tech-kern archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: ata(4) and NCQ

On Apr 26, 2011, at 9:22 AM, Eduardo Horvath <> wrote:

> On Tue, 26 Apr 2011, Manuel Bouyer wrote:
>> Actually I suspect the idea to have ATA drives show up as SCSI comes from
>> windows. But this was done so on windows because of a limitation of the
>> windows kernel: if you want to have a drive appears as ATA the controller
>> has to follow the pciide 1 specification. if you have a smarter controller,
>> the only way to have full control at the controller driver level is to
>> register it as a SCSI driver.
> I can speak to this.  (I probably shouldn't, but that's never stopped me 
> before.)  Solaris has this problem where making even a tiny a code change 
> involves huge amounts of overhead.  What typically happened is a bug is 
> found on a SCSI disk, which is fixed.  Then the same problem occurs on a 
> CDROM, ATA disk, etc.  Having one driver handle all devices solves this 
> problem so when something is fixed for SCSI disks it applies to 
> fibre-channel disks, CDROMs, etc.  Of course, this causes other problems 
> like complicating the sd driver much more than it would have to be to only 
> support SCSI disks instead of also CDROMs, removable devices, 
> fibre-channel disks, SAS disks, SATA disks, etc.

Which is one of the plus and minus sides of the geom layer in FreeBSD I 
suppose. It's a logical means of separating out the management logic common to 
all storage mediums from the medium specific logic.

I've been watching this discussion and one hing i was curious about was -- is 
it widely known that there are a handful of ATA <-> SCSI specs that have been 
released by IEEE over the past decade. I figured someone was aware of this on 
the list, but was kind of puzzled by the replies about FreeBSD, Linux and 
windows behavior when in fact the behavior is allowed and standard.


Home | Main Index | Thread Index | Old Index