The problem is that the SCSI layer likes to hand us scatter-gather
segments. Since they are all non-contiguous, I need to do each data
transfer to/from the device as multiple URBs.
Since I don't queue multiple bulk URBs, there is a time-gap between when
one ends and the next one begins. I refer to this wasted time as the turn
time. Note that this is an amount of real time elapsed, but not consumed
bandwidth.
To observe this, get an analyizer and look at a large read/write to the
device. Generally, just reading a file will cause this. You'll see lots
of bulk transfers which will appear to transfer all the data at once, but
with unusual pauses every 1-4KBytes. The pauses are the delay I'm
referring to.
Thus, this isn't a bulk-transfer problem as such. Bulk queuing will help
this problem, but until some issues in the protocol are worked out, I can't
safely implement it.
Matt Dharm
On Thu, Apr 19, 2001 at 03:17:36PM +0200, Martin Diehl wrote:
> On Tue, 10 Apr 2001, Matthew Dharm wrote:
>
> > I call this the "turn time" problem.. I've known about it for a while, and
> > it's much bigger than just this.
> >
> > On Tue, Apr 10, 2001 at 10:43:11PM +0200, Sancho Dauskardt wrote:
> > >
> > > I had major performance problems reading the SSFDC mapping table: that
> > > involed to usb_stor_bulk_msg() for each of the 4096 blocks and took around
> > > 17 sek (and only around 6 sek same device under Win98 ).. I guessed that
> > > this could possibly be a schedluing issue, and tried following:
>
> Hi Matthew,
>
> is this "turn time" problem you are talking about something specific for
> usb-storage class or more general? Could you give some indication please
> how to observe this - trigger, signature e.g.?
> I'm asking because I've done several performance (latency and throughput)
> measurements recently for bulk transfer and don't see any such problem or
> bottleneck.
> I could post some detailed numbers from the measurements if you are
> interested.
>
> Martin
--
Matthew Dharm Home: [EMAIL PROTECTED]
Maintainer, Linux USB Mass Storage Driver
M: No, Windows doesn't have any nag screens.
C: Then what are those blue and white screens I get every day?
-- Mike and Cobb
User Friendly, 1/4/1999
PGP signature
