TiBook, 1394, iPod working (sort of ;-)

TiBook, 1394, iPod working (sort of ;-)

[Ira Weiny]

This past week I finaly got around to getting up to date with Ben's tree.  I
built everything and except for a myserious hang on boot after upgrading to
MacOS 9.2 everything is working _great_.

So, I decided to check out what could be done with the iPod.  I have an iPod
connecting to Linux as a SBP2 drive.  All we need now for iPod support is
working HFS+.  But I don't know much about that so...

I have 2 questions.

1) How are things with HFS+?  I don't know much but have done some poking
around.  I notice that I can mount the iPod but the files it shows are not
correct.  But I have another HFS+ drive which seems to mount ok?  I have looked
at the volume info and have seen that the iPod version in the volume header is
"05" and the other drive is version "06"?  This seems odd since I would think
the earlier version would be better supported?  But I probably don't know what
I am talking about...

2) Does anyone know if there are any specs on the DB file for the iPod.  I
could work on reverse engineering it but if I could get some specs that would
be nice.  Anyone have inside connections at Apple?  Am I treading on thin ice
with licensing?

Thanks,
Ira Weiny
iweiny@acm.org


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

RE: TiBook, 1394, iPod working (sort of ;-)

[Halfmann, Klaus]

Hello Ira, you wrote:

> 1) How are things with HFS+?  I don't know much but have done
> some poking around.  I notice that I can mount the iPod but the files
it
> shows are not correct.  But I have another HFS+ drive which seems to
mount
> ok?  I have looked at the volume info and have seen that the iPod
version in the
> volume header is "05" and the other drive is version "06"?  This seems
odd
> since I would think the earlier version would be better supported?
But I
> probably don't know what
> I am talking about...

HFS+ Volumes are wrapped in a HFS-structure for security reasons, you
can
find out about his at some Website at Apple.

I once started doing a set of HFS+ Tools but got stuck due to personal
reasons.
The current state is that _reading_ works and writing works a little
bit.
I'd be happy someone would care about these. There is a (rotten) tarball

at penguinppc.org. I use(d) a CVS Repository at SuSE. Since I just
upgraded
to SuSE 7.3 I might try to build it again.

In case you are interested feel free to contact me.

	Klaus Halfmann http://klaus.halfmann.bei.t-online.de
			   (Mh, I really should update this page)

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: TiBook, 1394, iPod working (sort of ;-)

[Ethan Benson]

On Mon, Nov 26, 2001 at 05:33:51PM -0800, Ira Weiny wrote:
>
> This past week I finaly got around to getting up to date with Ben's tree.  I
> built everything and except for a myserious hang on boot after upgrading to
> MacOS 9.2 everything is working _great_.
>
> So, I decided to check out what could be done with the iPod.  I have an iPod
> connecting to Linux as a SBP2 drive.  All we need now for iPod support is
> working HFS+.  But I don't know much about that so...
>
> I have 2 questions.
>
> 1) How are things with HFS+?  I don't know much but have done some poking

for all appearances its abandoned.  the only sign of anything
happening with it is someone starting a sourceforge account for it,
all this has is the very pre-alpha development code which pretty much
just spews errors and panics the kernel, the same code that was
released about a year ago or so.

> around.  I notice that I can mount the iPod but the files it shows are not
> correct.  But I have another HFS+ drive which seems to mount ok?  I have looked

apple always does HFS+ in a wonky way, they basically create a
ordinary HFS filesystem, then put a huge hidden file with an HFS+
filesystem on it, thier drivers notice this and mount the image rather
then the wrapper.  when you tell linux to mount -t hfs whatever /mnt
you will see the HFS wrapper which won't have anything useful (usually
a README and a couple fake macos system files).

> at the volume info and have seen that the iPod version in the volume header is
> "05" and the other drive is version "06"?  This seems odd since I would think
> the earlier version would be better supported?  But I probably don't know what
> I am talking about...

i don't know much about this, its probably not significant, certainly
not to linux since it has zero support for any revision of HFS+,
outside of the userland hfsplusutils.

