Date: Sun, 09 Apr 2006 15:01:54 -0700 From: Maxim Sobolev <sobomax@FreeBSD.org> To: Bruce M Simpson <bms@spc.org>, Luigi Rizzo <rizzo@icir.org>, Mark Sergeant <msergeant@snsonline.net>, freebsd-current@FreeBSD.org Subject: Re: [DOCUMENT] Re: Intel Macs that boot FreeBSD? Message-ID: <44398452.6030007@FreeBSD.org> In-Reply-To: <20060408105014.GV80492@spc.org> References: <4433CFA1.90509@centtech.com> <2C74BB8F-271B-4505-9D94-B270B3A4ACBA@nordahl.net> <44348603.9070503@FreeBSD.org> <20060406190944.G56354@lexi.siliconlandmark.com> <4435B303.5040205@FreeBSD.org> <196C209A-A585-4AD5-B7C9-A27DD57ECD8C@snsonline.net> <20060408103224.GR80492@spc.org> <20060408034357.A44217@xorpc.icir.org> <20060408105014.GV80492@spc.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Some progress: further investigation revealed that cdboot has been hanging when trying to enable A20 line, apparently Mac simply don't have relevant legacy hardware. After commenting out appropriate routine call in cdboot I was able to get loader working just fine. The loader works normally, but unfortunately attempt to start kernel hangs up machine. I suspect that either Mac doesn't have real text mode or kernel also waits for some nonexisting legacy hardware indefinitely. -Maxim Bruce M Simpson wrote: > On Sat, Apr 08, 2006 at 03:43:58AM -0700, Luigi Rizzo wrote: >> Ii wonder if we can do more progress if we use >> a picobsd-like approach namely bypass btx and load the kernel >> directly. Reading your list we may well get stuck at the pci >> probe, but at least we should see the initial kernel messages. > > I would suggest that anyone who wants to get FreeBSD booting on the > mini-mac purchases a port 0x80 POST diag card (<$10 on eBay), and > instruments the early boot paths to output to such a device. > > If hacking kernel code then probably quicker to just use NFS and > PXE for starter bootstrap :-) > > Of course those of us who are lucky enough to have a PCI logic analyzer > (not me, though I bought an FPGA card explicitly for this purpose, > ENOTIME to learn hardware design language) will be able to do this > right away. > > It shouldn't be too difficult to implement PCI Configuration Register > Type 1 support. Ideally of course we'd support EFI natively w/o the shim. > > Regards, > BMS > >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?44398452.6030007>