NetBSD Problem Report #9383

Received: (qmail 28855 invoked from network); 10 Feb 2000 00:33:32 -0000
Message-Id: <200002100033.QAA29026@digital.clock.org>
Date: Wed, 9 Feb 2000 16:33:31 -0800 (PST)
From: "Erik E. Fair" <fair@digital.clock.org>
Reply-To: fair@netbsd.org
To: gnats-bugs@gnats.netbsd.org
Subject: kernel configs for SUN4, SUN4C, SUN4M
X-Send-Pr-Version: 3.95

>Number:         9383
>Category:       port-sparc
>Synopsis:       kernel configs for SUN4, SUN4C, SUN4M
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    abs
>State:          closed
>Class:          change-request
>Submitter-Id:   net
>Arrival-Date:   Wed Feb 09 16:36:00 +0000 2000
>Closed-Date:    Sun Jan 20 15:42:22 +0000 2008
>Last-Modified:  Sun Jan 20 15:42:22 +0000 2008
>Originator:     Erik E. Fair
>Release:        1.4.1
>Organization:
International Organization of Internet Clock Watchers
>Environment:

System: NetBSD digital.clock.org 1.4.2_ALPHA NetBSD 1.4.2_ALPHA (DIGITAL) #10: Mon Jan 10 22:38:56 PST 2000 fair@doomsday.clock.org:/usr/obj/sys/arch/alpha/compile/DIGITAL alpha

	Also applies to -current

>Description:
	NetBSD/sparc has a wonderful GENERIC kernel config and boot
	process that allows a single kernel to boot on any of the
	supported architectures: sun4, sun4c, and sun4m - something
	that Sun itself never achieved.

	Given that GENERIC has overhead in the MMU code and every
	device driver we know how to drive, perhaps we should
	provide a slightly more specific starting point for people
	who want to run a stripped kernel (one that contains only
	support for the architecture and devices that it is intended
	to run on), but are not as intimate with Sun SPARC hardware
	as those of us who are developers.

>How-To-Repeat:

>Fix:
	I suggest that, at minimum, we provide configs for SUN4,
	SUN4C, and SUN4M with support for devices that can run on
	that hardware.

	We might also want to provide specific kernel configs for
	the models of SPARCs that cannot be expanded, e.g. the SLC
	and ELC, since that should be a single-shot exercise.
>Release-Note:
>Audit-Trail:

From: David Brownlee <abs@mono.org>
To: "Erik E. Fair" <fair@digital.clock.org>
Cc: gnats-bugs@gnats.netbsd.org, port-sparc@netbsd.org
Subject: Re: port-sparc/9383: kernel configs for SUN4, SUN4C, SUN4M
Date: Thu, 10 Feb 2000 01:30:02 +0000 (GMT)

 On Wed, 9 Feb 2000, Erik E. Fair wrote:

 > 	I suggest that, at minimum, we provide configs for SUN4,
 > 	SUN4C, and SUN4M with support for devices that can run on
 > 	that hardware.
 > 
 > 	We might also want to provide specific kernel configs for
 > 	the models of SPARCs that cannot be expanded, e.g. the SLC
 > 	and ELC, since that should be a single-shot exercise.

 	I wrote a small Makefile and awk script to help automate this
 	a little while back:

 	ftp://ftp.netbsd.org/pub/incoming/abs/kernel-config-split/

 	SLC/ELC are not one shot as they would still have to be updated
 	as new pseudo devices, protocols, and potentially scsi devices
 	are added (though I agree they would be nice).

 The README is included below - what do people think?

         A common request is for standard kernel configs for specific machine
 	types eg: for sparc 'sun4m' vs 'sun4c' vs 'sun4', or for the i386
 	'eisa' vs 'pci', vs 'pcmcia', vs appropriate combinations.

         sys/arch/amiga/conf generates a DRACO config using cpp, but this
         makes the GENERIC config file a little harder to follow.

         I've come up with using '>xxx' tags in comments, and a small awk
         script.

         Essentially any line without a >xxx tag is always passed.
         Any line with any >xxx tag, must contain one of the desired tags.

 	ftp://ftp.netbsd.org/pub/incoming/abs/kernel-config/

 	    GENERIC         - Updated GENERIC file
 	    GENERIC.diff    - diffs from current GENERIC
 	    Makefile        - Makefile to generate SUN4 SUN4C and SUN4M
 	    split.awk       - Script to do the work
 	    SUN4            - Generated using tag:  sun4
 	    SUN4C           - Generated using tags: sun4c,sbus
 	    SUN4M           - Generated using tags: sun4m,sbus

 	As a more enthusiastic option you can specify a +tag which affects
 	all previous tags:
 	eg: sun4c,sbus,+fb -> sun4c, sbus, sun4c+fb, sbus+fb, fb

 	ftp://ftp.netbsd.org/pub/incoming/abs/kernel-config2/

 	    GENERIC         - Updated GENERIC file
 	    GENERIC.diff    - diffs from current GENERIC
 	    Makefile        - Makefile to generate SUN4{,C,M}{,_NOFB}
 	    splitconfig.awk - Version with +tag support
 	    SUN4            - Generated using tags: sun4,+fb
 	    SUN4C           - Generated using tags: sun4c,sbus,+fb
 	    SUN4M           - Generated using tags: sun4m,sbus,+fb
 	    SUN4_NOFB       - Generated using tag:  sun4
 	    SUN4C_NOFB      - Generated using tags: sun4c,sbus
 	    SUN4M_NOFB      - Generated using tags: sun4m,sbus

         a) I'd prefer to put split.awk somewhere MI - any problems with
            /sys/conf/splitconfig.awk ?
 	b) We cannot do this with 'explicit orphaning' (though I think
 	   that would be a good feature), as we need to be able to exclude
 	   CPU support (SUN4M etc).

