2.11 BSD on an 11/93 with an CMD SCSI Module

[Glen Slick]

On Wed, Mar 13, 2019 at 10:36 AM Josh Dersch via cctalk
<cctalk at classiccmp.org> wrote:
>
> On Wed, Mar 13, 2019 at 10:29 AM Zane Healy via cctalk <
> cctalk at classiccmp.org> wrote:
>
> >
> > > On Mar 13, 2019, at 10:15 AM, Bill Gunshannon via cctalk <
> > cctalk at classiccmp.org> wrote:
> > >
> > > Anybody had any luck with this configuration?
> > > I used VTServer to install it and everything works fine
> > > except for the very last step.  It installs completely
> > > and boots from within VTServer with ra(0,0,0)unix.  I
> > > issue the "dd if=/mdec/rauboot of=/def/rra0a count=1" and
> > > it completes successfully.  But the disk still won't boot.
> > > I have tried it three times now with the same results.
> > >
> > > Suggestions?
> > >
> > > bill
> >
> > Are all the patches applied?  IIRC, some of the later patches were around
> > SCSI support.  I know one of them is required for Viking QDT support, not
> > sure about CMD.  Mind you, I’ve never installed any flavor of UNIX on my
> > PDP-11’s.
> >
>
> Yes, you will need to apply the patches (I believe patch #448 is the one
> you need, but why not apply them all...), they fix the MSCP drivers & boot
> code for the Viking and CMD SCSI controllers.
>
> - Josh

Item 3 in Patch #441

Subject: libvmf,cpp,rauboot,mkstr(1) fixes, ld feature (#441)
Index: usr.lib/libvmf,lib/cpp,sys/mdec/rauboot.s,... 2.11BSD

Description:
1. The libvmf (virtual memory using tmp files) routines could crash
   or hang an application if too many locked segments were requested.

2. The C preprocessor (cpp) had a memory leak which could cause it
   to run out of memory _and_ also failed to check the return
   status of calloc() which would cause cpp to fail ungracefully.

3. The MSCP bootblock would not work with a CMD CQD220 adaptor due
   to the bootblock relying on the (undocumented?) behaviour of DEC
   adaptors.

4. In the mkstr(1) manpage the usage example given used lseek(2)
   incorrectly.

5. The linker (ld) does not accept the '-g' flag which on other
   systems says to retain debuggin symbols.

6. creat() is no longer a system call, thus it is an error to
   have creat(2) instead of creat(3).