This message is from the T13 list server.


One factor is that the drive starts transferring data ASAP, even if it is
still reading follow up sectors from the disk.  If the drive takes a big hit
in time (head/cylinder skew, retry on an error, mapped out defect), then the
interface on READs will usually shut down (burst terminated) for a period of
time.  

One designer I helped out once designed the host side assuming 1 command = 1
burst, and it worked in a lot of cases.  But he started having mysterious
data corruption on certain portions of the disk - the areas near mapped out
defects.

Jim


-----Original Message-----
From: Jim Castleberry [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, November 28, 2001 8:53 AM
To: Evans, Mark
Cc: 'Keith Clausen'; [EMAIL PROTECTED]
Subject: Re: [t13] Command completion of an UDMA OUT burst ...


This message is from the T13 list server.


On Wed, Nov 28, 2001 at 07:48:18AM -0800, Evans, Mark wrote:
> In addition, I
> think that most HDDs won't initiate burst termination before all data for
a
> command is transferred.  So, if there is a burst termination, it's because
> the host wanted to do something else.

In the few dozen devices from various manufacturers that I've had a reason
to look at in the lab, I don't think I've ever seen one that did NOT
terminate the burst before all data was transferred (not counting commands
transferring 1 sector or less).  And I've never seen one do a UDMA Pause
either.  They all seem to terminate for flow control rather than Pause.
YMMV.

jcastle

Subscribe/Unsubscribe instructions can be found at www.t13.org.
Subscribe/Unsubscribe instructions can be found at www.t13.org.

Reply via email to