I have an old BeOS Release-2 CD, circa 1998. When I try to mount it on Haiku, I can see a half dozen partitions on it in the DriveSetup application. Yet it doesn’t mount. I know it was meant to be used with a BeBox, but is there a way to mount it on Haiku/Intel?
Most likely this is because this CD contains sessions with obsoleted file systems which Haiku doesn’t support like ofs and hfs. Currently supported file systems can be seen here: http://cgit.haiku-os.org/haiku/tree/src/add-ons/kernel/file_systems
Thanks for your suggestion Diver. I executed a “dd” command on the raw drive, and put the output into a file. When I looked at it, I noticed that there was header info in the first few bytes:
- Apple_partition_map
- Apple_driver_43x7
- Mac_OSApple_HFS
- Mac_OSBe_BFS
- Apple_Extra_Free
Apparently, the different partitions were created to allow different OS types to load the CD …
I think this is one of those things where, when you get started, you can’t stop. I spent more time looking at the “dd” output. It implied that the old BeOS CDs were basically of the apple-type with an MBR in sector 0, and an Apple partition map starting at sector 1. So, I noticed an APM (Apple partition map) specification on the internet, and observed where in the map the partition start sector is defined. It seemed like it would be easy to use a hex editor to find the start sector for the BFS partition, “dd” it into a file, and then use the bfs_shell on it. Crazy.
Well Haiku seems to support Apple partition map http://cgit.haiku-os.org/haiku/tree/src/add-ons/kernel/partitioning_systems/apple maybe you can grep syslog for apple and see if there are any messages about it:
grep -i apple /var/log/syslog
Thanks again Diver for the links to the code. I looked at the syslog, and found repetitive “trying: partitioning_systems/apple/v1 …” entries in it. I looked at the code for which you provided a link …/kernel/partitioning_systems/apple/apple.cpp … and it seems to have a number of "// Todo: " statements in it …
Of course - it’s not worth too much more effort, except as a mental exercise. I’ll have big-endian/little-endian problems out the gazooobi. In retrospect - the bfs_shell wouldn’t work for the same reason. Looking at the apple.cpp source to which Diver’s link points, I see a Be_BFS, but not a Mac_OSBe_BFS type in the partition type-list … I imagine for the same reason, which is likely also the reason for the “trying … trying …” messages.
My initial reason for mounting the CD was so that I might snoop around in the source code - which as single bytes would be no problem …
Maybe I’ll just write a little parser program for the “dd” output …
Sigh.
Now that I’ve wasted everyone’s time, I may as well waste some more of it (I think by writing). Was the original Hobbit Processor for the BeBox big or little endian?
Hi Ronald!
You can found every bit of information about the BeBoxes here: http://bebox.nu/
AFAIK: The PowerPC processors was able to run in both endiannes mode, and the default was the big-endian, but FIXME!
You can read more about the PPC processors here: http://en.wikipedia.org/wiki/PowerPC
Bye and i wish You a good luck!
Z.
Thank you miqlas for the link to the Be company history site. I was wondering about the AT&T Hobbit processors in the original BeBoxes … whether they were big or little endian? (as implemented in BeBox)
From the bebox.nu:
"The Hobbit BeBox is named after the CPUs that were used in the very first BeBox prototype - two AT&T ATT92010 “Hobbit” CPUs (date code 9309A - manufactured in September 1993). In addition to the Hobbit chips, three AT&T DSP3210 DSPs (date code 9308s - manufactured in August 1993) were also used. The Hobbit hardware was abandoned in 1994 when AT&T ceased production of the chips. "
Sorry, i can’t found more info, but it can be a good startpoint for you to lurk on the net.
Z.
Miqlas: Thanks for the info. BeBox history reading is a little bit of a blast into the past - a kind of nostalgia, I guess. The real solution would be to find someone with a correct Mac or an old BeBox, have them load the old CD, and send me the resulting source code. But, that would be WAY too easy. Why ride in luxury when you can stroll down the flower path? While perusing a “BeOS” history site, I read that BeOS version 5 for Intel had a BFS driver that would allow for the use of either big or little endian file system meta-data, apparently for backward compatibility with the older stuff. Is that true? Outside of reading text files, I don’t suppose it was very useful. I looked through some of the Haiku BFS source, and saw some "//Todo; " comments wherein it was questioned whether or not Haiku should do the same thing. The flavor of the comments seemed to imply that it would be a PITB to implement, and not that worthwhile …
Hi Ronald!
I understand You, sometimes i also want to save the world, because i can’t stop 
Sadly i don’t have any ancient computer to help with that, but maybe we can help somehow. What kind of data do You need from this disk?
I’m just thinking about an another way: can we manage to boot this disk somehow in an emulated enviroment? AFAIK around this time the BeOS was able to run on real Mac hardware. I’m sure, there is no BeBox emulator, but maybe we can setup an virtualised environment with an Classic Mac emulator…
Bye!
Z.
Haiku does not support little-endian BFS on big-endian systems. And it is not possible to use BeOS drivers because the FS driver API was redesigned.
So, the easiest solution I can think of is booting BeOS R5, either natively or in a VM. From there you should be able to copy the data to a big-endian BFS partition, which you will be able to read from Haiku.