Interesting case. InnoStar like's to use Cross Block. Think of Cross Block like this. Drives that use Cross Block spread wear over multiple non-sequential areas with a pair, eg: Join by Page joins two sequential areas (block: 0x80 on NAND 1, block 0x80 on NAND 2). Cross Block joins pair blocks over non-sequential areas (block 0x80 on NAND 1, block 0x200 on NAND2). Block sorts must be done on each segment and then paired. One possible way to identify Cross Block is if you see block conflicts for every block within a dump.
The methodology for putting the NAND # in the proper order is Block Position % Total eg: Block 0 is in all blocks with remainder 0, Block 1 is in all blocks with remainder 1, Block 2 is in all blocks with remainder 2 and so on till Total and it loops again.
In this case the data wasn't organized by NAND chip but by halves of the dump... eg: (Dump Size / 4). Using the Cross Block step alone didn't work because the data wasn't organized using the equation above.
Initially I used Total Commander to split the dump into 4 equal parts and then use Cross Block however a certain genius I know suggested that using two Block Pairs could achieve the same result.
Parts = 4 is the number of total pair steps * 2
Mix = is the number of total pair steps * 2 / size of each interleave (ie: Join by 0x4000)
All information above might be wrong, this is just how I think of it.
Dump Size = 0x1000000000
1/2 Dump Size = 0x800000000
1/4 Dump Size = 0x400000000
Model
Layout
IS917 1094_32 Xor 3c58_128
Mix
Input
Order 0
Block Pair
Block 0x400000000
Page 0x200000
Block Pair
Block 0x800000000
Page 0x400000
Cross Block
Field Block Number
Block 0x200000
Parts 4
Step 1
Assembler
Block Number
Block Size 0x200000
Bank Space 0x40000
Image Cut
Banks 00 / 0x000 / 0x200f
Correction
Mix 4 / 0x4000
Scan Offset 0
Scan Back false
Scan Max 0
