cvs commit: src/sys/dev/ata ata-all.c ata-all.h ata-chipset.cata-dma.c ata-pci.c ata-pci.h

[Poul-Henning Kamp]

In message <20030408202308.GA49094 at tara.freenix.org>, Ollivier Robert writes:
>According to Poul-Henning Kamp:
>>    Tagged queueing:
>> 	Ability to reduce transaction latency by sending more than one
>> 	request to the drive at a time.  The request may be handled out
>> 	of order by the drive.
>> 	(A good thing!)
>
>Except for softupdates which rely on its own ordering of writes...

This actually has nothing to do with it.  If the drive only replies
to the transaction when the data is actually written, the biodone()
will reach softupdates and the right thing happen.

Soft updates does not really care if I/O transactions are reordered,
but it cares a lot about being able to trust that they are actually
written to stable storage, not just some cache, when drivers say so.

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.