Date: Fri, 27 Jul 2007 09:48:40 -0400 From: Kris Kennaway <kris@obsecurity.org> To: Attilio Rao <attilio@FreeBSD.org> Cc: Milos Vyletel <mv@rulez.sk>, current@freebsd.org, Peter Wemm <peter@wemm.org> Subject: Re: ULE status, invalid load, buildkernel times. Message-ID: <20070727134840.GA7305@rot26.obsecurity.org> In-Reply-To: <46A9C437.1080504@FreeBSD.org> References: <20070721174631.S561@10.0.0.1> <20070722114846.GA97996@rulez.sk> <20070722121631.GA8336@rulez.sk> <200707261750.19994.peter@wemm.org> <46A9C437.1080504@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--opJtzjQTFsWo+cga Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Fri, Jul 27, 2007 at 12:08:55PM +0200, Attilio Rao wrote: > The patch is not going to work as the slot for SI_ORDER_SECOND was=20 > alredy held by kern/subr_smp.c::mp_start() function. >=20 > Could you please try this comprehensive patch? > It has a fix for the mp_start / lapic_init confusion and some tricks=20 > with CNXT-ID bit which should exactly identify HTT: > http://people.freebsd.org/~attilio/machdep_lapic.diff >=20 > I have still to stress test it, so please use caution. I will update=20 > soon if I will have more information (a 'work for me' / 'don't work for= =20 > me' would be very appreciated though). This works for me on three systems: A real HTT system (kern.sched.topology is correctly set to 1) A dual core amd64 system (previously mis-probed as HTT, now correctly not) A quad core intel system (not previously mis-probed as HTT, but broken with Peter's patch). Kris --opJtzjQTFsWo+cga Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4 (FreeBSD) iD8DBQFGqfe4Wry0BWjoQKURAj+dAKC87LSt6to+/M4H0LNDemoXLL8J5gCg45me 2KRL/XEi7aKNdFWmwmVBK3c= =DPuL -----END PGP SIGNATURE----- --opJtzjQTFsWo+cga--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20070727134840.GA7305>