|
NAME
| |
VtBlock, VtCache, vtblockcopy, vtblockduplock, vtblockput, vtblockwrite,
vtcachealloc, vtcacheallocblock, vtcachefree, vtcacheglobal, vtcachelocal,
vtcachesetwrite, vtglobaltolocal, vtlocaltoglobal – Venti block
cache
|
SYNOPSIS
| |
#include <u.h>
#include <libc.h>
#include <venti.h>
typedef struct VtBlock
{
| |
uchar *data;
uchar type;
uchar score[VtScoreSize];
u32int addr;
...
|
} VtBlock;
VtCache* vtcachealloc(VtConn *z, ulong maxmem);
void vtcachefree(VtCache *c);
u32int vtglobaltolocal(uchar score[VtScoreSize])
void vtlocaltoglobal(u32int local, uchar score[VtScoreSize])
VtBlock* vtcacheallocblock(VtCache *c, int type, ulong size);
VtBlock* vtcachelocal(VtCache *c, u32int addr, int type);
VtBlock* vtcacheglobal(VtCache *c, uchar[VtScoreSize], int type,
ulong size);
void vtblockput(VtBlock *b);
void vtblockduplock(VtBlock *b);
int vtblockwrite(VtBlock *b);
void vtcachesetwrite(VtCache *c,
| |
| |
int (*write)(VtConn*, uchar[VtScoreSize], uint, uchar*, int));
|
|
VtBlock* vtblockcopy(VtBlock *b);
|
DESCRIPTION
| |
These functions provide access to a simple in-memory cache of
blocks already stored on a Venti server and blocks that will eventually
be stored on a Venti server.
A VtBlock represents a venti data block. Blocks stored on a venti
server, called global blocks, are named by the SHA1 hash of their
contents. This hash is recorded as the block’s score. Such blocks
are immutable. The cache also stores mutable blocks that have
not yet been written to a venti server. These blocks are called
local blocks, and have
special scores that are 16 zero bytes followed by a 4-byte big-endian
address. The address is an index into the internal set of cache
blocks.
The user-visible contents of a VtBlock are data, a pointer to
the data; type, the venti block type; score, the block’s score;
and addr, the block’s cache address.
Vtcachealloc allocates a new cache using the client connection
z (see venti-conn(3) and venti-client(3)), with maxmem bytes of
memory.
Vtcachefree frees a cache and all the associated blocks.
Vtglobaltolocal returns the local address corresponding to the
given local score. If passed a global score, vtglobaltolocal returns
the special constant NilBlock (~0). Vtlocaltoglobal is the opposite,
setting score to the local score for the cache address local.
Vtcacheallocblock allocates a new local block with the given type
and size.
Vtcachelocal retrieves the local block at address addr from the
cache. The given type must match the type of the block found at
addr.
Vtcacheglobal retrieves the block with the given score, dtype
and size from the cache, consulting the Venti server if necessary.
If passed a local score, vtcacheglobal invokes vtcachelocal appropriately.
The block references returned by vtcacheallocblock, vtcachelocal,
and vtcacheglobal must be released when no longer needed. Vtblockput
releases such a reference.
It is occasionally convenient to have multiple variables refer
to the same block. Vtblockduplock increments the block’s reference
count so that an extra vtblockput will be required in order to
release the block.
Vtblockwrite writes a local block to the Venti server, changing
the block to a global block. It calls the cache’s write function
to write the block to the server. The default write function is
vtwrite (see venti-client(3)); vtsetcachewrite sets it. Vtsetcachewrite
is used by clients to install replacement functions that run writes
in the background or perform other
additional processing.
Vtblockcopy copies a block in preparation for modifying its contents.
The old block may be a local or global block, but the new block
will be a local block.
The cache only evicts global blocks. Local blocks can only leave
the cache via vtblockwrite, which turns them into global blocks,
making them candidates for eviction.
If a new cache block must be allocated (for vtcacheallocblock,
vtcachelocal, vtcacheglobal, or vtblockcopy), but the cache is
filled (with local blocks and blocks that have not yet been released
with vtblockput), the library prints the score and reference count
of every block in the cache and then aborts. A full cache indicates
either that the cache is too
small, or, more commonly, that cache blocks are being leaked.
|
SOURCE
SEE ALSO
|
|
