Re: RFC: moving sysutils/fusefs-kmod to base system

From: Daniel O'Connor <doconnor_at_gsoft.com.au>
Date: Tue, 2 Sep 2008 19:12:30 +0930
On Tue, 2 Sep 2008, Volodymyr Kostyrko wrote:
> Aryeh M. Friedman wrote:
> > Unless I understand how the kernel does stuff there is no penalty
> > for having unused modules (except the size of the kernel that needs
> > to be loaded).   Keeping in mind that unless I am not reading stuff
> > corectly fusefs-kmod is the only FS related module that is not in
> > the base system.   Since any fundamental changes in the generic FS
> > API seems to break fusefs-kmod, and cause some very nasty effects
> > that are almost impossible to trace to fusefs-kmod (machine freezes
> > so no output or core dump)  it seems to make sense to move it to 
> > the base system  (after all we already do this with third party FS
> > code like x/zfs)  by moving it we force it to always compile
> > instead of breaking
>
> This can be done by documenting usage of make.conf PORTS_MODULES
> knob. Just a little notice in ports would suffice, not anybody out
> there compiles a new kernel daily.

<soapbox>
It would be nice if ports could put their kernel module source somewhere 
so that a buildkernel would build it.

This has several advantages
- You don't upgrade the port unless you want to when building a kernel.
- If the kernel API changes you find out because the port doesn't 
  compile then you can make an informed decision.
- You don't need a working network connection to rebuild your kernel.

</soapbox>

I did make some strawman patches for this but my make fu is weak so it 
wasn't very reliable :(

-- 
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

Received on Tue Sep 02 2008 - 07:42:44 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:34 UTC