HardPressed + floptical drive

[fadden@uts.amdahl.com (Andy McFadden)]

[ This is a repost of a message that originally went up on GEnie.  The
problem is that entering the "Folders" display in the HP CDev right
after inserting a floptical disk causes the CDev to return error $0028. ]


The problem with the HP "Folders" display and floptical drives is not
HP's fault, but rather it appears to be the RamFAST driver's fault.

In short, a DStatus call on a nonexistent device is returning error $28
(no device connected), something it should not be doing.  The DStatus
call looks like:

    pCount = 5
    devNum = 10 (9 is the floptical)
    code = 0 (GetDeviceStatus)
    list = [buffer]
    requestCount = 2

Normally this works fine.  And if you just call it, it WILL work fine.  You
have to make GS/OS calls in a certain sequence; I tried it from the
Exerciser and was able to get it to fail.


Here's what ya gotta do from the Exerciser:

  1. Boot GS/OS with nothing in the drive.
  2. Crank up the Exerciser.
  3. Insert a floptical disk into the drive.
  4. Call DInfo ($2c) on the floptical drive.  In my case it's device $09.
  5. Call DStatus ($2d) on the floptical drive.  I always used
     requestCount = 2; dunno if it matters or not.
  6. Call Volume ($08) on the device you got from step 1.  In my case this
     is .CVTECH.S7.F.  The disk will spin.
  7. Call DInfo ($2c) on the next device; in my case it's device $0a.  I
     usually get ".DEV2", but after doing steps 1-6 I get .CVTECH.S7.G.
     Somehow the RamFAST driver thinks I just grew an extra drive.
  8. Call DStatus ($2d) on that device.  It will fail with error $28 when
     the driver realizes that, no, there really isn't a device there.

If you don't do things EXACTLY as above, it won't happen.  I tried just
doing steps 7&8 and it worked fine, and always worked fine afterward.  The
person who first noticed this problem found something similar.

I can teach HP to be more lenient about errors, but I'd rather not if
this can be fixed without too much trouble.


[ A reply from Drew (CV Tech dude) indicated that the problem is likely
in the card firmware rather than the driver, making it difficult to fix.
I can work around the problem by ignoring the error message, though I'm
reluctant to do this because at one point my system crashed while I was
trying to repeat the bug.  A modified CDev will probably be part of v1.0.3. ]

-- 
fadden@uts.amdahl.com (Andy McFadden)

"Our UNIX is bigger than your UNIX"
[ These are my opinions, not Amdahl policies. ]