Strange file corruption with external usb drive

[Alexco]

Hello,

I am using WinUAE 1.5.3 to transfer some data to a harddrive which is normally installed in my A4000. It is attached via a Delock IDE-USB bridge to a Dell Laptop with WinXP SP3.
The HDD contains five partitions, 3 FFS and 2 SFS. All FFS are below 4GB.
I can boot in WinUAE from the drive without problems (using uaehf.device), but if I transfer data from a "virtual" drive (folder on my Laptop) to the HDD, some files get corrupted.
I already asked for help on a1k.org and Thomas told me to reduce MaxTransfer values. I did, but still files get corrupted, small files, like vnc.lib (127k, corrupt from byte 98305 to 101376) or eqmain.iff (70k, corrupt from byte 3585 to 16384) from AmigaOS3.9, as well as large .lha files.

I am using OS3.9BB2, Filemaster 3.1 for copy/compare, the drive is an ExcelStore Jupiter 80GB ATA (configured to 16 Heads and Master) and MaxTransfer tested down to 0x7ffe, but still no success.

Any further hints?

[Toni Wilen]

What kind of corruption? Does it change after each write?

Run in windowed mode and with -log parameter. Do you see any error messages in the log window when writing?

[Alexco]

Can´t say what kind of corruption. I am trying to set up some data files to see if there is some kind of pattern.
And I can´t see any errors while copy in progress, just a warning at the beginning:

Volume D:\ FS=NTFS maxlen=255 flags=000700FF
hfd open: 'HD_ExcelSto r Technology J80 P21O'
opening device '\\?\ide#diskst9120823as_____________________________3.adc___#5&1
6482f9&0&0.0.0#{53f56307-b6bf-11d0-94f2-00a0c91efb8b}'
device id string: 'ST9120823AS 3.ADC'
BPS=512 Cyls=14593 TPC=255 SPT=63 MediaType=12
device size 120034123776 (0x1bf2976000) bytes
4 MBR partitions found
0: num: 1 type: 07 offset: 32256 size: 32218389504, type not 0x76
1: num: 2 type: 07 offset: 32218421760 size: 87813089280, type not 0x76
non-empty MBR partition table detected, doing RDB check anyway
hd ignored, not empty and no RDB detected
opening device '\\?\usbstor#disk&ven_excelsto&prod_r_technology_j80&rev_p21o#6&2
4938293&0#{53f56307-b6bf-11d0-94f2-00a0c91efb8b}'
device id string: 'ExcelSto r Technology J80 P21O'
BPS=512 Cyls=10011 TPC=255 SPT=63 MediaType=12
device size 82348277760 (0x132c570000) bytes
no MBR partition table detected, checking for RDB
hd accepted (rdb detected at block 0)
opening device '\\.\C:'
device id string: 'ST9120823AS 3.ADC'
duplicate device, ignored
opening device '\\.\D:'
device id string: 'ST9120823AS 3.ADC'
duplicate device, ignored
opening device '\\.\E:'
device id string: 'TSSTcorp DVD+-RW TS-L632H D200'
Drive scan result: 3 Amiga formatted drives detected
WARNING: 'HD_ExcelSto r Technology J80 P21O' FSCTL_ALLOW_EXTENDED_DASD_IO return
ed 1
hd accepted (rdb detected at block 0)
HDF 'HD_ExcelSto r Technology J80 P21O' opened, size=80418240K mode=0 empty=1
hfd open: 'D:\Privat\Emulation\Amiga\Drives\Sys39.hdf'
HDF 'D:\Privat\Emulation\Amiga\Drives\Sys39.hdf' opened, size=204800K mode=0 emp
ty=1
Volume D:\ FS=NTFS maxlen=255 flags=000700FF

