Bug 26215

Summary: Incorrect way to start SVR4 emulation on boot
Product: Base System Reporter: Dima Sivachenko <dima>
Component: confAssignee: newton
Status: Closed FIXED    
Severity: Affects Only Me    
Priority: Normal    
Version: 3.3-RELEASE   
Hardware: Any   
OS: Any   
Attachments:
Description Flags
file.diff none

Description Dima Sivachenko 2001-03-30 12:40:01 UTC
svr4.ko requires streams.ko to be loaded.
/usr/sbin/svr4 takes care of this.
/usr/sbin/svr4 should be used in /etc/rc instead of 'kldload svr4'.

How-To-Repeat: 
Add svr4_enable="YES" in /etc/rc.conf and reboot.
An error occurs while loading svr4.ko.
Comment 1 Doug Barton freebsd_committer freebsd_triage 2001-04-05 06:07:41 UTC
Responsible Changed
From-To: freebsd-bugs->newton


Word is that Mr. Newton is knowledgeable in this area
Comment 2 newton freebsd_committer freebsd_triage 2001-04-09 23:40:01 UTC
In PR conf/26215, you wrote:

 > Release
 >     FreeBSD 3.3-RELEASE i386

Are you sure about this?  svr4 wasn't released until 4.0.

At the time it was released, the module was loaded from rc.i386, and it
used the svr4(8) script in the manner you've suggesed in the fix attached
to conf/26215.

Sheldon Hearn did a rearrangement of rc and rc.i386 on 18 Aug 2000, when
he moved svr4 and linux into /etc/rc and used kldload to load them instead
of the svr4 script.  At a similar time, it was noted that the module
dependency system was able to work out that svr4 required stremas, and 
the /usr/sbin/svr4 shell script was removed.

The fact that you've said you're running 3.3-RELEASE leaves me wondering
which version you're *really* running -- If it really is 3.3-RELEASE, then
you'll be using the old LKM interface, and "kldload svr4" in /etc/rc
won't really be appropriate, but you'd know that anyway because you're 
well-versed enough in rocket-science to run the module on an unsupported
OS :-)  

If you're really running 4.x, then this PR should be a non-event on a 
recent enough -current (say, 18 Aug 2000) because modules understand how
to load their own dependencies.  If you don't believe that's the case,
I'll blow the dust off my test box and put today's -current on it and 
work outwhere the problem lies, but I believe it should be a non-event.

If your 4.x is older than that, you might have problems because module
dependencies weren't quite working the way they do now.  If that's the case,
then yes, the suggested fix is correct.

So -- Can you please let me know what version of FreeBSD you're running?
That's really what I need to work out what (if anything) needs to be committed
to fix this.

   - mark

--------------------------------------------------------------------
I tried an internal modem,                    newton@atdot.dotat.org
     but it hurt when I walked.                          Mark Newton
----- Voice: +61-4-1620-2223 ------------- Fax: +61-8-82231777 -----
Comment 3 Dima Sivachenko 2001-04-10 09:47:05 UTC
I am running 4.3-BETA as of March 23.

--dima

PS: I just submitted this PR from machine running 3.3-RELEASE...


On Mon, Apr 09, 2001 at 03:40:01PM -0700, newton@FreeBSD.org wrote:
> In PR conf/26215, you wrote:
> 
>  > Release
>  >     FreeBSD 3.3-RELEASE i386
> 
> Are you sure about this?  svr4 wasn't released until 4.0.
> 
> At the time it was released, the module was loaded from rc.i386, and it
> used the svr4(8) script in the manner you've suggesed in the fix attached
> to conf/26215.
> 
> Sheldon Hearn did a rearrangement of rc and rc.i386 on 18 Aug 2000, when
> he moved svr4 and linux into /etc/rc and used kldload to load them instead
> of the svr4 script.  At a similar time, it was noted that the module
> dependency system was able to work out that svr4 required stremas, and 
> the /usr/sbin/svr4 shell script was removed.
> 
> The fact that you've said you're running 3.3-RELEASE leaves me wondering
> which version you're *really* running -- If it really is 3.3-RELEASE, then
> you'll be using the old LKM interface, and "kldload svr4" in /etc/rc
> won't really be appropriate, but you'd know that anyway because you're 
> well-versed enough in rocket-science to run the module on an unsupported
> OS :-)  
> 
> If you're really running 4.x, then this PR should be a non-event on a 
> recent enough -current (say, 18 Aug 2000) because modules understand how
> to load their own dependencies.  If you don't believe that's the case,
> I'll blow the dust off my test box and put today's -current on it and 
> work outwhere the problem lies, but I believe it should be a non-event.
> 
> If your 4.x is older than that, you might have problems because module
> dependencies weren't quite working the way they do now.  If that's the case,
> then yes, the suggested fix is correct.
> 
> So -- Can you please let me know what version of FreeBSD you're running?
> That's really what I need to work out what (if anything) needs to be committed
> to fix this.
> 
>    - mark
> 
> --------------------------------------------------------------------
> I tried an internal modem,                    newton@atdot.dotat.org
>      but it hurt when I walked.                          Mark Newton
> ----- Voice: +61-4-1620-2223 ------------- Fax: +61-8-82231777 -----
>
Comment 4 jkh freebsd_committer freebsd_triage 2001-04-15 08:18:37 UTC
State Changed
From-To: open->closed

A fix was committed to -current and -stable yesterday, thanks.