> 2) Does anyone know if there are any specs on the DB file for the iPod.  I
> could work on reverse engineering it but if I could get some specs that would
> be nice.  Anyone have inside connections at Apple?  Am I treading on thin ice
> with licensing?

don't hold your breath waiting for apple to help, im sure they want
nothing more then to keep this thing MacOSX only.  if you live in the
US (or other country with questionable laws regarding these matters)
you may well want to consult a lawyer before doing any reverse
engineering, the DMCA (or some other rediculous law) may or may not be
an issue with this.

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

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: TiBook, 1394, iPod working (sort of ;-)

[Ethan Benson]

On Tue, Nov 27, 2001 at 09:02:40AM +0100, Halfmann, Klaus wrote:
>
> HFS+ Volumes are wrapped in a HFS-structure for security reasons, you
> can
> find out about his at some Website at Apple.

uh, security reasons?  ive never heard that and can't believe it
provides any sort of real security.  the only motivation ive ever
read/heard about was to keep feebleminded macos users from being
confused and frightened by older versions of MacOS helpfully offering
to erase the HFS+ partitions/disks.

> I once started doing a set of HFS+ Tools but got stuck due to personal
> reasons.
> The current state is that _reading_ works and writing works a little
> bit.
> I'd be happy someone would care about these. There is a (rotten) tarball

there are plenty of people who seem to care, just none of them a) know
enough about filesystem hacking to help, or b) don't care enough to
want to hack on it.  ;-)

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

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: TiBook, 1394, iPod working (sort of ;-)

[Brad Boyer]

Ethan Benson wrote:
> On Mon, Nov 26, 2001 at 05:33:51PM -0800, Ira Weiny wrote:
> > 1) How are things with HFS+?  I don't know much but have done some poking
>
> for all appearances its abandoned.  the only sign of anything
> happening with it is someone starting a sourceforge account for it,
> all this has is the very pre-alpha development code which pretty much
> just spews errors and panics the kernel, the same code that was
> released about a year ago or so.

That would be me. I've almost got stuff working for a 2.4.x kernel,
which is what is in CVS for that project. I haven't had it crash
at all after my last block of rewrites, but it still doesn't work
properly. It seemed more important to fix the crashes, since I wasn't
going to get very far with it crashing all the time. I haven't had
much time to work on it recently.

http://sourceforge.net/projects/linux-hfsplus/

> i don't know much about this, its probably not significant, certainly
> not to linux since it has zero support for any revision of HFS+,
> outside of the userland hfsplusutils.

Well, according to Apple's documentation, the only version of HFS+
is version 4, so I have no idea what's up with that...

	Brad Boyer
	flar@allandria.com


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: TiBook, 1394, iPod working (sort of ;-)

[David A. Gatwood]

On Tuesday, November 27, 2001, at 12:02 AM, Halfmann, Klaus wrote:

> HFS+ Volumes are wrapped in a HFS-structure for security reasons, you
> can
> find out about his at some Website at Apple.

Umm... last I checked, they were wrapped in an "almost" HFS structure so
that older machines whose { firmware, ROMs } didn't support HFS+ could
still boot from them....  Did I miss something?  :-)


David


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: TiBook, 1394, iPod working (sort of ;-)

[Benjamin Herrenschmidt]

>
>> HFS+ Volumes are wrapped in a HFS-structure for security reasons, you
>> can
>> find out about his at some Website at Apple.
>
>Umm... last I checked, they were wrapped in an "almost" HFS structure so
>that older machines whose { firmware, ROMs } didn't support HFS+ could
>still boot from them....  Did I miss something?  :-)

No, you didn't ;) In fact, I even think recent MacOS (or is it MacOS X ?)
may create pure HFS+ volumes without the wrapper.

The main reason for the wrapper was for machines running non-HFS+ capable
filesystem to display something (where are all my files gone ?) instead
of proposing you to format the disk, and for the system to be able to
boot from HFS+ since OldWorld ROMs didn't understand it. Basically, the
wrapper contains a slim-fast MacOS System file which contains the necessary
HFS+ support to mount the real volume and continue booting from that.

