ST3320820AS datarecovery.

[rew]

Hi,

I have a ST3320820AS drive here (firmware 3.AAC) that I need to recover. I'm learning the seagate serial commands, and I'm getting somewhere....

(I try to learn on spare seagate drives, but when I'm confident I won't wreck things I might do them on the drive that needs to be recovered.)

I found that the bad drive can still be addressed pysically. It won't go into the /9 commandset, but in the /2 command set I can seek to a specific track, read a few blocks and then dump the buffer. I can then "see" the user's data. Good!

writing a program to send the commands, and decode the buffer dumps would work, but it would take around 10 years to read the whole drive..... Not ideal.

I have a few questions: Other drives allow me to set the baud rate to 115200:

Code:

> /T
T> B1152


But not this one. It will go as high as 19200, but nothing higher that I tried switches baud rate (I tried 38400 and 115200). Why won't this drive go higher?

Is there a way that I can tell the drive to do SATA but in "physical mode" such that I can quickly read all the physical blocks? I will then figure out the mapping to logical blocks later. As long as I have the raw physical data, I can continue from there...

Can I maybe initialize the physical mapping in such a way that it translates directly, and that I can read the drive over SATA?

Where do the physical to logical translations live? Can I try to read them from the serial port, and maybe have that help me recover the lost physical to logical mapping? Does anybody know the format that these are stored in?

The other drive that I tested had a big readbuffer. This one says it's got 10 (ehhh that's in hex... 16(dec)) read buffers, does anybody know why this one has so little allocated for read buffers? Maybe I can issue the command "read physical" over the serial port, and then read the buffer over sata? Does anybody know how to read the buffers, without issuing a read-disk request? Hmm. That won't work: Because the logical mapping doesn't work, the drive does not respond to sata commands.

After hitting "^Z" on the serial port, can I make the disk resume its normal program (sata/pata) without having to reset it? It seems as if ^C and ^R both reset it, and reset at least the PATA DMA mode....

[pepe]

First of all post terminal log!

[rew]

Code:

Reset
4096k x 16 DRAM
GALAXY - 1_Disk    S-6D   09-22-06_15:48

Buzz HM SFI
!
(P)SATA Reset
(H)SATA Reset
ESC(
T>
T>B1152
T>/2
2>S10,0
    Code - FE  Track 00010(0E799).2.000  Sns 000  Rty F7FF.50.80FF  Rtf 1800  LBA 12A8CBCD
2>S20,0
2>R0,10
2>
BT:0000
PC:0056
AP:0060
RL:0298
AL:029B
SL:02B3
MW:059C
AT:0640
TB:0C48:0020
SO:0C68:0039
FS:0CA2:0038
RD:0640:0010
WR:0650:0010
FM:0660:0018
AD:0660:001C
BA:0678:0003
ST:067B:0001
AS:067C:007F
DP:06FB:02A6
AC:06FB:02A6
SC:09A1:02A7
CT:3B60
uP:3FC0
MZ:3FC0

logbps :0200
codebps:0200B640,640
buffer 0640   comparing to 0640  RD:0640:10:00  WR:0650:10:00
Addr  0 1 2 3  4 5 6 7  8 9 A B  C D E F 10 1 2 3  4 5 6 7  8 9 A B  C D E F
0C8000 81451B00 82451B00 83451B00 84451B00 85451B00 86451B00 87451B00 88451B00
0C8020 89451B00 8A451B00 8B451B00 8C451B00 8D451B00 8E451B00 8F451B00 90451B00
0C8040 91451B00 92451B00 93451B00 94451B00 95451B00 96451B00 97451B00 98451B00
0C8060 99451B00 9A451B00 9B451B00 9C451B00 9D451B00 9E451B00 9F451B00 A0451B00
0C8080 A1451B00 A2451B00 A3451B00 A4451B00 A5451B00 A6451B00 A7451B00 A8451B00
0C80A0 A9451B00 AA451B00 AB451B00 AC451B00 AD451B00 AE451B00 AF451B00 B0451B00
0C80C0 B1451B00 B2451B00 B3451B00 B4451B00 B5451B00 B6451B00 B7451B00 B8451B00
0C80E0 B9451B00 BA451B00 BB451B00 BC451B00 BD451B00 BE451B00 BF451B00 C0451B00
0C8100 C1451B00 C2451B00 C3451B00 C4451B00 C5451B00 C6451B00 C7451B00 C8451B00
0C8120 C9451B00 CA451B00 CB451B00 CC451B00 FFFFFF0F CE451B00 CF451B00 D0451B00
0C8140 D1451B00 D2451B00 D3451B00 D4451B00 D5451B00 D6451B00 D7451B00 D8451B00


The dumped buffer looks like a FAT to me which souds entirely reasonable for where this disk is from.... Looks like I got myself the first 512 bytes of "userdata" from the drive.

Note that I got the prompt back at 9600 baud after the B1152 command.... For this session it was not useful to switch to 19200, which which would've worked.

Detecting the drive from the computer and the issuing a read request results in:

Code:

AT Er 01 Nwt Er 00 RdWr ffffffff.00.0070
ATA St d0 Er 04 Op 25 e,  14a61dd4,0100 0100 0000
Niwot:  00015bb4  ff  00015bb4.0.001 006f  3f00 006f 0000
AT Er 01 Nwt Er 00 RdWr ffffffff.00.0070
ATA St d0 Er 04 Op 25 e,  14a61dd4,0100 0100 0000
Niwot:  00015bb4  ff  00015bb4.0.001 006f  3f00 006f 0000
AT Er 01 Nwt Er 00 RdWr ffffffff.00.0070
ATA St d0 Er 04 Op 25 e,  14a61dd4,0100 0100 0000
Niwot:  00015bb4  ff  00015bb4.0.001 006f  3f00 006f 0000


Anybody willing to explain what I'm seeing here? "Er 04" is probably the ata error register, "op 25" probably the command: "read DMA ext"