Re: Any chance of release patch being committed?

2001-05-04 Thread Sheldon Hearn



On Fri, 04 May 2001 00:00:58 +0200, J Wunsch wrote:

 We are, but why would we use anything else than GENERIC for it,
 seriously?  I'd never roll a `release' for my current machine.

When you get commitment to a fixed hardware configuration for two years
and you want to leave your datacenter full of hosts ready to power up
and play off a custom CDROM, a release that uses a custom kernel is
desirable.

That said, I don't think that the release should pick up KERNCONF.  I
think a middle-ground option is to have the release pick up
RELEASE_KERNCONF.

Ciao,
Sheldon.

To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message



Re: Any chance of release patch being committed?

2001-05-04 Thread Terry Lambert

J Wunsch wrote:
 Terry Lambert [EMAIL PROTECTED] wrote:
 
  Isn't anyone but me and Walnut Creek cum BSDI cum Windriver
  Systems using make release?!?
 
 We are, but why would we use anything else than GENERIC for it,
 seriously?  I'd never roll a `release' for my current machine.

Examples abound:

1)  Licensed commercial drivers not present in GENERIC

2)  Cut down kernel for low memory systems

3)  Kernel without support for 686/586/386 in an
embedded system

4)  Kernel without most of the disk controllers, e.g.
a Compaq-specific Jump Start disk for FreeBSD

Etc.

It shouldn't matter anyway; it doesn't change the normal
default sysinstall behaviour, image, or image size at all
(the image is identical), unless KERNCONF is overridden.

If that's so unreasonable, I suggest you remove the ?
before the = in /usr/src/release/Makefile in the KERNCONF
assignment, which looks like:

KERNCONF?=GENERIC

Since it makes damn little sense to allow it to be overridden,
if the resulting CDROM will fail to install /kernel, and the
resulting system will not be able to boot without an explict
load kernel.MYKERNEL command to the boot loader.

Doing that would reduce functionality; making my change
would increase it... but make one or the other of them,
please: in it's current state, the code _pretends_ that
an override will work, when it will not.

I hate code that lies to me, even if by omission; don't you?

-- Terry

To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message



Re: Any chance of release patch being committed?

2001-05-04 Thread Makoto MATSUSHITA


tlambert2 Examples abound:

All of examples are just we want yet another *generic* kernel for
specific machines.

Why you don't replace src/sys/${arch}/conf/GENERIC file to your own
(customized) version? It is too simple, and matches your requirements.

-- -
Makoto `MAR' MATSUSHITA

To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message



Re: Any chance of release patch being committed?

2001-05-04 Thread Sheldon Hearn



On Fri, 04 May 2001 20:38:06 +0900, Makoto MATSUSHITA wrote:

 All of examples are just we want yet another *generic* kernel for
 specific machines.
 
 Why you don't replace src/sys/${arch}/conf/GENERIC file to your own
 (customized) version? It is too simple, and matches your requirements.

Because that's a repo-backed file.  Why make me manage CVS conflicts on
update?  Why not just allow something like RELEASE_KERNCONF?  What's the
real objection to it?  So far, all I've heard is Gee, I've never needed
that. :-)

Ciao,
Sheldon.

To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message



Re: Any chance of release patch being committed?

2001-05-04 Thread John Baldwin


On 04-May-01 Sheldon Hearn wrote:
 
 
 On Fri, 04 May 2001 00:00:58 +0200, J Wunsch wrote:
 
 We are, but why would we use anything else than GENERIC for it,
 seriously?  I'd never roll a `release' for my current machine.
 
 When you get commitment to a fixed hardware configuration for two years
 and you want to leave your datacenter full of hosts ready to power up
 and play off a custom CDROM, a release that uses a custom kernel is
 desirable.
 
 That said, I don't think that the release should pick up KERNCONF.  I
 think a middle-ground option is to have the release pick up
 RELEASE_KERNCONF.

That is fine, but do note that Terry didn't add a new KERNCONF variable, it was
already called that to begin with.  We don't use buildkernel/installkernel in
release, so it doesn't actually cause any conflicts.  But anyways.  Anyone have
any reasonable objections to Terry's patch?  If not I'll test it out for sanity
and then commit it.

 Ciao,
 Sheldon.

-- 

John Baldwin [EMAIL PROTECTED] -- http://www.FreeBSD.org/~jhb/
PGP Key: http://www.baldwin.cx/~john/pgpkey.asc
Power Users Use the Power to Serve!  -  http://www.FreeBSD.org/

To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message



Re: Any chance of release patch being committed?

2001-05-04 Thread J Wunsch

Sheldon Hearn [EMAIL PROTECTED] wrote:

 That said, I don't think that the release should pick up KERNCONF.  I
 think a middle-ground option is to have the release pick up
 RELEASE_KERNCONF.

That sounds best to me.

-- 
cheers, Jorg   .-.-.   --... ...--   -.. .  DL8DTL

http://www.sax.de/~joerg/NIC: JW11-RIPE
Never trust an operating system you don't have sources for. ;-)

To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message



Any chance of release patch being committed?

2001-05-03 Thread Terry Lambert

Hello?

I thought I had defended the reasons for my wanting the patch
so that the KERNCONF in /usr/src/release/Makefile actually
worked with sysinstall well enough that the patch should be
committed...

I haven't seen a commit go by... any chance of that change
being committed?

It doesn't really change the resulting code, unless KERNCONF
is set to something other than GENERIC.

Isn't anyone but me and Walnut Creek cum BSDI cum Windriver
Systems using make release?!?


Terry Lambert
[EMAIL PROTECTED]
---
Any opinions in this posting are my own and not those of my present
or previous employers.

To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message



Re: Any chance of release patch being committed?

2001-05-03 Thread J Wunsch

Terry Lambert [EMAIL PROTECTED] wrote:

 Isn't anyone but me and Walnut Creek cum BSDI cum Windriver
 Systems using make release?!?

We are, but why would we use anything else than GENERIC for it,
seriously?  I'd never roll a `release' for my current machine.

-- 
cheers, Jorg   .-.-.   --... ...--   -.. .  DL8DTL

http://www.sax.de/~joerg/NIC: JW11-RIPE
Never trust an operating system you don't have sources for. ;-)

To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message



Re: Any chance of release patch being committed?

2001-05-03 Thread Robert Watson


On Fri, 4 May 2001, J Wunsch wrote:

 Terry Lambert [EMAIL PROTECTED] wrote:
 
  Isn't anyone but me and Walnut Creek cum BSDI cum Windriver
  Systems using make release?!?
 
 We are, but why would we use anything else than GENERIC for it,
 seriously?  I'd never roll a `release' for my current machine. 

Actually, I'd love to see a release rolled using NEWCARD for
current.freebsd.org, so we can install using cardbus notebooks.  This came
up today actually... 

Robert N M Watson FreeBSD Core Team, TrustedBSD Project
[EMAIL PROTECTED]  NAI Labs, Safeport Network Services




To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message