cvs commit: src/sys/boot/i386/libi386 i386_module.c

[John Baldwin]

On 09-May-2003 Paul Richards wrote:
> On Fri, 2003-05-09 at 18:04, John Baldwin wrote:
>> On 09-May-2003 Paul Richards wrote:
>> > On Thu, May 01, 2003 at 10:25:37AM -0700, Peter Wemm wrote:
>> >> John Baldwin wrote:
>> >> > 
>> >> > On 30-Apr-2003 Peter Wemm wrote:
>> >> > > peter       2003/04/30 15:02:39 PDT
>> >> > > 
>> >> > >   FreeBSD src repository
>> >> > > 
>> >> > >   Modified files:
>> >> > >     sys/boot/i386/libi386 i386_module.c 
>> >> > >   Log:
>> >> > >   ACPI will always be present on AMD64 - it will never be an autodetect
>> >> > >   module.
>> >> > 
>> >> > I would require it in the kernel rather than load it as a module then.
>> >> 
>> >> Exactly.  But I was just getting annoyed with this warning during
>> >> development while I had it stubbed out.
>> > 
>> > You might still want it as a module even if it's mandatory so that
>> > you can choose which version to load from the bootloader when your
>> > testing new versions.
>> 
>> That would only work if the kernel linker supported that.  At
>> the moment it doesn't. :)
> 
> I don't follow, I do this all the time?
> 
> or do you mean that having it mandatory would not work if it's also a
> module?

Trying to kldload something that is already present in the kernel
either fails or results in a kernel panic.  If you are already
rebooting to test it, then you might as well build a new kernel
as a new module each time.

-- 

John Baldwin <jhb at FreeBSD.org>  <><  http://www.FreeBSD.org/~jhb/
"Power Users Use the Power to Serve!"  -  http://www.FreeBSD.org/