fdisk(8) vs gpart(8), and gnop

[Warren Block]

On Mon, 2 Jun 2014, Steven Hartland wrote:

>
> ----- Original Message ----- From: "Nathan Whitehorn" 
> <nwhitehorn at freebsd.org>
>> 
>> I think we basically don't have any lying disks anymore. The ATA code does 
>> a very good job of this -- most tell the truth, but in an odd way that gets 
>> reported up the stack. ada(4) has a quirks table for the ones that do not. 
>> If this is the only concern, then we should just stop telling people to 
>> worry about this.
>> 
>> My bigger concern is this pool upgrade one -- what if someone puts in a 4K 
>> disk in the future?
>
> Thats very much not the case I'm afraid, I try to add quirks for disk as
> they are reported but there's always going to be quite a few which are
> wrong until manufacturers stop making their FW lie :(
>

Both gpart and diskinfo show the correct values in the stripesize 
fields.  At least, I've yet to see it be wrong.  Maybe that is where ZFS 
should be getting the blocksize anyway.

(Of course, stripesize might only be correct due to the quirks you 
mention, in which case... never mind.)