mirror of
https://github.com/darlinghq/darling.git
synced 2026-01-17 10:19:43 -06:00
1144 lines
40 KiB
C
1144 lines
40 KiB
C
/*
|
|
* Copyright (c) 1999-2011 Apple Inc. All rights reserved.
|
|
*
|
|
* @APPLE_OSREFERENCE_LICENSE_HEADER_START@
|
|
*
|
|
* This file contains Original Code and/or Modifications of Original Code
|
|
* as defined in and that are subject to the Apple Public Source License
|
|
* Version 2.0 (the 'License'). You may not use this file except in
|
|
* compliance with the License. The rights granted to you under the License
|
|
* may not be used to create, or enable the creation or redistribution of,
|
|
* unlawful or unlicensed copies of an Apple operating system, or to
|
|
* circumvent, violate, or enable the circumvention or violation of, any
|
|
* terms of an Apple operating system software license agreement.
|
|
*
|
|
* Please obtain a copy of the License at
|
|
* http://www.opensource.apple.com/apsl/ and read it before using this file.
|
|
*
|
|
* The Original Code and all software distributed under the License are
|
|
* distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
|
|
* EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
|
|
* INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
|
|
* FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
|
|
* Please see the License for the specific language governing rights and
|
|
* limitations under the License.
|
|
*
|
|
* @APPLE_OSREFERENCE_LICENSE_HEADER_END@
|
|
*/
|
|
/* Copyright (c) 1998, 1999 Apple Computer, Inc. All Rights Reserved */
|
|
/* Copyright (c) 1995 NeXT Computer, Inc. All Rights Reserved */
|
|
/*
|
|
* Mach Operating System
|
|
* Copyright (c) 1987 Carnegie-Mellon University
|
|
* All rights reserved. The CMU software License Agreement specifies
|
|
* the terms and conditions for use and redistribution.
|
|
*/
|
|
/*
|
|
* Copyright (c) 1994 NeXT Computer, Inc. All rights reserved.
|
|
*
|
|
* Copyright (c) 1982, 1986, 1988 Regents of the University of California.
|
|
* All rights reserved.
|
|
*
|
|
* Redistribution and use in source and binary forms, with or without
|
|
* modification, are permitted provided that the following conditions
|
|
* are met:
|
|
* 1. Redistributions of source code must retain the above copyright
|
|
* notice, this list of conditions and the following disclaimer.
|
|
* 2. Redistributions in binary form must reproduce the above copyright
|
|
* notice, this list of conditions and the following disclaimer in the
|
|
* documentation and/or other materials provided with the distribution.
|
|
* 3. All advertising materials mentioning features or use of this software
|
|
* must display the following acknowledgement:
|
|
* This product includes software developed by the University of
|
|
* California, Berkeley and its contributors.
|
|
* 4. Neither the name of the University nor the names of its contributors
|
|
* may be used to endorse or promote products derived from this software
|
|
* without specific prior written permission.
|
|
*
|
|
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
|
|
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
|
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
|
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
|
|
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
|
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
|
|
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
|
|
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
|
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
|
|
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
|
|
* SUCH DAMAGE.
|
|
*
|
|
* @(#)mbuf.h 8.3 (Berkeley) 1/21/94
|
|
*/
|
|
/*
|
|
* NOTICE: This file was modified by SPARTA, Inc. in 2005 to introduce
|
|
* support for mandatory and extensible security protections. This notice
|
|
* is included in support of clause 2.2 (b) of the Apple Public License,
|
|
* Version 2.0.
|
|
*/
|
|
|
|
#ifndef _SYS_MBUF_H_
|
|
#define _SYS_MBUF_H_
|
|
|
|
#include <sys/cdefs.h>
|
|
#include <sys/appleapiopts.h>
|
|
|
|
#ifdef XNU_KERNEL_PRIVATE
|
|
|
|
#include <sys/lock.h>
|
|
#include <sys/queue.h>
|
|
|
|
/*
|
|
* Mbufs are of a single size, MSIZE (machine/param.h), which
|
|
* includes overhead. An mbuf may add a single "mbuf cluster" of size
|
|
* MCLBYTES/MBIGCLBYTES/M16KCLBYTES (also in machine/param.h), which has
|
|
* no additional overhead and is used instead of the internal data area;
|
|
* this is done when at least MINCLSIZE of data must be stored.
|
|
*/
|
|
|
|
/*
|
|
* The following _MLEN and _MHLEN macros are private to xnu. Private code
|
|
* that are outside of xnu must use the mbuf_get_{mlen,mhlen} routines since
|
|
* the sizes of the structures are dependent upon specific xnu configs.
