The Unix Heritage Society mailing list
 help / color / mirror / Atom feed
* [TUHS] Why is it always "fast boot" in 2.11BSD?
@ 2022-05-26 11:41 Matthias Bruestle
  2022-05-28 11:39 ` [TUHS] " Hans Rosenfeld
  0 siblings, 1 reply; 5+ messages in thread
From: Matthias Bruestle @ 2022-05-26 11:41 UTC (permalink / raw)
  To: tuhs

Hi,

I have noticed, that 2.11BSD is in all cases where I looked set
to "fast boot", which AFAIK means no fsck of at least /. I found
nobody talking about this or providing information about how to
change it to "slow boot" with a proper check, which is now normal.

Is there a reason why it is not possible to deactivate fast boot?
Or is it just that nobody bothered to do it?

Thanks
Matthias

-- 
When You Find Out Your Normal Daily Lifestyle Is Called Quarantine

^ permalink raw reply	[flat|nested] 5+ messages in thread

* [TUHS] Re: Why is it always "fast boot" in 2.11BSD?
  2022-05-26 11:41 [TUHS] Why is it always "fast boot" in 2.11BSD? Matthias Bruestle
@ 2022-05-28 11:39 ` Hans Rosenfeld
  2022-05-28 13:14   ` Matthias Bruestle
  0 siblings, 1 reply; 5+ messages in thread
From: Hans Rosenfeld @ 2022-05-28 11:39 UTC (permalink / raw)
  To: tuhs

On Thu, May 26, 2022 at 01:41:29PM +0200, Matthias Bruestle wrote:
> I have noticed, that 2.11BSD is in all cases where I looked set
> to "fast boot", which AFAIK means no fsck of at least /. I found
> nobody talking about this or providing information about how to
> change it to "slow boot" with a proper check, which is now normal.
> 
> Is there a reason why it is not possible to deactivate fast boot?
> Or is it just that nobody bothered to do it?

Interesting. When I played with 2.11BSD a few years ago I found that it
only did a "fast" boot if rebooted with "shutdown -rf". A cold boot
would always drop into single user and allow running fsck manually.

IIRC /boot or even one of the earlier bootstrap stages would check a
certain memory location for certain values, which the kernel(?) would
place there in the reboot syscall. As I wanted to always force a fast
boot, I had this in my simh.ini to force a fast boot:

d 157772 177777
d 157774 002400
d 157776 000000

If you're using a simulator I wouldn't be suprised if it does that
automatically nowadays, of perhaps 2.11BSD was patched a while back to
reverse that behaviour.

I don't remember what these values do, though. And you want to do the
opposite anyway. But perhaps this helps you find out more. 


Hans


-- 
%SYSTEM-F-ANARCHISM, The operating system has been overthrown

^ permalink raw reply	[flat|nested] 5+ messages in thread

* [TUHS] Re: Why is it always "fast boot" in 2.11BSD?
  2022-05-28 11:39 ` [TUHS] " Hans Rosenfeld
@ 2022-05-28 13:14   ` Matthias Bruestle
  2022-05-28 22:35     ` Matthias Bruestle
  0 siblings, 1 reply; 5+ messages in thread
From: Matthias Bruestle @ 2022-05-28 13:14 UTC (permalink / raw)
  To: tuhs

On Sat, May 28, 2022 at 01:39:03PM +0200, Hans Rosenfeld wrote:
> Interesting. When I played with 2.11BSD a few years ago I found that it
> only did a "fast" boot if rebooted with "shutdown -rf". A cold boot
> would always drop into single user and allow running fsck manually.

I use a quite basic simh.ini without fancy memory manipulations.
At the moment the behaviour is:
1) First boot => enter unix at : prompt => no fsck => root shell
2) ^D => "fast boot" => no fsck => multiuser
3) login + reboot => passed : prompt =>  no "fast boot" => fsck => multiuser
That is with the RA92 image provided with simh for the ESP32,
which contains a command to setup WiFi.

I found at minnie.tuhs.org/Distributions/Boot_Images/2.11_on_Simh
a version where Andru Luvisi has configured a 1MB RAM disk mounted
to /tmp for the temp file required by fsck.  Needed to compile a new
kernel, but with the RAM disk fsck runs nicely. I put a # in front of
rwhod in /etc/rc to prevent it beeing started, but then it never went
into multisuser. No idea why.