Mounting uaehf.device 0 (0):
FS: mounted virtual unit dh1 (D:\Privat\Emulation\Amiga\Drives\Work)
HD_ExcelSto r Technology J80 P21O:
RDB: RDSK detected at 0, FSHD=6, C=574416 S=280 H=1
Mounting uaehf.device 1 (0) (size=82348277760):
RDB: 'EDH0' dostype=444F5303. PartBlock=1
BlockSize: 512, Surfaces: 1, SectorsPerBlock 1
SectorsPerTrack: 280, Reserved: 2, LowCyl 2, HighCyl 4097
Buffers: 80, BufMemType: 00000000, MaxTransfer: 00007ffe, BootPri: 0
HD_ExcelSto r Technology J80 P21O:
Mounting uaehf.device 1 (1) (size=82348277760):
RDB: 'EDH1' dostype=444F5303. PartBlock=2
BlockSize: 512, Surfaces: 1, SectorsPerBlock 1
SectorsPerTrack: 280, Reserved: 2, LowCyl 4098, HighCyl 8193
Buffers: 80, BufMemType: 00000000, MaxTransfer: 00007ffe, BootPri: -1
HD_ExcelSto r Technology J80 P21O:
Mounting uaehf.device 1 (2) (size=82348277760):
RDB: 'EDH2' dostype=444F5303. PartBlock=3
BlockSize: 512, Surfaces: 1, SectorsPerBlock 1
SectorsPerTrack: 280, Reserved: 2, LowCyl 8194, HighCyl 12289
Buffers: 80, BufMemType: 00000000, MaxTransfer: 00007ffe, BootPri: -5
HD_ExcelSto r Technology J80 P21O:
Mounting uaehf.device 1 (3) (size=82348277760):
RDB: 'EDH3' dostype=53465300. PartBlock=4
BlockSize: 512, Surfaces: 1, SectorsPerBlock 1
SectorsPerTrack: 280, Reserved: 2, LowCyl 12290, HighCyl 249857
Buffers: 80, BufMemType: 00000000, MaxTransfer: 00007ffe, BootPri: 0
RDB: RDB filesystem 53465300 version 1.279
RDB: Filesystem loaded, 98400 bytes
HD_ExcelSto r Technology J80 P21O:
Mounting uaehf.device 1 (4) (size=82348277760):
RDB: 'EDH4' dostype=53465300. PartBlock=5
BlockSize: 512, Surfaces: 1, SectorsPerBlock 1
SectorsPerTrack: 280, Reserved: 2, LowCyl 249858, HighCyl 574415
Buffers: 80, BufMemType: 00000000, MaxTransfer: 00007ffe, BootPri: 0
RDB: RDB filesystem 53465300 version 1.279
RDB: 53465300 in FileSystem.resouce version 1.279
RDB: FS in FileSystem.resource is newer or same, ignoring RDB filesystem
HD_ExcelSto r Technology J80 P21O:
Mounting uaehf.device 2 (0):
FS: mounted HDF unit dh0 (0000-0c800000, D:\Privat\Emulation\Amiga\Drives\Sys39
hdf)
Mounting uaehf.device 3 (0):
FS: mounted virtual unit dh2 (D:\Privat\Emulation\Amiga\Drives\WDaten)
Thread 'hardfile' started (1124)
Thread 'hardfile' started (1144)
FS: work (flags=00000002) starting..
FS: wdaten (flags=00000002) starting..

[Toni Wilen]

To confirm it isn't Amiga-side issue, do same "corruption test" using hdf image. (use "Create hard disk image file" option in harddrive settings panel)

[Alexco]

strange...
I programmed a small asm test in AsmOne, which creates three times a simple pattern from 0 to $ffff and writes longs in form (pattern number) (counter) (e.g. $00020001, $00020002, ... $0002ffff...$0000ffff) to disk.
Creating the file does not trigger the effect regardless of writing to HDD or to an image.
If I copy the file from the image to the HDD, the file gets corrupted.
Sometimes an entire block is repeated, e.g. $00013dfe, $00013dff, $00013d80,$00013d81 instead of $00013e00, 00013e01.

I have attached the files as zip.