Date: Tue, 17 Mar 2015 09:51:50 -0600 From: Ian Lepore <ian@freebsd.org> To: Warner Losh <imp@bsdimp.com> Cc: Tim Kientzle <tim@kientzle.com>, freebsd-arm <freebsd-arm@freebsd.org> Subject: Re: How to use u-boot-beaglebone port? Message-ID: <1426607510.25614.3.camel@freebsd.org> In-Reply-To: <DC2A0B00-B528-45D8-862B-1F955818D6C2@bsdimp.com> References: <17B779D7-2962-4455-9062-51411F316648@gromit.dlib.vt.edu> <986F5E5D-C784-4BEF-81E3-49A9F27C0E8F@kientzle.com> <1426534773.95554.15.camel@freebsd.org> <DC2A0B00-B528-45D8-862B-1F955818D6C2@bsdimp.com>
index | next in thread | previous in thread | raw e-mail
On Tue, 2015-03-17 at 08:16 +0900, Warner Losh wrote: > > On Mar 17, 2015, at 4:39 AM, Ian Lepore <ian@FreeBSD.org> wrote: > > > > [*] Because armv6 != armv7 in this case. While armv6 is synonymous with > > armv7 for most purposes in freebsd, the rpi is the exception to that in > > that it really IS armv6, and that leads to the kernel being built with > > different cache maintenance routines that don't work on armv7. > > I thought all that was handled through a separate vector so we could, > in theory, do that. All of it is conditional, but that vectors back via a table. > It¢s been like that on arm since day one, though maybe that got broken > somewhere. > > Warner > Not broken, but rather "undone, by design." We now pass the proper arch on the command line so that v7 stuff gets built using v7 opcodes and with the right ACLE stuff defined. (This is actually all still in progress, some stuff still vectors through the old cpufuncs table.) We support RPi but we no longer let it cripple the newer boards/socs in pursuit of a mythical unified kernel that's never going to happen (at least, it's not going to happen in the form of v6+v7 in the same kernel). -- Ianhelp
Want to link to this message? Use this
URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1426607510.25614.3.camel>