"Gordon W. Ross" <gwr@mc.com> writes:

  > I have not been able to learn much about the DMA problem on the 3/50
  > because I have yet to catch my 3/50 in the act of data corruption.
  > However, I have studied some other driver soures for the Sun3 SCSI
  > controller and I have a change that might avoid the H/W problem.
  > (At least a comment in the Sprite code suggests it might...)

  > I can't tell if my fix helps or not, since my 3/50 does not appear
  > to have the data corruption problem, but I hope some brave souls
  > might step forward and volunteer to test my change.  Please send
  > some mail if you have an 3/50, a scratch disk, and guts. 8^)


I can do this. I have a methodology which seems to work without corrupting
the disk. I have a disk with a small partition (I split off some swap
space) which is "disposable". I boot single user, mount everything
read only. Then I can test the IO to the small scratch partition.

One thing I am not keen on is recompiling since I assume it would
take approximately forever with si_options = 0. So, if you can
give me a precompiled kernel, I am happy to check it out.

Ian