Date: Mon, 16 Mar 2015 13:04:21 +0800 From: Jia-Shiun Li <jiashiun@gmail.com> To: "freebsd-arm@freebsd.org" <freebsd-arm@freebsd.org> Subject: Re: [Differential] [Accepted] D2018: Introduce Annapurna Labs AHCI support Message-ID: <CAHNYxxPaMua8qCmJa-VQv5GxZOiV4EMynf47ky0WE0qy2enNJw@mail.gmail.com> In-Reply-To: <65260d32d448683b765c81e2761126e7@localhost.localdomain> References: <differential-rev-PHID-DREV-lp4daqpuzz2y5vyrxipd-req@FreeBSD.org> <65260d32d448683b765c81e2761126e7@localhost.localdomain>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, Mar 14, 2015 at 9:29 AM, imp (Warner Losh) < phabric-noreply@freebsd.org> wrote: > imp accepted this revision. > imp added a comment. > This revision is now accepted and ready to land. > > Thanks for the comments. Still not sure what an Annapurna is.... > > REVISION DETAIL > https://reviews.freebsd.org/D2018 > > To: zbb, mav, smh, imp > Cc: freebsd-arm > Probably SoC from Annapurna Labs. https://gigaom.com/2015/01/22/confirmed-amazon-is-buying-annapurna-labs/ The init part probably should have been done in boot loader when initing hardware. The AHCI standard was designed in the way that system designers have room to adjust. E.g. use PI to indicate only two ports were connected on a notebook when silicon supports 6, so a generic driver can detect it and not to assume it can use all 6 ports. Not sure about Annapurna's implementation. Intel's ICH/PCH implement these read-only registers as write once registers after powered on, so BIOS will always be the only one modifying, and drivers see them as read-only. It is also possible that implementations choose to init by values in external ROM/flash, internal fuse, or strapping pins. It will need docs from Annapurna, spec or errata, to see if it disagree with AHCI in any ways, and if there is better alternative to quirks. -Jia-Shiun.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAHNYxxPaMua8qCmJa-VQv5GxZOiV4EMynf47ky0WE0qy2enNJw>