Date: Sat, 19 Mar 2016 20:00:34 -0500 From: Justin Hibbits <chmeeedalf@gmail.com> To: Thomas Rix <trix@juniper.net> Cc: "freebsd-ppc@freebsd.org" <freebsd-ppc@freebsd.org> Subject: Re: Book-E SMP Message-ID: <20160319200034.104e126f@zhabar.knownspace> In-Reply-To: <D312FBCA.19AB0%trix@juniper.net> References: <20160319121137.61abbfee@zhabar.knownspace> <D312EED0.19AA7%trix@juniper.net> <20160319135811.489c45b1@zhabar.knownspace> <D312FBCA.19AB0%trix@juniper.net>
index | next in thread | previous in thread | raw e-mail
You'll need a netboot setup for starters, just a kernel will do. Then:
* It looks like initial SMP support was introduced in 2009, r192532
(FreeBSD8 era, IIRC). If you can test and start the bisect there,
that would be ideal, though I'm not sure if it would build on a
recent head (might need to setup a VM for that).
* If you see an 'AP launched!' or similar message, everything is fine.
It will panic if it fails, or at least print "SMP: CPU * didn't wake
up.".
* If you see "SMP: CPU * already out of hold-off state!" you'll
need to do something else, but I'm not sure (post if you see it, we
don't handle that case at all yet).
* Realistically, you should just need to bisect over the revisions of
sys/powerpc/booke/locore.S and sys/powerpc/booke/pmap.c, which knocks
the revision search space from 100k to a couple dozen to get the
ballpark beginning.
- Justin
On Sat, 19 Mar 2016 19:53:18 +0000
Thomas Rix <trix@juniper.net> wrote:
> Ok, send me the instructions on the baseline setup and the test(s)
> you are using to do the bisect.
> T
> ---
> Tom Rix
> Sr. Staff Compiler Engineer
> trix@juniper.net
>
>
>
>
>
> On 3/19/16, 11:58 AM, "Justin Hibbits" <chmeeedalf@gmail.com> wrote:
>
> >That would be perfect.
> >
> >- Justin
> >
> >On Sat, 19 Mar 2016 18:56:44 +0000
> >Thomas Rix <trix@juniper.net> wrote:
> >
> >> I have p2020, is this good enough to do a) ?
> >> T
> >>
> >> ---
> >> Tom Rix
> >> Sr. Staff Compiler Engineer
> >> trix@juniper.net
> >>
> >>
> >>
> >>
> >>
> >> On 3/19/16, 10:11 AM, "owner-freebsd-ppc@freebsd.org on behalf of
> >> Justin Hibbits" <owner-freebsd-ppc@freebsd.org on behalf of
> >> chmeeedalf@gmail.com> wrote:
> >>
> >> >Currently SMP is broken on Book-E hardware, and I'm uncertain as
> >> >to why. Has anyone had SMP working with stock FreeBSD on Booke-E
> >> >hardware in the past? The only SMP book-E hardware I own requires
> >> >code support that post-dates anything which may have broken it.
> >> >Would anyone be willing to either:
> >> >
> >> >a) Test the introduction of SMP, and bisect from there to now, on
> >> >e500v2 hardware, or
> >> >b) loan me hardware to do this?
> >> >
> >> >I'm looking for dual-core e500v2, as that was the only hardware
> >> >FreeBSD booted on at the time SMP was introduced.
> >> >
> >> >Thanks,
> >> >Justin
> >> >_______________________________________________
> >> >freebsd-ppc@freebsd.org mailing list
> >> >https://lists.freebsd.org/mailman/listinfo/freebsd-ppc
> >> >To unsubscribe, send any mail to
> >> >"freebsd-ppc-unsubscribe@freebsd.org"
> >>
> >
>
help
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20160319200034.104e126f>