geom_part_mbr vs user

[Andriy Gapon]

on 03/11/2010 08:44 Andrey V. Elsukov said the following:
> On 02.11.2010 20:02, Andriy Gapon wrote:
>> But I still believe that geom_part_mbr "thinking" that it is smarter than me is
>> incorrect.  Yeah, keep using CHS params for some defaults, etc, but do obey what I
>> explicitly specify if doesn't violate fundamental constraints like media size or
>> slice overlaps.
> 
> I think it was done for maximum compatibility. Some utilities assume that partitions
> will be aligned on cylinder boundaries.


Possibly.
Those utilities likely need some modernization too.

Strictly speaking modern (ATA) drives don't even report any CHS numbers.
So we mostly just make up some odd [non-even :-)] numbers and then base pretty
important decisions on those made up numbers.
-- 
Andriy Gapon