ref: 26e00d014cebde8bc07725700b5ac9c0cade8171
dir: /sys/doc/fs/p2/
.SH
The server processes
.PP
The main file system algorithm is a set
of identical processes
named
.CW srv
that honor the
9P protocol.
Each file system process waits on
a message queue for an incoming request.
The request contains a 9P message and
the address of a reply queue.
A
.CW srv
process parses the message,
performs pseudo-disk I/O
to the corresponding file system block device,
formulates a response,
and sends the
response back to the reply queue.
.PP
The unit of storage is a
logical block
(not physical sector) of data on a device:
.Ex
.TA 0.5i 1i 1.5i 2i 2.5i 3i 3.5i 4i 4.5i 5i 5.5i
enum
{
RBUFSIZE = 8*1024
};
typedef vlong Off;
typedef
struct
{
short pad;
short tag;
Off path;
} Tag;
enum
{
BUFSIZE = RBUFSIZE - sizeof(Tag)
};
typedef
struct
{
uchar data[BUFSIZE];
Tag tag;
} Block;
.Ee
All devices are idealized as a perfect disk
of contiguously numbered blocks each of size
.CW RBUFSIZE .
Each block has a tag that identifies what type
of block it is and a unique id of the file or directory
where this block resides.
The remaining data in the block depends on
what type of block it is.
.PP
The
.CW srv
process's main data structure is the directory entry.
This is the equivalent of a UNIX i-node and
defines the set of block addresses that comprise a file or directory.
Unlike the i-node,
the directory entry also has the name of the
file or directory in it:
.Ex
enum
{
NAMELEN = 56,
NDBLOCK = 6,
NIBLOCK = 4,
};
.Ee
.Ex
typedef
struct
{
char name[NAMELEN];
short uid;
short gid;
ushort mode;
short wuid;
Qid qid;
Off size;
Off dblock[NDBLOCK];
Off iblocks[NIBLOCK];
long atime;
long mtime;
} Dentry;
.Ee
Each directory entry holds the file or directory
name, protection mode, access times, user-id, group-id, and addressing
information.
The entry
.CW wuid
is the user-id of the last writer of the file
and
.CW size
is the size of the file in bytes.
The addresses of the first 6
blocks of the file are held in the
.CW dblock
array.
If the file is larger than that,
an indirect block is allocated that holds
the next
.CW BUFSIZE/sizeof(Off)
block addresses of the file.
The indirect block address is held in
.CW iblocks[0] .
If the file is larger yet,
then there is a double indirect block that points
at indirect blocks.
The double indirect address is held in
.CW iblocks[1]
and can point at another
.CW (BUFSIZE/sizeof(Off))\u\s-2\&2\s+2\d
blocks of data.
This is extended through a quadruple indirect block at
.CW iblocks[3]
but the code is now parameterised to permit easily changing the
number of direct blocks and the depth of indirect blocks,
and also the maximum size of a file name component.
The maximum addressable size of a file is
therefore 7.93 petabytes at a block size of 8k,
but 7.98 exabytes (just under $2 sup 63$ bytes) at a block size of 32k.
File size is restricted to $2 sup 63 - 1$ bytes in any case
because the length of a file is maintained in a
(signed)
.I vlong .
These numbers are based on
.I fs64
which has a block size of 8k and
.CW sizeof(Off)
is 8.
.PP
The declarations of the indirect and double indirect blocks
are as follows.
.Ex
enum
{
INDPERBUF = BUFSIZE/sizeof(Off),
};
.Ee
.Ex
typedef
{
Off dblock[INDPERBUF];
Tag ibtag;
} Iblock;
.Ee
.Ex
typedef
{
Off iblock[INDPERBUF];
Tag dibtag;
} Diblock;
.Ee
.PP
The root of a file system is a single directory entry
at a known block address.
A directory is a file that consists of a list of
directory entries.
To make access easier,
a directory entry cannot cross blocks.
In
.I fs64
there are 47 directory entries per block.
.PP
The device on which the blocks reside is implicit
and ultimately comes from the 9P
.CW attach
message that specifies the name of the
device containing the root.