On Mar 1, 2006, at 8:55 AM, Jens Axboe wrote:
On Wed, Mar 01 2006, Mark Lord wrote:
NCQ vs. TCQ: NCQ has a much more efficient low-level protocol,
making the host-side (controller, operating-system) quite a bit
simpler than with NCQ.
Or in laymens terms - TCQ sucks and NCQ doesn't :-)
NCQ has many more advantages than TCQ, apart from both a more
efficient
low level protocol and ease of implementation. TCQ basically just
allows
the drive to do some reordering, it still serializes everything and
requires too many interrupts.
The problem with TCQ is that the host can't disconnect on writes
after sending the data to the drive but before receiving the status.
The host can only disconnect between sending the command and moving
the data. Consequently TCQ is useless for writes, which is where you
really need it. It works OK for reads. TCQ was really invented as a
way to allow CD-ROM drives to play nice on the same ATA bus as disks.
The reason you need write queuing is for data integrity reasons, not
for performance. ATA disks effectively get command-queuing on writes
even without TCQ and NCQ - they simply park the data in a volatile
RAM cache, tell the host that the data is saved on persistent
storage, and then asynchronously write the queued data to the
physical media. The drive reorders those writes and will gather
sequential writes.
However, note that all filesystems that make even a pretense of
trying to maintain filesystem integrity after a power failure (note
that the Windows NT implementation of FAT32 does not attempt to
maintain filesystem integrity after a power failure) depend on
knowing when data makes it to persistent storage, so they can order
their writes correctly. ATA disk write caching breaks this guarantee.
To restore filesystem integrity on a careful-write filesystem like
most unix filesystems, you have to disable write-caching in the
drive. This causes such a drastic loss of performance (you basically
get only one sequential write per disk revolution), that you must
then implement command-queuing to allow the drive to gather
sequential writes to make the system usable.
As an alternative, if you have a journalling filesystem, you can
leave the disk cache enabled but selectively write-through your
metadata using force-unit-access (FUA).
Regards,
-Steve
--
Steve Byan <smb@xxxxxxxxxxx>
Software Architect
Egenera, Inc.
165 Forest Street
Marlboro, MA 01752
(508) 858-3125
-
: send the line "unsubscribe linux-ide" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
