Subject: Re: Tapeboot problem SOLUTION
To: None <Ian.Dall@dsto.defence.gov.au, eichin@kitten.gen.ma.us, tom@sdf.com>
From: Gordon W. Ross <gwr@mc.com>
List: port-sun3
Date: 12/11/1995 12:35:00
An apology is in order.

It has been pointed out that my MakeBootTape script is broken.

I must confess that I changed that script just before release
when I found out the files would all be distributed in gzip-ed
form.  Unfortunately, I broke it.  I'm sorry...

The script appears to work if you change each occurance of
	bs=
to
	obs=
which is required due to taking input from a pipe.

> Date: Mon, 11 Dec 1995 09:55:24 +1030
> From: Ian Dall <dall@HFRD.DSTO.GOV.AU>
> References: <9512080139.AA01757@copper.bc.ca>
> 	<Pine.BSF.3.91.951208040004.19411A-100000@misery.sdf.com>

> [...] Consider the command
> 
>   gunzip < netbsd-rd.gz | dd bs=8k of=$T
> 
> due to the pipe, dd will get short reads, [...]

Yes.  That's the bug.  I added the pipe but did not change
the bs=8k to obs=8k which is required due to the pipe.

> In short, the bs=8k doesn't do much.

The 8k block size is to reduce the SCSI command overhead.
Without that, the tape might not be able to keep streaming.

> On the other hand, reading from a file:
> 
>  gunzip netbsd-rd.gz; dd if=netbsd-rd bs=8k conv=sync
> 
> works because the only short read happens at the end of the file and
> the conv=sync pads that to 8k which is fine. This is the solution I
> adopted, though I am not at all sure dd is doing the right thing.

Right.  If the files are not compressed, making the tape is easier.

Thanks for helping to identify and correct this error.
Please accept my sincere apology for the error.

Gordon Ross