ID              : 00.
SenseFlags              :
Sense Key               : Illegal request.
Additional sense length : 10.
Cmd information bytes   : 00 00 00 00.
Drive specific flags    :
Extended sense code ASC : 0x20.
Extended sense code ASCQ: 0x00.
ASC,ASCQ encoded error is invalid command operation code.
Field replaceable unit code FRUC 0x00.
Additional flags        :
	Error data output.
	Byte error.
	Bit position 0x00.
Byte position           : 0x00 0x00.