SAT definition of ARRE bit in mode page

James C Hatfield james.c.hatfield at seagate.com
Mon Feb 8 11:05:20 PST 2010


Formatted message: <a href="http://www.t10.org/cgi-bin/ac.pl?t=r&f=r1002082_f.htm">HTML-formatted message</a>

Bill,
(As I understand ATA HDDs): If a sector is successfully recovered, then
there is no need to actually reallocate it. Reallocations only occur for
unsuccessful recoveries.
Also, given that DATA SET MANAGEMENT(Trim) may apply to HDDs, if a sector is
Trimmed, then the device does not have to wait for a new Write from the host
in order to perform the reallocation, as the actual data in the reallocated
location is irrelevant (subject to DRAT rules, of course).
Thank You !!!
-----------------------------------------------------------------
Jim Hatfield
Seagate Technology LLC
  e-mail:  James.C.Hatfield at seagate.com
  s-mail:  389 Disc Drive;  Longmont, CO 80503 USA
  voice:  720-684-2120
  fax....: 720-684-2766
On Mon, Feb 8, 2010 at 10:45 AM, <Bill.Martin at emulex.com> wrote:
>  In SAT, it was determined during letter ballot resolution to make ARRE
> set to zero and AWRE set to one in the Read-write error recovery mode page.
> The reasoning was as follows (from 06-121r5):
>
>
>
> Discussed: What is the reasoning behind the specified values for ARRE and
> AWRE? LB comments from HPQ and LSI as well as 05-241r2 suggest AWRE should
> be "one", and ARRE should be "zero". My understanding is that an ATA device
> will auto-reallocate on an unrecoverable READ error, but that the
> reallocation is deferred until the next time that same logical sector is
> written. An ATA device does not check data consistency on a write, and
there
> is no write-and-verify equivalent, so I don't see that an ATA device will
> auto-reallocate due to an error on a write command. But..., the data that
> ends up in the alternate sector is the data from the write, not the data
> from the read, so AWRE is one and ARRE is zero.
>
>
>
> This discussion is based around an unrecoverable read error; however from
> SBC-3r21, ARRE is based on a recovered read error.  The following is from
> 6.4.5 Read-Write Error Recovery mode page:
>
>
>
> An ARRE bit set to one specifies that the device server shall enable
> automatic reassignment of defective logical blocks during read operations.
> All error recovery actions required by the error recovery bits (i.e., the
> EER bit, the PER bit, the DTE bit, and the DCR bit) shall be processed. The
> automatic reassignment shall then be performed only if the device server
> successfully recovers the data. The recovered data shall be placed in the
> reassigned logical block. The device server shall report any failures that
> occur during the reassignment operation. Error reporting as specified by
the
> error recovery bits (i.e., the EER bit, the PER bit, the DTE bit, and the
> DCR bit) shall be performed only after completion of the  reassignment
> operation. See the REASSIGN BLOCKS command (see 5.19) for error procedures.
>
>
>
> This clearly states that the reassignment is ONLY performed if the device
> server successfully recovers the data.
>
>
>
> I am looking for clarification on how ATA devices operate.  If an ATA
> device recovers read data will it reassign the block during the read
> operation to a different location on the media?  If it will make this
> reassignment, then I will bring in a proposal for SAT-3 to change the bit
> settings in this mode page.
>
>
>
> Bill Martin
>
> Emulex
> Office of Technology
> Industry Standards
> 916 772-3658
> 916 765-6875 (Cell)
> bill.martin at emulex.com
>



More information about the T10 mailing list