I have submitted https://reviews.freebsd.org/D23376 for avoiding the PowerMac threads-stuck-sleeping issue

[Mark Millard]

On 2020-Jan-27, at 15:36, Dennis Clarke via freebsd-ppc <freebsd-ppc at freebsd.org> wrote:

> On 2020-01-27 14:54, Francis Little wrote:
>> Well that seems to do the trick!
>> PowerMac G5 Quad, running the r357172 kernel, plus this patch.
>> All fans normal, no odd hanging threads, compiling ports etc with all 4
>> cores, seems stable.
>> Will continue testing.
> 
> Really?  Working well?  I know that you and I have very similar hardware
> and yet there are differences also. I will give it a try.
> 
> Did you use a snapshot or compile from source ?

He built a kernel that was based on applying a
source code patch that I put out for review, see
https://reviews.freebsd.org/D23376 . There is no
snaphot. None of the code changes are part of
FreeBSD at this point.

Even if the techniques involved are all deemed okay
as they are, the source code is going to need a round
of conversion to style(9) criteria to be acceptable
for check-in.

(I've seen references to style(9) but am not familiar
with the content and did not add that to my burden
back when I was trying to figure out how to avoid
this threads-stuck-sleeping issue and, before that,
investigating avoiding the usefdt mode failures, both
being PowerMac-context problems. I'd be amazed if how
I coded happened to fit an unknown-to-me coding
standard.)

I've been running with the changes since 2019-May or
so, building and using a few versions of FreeBSD
since then. My current context is head -r356426 based
(so older than Francis Little's head -r357172 context).

===
Mark Millard
marklmi at yahoo.com
( dsl-only.net went
away in early 2018-Mar)