I mean, presumably this is all device-side on the stick.
It mounts the two partitions separately, and presents them separately. It doesn’t need to pass through the actual SD card to the host system at all.
If you dd the decoy partition you only ever see the decoy partition. Because the decoy partition is presented as a complete device, not just a partition, to the host system.
Why would this be incompatible with linux? If this is done the way it seems it is done, it’s not a filesystem trick.
It’s two partitions on one SD card being presented as two separate devices by the microcontroller that has actual direct access.
The real vulnerability would be that you could just crack it open to find the SD card.
I mean, presumably this is all device-side on the stick.
It mounts the two partitions separately, and presents them separately. It doesn’t need to pass through the actual SD card to the host system at all.
If you dd the decoy partition you only ever see the decoy partition. Because the decoy partition is presented as a complete device, not just a partition, to the host system.
Why would this be incompatible with linux? If this is done the way it seems it is done, it’s not a filesystem trick.
It’s two partitions on one SD card being presented as two separate devices by the microcontroller that has actual direct access.
The real vulnerability would be that you could just crack it open to find the SD card.