Date: Sun, 27 Aug 2017 08:51:09 -0700 From: Conrad Meyer <cem@freebsd.org> To: Ed Schouten <ed@nuxi.nl> Cc: src-committers <src-committers@freebsd.org>, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: Re: svn commit: r322942 - head/sys/kern Message-ID: <CAG6CVpWG9vS2wJsscRr2dOH=%2B_6HpXVRRqN1mrZffVfzbWsROw@mail.gmail.com> In-Reply-To: <CABh_MKknHyosfENq8SPzSOAhGFKL_tALfNdn5=eJqa29yjFqSA@mail.gmail.com> References: <201708270514.v7R5EmDK064785@repo.freebsd.org> <CABh_MKknHyosfENq8SPzSOAhGFKL_tALfNdn5=eJqa29yjFqSA@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, Aug 27, 2017 at 1:18 AM, Ed Schouten <ed@nuxi.nl> wrote:
> Hi Conrad,
>
> [ Context: I don't understand anything about CPU topologies! ]
>
> 2017-08-27 7:14 GMT+02:00 Conrad Meyer <cem@freebsd.org>:
>> + while (top->cg_children == 1) {
>> + top = &top->cg_child[0];
>> + top->cg_parent = NULL;
>> + }
>
> This only removes layers from the top if they only have a single
> child. Question: is it possible/likely to have CPU topologies where
> one of the middle layers only has one child? If so, would it make
> sense to use tree traversal here to do the flattening?
Hi Ed,
It is probably possible for middle layers to have a single child, and
your proposal is a good one. This is just the canned patch from Jeff
I already had lying around :-).
Best,
Conrad
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAG6CVpWG9vS2wJsscRr2dOH=%2B_6HpXVRRqN1mrZffVfzbWsROw>