BUFFERCACHE(9) Kernel Developer's Manual BUFFERCACHE(9)

NAME

buffercache, bread, breadn, bwrite, bawrite, bdwrite, getblk, geteblk, incore, allocbuf, brelse, biodone, biowaitbuffer cache interfaces

SYNOPSIS

#include <sys/buf.h>

int
bread(struct vnode *vp, daddr_t blkno, int size, struct kauth_cred *cred, int flags, buf_t **bpp);

int
breadn(struct vnode *vp, daddr_t blkno, int size, daddr_t rablks[], int rasizes[], int nrablks, struct kauth_cred *cred, int flags, buf_t **bpp);

int
bwrite(buf_t *bp);

void
bawrite(buf_t *bp);

void
bdwrite(buf_t *bp);

buf_t *
getblk(struct vnode *vp, daddr_t blkno, int size, int slpflag, int slptimeo);

buf_t *
geteblk(int size);

buf_t *
incore(struct vnode *vp, daddr_t blkno);

void
allocbuf(buf_t *bp, int size, int preserve);

void
brelse(buf_t *bp, int set);

void
biodone(buf_t *bp);

int
biowait(buf_t *bp);

DESCRIPTION

The buffercache interface is used by each filesystems to improve I/O performance using in-core caches of filesystem blocks.

The kernel memory used to cache a block is called a buffer and described by a buf structure. In addition to describing a cached block, a buf structure is also used to describe an I/O request as a part of the disk driver interface.

FUNCTIONS

bread(vp, blkno, size, cred, flags, bpp)
Read a block corresponding to vp and blkno. The buffer is returned via bpp. The units of blkno are specifically the units used by the VOP_STRATEGY() routine for the vp vnode. For device special files, blkno is in units of DEV_BSIZE and both blkno and size must be multiples of the underlying device's block size. For other files, blkno is in units chosen by the file system containing vp.

If the buffer is not found (i.e. the block is not cached in memory), bread() allocates a buffer with enough pages for size and reads the specified disk block into it using credential cred.

The buffer returned by bread() is marked as busy. (The B_BUSY flag is set.) After manipulation of the buffer returned from bread(), the caller should unbusy it so that another thread can get it. If the buffer contents are modified and should be written back to disk, it should be unbusied using one of variants of bwrite(). Otherwise, it should be unbusied using brelse().

breadn(vp, blkno, size, rablks, rasizes, nrablks, cred, flags, bpp)
Get a buffer as bread(). In addition, breadn() will start read-ahead of blocks specified by rablks, rasizes, nrablks.
breada(vp, blkno, size, rablkno, rabsize, cred, flags, bpp)
Same as breadn() with single block read-ahead. This function is for compatibility with old filesystem code and shouldn't be used by new ones.
bwrite(bp)
Write a block. Start I/O for write using VOP_STRATEGY(). Then, unless the B_ASYNC flag is set in bp, bwrite() waits for the I/O to complete.
bawrite(bp)
Write a block asynchronously. Set the B_ASYNC flag in bp and simply call VOP_BWRITE(), which results in bwrite() for most filesystems.
bdwrite(bp)
Delayed write. Unlike bawrite(), bdwrite() won't start any I/O. It only marks the buffer as dirty (B_DELWRI) and unbusy it.
getblk(vp, blkno, size, slpflag, slptimeo)
Get a block of requested size size that is associated with a given vnode and block offset, specified by vp and blkno. If it is found in the block cache, make it busy and return it. Otherwise, return an empty block of the correct size. It is up to the caller to ensure that the cached blocks are of the correct size.

If getblk() needs to sleep, slpflag and slptimeo are used as arguments for cv_timedwait().

geteblk(size)
Allocate an empty, disassociated block of a given size size.
incore(vp, blkno)
Determine if a block associated to a given vnode and block offset is in the cache. If it is there, return a pointer to it. Note that incore() doesn't busy the buffer unlike getblk().
allocbuf(bp, size, preserve)
Expand or contract the actual memory allocated to a buffer. If preserve is zero, the entire data in the buffer will be lost. Otherwise, if the buffer shrinks, the truncated part of the data is lost, so it is up to the caller to have written it out first if needed; this routine will not start a write. If the buffer grows, it is the callers responsibility to fill out the buffer's additional contents.
brelse(bp, set)
Unbusy a buffer and release it to the free lists.
biodone(bp)
Mark I/O complete on a buffer. If a callback has been requested by B_CALL, do so. Otherwise, wakeup waiters.
biowait(bp)
Wait for operations on the buffer to complete. When they do, extract and return the I/O's error value.

CODE REFERENCES

The buffer cache subsystem is implemented within the file sys/kern/vfs_bio.c.

SEE ALSO

intro(9), vnode(9)

Maurice J. Bach, The Design of the UNIX Operating System, Prentice Hall, 1986.

Marshall Kirk McKusick, Keith Bostic, Michael J. Karels, and John S. Quarterman, The Design and Implementation of the 4.4BSD Operating System, Addison Wesley, 1996.

BUGS

In the current implementation, bread() and its variants don't use a specified credential.

Because biodone() and biowait() do not really belong to buffercache, they shouldn't be documented here.

November 11, 2009 NetBSD 6.1