renegade-project/linux-next
2
0
Fork 0
mirror of https://github.com/edk2-porting/linux-next.git synced 2024-12-22 20:23:57 +08:00
Code
391fc09a14
linux-next/fs/sysv/INTRO
Linus Torvalds 1da177e4c3 Linux-2.6.12-rc2 
Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.

Let it rip!
2005-04-16 15:20:36 -07:00

183 lines
7.0 KiB
Plaintext
Raw Blame History

This is the implementation of the SystemV/Coherent filesystem for Linux.
It grew out of separate filesystem implementations
Xenix FS Doug Evans <dje@cygnus.com> June 1992
SystemV FS Paul B. Monday <pmonday@eecs.wsu.edu> March-June 1993
Coherent FS B. Haible <haible@ma2s2.mathematik.uni-karlsruhe.de> June 1993
and was merged together in July 1993.
These filesystems are rather similar. Here is a comparison with Minix FS:
* Linux fdisk reports on partitions
- Minix FS 0x81 Linux/Minix
- Xenix FS ??
- SystemV FS ??
- Coherent FS 0x08 AIX bootable
* Size of a block or zone (data allocation unit on disk)
- Minix FS 1024
- Xenix FS 1024 (also 512 ??)
- SystemV FS 1024 (also 512 and 2048)
- Coherent FS 512
* General layout: all have one boot block, one super block and
separate areas for inodes and for directories/data.
On SystemV Release 2 FS (e.g. Microport) the first track is reserved and
all the block numbers (including the super block) are offset by one track.
* Byte ordering of "short" (16 bit entities) on disk:
- Minix FS little endian 0 1
- Xenix FS little endian 0 1
- SystemV FS little endian 0 1
- Coherent FS little endian 0 1
Of course, this affects only the file system, not the data of files on it!
* Byte ordering of "long" (32 bit entities) on disk:
- Minix FS little endian 0 1 2 3
- Xenix FS little endian 0 1 2 3
- SystemV FS little endian 0 1 2 3
- Coherent FS PDP-11 2 3 0 1
Of course, this affects only the file system, not the data of files on it!
* Inode on disk: "short", 0 means non-existent, the root dir ino is:
- Minix FS 1
- Xenix FS, SystemV FS, Coherent FS 2
* Maximum number of hard links to a file:
- Minix FS 250
- Xenix FS ??
- SystemV FS ??
- Coherent FS >=10000
* Free inode management:
- Minix FS a bitmap
- Xenix FS, SystemV FS, Coherent FS
There is a cache of a certain number of free inodes in the super-block.
When it is exhausted, new free inodes are found using a linear search.
* Free block management:
- Minix FS a bitmap
- Xenix FS, SystemV FS, Coherent FS
Free blocks are organized in a "free list". Maybe a misleading term,
since it is not true that every free block contains a pointer to
the next free block. Rather, the free blocks are organized in chunks
of limited size, and every now and then a free block contains pointers
to the free blocks pertaining to the next chunk; the first of these
contains pointers and so on. The list terminates with a "block number"
0 on Xenix FS and SystemV FS, with a block zeroed out on Coherent FS.
* Super-block location:
- Minix FS block 1 = bytes 1024..2047
- Xenix FS block 1 = bytes 1024..2047
- SystemV FS bytes 512..1023
- Coherent FS block 1 = bytes 512..1023
* Super-block layout:
- Minix FS
unsigned short s_ninodes;
unsigned short s_nzones;
unsigned short s_imap_blocks;
unsigned short s_zmap_blocks;
unsigned short s_firstdatazone;
unsigned short s_log_zone_size;
unsigned long s_max_size;
unsigned short s_magic;
- Xenix FS, SystemV FS, Coherent FS
unsigned short s_firstdatazone;
unsigned long s_nzones;
unsigned short s_fzone_count;
unsigned long s_fzones[NICFREE];
unsigned short s_finode_count;
unsigned short s_finodes[NICINOD];
char s_flock;
char s_ilock;
char s_modified;
char s_rdonly;
unsigned long s_time;
short s_dinfo[4]; -- SystemV FS only
unsigned long s_free_zones;
unsigned short s_free_inodes;
short s_dinfo[4]; -- Xenix FS only
unsigned short s_interleave_m,s_interleave_n; -- Coherent FS only
char s_fname[6];
char s_fpack[6];
then they differ considerably:
Xenix FS
char s_clean;
char s_fill[371];
long s_magic;
long s_type;
SystemV FS
long s_fill[12 or 14];
long s_state;
long s_magic;
long s_type;
Coherent FS
unsigned long s_unique;
Note that Coherent FS has no magic.
* Inode layout:
- Minix FS
unsigned short i_mode;
unsigned short i_uid;
unsigned long i_size;
unsigned long i_time;
unsigned char i_gid;
unsigned char i_nlinks;
unsigned short i_zone[7+1+1];
- Xenix FS, SystemV FS, Coherent FS
unsigned short i_mode;
unsigned short i_nlink;
unsigned short i_uid;
unsigned short i_gid;
unsigned long i_size;
unsigned char i_zone[3*(10+1+1+1)];
unsigned long i_atime;
unsigned long i_mtime;
unsigned long i_ctime;
* Regular file data blocks are organized as
- Minix FS
7 direct blocks
1 indirect block (pointers to blocks)
1 double-indirect block (pointer to pointers to blocks)
- Xenix FS, SystemV FS, Coherent FS
10 direct blocks
1 indirect block (pointers to blocks)
1 double-indirect block (pointer to pointers to blocks)
1 triple-indirect block (pointer to pointers to pointers to blocks)
* Inode size, inodes per block
- Minix FS 32 32
- Xenix FS 64 16
- SystemV FS 64 16
- Coherent FS 64 8
* Directory entry on disk
- Minix FS
unsigned short inode;
char name[14/30];
- Xenix FS, SystemV FS, Coherent FS
unsigned short inode;
char name[14];
* Dir entry size, dir entries per block
- Minix FS 16/32 64/32
- Xenix FS 16 64
- SystemV FS 16 64
- Coherent FS 16 32
* How to implement symbolic links such that the host fsck doesn't scream:
- Minix FS normal
- Xenix FS kludge: as regular files with chmod 1000
- SystemV FS ??
- Coherent FS kludge: as regular files with chmod 1000
Notation: We often speak of a "block" but mean a zone (the allocation unit)
and not the disk driver's notion of "block".
Bruno Haible <haible@ma2s2.mathematik.uni-karlsruhe.de>
View Git Blame Copy Permalink