Ben.


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

libXrender bug crashes OpenOffice

[Jason E. Stewart]

Hey All,

While running Kevin Hendrix's build of OpenOffice I kept getting
Abort's for no reason. We've finally limited down to the Xrender
lib. When it's not on my system, OpenOffice runs fine, with it I get
crashes.

Kevin found a bug in the library and fixed it and OpenOffice runs fine
on his system with libXrender installed. I downloaded the latest xlibs
source from sid and checked, but his patch hasn't yet been
included.

Will it be included soon?

Thanks,
jas.

--- xc/lib/Xrender/Xrender.c.prev	Wed Oct 10 22:38:25 2001
+++ xc/lib/Xrender/Xrender.c	Wed Oct 10 22:39:09 2001
@@ -193,7 +193,7 @@
 	       rep.numVisuals * sizeof (xPictVisual));
     xData = (void *) Xmalloc (rlength);

-    if (!xri || !xFormat)
+    if (!xri || !xData)
     {
 	if (xri) Xfree (xri);
 	if (xData) Xfree (xData);

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: libXrender bug crashes OpenOffice

[Michel Dänzer]

On Tue, 2001-11-27 at 23:37, Jason E. Stewart wrote:

> While running Kevin Hendrix's build of OpenOffice I kept getting
> Abort's for no reason. We've finally limited down to the Xrender
> lib. When it's not on my system, OpenOffice runs fine, with it I get
> crashes.
>
> Kevin found a bug in the library and fixed it and OpenOffice runs fine
> on his system with libXrender installed. I downloaded the latest xlibs
> source from sid and checked, but his patch hasn't yet been
> included.
>
> Will it be included soon?

Err... and what does that have to do with Linux/PPC development? ;)


--
Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer
XFree86 and DRI project member   /  CS student, Free Software enthusiast

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: TiBook, 1394, iPod working (sort of ;-)

[Ira Weiny]

On Tue, 27 Nov 2001 23:18:34 +0100
Benjamin Herrenschmidt <benh@kernel.crashing.org> wrote:

>
> No, you didn't ;) In fact, I even think recent MacOS (or is it MacOS X ?)
> may create pure HFS+ volumes without the wrapper.
>
> The main reason for the wrapper was for machines running non-HFS+ capable
> filesystem to display something (where are all my files gone ?) instead
> of proposing you to format the disk, and for the system to be able to
> boot from HFS+ since OldWorld ROMs didn't understand it. Basically, the
> wrapper contains a slim-fast MacOS System file which contains the necessary
> HFS+ support to mount the real volume and continue booting from that.
>

I think this makes sence because with Linux the HFS+ partition on my HD shows
up with the "where are all my files gone?" thing.  (Which BTW dropped me into
xmon when I vim'ed it...)  But the iPod did not show up with anything
intelligable.

Must be that the iPod is "pure" HFS+ when the other is a wrapper.

To quote a pointy eared character "Fascinating."

Thanks for the info,
Ira Weiny
iweiny@acm.org


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: libXrender bug crashes OpenOffice

[Jason E. Stewart]

"Michel Dänzer" <michdaen@iiic.ethz.ch> writes:

> On Tue, 2001-11-27 at 23:37, Jason E. Stewart wrote:
>
> > While running Kevin Hendrix's build of OpenOffice I kept getting
> > Abort's for no reason. We've finally limited down to the Xrender
> > lib. When it's not on my system, OpenOffice runs fine, with it I get
> > crashes.
> >
> > Kevin found a bug in the library and fixed it and OpenOffice runs fine
> > on his system with libXrender installed. I downloaded the latest xlibs
> > source from sid and checked, but his patch hasn't yet been
> > included.
> >
> > Will it be included soon?
>
> Err... and what does that have to do with Linux/PPC development? ;)

Perhaps I was mistaken, but I thought that the bug only existed on
PPC. Is this too X-specific and should be posted elsewhere?

jas.