 							Thu Feb 10 2000
 							abs@netbsd.org



From: matthew green <mrg@eterna.com.au>
To: fair@digital.clock.org
Cc: gnats-bugs@gnats.netbsd.org
Subject: re: port-sparc/9383: kernel configs for SUN4, SUN4C, SUN4M 
Date: Fri, 18 Feb 2000 09:14:22 +1100

 you will find old SUN4 SUN4C and SUN4M configs in CVS i believe...
 we got rid of them for some reason. 

From: "Erik E. Fair" <fair@clock.org>
To: matthew green <mrg@eterna.com.au>
Cc: gnats-bugs@gnats.netbsd.org
Subject: re: port-sparc/9383: kernel configs for SUN4, SUN4C, SUN4M
Date: Tue, 22 Feb 2000 02:25:04 -0800

 We probably nuked the architecture specific kernels because GENERIC 
 boots on everything that we support in the 32-bit world. However, 
 GENERIC is not as efficient (space or time) as an arch-specific 
 kernel. This is why GENERIC makes for a terrific install kernel or 
 for an external bootable disk, but is not the best basis for a 
 production kernel.

 Also, the last time I tried to build a stripped 1.4.2_ALPHA sun4 
 kernel, it failed; the graphics devices seem to have Sbus on the 
 brain (even the VME-only ones). This needs to be fixed, and when I 
 reproduce the problem, I will send-pr it separately.

 	Erik <fair@clock.org>

From: matthew green <mrg@eterna.com.au>
To: "Erik E. Fair" <fair@clock.org>
Cc: gnats-bugs@gnats.netbsd.org
Subject: re: port-sparc/9383: kernel configs for SUN4, SUN4C, SUN4M 
Date: Tue, 22 Feb 2000 22:22:20 +1100

    We probably nuked the architecture specific kernels because GENERIC 
    boots on everything that we support in the 32-bit world. However, 
    GENERIC is not as efficient (space or time) as an arch-specific 
    kernel. This is why GENERIC makes for a terrific install kernel or 
    for an external bootable disk, but is not the best basis for a 
    production kernel.

 nah, they were nuked for:

 revision 1.5
 date: 1996/05/15 23:57:52;  author: thorpej;  state: dead;  lines: +1 -1
 These haven't been kept up to date.


 so, if they are to return, they should probably be automatically generated
 somehow...


    Also, the last time I tried to build a stripped 1.4.2_ALPHA sun4 
    kernel, it failed; the graphics devices seem to have Sbus on the 
    brain (even the VME-only ones). This needs to be fixed, and when I 
    reproduce the problem, I will send-pr it separately.

 not being a sysadmin anymore, there is only one `production' NetBSD
 box i run and it's still running 1.3-branch (pretty much top-of).. all
 my other boxes run -current :-)

From: "Charles M. Hannum" <abuse@spamalicious.com>
To: fair@netbsd.org, mrg@netbsd.org, abs@netbsd.org
Cc: gnats-bugs@netbsd.org
Subject: Re: port-sparc/9383: kernel configs for SUN4, SUN4C, SUN4M
Date: Tue, 6 Jul 2004 23:01:10 +0000

 Everyone involved in this PR is a "developer".  It's been over 4 years.  You 
 should either implement or close it.
Responsible-Changed-From-To: port-sparc-maintainer->abs 
Responsible-Changed-By: abs 
Responsible-Changed-When: Tue Jul 6 23:05:48 UTC 2004 
Responsible-Changed-Why:  
I'll see if I can finally fix this 

From: David Brownlee <abs@NetBSD.org>
To: "Charles M. Hannum" <abuse@spamalicious.com>
Cc: fair@netbsd.org, mrg@netbsd.org, gnats-bugs@netbsd.org
Subject: Re: port-sparc/9383: kernel configs for SUN4, SUN4C, SUN4M
Date: Wed, 7 Jul 2004 00:06:11 +0100 (BST)

 On Tue, 6 Jul 2004, Charles M. Hannum wrote:

 > Everyone involved in this PR is a "developer".  It's been over 4 years.  You
 > should either implement or close it.

  	Agreed. Amusingly enough I raised this on another thread recently.
  	I'll take the PR.

 -- 
  		David/absolute       -- www.NetBSD.org: No hype required --
State-Changed-From-To: open->closed
State-Changed-By: abs@narn.netbsd.org
State-Changed-When: Sun, 20 Jan 2008 15:42:22 +0000
State-Changed-Why:
not going to fix


>Unformatted:

Home	PR Database Search