Re: Linux partitions
From: wolfgang kern (nowhere_at_nevernet.at)
Date: Wed, 12 May 2004 12:45:15 +0200
| > and where to find what's behind this partitions
| > (BootRecord-interpretation and DirStructures).
| > I don't like to read many books,
| > I'd prefer compact tables with the needed info.
| Ralf Brown's has a compact table under "int 19h - bootstrap loader"
| (there is a table near it which lists all of the types)...those that
| relate to Linux in my now rather old version of Ralf's list are:
| 81h Linux
| 82h Linux Swap partition
| 83h Linux native file system (ext2fs/xiafs)
| [ Copied from Ralf Brown's Interrupt List ]
Yes my RBIL (DnLd end of 2003) shows the same,
but I thought there 'must' be more.
| One thing to pay attention to, wolfgang...technically, the byte in the
| MBR is the _OS indicator_ and NOT a "partition type indicator"...
| So, looking at this, you'll have to use some other means to determine
| the specific FS...actually reading some sectors and looking to see if
| certain bytes and formats are being used...
Right, I realised this.
I always check the bootsector from a partition as well.
May I rely on at least the partition entries are uniform structures?
If so I could skip over unknown/unsupported parts in the chain.
| ...implement a "cheap hack" (rather than teaching the FS drivers to
| "hide" things? Nah, we'll just change its number! That'll "hide" it!
KESYS isn't much different here ;)
1994 I did not check if '4B'="K" is already in use by someone else.
It's more by coincidence that this partition-type is still 'owned' by me.
The identification-story is found in the boot-record anyway,
which is part of the partition-sector on KESYS formatted media.
| ...especially because if you
| misdiagnose the FS and then start _writing_ anything, you're just
| going to screw things up big-time that you might ruin it permanently
That's the point.
| They probably should have set up some "authority" by which you
| "register" the OS indicator, who keeps track of it and prevents
| clashes and abuse of the system...but, well, they didn't...
And now it's too late.
| ...the documentation for an FS might, in fact,
| helpfully even suggest how Linux itself works it out or whatever (and,
| of course, if you copy the "official" method of detection, then it
| should be trustworthy, as that's what the OS itself probably uses
While I fixed some bugs in my M$-FAT-export module, I got the idea
to add HPFS, NTFS and Linux (at least) import-modules to KESYS.
81h old linux
83h primary /reisnerFS /JFS/ XFS
85h ext /ext2 /ext3
FAT: (the hidden, and 0e weren't supported yet)
01 FAT12 <16MB 11 hidden
04 FAT16 <32MB 14 hidden
06 FAT16 <8GB 16 hidden
0b FAT32 1b hidden
0c FAT32 LBA 1c hidden
0e FAT16 LBA 1e hidden
0f FAT32 '0b'ext 1f hidden