sp193
Developer
Uh, so there were a few booboos that I won't add into the changelog above because they were all introduced by me. But FYI...
LaunchELF build 2017/06/18:
USB disks will only be mounted if they either contain a supported partition (i.e. FAT12/FAT16/FAT32). Mounting the whole disk will no longer be done, unless the disk is deemed to not have a valid MBR.
Previously, it would attempt to mount the disk, even if it had a valid MBR. This also led to cases of USB disks being improperly accessed, if they were formatted with some unsupported filesystem like exFAT.
I didn't find any standard that specified how the operating system can determine whether the disk contains a MBR or a VBR, which both have the same magic value in the exact same position (why???). But I did some experiments with Windows 10, which led me to conclude that Windows will at least check that for every valid (ID is not zero) partition entry:
There might have been some check against the size of the partition, but I did not manage to identify a pattern.
Link: https://www.sendspace.com/file/xyyypr
LaunchELF build 2017/06/18:
- Fixed USB support on disks with no partition table (thumb drives).
- Fixed formatting of PFS partitions.
USB disks will only be mounted if they either contain a supported partition (i.e. FAT12/FAT16/FAT32). Mounting the whole disk will no longer be done, unless the disk is deemed to not have a valid MBR.
Previously, it would attempt to mount the disk, even if it had a valid MBR. This also led to cases of USB disks being improperly accessed, if they were formatted with some unsupported filesystem like exFAT.
I didn't find any standard that specified how the operating system can determine whether the disk contains a MBR or a VBR, which both have the same magic value in the exact same position (why???). But I did some experiments with Windows 10, which led me to conclude that Windows will at least check that for every valid (ID is not zero) partition entry:
- The starting LBA must not be zero.
- The starting LBA must exist within the disk.
There might have been some check against the size of the partition, but I did not manage to identify a pattern.
Link: https://www.sendspace.com/file/xyyypr
