On Sun, 17 Oct 1999, Anders Magnusson wrote:

> I have finally gotten some spare time, and have focused on getting the
> VS3100 SCSI DMA engine running. And it seems to work for me now
> on a MicroVAX 3100 anyway :-)

I've not stress tested the SCSI DMA stuff yet (must run some kernel and
package builds :-) but it looks to be performing well.  The new kernel
booted into my 1.4.1 environment without any unexpected problems.

Reads on my VS3100/M76 with a really horrid RZ22 drive:

  futplex# dd if=/dev/rsd0c of=/dev/null bs=64k count=512
  512+0 records in
  512+0 records out
  33554432 bytes transferred in 102 secs (328965 bytes/sec)

Reads on my VS3100/M38 with a slightly less horrid RZ24 drive:

  meme# dd if=/dev/rsd0c of=/dev/null bs=64k count=512
  512+0 records in
  512+0 records out
  33554432 bytes transferred in 54 secs (621378 bytes/sec)

I've not tested my VS3100/M30 with SCSI/MFM controller yet, so I don't
know if that behaves or not.  Still, the above results are really really
good.  The VAX port is really coming together nicely now.  Will it be
possible to get the SCSI DMA changes backported for the 1.4.2 release, or
must we wait until 1.5 before the changes will be present in an official
release?

Also, and someone else mentioned this already, why the change from ln0 to
le0?  I know we used to have le0 as the ethernet device name a year or
more back before you decided to use ln0.  Why the change back to le0?

Thanks for all the hard work on this.  It's greatly appreciated.

-brian.
--- Brian Chase | bdc@world.std.com | http://world.std.com/~bdc/ -----
Never eat any food which might accidentally break inside your mouth!  -- K.