This would be linux-utils, but be warned the fdisk code is competely FUBAR. I gave up on it when I realized that some mysterious segfaults ware caused by unaligned pointer arithmetic and/or conversion of some data to little endianness, and changing those would likely have broken fdisk for other architectures.
A testcase which always "worked" for me on IP22: create a DVH with parted, and then try to read/change it with fdisk. Apparently parted poisons the DVH for fdisk. :-)
Could you provide me a link/contact to linux-util to submit such a patch? The change is that while the length of the partition properly takes into account the blocksize, it still sets the block size to 512 no matter what the actual block size is. The change is to set this value to sector_size instead of hard coding 512.
Thanks for your assistance in this matter.
eyt* -- Eric Y. Theriault http://www.eyt.ca
