RISC iX - Cloned!

[IanJeffray]

I've added RISC iX partition support to Partition Manager 0.88, currently read-only.

Went pop with two images on R140 but does something that looks reasonable on A540 with a single image.

Doesn't seem too happy with the Simtec podule - only shows four of the 8 partitions, and not very well.

[sirbod]

On further investigation having compared a drive formatted with st506form2 and the 4corn image, the Unix partition table is slightly different. The magic value (first word of the partition sector) differs as does the offset of the partitions in the structure.

If the non-ADFS type ID is 1, the magic value is "trap" and the partitions start at offset +4
If the non-ADFS type ID is 2, the magic value is " seJ" and the partitions start at offset +8

That implies RISC iX handles ADFS and SCSI differently and the partition format is possibly different.

EDIT: Type 1 the Unix partition address/sizes is specified in cylinders and type 2 it looks like LBA
EDIT2: Forgot to swap the byte order round on the magic values, which I've now corrected above

[IanJeffray]

EDIT: Type 1 the Unix partition address/sizes is specified in cylinders and type 2 it looks like LBA

That probably explains why you're showing the iX partitions as the wrong size?
For reference, the swap is actually 20MB and root is 167MB for this particular image.

[philpem]

On further investigation having compared a drive formatted with st506form2 and the 4corn image, the Unix partition table is slightly different. The magic value (first word of the partition sector) differs as does the offset of the partitions in the structure.

If the non-ADFS type ID is 1, the magic value is "trap" and the partitions start at offset +4
If the non-ADFS type ID is 2, the magic value is "Jes " and the partitions start at offset +8

That implies RISC iX handles ADFS and SCSI differently and the partition format is possibly different.

EDIT: Type 1 the Unix partition address/sizes is specified in cylinders and type 2 it looks like LBA

Cylinders vs. LBA would make sense - given that ST506 drives are addressed in cylinder-head-sector form, and SCSI is an LBA (sector number, numbered from zero).

[dinang]

Good to see Drobe back up and running.

On another note, I'm struggling to get X running in 256 colour mode. Has anyone got this up and running?

EDIT: Now figured how to do this - just had to modify the startx script to call Xarm with the -c8 parameter in the line that runs xinit.

Thanks,

Rob

Sorry to dig up an old thread. Thanks for this hint Rob. Just in case anyone else is also trying this, it is possible to get 256 colour mode without altering any scripts. I'm running RISCiX 1.2.1c (https://www.4corn.co.uk/articles/riscix121c/) in Arculator 2.1 (https://github.com/sarah-walker-pcem/arculator). After following the instructions on 4corn to configure the RISCiX image in Arculator 2.1 (I used an A540 emulation model with 16MB RAM, RISC OS 3.1, Acorn AKA31 SCSI podule with the RISCiX disk image mapped to SCSI device 0, Multisync monitor), I used your suggestion to disable automatic monitor type detection in RISC OS's !Configure app (selecting SVGA monitor type and leaving automatic mode selection enabled). After making all the necessary changes on the RISC OS side, I hard reset Arculator before attempting to enter RISCiX.

Once at the RISCiX console prompt to select 256 colours I ran startx using the following command line:

#startx -- -svga -c8

The -svga switch seems unnecessary. Once in X windows (motif), I used the Utilities->Colour Map to test the 256 colour palette.

It is fantastic to be able to play with RISCiX - something I wasn't able to do back in the day when I had my old A440/1 or A4000 at Uni. To be able to do this at no cost in an emulator is even more amazing (thanks @Sarah Walker). Thanks also to everyone on this thread for giving me loads of hints and tips to get going.

Cheers

Dinan

[sirbod]

On further investigation having compared a drive formatted with st506form2 and the 4corn image, the Unix partition table is slightly different. The magic value (first word of the partition sector) differs as does the offset of the partitions in the structure.

If the non-ADFS type ID is 1, the magic value is "trap" and the partitions start at offset +4
If the non-ADFS type ID is 2, the magic value is " seJ" and the partitions start at offset +8

That implies RISC iX handles ADFS and SCSI differently and the partition format is possibly different.

EDIT: Type 1 the Unix partition address/sizes is specified in cylinders and type 2 it looks like LBA

Some other observations I've noticed while enhancing Partition Manager to allow the FileCore volume to be reformatted without affecting the RISCiX install:

• The RISCiX area is embedded within the FileCore volume
• The last sector of the RISCiX partitions coincides with the last sector of the FileCore volume
• The disc end marker in the FileCore map is adjusted to the start of the RISCiX area. This has the knock-on effect of marking the RISCiX area as used space when viewed in RISC OS
• Unix partition 6 (possibly depending on the format utility used) covers the region up to the RISCiX start cylinder
• Unix partition 7 covers the full FileCore area

Why it's implemented this way is probably lost to time. You have to wonder...did RISCiX use FileCore as the filesystem during its early development?