Date: Wed, 22 Apr 2009 23:03:04 +0100 From: Ceri Davies <ceri@submonkey.net> To: John Baldwin <jhb@freebsd.org> Cc: svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org Subject: Re: svn commit: r191405 - in head/sys: amd64/amd64 i386/i386 Message-ID: <20090422220304.GB54875@submonkey.net> In-Reply-To: <200904221759.04446.jhb@freebsd.org> References: <200904222140.n3MLebn3068260@svn.freebsd.org> <200904221759.04446.jhb@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--G4iJoqBmSsgzjUCe Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Wed, Apr 22, 2009 at 05:59:04PM -0400, John Baldwin wrote: > On Wednesday 22 April 2009 5:40:37 pm John Baldwin wrote: > > Author: jhb > > Date: Wed Apr 22 21:40:37 2009 > > New Revision: 191405 > > URL: http://svn.freebsd.org/changeset/base/191405 > >=20 > > Log: > > Adjust the way we number CPUs on x86 so that we attempt to "group" all > > logical CPUs in a package. We do this by numbering the non-boot CPUs > > by starting with the first CPU whose APIC ID is after the boot CPU and > > wrapping back around to APIC ID 0 if needed rather than always starti= ng > > at APIC ID 0. While here, adjust the cpu_mp_announce() routine to li= st > > CPUs based on the mapping established by assign_cpu_ids() rather than > > making assumptions about the algorithm assign_cpu_ids() uses. >=20 > An example is probably in order for this to make sense. Suppose you have= a=20 > system with two quad-core CPUs. Package 0 has CPUs numbered 0, 1, 2, and= 3. =20 > Package 1 has CPUs numbered 4, 5, 6, and 7. With the old code, if packag= e 0=20 > won the election to be the boot processor, then CPU 0 would be the BSP an= d=20 > the logical IDs would match the APIC IDs. However, if package 1 won the= =20 > election during POST, then CPU 0 would be APIC ID 4 on package 0 followed= by=20 > CPU 1 being APIC ID 0, CPU 2 being APIC ID 1, etc. Thus, when CPU 0 was = the=20 > boot CPU you had a nice grouping where CPUs 0-3 were a single package and= =20 > CPUs 4-7 were another package. However, when CPU 4 was the boot CPU, CPU= s 0=20 > and 5-7 where one package, and CPUs 1-4 where the second package. The ef= fect=20 > of this patch is to change the case when CPU 4 is the boot CPU such that = CPUs=20 > 0-3 are now all from CPU 4's package (APIC IDs 4-7), and CPUs 4-7 are fro= m=20 > the other package (APIC IDs 0-3). What this means, in turn, is that in b= oth=20 > cases you now always have CPUs 0-3 as one package and CPUs 4-7 as another= =20 > package regardless of which CPU wins the boot-time election. What if I have HT turned on? Do they bunch up with the "real" CPU IDs or all together at the end? Ceri --=20 That must be wonderful! I don't understand it at all. -- Moliere --G4iJoqBmSsgzjUCe Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (FreeBSD) iD8DBQFJ75QYocfcwTS3JF8RAj4hAKC6hCKSn53TlPljqDR/w/PAPxRyxACdE7g8 woYCJRMhkvG1mVEJ2tMfF7Q= =YGP1 -----END PGP SIGNATURE----- --G4iJoqBmSsgzjUCe--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20090422220304.GB54875>