svn commit: r361352 - in head/sys: amd64/amd64 i386/i386

[John Baldwin]

On 5/22/20 8:37 AM, Mark Johnston wrote:
> On Fri, May 22, 2020 at 06:10:59PM +0300, Konstantin Belousov wrote:
>> On Fri, May 22, 2020 at 01:18:55AM +0000, Mark Johnston wrote:
>>> Author: markj
>>> Date: Fri May 22 01:18:55 2020
>>> New Revision: 361352
>>> URL: https://svnweb.freebsd.org/changeset/base/361352
>>>
>>> Log:
>>>   Fix the build after r361033 when ACPI is disabled.
>> What is the sense in doing this for amd64 ?  I doubt that we can boot
>> with ACPI disabled.
> 
> It is possible to boot in virtualized environments without ACPI.  I'm
> not sure why it is especially useful, but I wanted to avoid regressing a
> stable branch since at least one person noticed the breakage fairly
> quickly, and it is easy to fix.

I agree it mostly seems dubious.  I believe at one point that NetApp
at least was using FreeBSD/amd64 on gear that didn't have a "real" BIOS
and didn't have ACPI.  That was probably 10 years ago though when I last
heard that, so it may no longer be true.  That said, even the !ACPI code
we have now is a runtime check so it should still fallback if ACPI isn't
detected at runtime or is disabled via a hint.  I'm not sure how much we
should go out of our way to permit compiling without ACPI.  On any modern
consumer hardware you are going to want ACPI.

-- 
John Baldwin