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: libXrender bug crashes OpenOffice

[Michel Dänzer]

On Wed, 2001-11-28 at 00:47, Jason E. Stewart wrote:
>
> "Michel Dänzer" <michdaen@iiic.ethz.ch> writes:
>
> > On Tue, 2001-11-27 at 23:37, Jason E. Stewart wrote:
> >
> > > While running Kevin Hendrix's build of OpenOffice I kept getting
> > > Abort's for no reason. We've finally limited down to the Xrender
> > > lib. When it's not on my system, OpenOffice runs fine, with it I get
> > > crashes.

BTW how can I reproduce those crashes? No 'success' so far.


> > > Kevin found a bug in the library and fixed it and OpenOffice runs fine
> > > on his system with libXrender installed. I downloaded the latest xlibs
> > > source from sid and checked, but his patch hasn't yet been
> > > included.
> > >
> > > Will it be included soon?
> >
> > Err... and what does that have to do with Linux/PPC development? ;)
>
> Perhaps I was mistaken, but I thought that the bug only existed on
> PPC. Is this too X-specific and should be posted elsewhere?

At least I didn't see anything PPC specific in the fix you posted.

In particular though, you should submit this in a bug against the Debian
xlibs package for it to be included there.

AFAIK the fix has already been applied to the XFree86 CVS trunk, so it
will be in the 4.2.0 release. Don't know about the 4.1 branch.


PS: No, you can't build only libXrender.

--
Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer
XFree86 and DRI project member   /  CS student, Free Software enthusiast

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: TiBook, 1394, iPod working (sort of ;-)

[Ethan Benson]

On Tue, Nov 27, 2001 at 09:28:47AM -0800, Brad Boyer wrote:
>
> Well, according to Apple's documentation, the only version of HFS+
> is version 4, so I have no idea what's up with that...

well it also doesn't document things like the permission bits, it just
says they are unused and reserved, obviously this is not the case
given OSX, apple just has not updated this document.

that means theres a bit of reverse engineering involved as well since
apple is using a newwer version of HFS+ which is not entirely
documented.

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

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: TiBook, 1394, iPod working (sort of ;-)

[Timothy A. Seufert]

At 3:58 PM -0900 11/27/01, Ethan Benson wrote:
>On Tue, Nov 27, 2001 at 09:28:47AM -0800, Brad Boyer wrote:
>>
>>  Well, according to Apple's documentation, the only version of HFS+
>>  is version 4, so I have no idea what's up with that...
>
>well it also doesn't document things like the permission bits, it just
>says they are unused and reserved, obviously this is not the case
>given OSX, apple just has not updated this document.
>
>that means theres a bit of reverse engineering involved as well since
>apple is using a newwer version of HFS+ which is not entirely
>documented.

The documentation is the source now, it would seem.  (just like much
of Linux :)  See:

   xnu-9-1/bsd/hfs/

in the Darwin kernel (aka "xnu") source.  hfs_format.h defines the
on-disk layout, including the structure used for storage of
permission bits.

 From the information in that file it looks like the HFS+ version is
still 4, BTW.

Looks like it should even be possible to figure out how Apple did the
hard- and soft-link hacks.  Maybe one day Linux will be able to boot
directly from HFS+ just as Darwin does.  That would be perverse.  :)
--
Tim Seufert

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: TiBook, 1394, iPod working (sort of ;-)

[Ethan Benson]

On Wed, Nov 28, 2001 at 02:40:09AM -0800, Timothy A. Seufert wrote:
>
> Looks like it should even be possible to figure out how Apple did the
> hard- and soft-link hacks.  Maybe one day Linux will be able to boot
> directly from HFS+ just as Darwin does.  That would be perverse.  :)

damn you!  now i have to clean up a big mess of puke...

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

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: TiBook, 1394, iPod working (sort of ;-)

[Benjamin Herrenschmidt]

>well it also doesn't document things like the permission bits, it just
>says they are unused and reserved, obviously this is not the case
>given OSX, apple just has not updated this document.
>
>that means theres a bit of reverse engineering involved as well since
>apple is using a newwer version of HFS+ which is not entirely
>documented.

Darwin source contains the full implentation of HFS+ used by Apple.

Ben.


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/