K6602743
Rev.2
02.12.’02
- 95 -
6.3.2.12 Note For Write Cache and Auto Reallocation
(1) Loss of data in write cache
Write cache is a performance enhancement whereby the device reports as completion the write commands
to the host as soon as the device has received all of the data into its cache buffer memory. This means that
there is a possibility that power off even after write command completion might cause the loss of the data
that the device has not written onto the media.
Therefore it is recommended that some other command except write command shall be executed before
powering the device off.
(2) Error Report and Auto Write Reallocation
In case of write cache mode, the device reports the write command completion after receiving all data from
host immediately. After this command completion, the device automatically reallocates the error sector when
the device cannot recover the error in write operation. By this auto reallocation, the unrecoverable error
sector is reassigned to a spare sector, and the data of the error sector are written on the spare sector. If the
device cannot recover the data by this auto write reallocation, the device reports the error as follows:
a) The error occurred when the command execution is on going, the error is reported for
the current command.
b) The error occurred when the command execution is not on going, the error is reported for
by the next command.
In case of non-write cache mode, the device reports the write command completion after the completion of
write operation on the media. If an error occurred during write operation on the media, the device
automatically reallocates the error sector when the device cannot recover the error in write operation and
reports the command completion.
The Auto Write Reallocation cannot be disabled.
(3) Read Auto Reallocation
Non recovered read errors:
When a read operation fails after error recovery is fully carried out, an error is reported to the host. This error
location is registered internally as a candidate for the read reallocation. When the error location is specified
as a termite of subsequent write operation, the error location is reallocated automatically.
Recovered read errors:
When a read error operation for a sector failed once and recovered at the certain retry step, the recovered
sector of the data is reallocated automatically.
