Date: Wed, 10 Jul 2013 12:15:14 +0300 From: Konstantin Belousov <kostikbel@gmail.com> To: Bruce Evans <brde@optusnet.com.au> Cc: Baptiste Daroussin <bapt@FreeBSD.org>, Adrian Chadd <adrian@FreeBSD.org>, Dimitry Andric <dim@FreeBSD.org>, Andrew Turner <andrew@fubar.geek.nz>, freebsd-arch@FreeBSD.org Subject: Re: Adding a MACHINE_ARCH note Message-ID: <20130710091514.GA91021@kib.kiev.ua> In-Reply-To: <20130710182855.Y1991@besplex.bde.org> References: <20130709090744.0e497e7e@bender.Home> <32F979BD-FB5C-4111-9586-4C5E7C6DFA71@bsdimp.com> <20130709234837.559e3769@bender.Home> <CAJ-Vmo=iV8BsGriFRgNuP-ZJdQhpmBLhjAkz-nSVRS0HPKSyOQ@mail.gmail.com> <CAGE5yCpJmRDvnaYtozj4bCqNoQXH=1e96HPJAqwJuRdn4H9BZA@mail.gmail.com> <CAJ-Vmo=tmGDW3Ubw9nr5rb30bXr1dcJUkKLOU7L=_bx29zvEhw@mail.gmail.com> <CAGE5yCq9gQERDkbi4wu=6tNUap24ZR7sL7aF%2BzmEO0eT6nxPsA@mail.gmail.com> <F79E2F76-A234-499A-ABB7-1ABA62283E9D@FreeBSD.org> <20130710070319.GX91021@kib.kiev.ua> <20130710182855.Y1991@besplex.bde.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--jK4jXyj6WYQn7qqq Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Wed, Jul 10, 2013 at 06:43:22PM +1000, Bruce Evans wrote: >=20 > ISTR reading a claim that pure 64-bit mode was mainly a marketing ploy > by AMD. It distinguished them from Intel and let them get to market > faster. The result was a marketing win but a not so good instruction > set architecture. This was contrasted with sparc64 (Sun?), where > 32-bit applications remained the default and the differences were > smaller (no mode switch?). Sparcv8 did not have an issue with the very limited size of the register file. They did not need to change instruction encoding to allow bigger set of the registers. Amd had to change encoding, and indeed I remember reading somewhere that they decided to not allow the REX prefix for compatibility mode to push the "full" 64bit tagline. Solaris port to amd64 is very similar to sparcv9: almost all usermode binaries are 32 bit, libraries are supplied in both 32 and 64 bit versions, compiler defaults to -m32. Only several binaries are 64 bit, most notable are procfs tools and debugger. There is no reason why such approach could not been taken for the FreeBSD port, except convenience and avoiding the work for properly support all kernel interfaces in compat 32bit and 64bit variants from the beginning. We still have almost no compat32 for the management interfaces, esp. network. cc -m32 works in HEAD. I believe that multiarch for ports is not even thought about, would be very happy to be wrong there. --jK4jXyj6WYQn7qqq Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.20 (FreeBSD) iQIcBAEBAgAGBQJR3SYiAAoJEJDCuSvBvK1B3XcQAIf5OI4DGGAnorMdz4wz/615 oWP8NmldYA834fnAnxtWjvSbDorcJolGKAQ05Q39cPfUFsuiKwN5n+ZZF8gveG3w ltWNYN5J9o09K/FDuQOK9yXV1VsQArtfW05cgxvdt8Gu+q6FbsL+LeSvULxJ6jL+ HoI90UO3fjw4GD+SLAaiDwAYOyaoX13gQiRUqXDqFQd0Y3E3JYk7lkx0WzOupgCy cBXPS+w6QUVWfL9ynNDIwcwWrt0judQF56s2sp3ilYj9iENDysTttPhUnP+JJit/ SztVS945Wh6yvjYhTdTgcd6Ckx1OpCqxd8j5WEGpgPGkXe9C+MpR/KkL0F0SZQ8G HQatqHr5+m17s/MxBrIXPdytBA49j7TnN2kVkYmOBz2kvxUn/25fu0Y33j5zc7Vk 2XnHja9/tkjtSwve7il3urgXtx3A1NtMZtW0sYcE5BnAD6vSUB3RB4qHynxjbbB/ 1TK1Dhtxm31/xO4UxM+gBr0epYm8xf1IEURxGUxP7klx4Q2rkegWih9GGRf2Spkj F9LkGcUOojuKTWR7p6ok6qmtBdee6jUk8F3mNWz0DMdow0x+AennEAx8S2oesyD+ VCo3KDaenAHZebgcpXzgpE51zSxz1hzFsOBHzHOM72QFMJABQgxpgUoUqVq9WKrU mrt82sgru2mv42v6Hkfz =jBHK -----END PGP SIGNATURE----- --jK4jXyj6WYQn7qqq--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20130710091514.GA91021>