| Author |
Message |
Bolo
Joined: 05 Dec 2017
Posts: 727
Location: 506F6C616E64
|
 Posted: Tue May 12, 2020 19:10 Post subject: [26 GB] microSD Fake 256Gb - 0xad de 14 ab 42
|
|
Are you able to make adjust layout for even blocks for this case ?
- Flash type : microSD
- Flash name: fake 256Gb
- Totall size: 8GB
- ID: ad de 14 ab 42
- Banks: 1
- Channel: 1
- Sector: 1104
- XOR for even blocks (15 sectors !!): 8855F9, 64p
- Layout for odd blocks (16 sectors): Unknown 1104 v2 Xor 8855_256 q
NOTES:
1) Dump is partially fixed with ECC for odd blocks (16 sectors - screenshot), for 15 sectors needs to adjust layout - otherwise there are 0xFF gaps at end (screenshot). I splitted dump into odd and even blocks using Core V5 and put into container too. If it's necessary to make model I can add empty sector at the end for each page
2) There are totally over 700 marked position of software Bad Bytes in even blocks only - marked them all and included in job.txt
3) There are similar case here - http://www.flash-extractor.com/forum/viewtopic.php?t=11265 so I think that solution can be used there too
4) After analyze I found that this chip need to be readed without VCC rail connected (voltage only from Data buffers), 3V and with Slow timing (3-1-4) - then quality becomes perfect ! There are no special/custom addr/regs for RR in OTP registers.
5) Container contains:
FE.zip - full fixed dump readed by FE with Job.txt and all BB positions
FE_ODD.zip - extracted odd blocks
FE_EVEN.zip - extracted even blocks
FE_BEFORE_ECC.zip - full dump without ECC fix with Job.txt and all BB positions
Dumps uploading
P.S
On this card will be mainly videos so I'm not sure if you got correct checkpoints but maybe on Arvika case there are photos - they will use same solution.
|
|
|
Bolo
Joined: 05 Dec 2017
Posts: 727
Location: 506F6C616E64
|
|
Posted: Tue May 12, 2020 22:14 Post subject:
|
|
Upload done
P.S
Also Back card photo attached
|
|
|
Сергей
Joined: 26 Aug 2005
Posts: 20281
Flash-Extractor developer
|
|
Posted: Mon May 18, 2020 22:36 Post subject: Re: microSD Fake 256Gb - 0xad de 14 ab 42
|
|
| Bolo wrote: | | [*]XOR for even blocks (15 sectors !!): 8855F9, 64p |
There is Xor before ECC
ECC map is green, so XOR is OK
("Before ECC Fix" dump not have red blocks)
Use Block Cut 0x3C00/0x400
Block Size will change 0x400000 -> 0x3C0000
But all pages have 1 red sector
So, all files will be broken
|
|
|
Bolo
Joined: 05 Dec 2017
Posts: 727
Location: 506F6C616E64
|
|
Posted: Mon May 25, 2020 10:14 Post subject:
|
|
Are you able to check my idea ?? After looking to dump - especially to 0x50804400. You can see there:
Page0 + + + + + + + - + + + + + + + N
Page1 + + + + + + + - + + + + + + + N
Page2 + + + + + + + - + + + + + + + N
Page3 + + + + + + + - + + + + + + + N
Pagen + + + + + + + - + + + + + + + N
- is bad sector (maybe not used by controller or filled with garbage)
N is not written sector (thats why we got 15 sectors)
+ normal sector in this case empty
As I can see there are bad sectot on 8th position on every page as well there are no 16th sector.
Maybe solution are to cut those sectors out so in fact we will create layout with 14 sectors ((0x400)*7/0x400/(0x400)*7/0x400) for even blocks so correct Block Cut schould be 0x1c00/0x400, after this it loos almost perfect (picture attached)
I can reupload more fixed dump if you need
|
|
|
Сергей
Joined: 26 Aug 2005
Posts: 20281
Flash-Extractor developer
|
|
Posted: Tue May 26, 2020 13:50 Post subject:
|
|
Use Auto find mix to find pages with FAT tables
You will see that red sectors is filled with data (tables)
We can not cut them
|
|
|
Bolo
Joined: 05 Dec 2017
Posts: 727
Location: 506F6C616E64
|
|
Posted: Mon Jun 01, 2020 12:51 Post subject:
|
|
|
You can close this case - client cancelled task
|
|
|
|
