Re: [Gptfdisk-general] sgdisk and Hybrid without prepend of EFI partition support

[Antoine L. <Ant...@us...>]

Mario Fetka wrote:
> small correction
> fidsk sees it like that when i use the pipe version
>
> --------------------------------------------------------------------------
>    Device Boot      Start         End      Blocks   Id  System
> /dev/sda1   *        2048      616447      307200   83  Linux
> /dev/sda4               1        2047        1023+  ee  GPT
>
> Partition table entries are not in disk order
>
> ------------------------------------------------------
>
> but sgdisk -h 1
>
> creates
>
> -------------------------------------------------------------
> Command (m for help): p
>    Device Boot      Start         End      Blocks   Id  System
> /dev/sda1               1        2047        1023+  ee  GPT
> /dev/sda2            2048      616447      307200   83  Linux
>
>
>
> and with this the u-boot of the device finds the GPT as first part and this
> is not working
> my request was the posiblity to change this also for the sgdisk so that i
> can say the ee GPT disk is not part 1 in mbr
>
This can be awfully complex, and I wonder how can semi-complex cases
could be sorted (and of course, if things are not crystal clear at
design time, the probabilities for bugs + misuses are increasing
steadily...)

For an example, taking your case slightly edited, consider

GenLSPro ~ # sgdisk -p /dev/sda
Disk /dev/sda: 195371568 sectors, 93.2 GiB
Logical sector size: 512 bytes
Disk identifier (GUID): xxxx-yyyyy...
Partition table holds up to 128 entries
First usable sector is 34, last usable sector is 195371534
Partitions will be aligned on 2048-sector boundaries
Total free space is 2541 sectors (1240.7 KiB)

Number  Start (sector)    End (sector)  Size       Code  Name
   1            2048          616447   300.0 MiB   8300  Linux filesystem
   2          616448       168388607   80.0 GiB    8300  Linux filesystem
   3       168388608       170485759   1024.0 MiB  8200  Linux swap
   4       170485760       195371007   11.9 GiB    8300  Linux filesystem

(I rounded down sd4's size.) There are nine distinct areas across the disk:
Number  Start (sector)    End (sector)  Size       Code  Name
 a                 0               1                     MBR
 b                 1              33                     GPTable
 c                34            2047                     free space
   1            2048          616447   300.0 MiB   8300  Linux filesystem
   2          616448       168388607   80.0 GiB    8300  Linux filesystem
   3       168388608       170485759   1024.0 MiB  8200  Linux swap
   4       170485760       195371007   11.9 GiB    8300  Linux filesystem
 d         195371008       195371534                     free space
 e         195371535       195371567                     alternate GPTable

Now, on what criteria could it be decided that using `sgdisk -H 1` makes area {a+b+c} to be covered under entry #4 typed EE, while the area {2+3+4+d+e} would not be covered? Why not {a+b} instead? Why to not have two EE entries (like some hybrid-making tools are already doing?)

Same questions with `sgdisk -H 2`: should the #4 EE entry covers {a+b+c}, or {a+b+c+1}? Why?

And what to do when the command is `sgdisk -H 4:2` ?
I guess the best answer is "Overtricky, giving up"...


Antoine