Re: [Bug-xorriso] fdisk complains: GPT PMBR size mismatch

[Thomas Schmitt]

Hi,

the reason why gdisk complains about ISOs which were not made with
  -compliance no_emul_toc
is that the disk GUID and partition GUIDs differ between main GPT and
backup GPT.
(The backup GPT is written to the end of the GPT "disk" and should of
 course have the same GUIDs as the main GPT at the start of the "disk".)

This happens with multi-session emulation because new GUIDs are
computed when the superblock of session 1 gets written.
So the backup GPT contains the GUIDs of a (yet useless) GPT that is
64 KiB away from the location of the real main GPT.

The complaints of gdisk about bad CRC vanish, although the CRC
computation has not been changed. So i assume this was inner confusion
of gdisk.


> With -padding 0 and -compliance no_emul_toc neither fdisk nor gdisk
> print any errors.

That's good to know.

I seem to have fixed the GUID issue now. So the -compliance command
should not be necessary any more with the next release.

The padding issue still needs a good automatic solution.
Maybe fallback to padding 0 if GPT is desired. I will have
to consider the wider consequences of such a move.

Currently i am entangled in a big libisofs changeset which needs
lots of testing. The GUID fix will probably have to wait a few
more days before it arrives in the bzr repository.


Thanks again for testing. I promise to care for any problem report.


Have a nice day :)

Thomas