Date: Thu, 6 Feb 2014 13:56:54 -0700 From: Warner Losh <imp@bsdimp.com> To: Kamil Choudhury <Kamil.Choudhury@anserinae.net> Cc: "freebsd-hackers@freebsd.org" <freebsd-hackers@freebsd.org>, George Mitchell <george+freebsd@m5p.com> Subject: Re: opteron a1100 arm Message-ID: <DD4A24EA-9EB6-4A66-8E84-A7D6959514E0@bsdimp.com> In-Reply-To: <F9A7386EC2A26E4293AF13FABCCB32B3015127368B@janus.anserinae.net> References: <alpine.BSF.2.00.1401311911120.2427@wojtek.tensor.gdynia.pl> <1391538649.19169.79261269.3C5F49D1@webmail.messagingengine.com> <CAFU734xXWyc_TqBJ7e4MhD2nB01BAejR_1vT9%2B_5Ar5mJncncA@mail.gmail.com> <493DEB39-C4B4-409E-B8B2-B1B11E013754@netgate.com> <60555.1391549390@critter.freebsd.dk> <23B18B88-D888-46B3-99F6-905F86E20FAF@netgate.com> <lcsqov$ui9$1@ger.gmane.org> <DCE7B22A-D51F-471A-B446-4B76D87E775D@netgate.com>, <52F2461B.1070405@m5p.com> <F9A7386EC2A26E4293AF13FABCCB32B3015127368B@janus.anserinae.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Feb 5, 2014, at 7:34 AM, Kamil Choudhury wrote: >=20 > George wrote:=20 >> This discussion is very entertaining. But let's not lose sight of = the >> main point: ARM processors are not going away. It is to FreeBSD's >> detriment if we don't have Tier 1 support for a processor that, by = some >> guesses, already comprises the majority of new installations by = count. >=20 > Moving on to specifics, what concrete steps would be necessary to = actually=20 > get FreeBSD support for this architecture?=20 You'll need a new pmap for arm, and a bunch of other things. Plus you'll = need to get ACPI working on arm64, and you may also need FDT + ACPI = coexistence depending on your development route. Plus you'll need arm64 = (aarch64) support in llvm and/or some external toolchain (which would = mean you'd need to beef up the external toolchain support for freebsd's = build). And then you'll need new drivers. And people to test it and = hammer it and fix the stability issues that pop up. And that's not even = considering the issues with pkg support etc. It is a heavy lift to get all the way there, although a 'hack' port that = gets to single user with a ram-disk or nfs root and a serial port = working wouldn't be too hard. So it's like Ian said: we need bodies working on this... So far, I'm = only aware of one part timer. Warner
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?DD4A24EA-9EB6-4A66-8E84-A7D6959514E0>