NAME¶

mdchain, md_initm, md_done, md_append_record, md_next_record, md_get_uint8, md_get_uint16, md_get_uint16be, md_get_uint16le, md_get_uint32, md_get_uint32be, md_get_uint32le, md_get_int64, md_get_int64be, md_get_int64le, md_get_mem, md_get_mbuf, md_get_uio — set of functions to dissect an mbuf chain to various data types

SYNOPSIS¶

options LIBMCHAIN kldload libmchain

#include <sys/param.h>
#include <sys/uio.h>
#include <sys/mchain.h>

void
md_initm(struct mdchain *mdp, struct mbuf *m);

void
md_done(struct mdchain *mdp);

void
md_append_record(struct mdchain *mdp, struct mbuf *top);

int
md_next_record(struct mdchain *mdp);

int
md_get_uint8(struct mdchain *mdp, uint8_t *x);

int
md_get_uint16(struct mdchain *mdp, uint16_t *x);

int
md_get_uint16be(struct mdchain *mdp, uint16_t *x);

int
md_get_uint16le(struct mdchain *mdp, uint16_t *x);

int
md_get_uint32(struct mdchain *mdp, uint32_t *x);

int
md_get_uint32be(struct mdchain *mdp, uint32_t *x);

int
md_get_uint32le(struct mdchain *mdp, uint32_t *x);

int
md_get_int64(struct mdchain *mdp, int64_t *x);

int
md_get_int64be(struct mdchain *mdp, int64_t *x);

int
md_get_int64le(struct mdchain *mdp, int64_t *x);

int
md_get_mem(struct mdchain *mdp, caddr_t target, int size, int type);

int
md_get_mbuf(struct mdchain *mdp, int size, struct mbuf **m);

int
md_get_uio(struct mdchain *mdp, struct uio *uiop, int size);

DESCRIPTION¶

These functions are used to decompose mbuf chains to various data types. The mdchain structure is used as a working context and should be initialized through a call of the mb_initm() function. It has the following fields:

md_top

(struct mbuf *) A pointer to the top of the parsed mbuf chain.

md_cur

(struct mbuf *) A pointer to the currently parsed mbuf.

md_pas

(int) Offset in the current mbuf.

The md_done() function disposes of an mbuf chain pointed to by the mdp->md_top field and sets the field to NULL.

The md_append_record() appends a new mbuf chain using m_nextpkt field to form a single linked list of mbuf chains. If the mdp->md_top field is NULL, then this function behaves exactly as the md_initm() function.

The md_next_record() function extracts the next mbuf chain and disposes the current one, if any. For a new mbuf chain it calls the md_initm() function. If there is no data left the function returns ENOENT.

All md_get_*() functions perform an actual copy of the data from an mbuf chain. Functions which have le or be suffixes will perform conversion to the little- or big-endian data formats.

md_get_mem() function copies size bytes of data specified by the source argument from an mbuf chain. The type argument specifies the method used to perform a copy, and can be one of the following:

MB_MSYSTEM

Use the bcopy() function.

MB_MUSER

Use the copyin(9) function.

MB_MINLINE

Use an “inline” loop which does not call any function.

If target is NULL, an actual copy is not performed and the function just skips the given number of bytes.

RETURN VALUES¶

All int functions return zero if successful, otherwise an error code is returned.

Note: after failure of any function, an mbuf chain is left in the broken state and only the md_done() function can safely be called to destroy it.

EXAMPLES¶

struct mdchain *mdp;
struct mbuf *m;
uint16_t length;
uint8_t byte;

receive(so, &m);
md_initm(mdp, m);
if (md_get_uint8(mdp, &byte) != 0 ||
    md_get_uint16le(mdp, &length) != 0)
	error = EBADRPC;
mb_done(mdp);

SEE ALSO¶

mbchain(9), mbuf(9)

AUTHORS¶

This manual page was written by Boris Popov <bp@FreeBSD.org>.