Driver Update Disk discussion

[Daniel O'Connor]

On Fri, 25 Feb 2005 14:14, M. Warner Losh wrote:
> : I think PORTS_MODULES is a little suboptimal..
>
> How so?

If you upgrade your ports tree and then rebuild your kernel you may upgrade a 
port KLD without wanting to.

> : Fx5200 Go). It also means that the kernel build/install does not result
> : in non kernel things being altered.
>
> How do your patches work?

http://www.dons.net.au/~darius/port-kld.diff
http://www.dons.net.au/~darius/port-makefile.txt

(Rename the later to Makefile and put it in /usr/local/kld)

It hooks into /usr/src/sys/modules/Makfile and looks for module directories to 
build in a certain directory. It [is supposed to] acts like another 
subdirectory of /usr/src/sys/modules.

> Do they work with multiple kernel trees? 

I am pretty sure it does but I haven't explicitly tested it - certainly the 
source directory doesn't get dirtied by builds.

> I guess I'm a little unclear why this is better or worse than
> PORTS_MODULES.  I guess I'm missing the explicit step, since I only
> ever update the parts of my ports tree that I'm upgrading with
> portupgrade.

Ahh.. I NFS mount my ports tree between a bunch of machines they all have 
WRKDIRPREFIX, and the ports tree is updated each day using cvsup.

I don't think this is an uncommon setup but I don't have any evidence.

> : Also speaking of KLD ports.. I really wish they wouldn't install
> : into /boot/modules (I patch so they don't) as it is a really good way to
> : shoot yourself in the foot during an upgrade :(
>
> Usually this is only a problem when tracking or jumping to current,
> but I understand...

Yeah usually, but I have had it happen in -stable too (very rare). Even so it 
can be a pretty big waste of time as a developer when you're tracking 
current ;)

-- 
Daniel O'Connor software and network engineer
for Genesis Software - http://www.gsoft.com.au
"The nice thing about standards is that there
are so many of them to choose from."
  -- Andrew Tanenbaum
GPG Fingerprint - 5596 B766 97C0 0E94 4347 295E E593 DC20 7B3F CE8C
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-hackers/attachments/20050225/9f68ac6b/attachment.bin