From:    Kevin Chadwick via Dng <dng@???>

Date:    Wed, 17 Sep 2025 10:56:26 +0100 (GMT+01:00)

> Maybe I have lost the thread of UEFI not being harmful ...

The original question is here.
https://lists.dyne.org/lurker/message/20250912.165516.4855da4a.en.html

Yes, I didn't ask whether UEFI is harmful. Asked about the "free
space preceding the EFI System Part". That assumes GPT & the UEFI
System Part exist. The free space is created when partitioning with
GPT. A partitioning scheme is established before the boot process is
established.

I'm learning about 64 bit systems. Corrections welcome.

A few additional observations from recent efforts.

* When setting up GPT, gparted creates a free space part of at least
one MiB. If you set 0 MiB, gparted immediately revises to 1 MiB. As
Fraser commented, the part can be larger than 1 MiB.

* If you choose the legacy boot process, the free space part has a
role as in a traditional MBR.

* If you choose the UEFI boot process, the free space becomes a
"protective MBR". Discussed in the Wikipedia GUID article cited
below.

* I've read remarks that UEFI in some early Apple implementations do
not conform to the UEFI specification. Here I have a Mac Mini A1283,
(c) 2008. Model identifier: Macmini3.1. Partitoned with GPT, 2 MiB
initial free space as suggested by Fraser. With the Devuan Daedalus
installer, installed GRUB2 . No problem with GRUB2/UEFI booting. If
the Apple UEFI implementation had a bug, someone found a workaround.

I found these articles helpful.

https://en.wikipedia.org/wiki/Master_boot_record
https://en.wikipedia.org/wiki/GUID_Partition_Table
https://en.wikipedia.org/wiki/GNU_GRUB

Regards,                        ... P.

-- 
mobile:  +1 778 951 5147 <= still working
VoIP:    +1 778 508 0020 <= new & working
VoIP:    +1 604 xxx 0140 <= defunct
projects: en.wikibooks.org/wiki/User:PeterEasthope