On Apr 17, 2005, at 11:39 AM, Manuel Bouyer wrote:
>>
>>> Yes, this is quite possible
>>> But the kernel should spin up the drives and process with the normal
>>> identification and negotiation then. I have a 1.6.2 system with 
>>> drives
>>> configured to not spin up at poweron, and it works as expected.
>> Well, in 1.6 it right under waiting 2 seconds for devices to settle,
>> in 2.0 It runs thru atabus drives, then atapi, then the scsi, I only
>
> This is normal. It's related to a change to the IDE subsystem, not 
> SCSI.
>
>> see the initial abbreviated drive info, the Check Condition line etc,
>> No: sd1: 17366 MB, 8154 cyl ...
>> No: sd2: sync (50.00ns ) ... on the first boot, when spinup is needed
>> -- the kernel
>> is getting this info, but not displaying it?
>
> No, it's probably not getting it, or it would print it.
> Maybe it comes later, when the disk is first accessed ?
> Can you try the attached program ?
> Usage: ./tst /dev/rsd1d 10000
> It will print the speed at which the bus can move data from the 
> drive's cache.

If drive needs spin-up then  ~2.14 MB/s
if not  ~7.49 MB/s

> Do you have an idea if the drive could initiate the negotiaition ?
> There may be a jumper for this, but it depends on the drive model.
>
Decided to try more jumper settings before responding.  There's a 
jumper marked
sync on the converter, set that, took off 1 on the drives called "Dis 
Ti Sy"
and I see "sd1: async, 8-bit-transfers", but Aborted Commands every 
where -- can't use, which I think I had discovered before, putting it 
back on, its 2.14/7.50 again.
There's something called "SP SYNC (for 68DF& 80pin)" enabled for 1 
drive, still 2.14/7.50
I think I had determined that it didn't seem to make a difference some 
time before... it is a 50pin controller  -- Basically the answer seems 
to be NO, I'm guessing Dis Ti Sy may be DISsable negoTIation SYncronous 
, so I can only turn it off?

I haven't found any docs that explain this specific drive's jumper 
settings, some docs about the drive family assembly specs, that mention 
the model, but nothing that has the jumper block this drive has -- 
maybe lost when Hitachi took over support from IBM, maybe they never 
existed, or were document only as part of some IBM system, that IBM no 
longer supports.

Kinda of slow isn't it?  I'll see if adapter config changes anything.