bug squashed in up.c

[utzoo!decvax!ucbvax!ucsfcgl!tef]

There's a bug in the 4.1bsd `up' disk driver
(and maybe other disk drivers as well-- I haven't
checked anything other than up.c).

The way the upslave probe code was originally
set up, it ALWAYS thought a slave drive was
present, even if it wasn't.  Thus if you set
up your config file to check for two slave
drives but really only had 1 present, the autoconfigure
code still told the up driver to attach 2 drives.

The fix is to add a new line of code after you
load the drive number into the cs2 register.
The added line (in upslave()) is marked with
a "+" below:

	upaddr->upcs1 = 0;	/* conservative */
	upaddr->upcs2 = ui->ui_slave;
+	upaddr->upcs1 = UP_NOP|UP_GO;
	if (upaddr->upcs2&UPCS2_NED) {


This fix was tested on a vax750 with emulex sc21v controller.

tom ferrin