Subject: Re: two disks, two controlers, same bad block
To: None <tech-kern@NetBSD.org>
From: George Georgalis <george@galis.org>
List: tech-kern
Date: 03/27/2006 13:10:04
On Mon, Mar 27, 2006 at 07:47:22PM +0200, Martin Husemann wrote:
>On Mon, Mar 27, 2006 at 12:39:30PM -0500, Charles M. Hannum wrote:
>> The correct solution for this is to just use LBA48 for that sector
>> number and put a comment in the code stating the reason.
>
>Yes, we should do that.

how much of an impact will that have on resources; "for every
block read or write check if it requires LBA48," sounds kinda
heavy.

just reading on badsect, neat program. while it is not 100% the
right solution, because it's a firmware issue not a 'bad block' --
it seems like a good and simple fix because, effectively, the block
is bad.

that said. what a PITA to make a BAD directory each time a seagate
disk is deployed.  instead of processing drive specific quirks at
runtime, why not make a one time quirk for newfs (or disklabel or
fdisk...) to reserve that seagate block from future use?

also I notice, there is no problem with adjacent blocks. but the
solutions discussed refer to sectors, are they synonymous?

// George


-- 
George Georgalis, systems architect, administrator <IXOYE><
http://galis.org/ cell:646-331-2027 mailto:george@galis.org