Date: Sun, 14 Sep 1997 21:20:24 -0400
From: allisonp@world.std.com (Allison J Parent)
Message-Id: <199709150120.AA07732@world.std.com>
To: pups@minnie.cs.adfa.oz.au
Subject: Re: Bootstrap Idea
Sender: owner-pups@minnie.cs.adfa.oz.au
Precedence: bulk

<There are a few good tricks to get PDP11 with odd configurations to load. T

Boots for this list were found in the RT11 pocket guide, the longest was 
31words. Most about 20.

rx01
rx02
tm11
tju16
rp02/3
rjs03/4
rk11
rf11
dectape(tu60)
dectapeII(tu58)
RL01/2
Rk06

<first is the venerable paper tape bootstrap. It consists of only 8 instruct
<(28 bytes) and works on serial ports. It then loads the 'absolute loader'

The tu58 is also short and can load any program with code in block 0
(default boot block).  the tu58 can also simulate a console driving ODT.
The best part if TU58 is serial interface and can be hooked up to any 
unix/linux/dos/cpm/Whatever box that has a serial port.

<On the subject of bad blocks, V6 and V7 offered no bad block strategies.
<The DEC spec for RK05's was 200 tracks by 12 sectors by 2 surface plus 3
<bad block tracks for 4800 blocks plus 72 spare. The media was generally pre
<good, and all Unix versions used 4872 block filesystems.

the above describes a strategy.  

Well, transparent or system based the blocks are allocated.  CP/M does not 
do block replacement but utilities can mark bad blocks by allocating them, 
or revector them at the hardware level.  I believe mkfs does this.  Some 
disks like the RQDXn do this at the hardware level.

Allison