Kill old, non-INTRNG code on ARM

[Warner Losh]

On Sat, Aug 11, 2018 at 11:40 AM, Warner Losh <imp at bsdimp.com> wrote:

>
>
> On Sat, Aug 11, 2018, 11:11 AM Ian Lepore <ian at freebsd.org> wrote:
>
>> On Sat, 2018-08-11 at 08:52 -0600, Warner Losh wrote:
>> > One last thing, even though we're in a slush, is to kill the pre-
>> > INTRNG
>> > code.
>> >
>> > It's required for armv6 and v7 for some time (only the Amlogic 8726
>> > boards
>> > remain broken there). And we're down to two families of chips which
>> > support
>> > armv5: rt1310 and the Marvel gear. Both of them support INTRNG.
>> >
>> > I propose we eliminate the INTRNG option and unifdef the code to
>> > remove the
>> > undefined case.
>> >
>> > Comments?
>> >
>>
>> Wait a sec, the INTRNG option can't go away, it's used by multiple
>> arches these days. What we can do is make it on-by-default for arm*
>> arches, and eliminate any #if[n]def INTRNG in arm-specific code.
>>
>
> Yes. That's a better way of saying what I had in mind. GC the code that
> will never compile on arm so we don't have to support it on 12. Or rather
> trip over it...
>

OK. I misread things. Marvell still isn't converted to INTRNG.

I have a more modest patch that moves things around a little in the config
files to help...

Warner