A request for unnested UFS implementation in MBR
jude.obscure at yandex.com
Sun Jul 8 01:17:44 UTC 2018
On 07/08/18 06:05, RW via freebsd-questions wrote:
> On Sun, 8 Jul 2018 04:52:11 +0530
> Manish Jain wrote:
>> On 07/08/18 04:44, Polytropon wrote:
>>> They don't. With GPT, there is no need for BSD labels anymore.
>> All I am saying is exactly the same possibility for MBR.
>> We can create a UFS implementation, perhaps named ufs, that gets
>> recorded directly in MBR table. Right now the implementation is
>> If someone could just touch a few things, it improves things for
>> eternity when we do not have bother about the extra layer (BSD). Any
>> extra filesystems the user needs should be found in the EBR, not in
>> the BSD.
>> Why should a PC have multiple nesting schemas ? It only pains the
>> user in the future when the need for the extra nest was only in the
>> past (when there presumably was no EBR nest).
> I think it did exist, but BSD avoided the mistake made by Linux in
> adopting the EBR kludge.
> If you need multiple OSs instances on a drive, it's self-evidently
> better to label their partitions hierarchically rather then number
> them in a flat space.
I submit to the wisdom of choices BSD made, and continues to make. Which
is why I use FreeBSD - very proudly - as my only OS.
Noone can legitimately claim better claim better solutions than
envisaged by Berkeley.
But in my humble opinion, the right way to treat MBR is simply as GPT
(with the number of partitions permitted not 128, but 3 - with 128
non-bootable extras relegated to a system the MBR schema itself provides
Who gains today for the loss of the user who cannot create UFS
partitions in MBR ?
Thanks and a hope that someone will see a point in my humble request.
More information about the freebsd-questions