Re: What's the plan for powerpc64 in FreeBSD 16

----- Original Message -----
> From: "Joseph Mingrone" <[email protected]>
> To: "Minsoo Choo" <[email protected]>
> Cc: "[email protected]" <[email protected]>, "Lance Albertson via RT" <[email protected]>
> Sent: Saturday, November 22, 2025 2:56:35 PM
> Subject: Re: What's the plan for powerpc64 in FreeBSD 16

> On Mon, 2025-11-17 at 18:24, Minsoo Choo <[email protected]> wrote:
> 
>> Resending this message since I forgot to "replay all" :(
> 
>> On Monday, November 17th, 2025 at 1:13 PM, Tomek CEDRO <[email protected]> wrote:
> 
>>> On Mon, Nov 17, 2025 at 5:58 PM Warner Losh [email protected] wrote:
> 
>>> I guess the same as for ARMv7? Do not delete the code, so someone can
>>> still use/work with it when needed, but mark "unsupported" or "tier 3"
>>> or similar, then focus on new hardware platforms that lags behind
>>> already?
> 
>>> What does "retire" exactly mean? I hope its just lack of official
>>> support but not the source code removal right? :-P
> 
>>> --
>>> CeDeROM, SQ7MHZ, http://www.tomek.cedro.info
> 
>> The problem is not just letting the code to sit in the source tree. Assume
>> someone is making a change that needs modification to all architecture (maybe
>> build system or very low-level stuff). How do they know that their code work on
>> real hardware if there is no hardware or developers to work on a specific
>> platform (in this case, big-endian powerpc)? Assume this patch is crucial for
>> FreeBSD's evolution. Does powerpcbe worth enough to stall the review process
>> due to testing/verification issue? If is it "unsupported" there is no reason to
>> keep the code in the upstream. Someone else can fork the repository and
>> maintain their own powerpcbe port. As long as it sits in the source tree, there
>> is minimum effort needed (such as @releng) to check if the code works.

While I agree, I do want to ensure everyone understands that powerpc64be != powerpc64le; it's very easy to inadvertently lump them together for this sort of discussion.  I expect over time we will end up in a state where powerpc64be is broken and powerpc64le is functioning normally; the latter should block, the former should not in my opinion.

> A group at the University of Oregon has some POWER systems available to us.  If
> you are interested, please contact Lance Albertson, who is copied here, as soon
> as possible.  He's been asking whether anyone is using the hardware and says we
> risk losing access if no one shows interest.

For all those folks that would like to help and don't have hardware access, this is a great alternative to the Raptor bare metal boxes.  Especially for items like application development (e.g. furthering the Go port Raptor put together) these machines are a great resource -- I know Lance as well, so if there's any issues getting access set up just let me know.

Thanks!