Three comments for final draft for Fuji Software Defect Management

keiji_katata at post.pioneer.co.jp keiji_katata at post.pioneer.co.jp
Wed Jan 8 23:57:47 PST 2003


* From the T10 Reflector (t10 at t10.org), posted by:
* keiji_katata at post.pioneer.co.jp
*

Hello all,

I got three comments from several FW engineers.

1. Fatal write error should be stored in DBI.

In case of fatal read error, Type 3 is stored in DBI. So it is natural that
fatal write error stores Type 4 in DBI. This is not described clearly in
section "7.5 Four types of defect level".


2. Verify command should be terminated by fatal error.

If fatal error happened, verify command is terminated. I think it is
natural. This is not described in section "7.6 Enhanced Defect Reporting".


3. DBIF (DBI full) flag should be cleared by Write command of DRT-DM mode.

In small DBI cache model of DRT-DM mode, device may copy DBI (defective
block information) from RDBI to WDBI. Then some of DBIs should be pushed
out from WDBI memory. At the end of a verify command that caused DBI full,
DBIF=1 is in the newest (latest) WDBI memory. For example, drive has 10
WDBI memory. The drive has terminated a verify command when drive found 10
defective blocks. Then the latest (10th) WDBI has DBIF=1. By the write
command, if a DBI is copied from RDBI to WDBI, the oldest (1st) WDBI data
is pushed out. So WDBI of DBIF=1 is moved from 10th to 9th. The new DBI
|from RDBI is in 10th.
The 10th DBI breaks the consistency of DBIF=1. The first 9 DBIs have
relationship wtih DBIF=1. But a DBI does not have any. So I think that by
the write command, DBIF should be cleared.

Please send any comment to me.

Best regards,

Keiji Katata
PIONEER CORP.


*
* For T10 Reflector information, send a message with
* 'info t10' (no quotes) in the message body to majordomo at t10.org




More information about the T10 mailing list