From mboxrd@z Thu Jan 1 00:00:00 1970 Date: Sat, 27 May 2000 19:15:58 -0800 From: Ethan Benson To: "Michael A. Peters" Cc: linuxppc-dev@lists.linuxppc.org Subject: Re: ybin and kernel level Message-ID: <20000527191558.C29071@plato.local.lan> References: <392C4F17.61D04687@mac.com> <20000524225002.O383@plato.local.lan> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="uBaQ+BsiqpBBIRFI" In-Reply-To: ; from mpeters@mac.com on Sat, May 27, 2000 at 10:33:04AM -0700 Sender: owner-linuxppc-dev@lists.linuxppc.org List-Id: --uBaQ+BsiqpBBIRFI Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: quoted-printable On Sat, May 27, 2000 at 10:33:04AM -0700, Michael A. Peters wrote: > I saw it noted in the preliminary yBin/yaBoot documentation that a=20 > newer kernel was recommended, suggesting a 2.2.15pre3 or later. >=20 > Is there anything specific about the 2.2.14 that won't allow it to work? well it will work if you apply patches by BenH, there are changed to prom.c and such to make it compatible with yaboot. =20 > I'm using a 2.2.14 kernel built from standard source (ftp.kernel.org)=20 > with the 2.3.50 usb backport. you would need BenH patches (no longer avaiable AFAIK) > I tried that backport with a 2.2.15 kernel, and it boots (BootX), but=20 > when booting the typical USB info doesn't scroll by the screen and no=20 > usb devices are available (for example, upon starting X, error 19,=20 > can't open usb mouse device) 2.2.15 Linus is pretty close, but there is still alot of PPC stuff that needs patching, i still reccommend the Paul tree for newworlds. i have a final (no pre) 2.2.15 with paul's patches, in case he has already thrown away 2.2.15 in favor of unstable 2.2.16pres, if that happens (like it has for the last 4 kernels i have kept track of) i will make a unified diff to the linus tree avialable. > Part of me thinks that may be BootX related, I had to upgrade BootX=20 > when switching to my current kernel due to usb weirdness, but I don't=20 > know for sure. Bootx does all kinds of strange things on newworlds anymore. especially on OS9. (it gets worse the newwer the hardware it would appear)=20 > I know the FAQ is up, but its not quite finished yet- (but what is=20 > done is very good so far) compliments to Eric for the great work on the FAQ. i think he is working on partitioning help now. =20 > The way I made a boostrap partition was to first create apple hfs=20 > partition. I then wrote down the start block and length in blocks,=20 > and (using pdisk) deleted the partition, using a C {create by=20 > specified type), entering Apple_Bootstrap when it asked for file=20 > system type. >=20 > Is that correct? yup that should work fine, i typically just create a single HFS partition that spans all the space i will need for all the linux partitions and delete it in pdisk/mac-fdisk creating all the linux partitions i need. the main problem i see with using apples fdisk for only the bootstrap partition is i don't think it lets your create it 800K, anything larger then that is just a huge waste of space (even with 800K partition 740 of it is wasted) > *** >=20 > The drive my bootstrap partition will be located on is internal scsi,=20 > attached to an adaptec 2930u controller. do you have macos on an internal IDE? or is there even an internal IDE at all? if not default OF settings will probably work but it might be slow. if that is the case you will have to set the boot-device yourself unfortunatly. (is anyone still working on reenabling /dev/nvram on newworld?) > I'm fairly sure I can find the full path to the device in the open=20 > firmware, but interestingly enough- Apple System Profiler does NOT=20 > see the scsi card or give any info on it. interesting, yes you can find it in OF, just do: dev / ls and walk the tree like so: dev /pci@XXXXXXXX ls until you find it, it sucks but thats the price you pay for scsi i suppose (after the dent in the wallet scsi makes ;-)) > It does not appear in the pci devices, although volumes attached to=20 > the card ARE seen by System Profiler, and seen on bus 1 (opposed to=20 > the built in bus 0) are you sure it has an OF ROM? for it to be bootable it must have an OF bootrom on the card, otherwise it probably just uses a MacOS extension to make it visable to macos. linux could do the same but it would be impossible to boot from it. (from either OS since OF must be able to read it) --=20 Ethan Benson http://www.alaska.net/~erbenson/ --uBaQ+BsiqpBBIRFI Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.1 (GNU/Linux) Comment: For info see http://www.gnupg.org iEYEARECAAYFAjkwj24ACgkQJKx7GixEevz6CwCdFzDvsPztcBA8XEDXwNOdgLdI haYAn0+p0pkSEnHOat8yJsi8y3psCb1t =xz5+ -----END PGP SIGNATURE----- --uBaQ+BsiqpBBIRFI-- ** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/