> IIRC /boot or even one of the earlier bootstrap stages would check a
> certain memory location for certain values, which the kernel(?) would
> place there in the reboot syscall. As I wanted to always force a fast
> boot, I had this in my simh.ini to force a fast boot:
> 
> d 157772 177777
> d 157774 002400
> d 157776 000000

Seems to be related to bootdev here:
73Boot from ra(0,0,0) at 0172150
: unix
Boot: bootdev=02400 bootcsr=0172150

But it looks always like this, with and without fastboot.

Thanks
Matthias

-- 
When You Find Out Your Normal Daily Lifestyle Is Called Quarantine

^ permalink raw reply	[flat|nested] 5+ messages in thread

* [TUHS] Re: Why is it always "fast boot" in 2.11BSD?
  2022-05-28 13:14   ` Matthias Bruestle
@ 2022-05-28 22:35     ` Matthias Bruestle
  2022-05-29 12:12       ` Matthias Bruestle
  0 siblings, 1 reply; 5+ messages in thread
From: Matthias Bruestle @ 2022-05-28 22:35 UTC (permalink / raw)
  To: tuhs

Just had a look at an image patched to 469. There are awesome
improvements. There is top, swap space and a clean flag. I will try to
get it to 473. The clean flag is really useful, because the complete
fsck of the RA92 image on the ESP32 take at least an hour. On the
first glance it looks like someone investet a lot of work into
a proper startup.

Matthias

-- 
When You Find Out Your Normal Daily Lifestyle Is Called Quarantine

^ permalink raw reply	[flat|nested] 5+ messages in thread

* [TUHS] Re: Why is it always "fast boot" in 2.11BSD?
  2022-05-28 22:35     ` Matthias Bruestle
@ 2022-05-29 12:12       ` Matthias Bruestle
  0 siblings, 0 replies; 5+ messages in thread
From: Matthias Bruestle @ 2022-05-29 12:12 UTC (permalink / raw)
  To: tuhs

[-- Attachment #1: Type: text/plain, Size: 1850 bytes --]

On Sun, May 29, 2022 at 12:35:20AM +0200, Matthias Bruestle wrote:
> Just had a look at an image patched to 469. There are awesome
> improvements. There is top, swap space and a clean flag. I will try to
> get it to 473. The clean flag is really useful, because the complete
> fsck of the RA92 image on the ESP32 take at least an hour. On the
> first glance it looks like someone investet a lot of work into
> a proper startup.

So I took the 2.11BSD pl469 (is there a custom how to specify the pl?)
from PiDP11. It is a bit larger than an RA72. I truncated it to the RA72
size, because simh would only use it r/o. I haven't noticed a problem
because of this and the partitions look good, although I would have
partitioned it differently. 8MB for /, 410MB for /usr and 550MB for
/home. / is already 44% full.

Patching went well. Pasted a shar of a tar, because I am still not able
to read a tape. (Tips on simh configuration and device?) Haven't looked
at pl469, but I can say pl473 is really nice. It boots after a few seconds
on its own just like a modern Linux. It is doing also a fsck just like a
modern Linux depending on the state of the fs. Initializing everything
down to the multiuser state and login. AFAIK 2BSD was used in the past
for doing work. I would have expected something like it is now from old
2.11BSD versions. Maybe the automatic booting with timeout is a newer
invention and top a newer software, but the features, which were there
should have been working in a conventient way. In case someone here
helped making the patches: Thanks!

With my port of esppdp (based on simh 4) to the M5Stack Core2 and the
wifid daemon from esppdp I have now a nice running PDP-11 in a box of
5.5*5.5*1.5cm. Screenshot of top running on it via a telnet login.

Matthias

-- 
When You Find Out Your Normal Daily Lifestyle Is Called Quarantine

[-- Attachment #2: 2.11BSDpl473-top.png --]
[-- Type: image/png, Size: 13638 bytes --]

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2022-05-29 12:20 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-05-26 11:41 [TUHS] Why is it always "fast boot" in 2.11BSD? Matthias Bruestle
2022-05-28 11:39 ` [TUHS] " Hans Rosenfeld
2022-05-28 13:14   ` Matthias Bruestle
2022-05-28 22:35     ` Matthias Bruestle
2022-05-29 12:12       ` Matthias Bruestle

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).