home *** CD-ROM | disk | FTP | other *** search
- Pl8 mainly fixes the problems in pl7 :-) . Main changes listed below.
-
- 1. ICD software uses a totally nonstandard (and downright idiotic) meaning
- for the device number in physical mode Rwabs. This meant pl7 and ICD would
- end up minit'ing the wrong drive if you were unlucky :-( Pl8 uses the
- correct (but non standard) ICD form (thanks to Scott Kolodzieski for helping
- with tests to sort this one out).
-
- 2. The 'force' option to minit has been removed. ICD software refuses to
- let you have partitions bigger than 32Mb and 512 byte sectors. It refuses
- in a rather nasty way by absolutely denying all access to the partition!
- This forces rebuilding the partition (which is very destructive) or using
- a disk editor (which is not, but trickier). If you did use the '-f' option
- and want to recreate a new Minixfs partition then either rebuild the partition
- (e.g. using the icdfmt option) or edit the sector size back to the original
- value. If you can't do this then let me know and I'll write a short program
- to reset the sector size automatically.
-
- 3. AHDI (in common with some other software) checks you haven't gone past
- the end of a partition from the filesystem size in the boot sector. AHDI
- had just fixed the lrecno bug, this checking effectively makes this fix
- unusable because there is no way to set the size to more than 65535 logical
- sectors. Note to hard disk software writers: don't check the partition
- boundaries with the boot sector; use the partition table. Anyway Minixfs
- now does a check to see if it can access the last sector of a partition,
- if it can't it switches to physical mode.
-
- 4. When minit was used to initialise a partition just a bit bigger than a
- multiple of 8Mb it messed up in a rather painful way (oddly enough this
- bug had survived since well before 0.55). (thanks Juergen!)
-
- 5. Protection has been given yet another overhaul. I've developed a new method
- based on the 'null disk' idea which most hard disk software seemed to like.
- Keep the logical sector size the same but add a pseudo root directory after the
- inode table but before the first data zone. This uses less than two logical
- sectors so is a small price to pay for proper protection . I hope that this will
- finally satisfy all hard disk software and sector sizes. If by some miracle it
- does, I'll make it the only protection method and on by default. So far ICD, HDD
- and AHDI seem to like this with a variety of sector sizes. I would like to hear
- any reports about this method not working: try (after minit'ing a partition with
- the -P option) copying a file to a Minix partition under TOS (without MiNT
- running), you should get an error message.
- NB. This protection method is such that it is not easy to protect an already
- existing filesystem if the sector size is not 512 bytes. If there is any demand
- I may add an option to fsck to do this (it does involve reorganising the
- filesystem a bit). Older versions of Minixfs fsck will abort with an error
- message when used with this protection method and sector sizes of 1K or more.
- Minix fsck (as opposed to Minixfs fsck distributed here) will signal a warning
- if used with this protection method, but carry on regardless: this is normal.
-
- 6. Minit -t should now correctly report the information about access modes and
- XHDI info.
-
- My thanks to the many who supplied bug reports, fixes and conducted tests on
- their systems. These have helped enormously, keep them coming!
-
- Have fun, Steve.
-