[lkml]   [2000]   [Sep]   [30]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: reading 1 hardsector size, not one block size

Greetings All,

WARNING: More infor than you every really wanted to know about the world
of storage and the concerns + issues, but read the whole thing or you will
never get the point.

On Sat, 30 Sep 2000, Theodore Y. Ts'o wrote:

> Date: Fri, 29 Sep 2000 22:16:53 -0700 (PDT)
> From: Andre Hedrick <>
> Basically you can de-stroke a drive with what you let the OS/FS report.
> Once this is done there is no way any FS can get to the stuff beyond what
> it knows about.
> I'm not sure what you mean by "de-stroke" here.
> 100 sectors total drive size.
> OS/FS knows about 90 sectors and can only do up to 90.
> Bypass Layer with direct access will see the stuff from 91-100.
> OK, fine, with your special hacks it's possible to use some extra
> sectors. But is it actually safe to use sectors 91-100? These wouldn't
> happen to be the blocks used for doing bad block sparing, would they?


The 100 sectors are what you could ever see with out vender specific
access, we will not go there ... today ...

Same drive but 110 sectors, with 10 reserved for slip/vender specials,
thus you can only ever get access to 100 sectors. This is no matter what
you do, or me.

The OS can "de-stroke", oops need to explain.

de-stroke: to reduce the total size or capacity in its appearence or
presentation to the user.

Yes all OEM's do this because they need everything to look the same.
This is why you may get a replacement drive that is in reality a 30GB
drive but because you had a 15GB drive originally, that is the capacity of
the new drive. It has been "de-stroked" by half.

Now lets install the Linux on this 100 sector drive and tell it that it is
only 90 sectors in size. Linux can only access that which it knows about
via the FS. Thus accessing the drive by the FS will restrict you access
to all regions of the drive.


If we match head and tails (we will need to soon).

D(0) and D(100) need to match.
This is a backup sector, in case D(0) gets wacked, but a Broken Pane.
The BIOS/OS could recover with a redirected look.

Now O(101)->O(110) is were DIAG records by the OEM are keep and there is
no way to get then with out access the OEM subset of the MAN vender unique
access commands.

Now back to D(91)->D(99) are buffer sectors or OEM public data access for
what ever they want.


Andre Hedrick
The Linux ATA/IDE guy

If you are confuse, good, this is where they want us to be, but I
understand the issues, and will continue to help decode the lingo.

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
Please read the FAQ at

 \ /
  Last update: 2005-03-22 12:38    [W:0.099 / U:0.256 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site