Didn't try the first one with Linux, Windows definitely could only see 27GiB, no matter what way I tried to look at the drive, nor did I try anything clever with putting the iFlash in some other kind of adaptor, only in the iPod (not at all sure what the computer would see if one did that, given that the ipod, to the limited degree I understand it, does some weirdness that obfuscates what the drive actually is and presents something quasi-fictional to the computer?).
Yeah, you'd directly plug the hard drive into the PC. Given that the Apple firmware is obfuscating what is exported over USB, we can't be sure if the failure is due to Apple's firmware or a limitation of the iFlash adapters themselves.
I don't Recall what rockbox supposedly reported when installed on that "27GiB" drive, but it was before I considered the Rockbox storage code rework to be "stable".
It's worth noting that _if_ the iFlash adapters do the right thing (ie properly handle >32bit addresses), then Rockbox is expected to JustWork(tm). Given that the ipod6g/7g partitions and formats the drive using 4K logical sectors, this means we could theoretically support a 16TiB FAT32 volume on those devices, even with MBR partitioning.
2200GB seems to translate to 2048.9GiB, which would surely be almost exactly 2TiB, so presumably would be right on the margin for exceeding the limit?
Yeah, the point is >2048.0GiB requires >32-bit addressing (with 512-byte sectors) My guess is that the reason the capacity is being reported wrong is that something in the middle (ie apple's disk-obfuscating firmware or the iflash firmware... or both) uses a 32-bit value for the capacity instead of 64-bit (technically 48-bit is all that's needed) which overflows.
Rockbox's core storage and filesystem code (plus the ATA drivers) should all be 64-bit-address clean now. The native SD stuff still needs work to support the large capacities that SDUC enables, but that probably won't happen unless someone donates a >2TiB SDUC card or three to the cause once they become available.