Bill Studenmund <wrstuden@loki.stanford.edu> wrote:

>On Thu, 9 Apr 1998, semios doz wrote:
>
>> If you're getting a SCSIRead(), #5 error then the netbsd/mac68k 
>> <http://www.netbsd.org/Ports/mac68k/faq/faq-4.html#ss4.12> faq explains 
>> why.  Basically, installer has a bug that will only allow it to install 
>> to the first gig of any hard disk.  This was the problem I had installing 
>> it onto my system.  I repartitioned the disk, so that the NetBSD 
>> partitions lie within the first gig, and now it works fine.
>
>We should make it a little clearer. The SCSIRead #5 error is not the
>problem. It always seems to do that.
>
I think this is caused by some sort of Phase error in SCSI processing.  
There are some interesting comments about this in the MkLinux code I've 
been looking at, and there is some special code in there to deal with 
different versions of SCSI Manager found in the various Macs.  (One place 
in this code indicates these errors should be ignored and they aren't 
even reported to the user!)  Hopefully this problem will disappear when I 
get these routines installed into Mkfs.

>The big-drive problem is seperate, though I thought it'd been fixed. I
>know it was at one time.... The "big sets fail but small ones work" makes
>it sound more like a memory leak or some such. ??

Yeah, the big-drive problem has been fixed for a long time.  *BUT* that 
doesn't mean that Mkfs is not without error in the way it formats the BSD 
Filesystem.  That code in Mkfs hasn't changed since the early days of the 
old MacBSD/Alice effort when all supported disks were much smaller than 
1Gig.  I'm quite sure there have been many, many bugs squashed in "newfs" 
since those days but unfortunately those changes haven't been rolled into 
Mkfs.  I'm working on fixing this oversight in a new version, but it's 
low priority and slow going right now.

-bob