Comment 2 for bug 1210327

Another possibility would be to read the image off the device following the write and generate a md5 for that - possibly faster due to more efficiency reading the entire image vs small chunks?