[Simh] {Spam?} transferring files to and from v6 unix running in pdp11 emulator

[Will Senn]

On 11/20/15 1:22 PM, Anders Magnusson wrote:
> Den 2015-11-20 kl. 20:12, skrev Will Senn:
>>
>>
>> On 11/20/15 10:51 AM, Anders Magnusson wrote:
>>> Den 2015-11-20 kl. 17:35, skrev Will Senn:
>>>> Hi,
>>>>
>>>> I have searched and searched and have not found a satisfactory 
>>>> answer to this question:
>>>>
>>>> How can I efficiently copy files from my host system to unix 
>>>> version 6 running in the pdp11 emulator and from unix 6 to my host 
>>>> system?
>>> I usually do this by just tar'ing to/from a disk file, something like:
>>>
>>> on pdp11: tar cf /dev/rxp0c myfiles
>>> on host: tar xf simh-disk-file
>>>
>>> Works both ways :-)
>>>
>>> -- Ragge
>> Hi,
>>
>> This sounded promising, but as it turns out there appear to be 2 
>> problems... 1. v6 doesn't have tar, and 2. it looks like v6 doesn't 
>> have rx device support (floppies). Are you using v6 with tweaks, or 
>> v7, or some other variant of unix?
> Hm, it was some time ago, but I think it was tar.  Otherwise try cpio?
> You do not have to use floppies.  rxp is the raw device for rp disks 
> (avoids caching).  Just work against that file.
>
> -- Ragge
>

Thanks. I combined a number or responses and came up with the following...

--------
--From Unix v6 on PDP 11/40 SimH to Host:
--------
create and attach an additional rk device:
simh> attach rk3 rk3
co

get a file of interest and note it's size in bytes:
# ls -l /etc/rc
-rw-rw-r--  1 bin        90 Oct 10 12:32 /etc/rc

look at the od dump of the file for comparison later:
# od -c /etc/rc
0000000  r  m     -  f     /  e  t  c  /  m  t  a  b \n
0000020  /  e  t  c  /  u  p  d  a  t  e \n  /  e  t  c
0000040  /  m  o  u  n  t     /  d  e  v  /  r  k  1
0000060  /  u  s  r  /  s  o  u  r  c  e \n  /  e  t  c
0000100  /  m  o  u  n  t     /  d  e  v  /  r  k  2
0000120  /  u  s  r  /  d  o  c \n \n
0000132

write the file to the rk device (the sync may not be needed, but the 
result looks cleaner, also it doesn't apper that you can specify bs=1, 
device errors out):
# dd if=/etc/rc of=/dev/rrk03 conv=sync
0+1 records in
1+0 records out

exit the sim and then on the host, read from the rk image using bs=1 and 
count from the ls output:
$ dd if=rk3 of=rc bs=1 count=90
90+0 records in
90+0 records out

$ od -c rc
0000000    r   m       -   f       /   e   t   c   /   m   t   a b  \n
0000020    /   e   t   c   /   u   p   d   a   t   e  \n   /   e t   c
0000040    /   m   o   u   n   t       /   d   e   v   /   r   k 1
0000060    /   u   s   r   /   s   o   u   r   c   e  \n   /   e t   c
0000100    /   m   o   u   n   t       /   d   e   v   /   r   k 2
0000120    /   u   s   r   /   d   o   c  \n \n
0000132

This looks perfect, but I'm skeptical.

--------
-- From Host to Unix v6 on PDP 11/40 SimH Host:
--------
make a minor edit to the rc file (change m to n in the word mtab) and 
note it's size in bytes:
$ ls -l rc
-rw-r--r--  1 wsenn  staff  90 Nov 20 16:15 rc

it better be 90, unless I did something other than changing a letter

$ dd if=rc of=rk3 conv=sync
0+1 records in
1+0 records out
512 bytes transferred in 0.000037 secs (13854733 bytes/sec)
note the count of blocks

with the number of blocks on hand, fire up the simulation
# dd if=/dev/rrk3 of=rc.dd count=1

because of the fact that I can't specify bs=1, the result is padded:
od -c rc.dd
0000000  r  m     -  f     /  e  t  c  /  n  t  a  b \n
0000020  /  e  t  c  /  u  p  d  a  t  e \n  /  e  t  c
0000040  /  m  o  u  n  t     /  d  e  v  /  r  k  1
0000060  /  u  s  r  /  s  o  u  r  c  e \n  /  e  t  c
0000100  /  m  o  u  n  t     /  d  e  v  /  r  k  2
0000120  /  u  s  r  /  d  o  c \n \n \0 \0 \0 \0 \0 \0
0001000

so, I read from the dd file with the number of bytes still in hand:
dd if=rc.dd of=rc bs=1 count=90
90+0 records in
90+0 records out

then diff the file against the original:
diff rc /etc/rc
1c1
* rm -f /etc/ntab
---
. rm -f /etc/mtab

Woohoo and whew! A lot of work, but file transfer in both directions. 
Does this approach seem sound to you or do you see some refinements that 
are possible/necessary?

Thanks,

Will