[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: reading linux from MacOs



On Wed, Jul 04, 2001 at 11:40:16PM +0200, Michel Lanners wrote:

> - HFS or HFS+ partitions: native under MacOS; directing mounting under
>   Linux only possible for HFS (but _dangerous_!! for writing); useable
>   with hfsutils or hfsplus;

well if you call hfsutils usable... outside of shell scripting its a
really painful set of utilities to deal with.

> - ext2 partitions: native under Linux; mountable with MountX under MacOS
>   (no idea however about useability, stability and compatibility);

when i last tried (and MountX has not changed since then) it was quite
unstable, and readonly (as it should be, read-write is too dangerous
IMO). 

keep in mind that using this to some degree opens you to security
problems, you lose all file permission protection, thus if this became
popular you might just find one of those nifty OTDs [1] that grabs
/etc/shadow and mails it somewhere.  

MacOS is a No Security OS.

> - I could think about an MSDOS-formatted exchange partition; although
>   you'd standardize on the worst standard ;-). Also, I don't know
>   whether it will actually work; never tried. Should be mountable under
>   both MacOS and Linux.

this will only work if you use x86 BIOS partition tables instead of
Apple partition tables.  MacOS does not seem to have any partition
type registered for DOS filesystems so it simply ignores any DOS
partitions (if you make them Apple_HFS macos tries to read them as
hfs, fails and helpfully offers to erase it). 

if you really wanted to do that it would be possible, just use a dos
formatted bootstrap partition (type 0x41 PReP Boot) adding
fstype=msdos to /etc/yaboot.conf, you also have to compile your own
copy of yaboot with the config file name changed to yaboot.cnf instead
of yaboot.conf (/etc/yaboot.conf remains, since thats used by ybin).
ybin is also able to configure OpenFirmware to understand this strange
setup.  

note though that this is not all that well tested, i did experiment
with it a long time ago. ybin has supported this configuraton forever.  

> - Fileserver partitions, like AFP (Appleshare or netatalk), NFS (no idea
>   about MacOS support), SMB (idem about Apple support).

no NFS support in MacOS except for a rather expensive program (last i
heard).  

> Err... you'll find lots of people that disagree. Current Apple hardware
> tends to have a rather clean design and relatively few hardware bugs.
> However most of the 68k designs were a complete mess of various legacy
> things patched around a 'current' processor. They refused modern designs
> like DMA for ages. Talk about braindead here... You could argue that
> some of their 68k designs contain more braindamage than any PC hardware.

and OldWorld macs suck horribly too.  (all the hardware problems we
have on penguinppc do nothing but amaze me that the damn thing is
still alive). 

> Some of their 'high quality' was bought by staying away from
> top-performance hardware. You can argue whether that was a good or a bad
> decision....

its all relative, top performance hardware doesn't stay top
performance long enough for you to get the box open anyway.  

> The PowerPC architecture is, IMO, a much better design than what's
> currently available for i386 (less legacy). But the comparison is
> unfair, PPC is a new design and much younger than i386. What can be said
> against Intel is that they should have gotten rid of lots of legacy
> stuff (at the price of compatibility) a long time ago.

MS won't let them, their junk OS still depends on all that legacy
crud.  

of course so does MacOS, apple just cheated by taking all the legacy
cruft and sticking it into a file piggybacked on an OpenFirmware
bootloader which loads the cruft into RAM at bootup.  

it would be an interesting project to see if it would have been
possible to build an OpenFirmware x86 box and load a 16 bit BIOS image
into RAM for the MS OSes...

-- 
Ethan Benson
http://www.alaska.net/~erbenson/

Attachment: pgpLdgWSrCClu.pgp
Description: PGP signature


Reply to: