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

[Daniel Martin]

On 9 April 2014 21:41, Thomas Schmitt <address@hidden> wrote:
> Hi,
>
> i can reproduce the gdisk complaints with a script that i derived from
> yours and gdisk 0.8.1. The script just packs up some data files in
> the same way as yours and announces them as boot images. (It would
> not boot, of course. But gdisk does not know. :))
>
> Further i can avoid them by using the mkisofs emulation.
> You are blazing the trail. It's about time that somebody tests.
> I am heavily thankful.

Thank you for the fast and detailed replies.

> The decisive difference this time seems to be the fact that
> the mkisofs emulation by default disables the emulation of
> multi-session on disk files.
>
> I was able to get my mock-up ISO acceptable for gdisk by adding
> the command
>   -compliance no_emul_toc
> to disable the multi-session emulation:
>
>     ${XORRISO} \
>         -compliance no_emul_toc \
>         -padding 0 \
>         ...

Confirmed, -compliance no_emul_toc makes gdisk happy.

> The multi-session emulation reserves 64 kB at the image start
> for a superblock that may be overwritten by future sessions.
> The first session begins after these 64 kB by its own System Area
> and superblock.
> The reason for this is that when more sessions can be appended,
> the first superblock can be updated without damaging the original
> info for session 1. This allows to retrieve the chain of all
> session superblocks from 1 to end.

As I'm not going to add any sessions I don't have any problems when
disabling it.

> But obviously it sabotages GPT.
> The mkisofs emulation disables it to save space.
>
> With the -compliance setting, i get this output from gdisk:
> (My fdisk warns of GPT and shows MBR partition info.)
> ----------------------------------------------------------------------
> GPT fdisk (gdisk) version 0.8.1
>
> Partition table scan:
>   MBR: protective
>   BSD: not present
>   APM: not present
>   GPT: present
>
> Found valid GPT with protective MBR; using GPT.
> Disk /home/thomas/test.iso: 436 sectors, 218.0 KiB
> Logical sector size: 512 bytes
> Disk identifier (GUID): CA0A0836-D6BF-4624-A6FB-7B40D37543B5
> Partition table holds up to 248 entries
> First usable sector is 64, last usable sector is 372
> Partitions will be aligned on 64-sector boundaries
> Total free space is 1 sectors (512 bytes)
>
> Number  Start (sector)    End (sector)  Size       Code  Name
>    1              64             127   32.0 KiB    0700  Gap0
>    2             128             255   64.0 KiB    EF00  EFI boot partition
>    3             256             371   58.0 KiB    0700  Gap1
> ----------------------------------------------------------------------
>
> With default setting i get complaints:
> ----------------------------------------------------------------------
> ...
> Warning! Main and backup partition tables differ! Use the 'c' and 'e' options
> on the recovery & transformation menu to examine the two tables.
> ...
> Warning! One or more CRCs don't match. You should repair the disk!
> ...
>   GPT: damaged
>
> ****************************************************************************
> Caution: Found protective or hybrid MBR and corrupt GPT. Using GPT, but disk
> verification and recovery are STRONGLY recommended.
> ****************************************************************************
> ...
> ----------------------------------------------------------------------
>
> I will examine why the GPT is so messed up with the native default.
>
> If gdisk likes your ISO, then please let fdisk tell its opinion
> and forward it to me.

With -padding 0 and -compliance no_emul_toc neither fdisk nor gdisk
print any errors. Using -comliance without -padding gives the "GPT
PMBR size mismatch" warning with fdisk, but it's fine for gdisk.
I've attached the initial 34KiB of both (both with -compliance, one
with and one without -padding).


Thanks,
    Daniel Martin

img_pad_zero.bin
Description: Binary data

img_wo_pad.bin
Description: Binary data