cvs commit: src/sbin/fdisk fdisk.c

[Poul-Henning Kamp]

In message <20030331173658.GA17388 at elvis.mu.org>, Alfred Perlstein writes:
>* Poul-Henning Kamp <phk at phk.freebsd.dk> [030331 01:34] wrote:
>> In message <20030331140740.R17363 at gamplex.bde.org>, Bruce Evans writes:
>> 
>> >This is more broken than I thought.  fdisk now defaults to a wrong geometry
>> >in all cases where the BIOS geometry differs from the firmware geometry.
>> >
>> >The default geometry was H=1/S=1 or whatever is returned by DIOCGDINFO
>> >until this was broken in -current.  This is still non-broken in RELENG_4.
>> 
>> I don't particularly care for what numbers we use, but I do not
>> want a lot of elaborate DWIM code since in this particular case the
>> user had better know exactly what she is doing anyway.
>
>The problem is that each time you muck with these systems it becomes
>harder to get it to DWIW, nevermind DWIM.  Still wish we had the Solaris
>format tool.

If I had a (quasi-)sensible[1] on-disk format, writing such a tool would
be trivial.  The reason it is not trivial is because both the MBR and
BSD (as used by us) have major "issues" which you have to work around and
guess at as best you can.

Poul-Henning

[1] SUNLABELS suck in general, but since sun has control over the
firmware, this doesn't affect the users much.

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.