|
|
*/
|
|
#define _MLEN (MSIZE - sizeof(struct m_hdr)) /* normal data len */
|
|
#define _MHLEN (_MLEN - sizeof(struct pkthdr)) /* data len w/pkthdr */
|
|
|
|
#define NMBPBGSHIFT (MBIGCLSHIFT - MSIZESHIFT)
|
|
#define NMBPBG (1 << NMBPBGSHIFT) /* # of mbufs per big cl */
|
|
|
|
#define NCLPBGSHIFT (MBIGCLSHIFT - MCLSHIFT)
|
|
#define NCLPBG (1 << NCLPBGSHIFT) /* # of cl per big cl */
|
|
|
|
#define NMBPCLSHIFT (NMBPBGSHIFT - NCLPBGSHIFT)
|
|
#define NMBPCL (1 << NMBPCLSHIFT) /* # of mbufs per cl */
|
|
|
|
#define NCLPJCLSHIFT ((M16KCLSHIFT - MBIGCLSHIFT) + NCLPBGSHIFT)
|
|
#define NCLPJCL (1 << NCLPJCLSHIFT) /* # of cl per jumbo cl */
|
|
|
|
/*
|
|
* Macros for type conversion
|
|
* mtod(m,t) - convert mbuf pointer to data pointer of correct type
|
|
* dtom(x) - convert data pointer within mbuf to mbuf pointer (XXX)
|
|
*/
|
|
#define mtod(m, t) ((t)m_mtod(m))
|
|
#define dtom(x) m_dtom(x)
|
|
|
|
/* header at beginning of each mbuf: */
|
|
struct m_hdr {
|
|
struct mbuf *mh_next; /* next buffer in chain */
|
|
struct mbuf *mh_nextpkt; /* next chain in queue/record */
|
|
int32_t mh_len; /* amount of data in this mbuf */
|
|
caddr_t mh_data; /* location of data */
|
|
short mh_type; /* type of data in this mbuf */
|
|
short mh_flags; /* flags; see below */
|
|
};
|
|
|
|
/*
|
|
* Packet tag structure (see below for details).
|
|
*/
|
|
struct m_tag {
|
|
u_int64_t m_tag_cookie; /* Error checking */
|
|
#ifndef __LP64__
|
|
u_int32_t pad; /* For structure alignment */
|
|
#endif /* !__LP64__ */
|
|
SLIST_ENTRY(m_tag) m_tag_link; /* List of packet tags */
|
|
u_int16_t m_tag_type; /* Module specific type */
|
|
u_int16_t m_tag_len; /* Length of data */
|
|
u_int32_t m_tag_id; /* Module ID */
|
|
};
|
|
|
|
#define M_TAG_ALIGN(len) \
|
|
(P2ROUNDUP(len, sizeof (u_int64_t)) + sizeof (struct m_tag))
|
|
|
|
#define M_TAG_VALID_PATTERN 0xfeedfacefeedfaceULL
|
|
#define M_TAG_FREE_PATTERN 0xdeadbeefdeadbeefULL
|
|
|
|
/*
|
|
* Packet tag header structure (at the top of mbuf). Pointers are
|
|
* 32-bit in ILP32; m_tag needs 64-bit alignment, hence padded.
|
|
*/
|
|
struct m_taghdr {
|
|
#ifndef __LP64__
|
|
u_int32_t pad; /* For structure alignment */
|
|
#endif /* !__LP64__ */
|
|
u_int64_t refcnt; /* Number of tags in this mbuf */
|
|
};
|
|
|
|
/* Values for pftag_flags */
|
|
#define PF_TAG_GENERATED 0x000001 /* pkt generated by PF */
|
|
#define PF_TAG_FRAGCACHE 0x000002
|
|
#define PF_TAG_TRANSLATE_LOCALHOST 0x000004
|
|
#define PF_TAG_FLOWHASH 0x000100 /* valid flowhash value */
|
|
#define PF_TAG_HDR_INET 0x000200 /* hdr points to IPv4 */
|
|
#define PF_TAG_HDR_INET6 0x000400 /* hdr points to IPv6 */
|
|
#define PF_TAG_TCP 0x000800 /* payload is TCP */
|
|
#define PF_TAG_FLOWADV 0x010000 /* local flow advisory */
|
|
#define PF_TAG_QUEUE1 0x100000 /* queue-specific */
|
|
|
|
#define IF_PKTSEQ_SHIFT 4
|
|
|
|
/* PF mbuf tag */
|
|
struct pf_mtag {
|
|
void *pftag_hdr; /* saved hdr pos in mbuf, for ECN */
|
|
unsigned int pftag_rtableid; /* alternate routing table id */
|
|
union {
|
|
struct {
|
|
u_int32_t qid;
|
|
union {
|
|
u_int8_t val8[4];
|
|
u_int16_t val16[2];
|
|
u_int32_t val32;
|
|
} __qpriv_u; /* for queue-specific use */
|
|
} __pf_data;
|
|
u_int64_t pktseq;
|
|
} __pfifseq_u; /* Used for pf or interface bandwidth measurement */
|
|
#define pftag_qid __pfifseq_u.__pf_data.qid
|
|
#define pftag_qpriv8 __pfifseq_u.__pf_data.__qpriv_u.val8
|
|
#define pftag_qpriv16 __pfifseq_u.__pf_data.__qpriv_u.val16
|
|
#define pftag_qpriv32 __pfifseq_u.__pf_data.__qpriv_u.val32
|
|
#define pftag_pktseq __pfifseq_u.pktseq
|
|
u_int32_t pftag_flowhash;
|
|
u_int16_t pftag_tag;
|
|
u_int16_t pftag_routed;
|
|
u_int32_t pftag_flags; /* PF_TAG flags */
|
|
};
|
|
|
|
/* TCP specific mbuf tag */
|
|
struct tcp_mtag {
|
|
u_int tm_tso_segz; /* TSO segment size (actual MSS) */
|
|
u_int16_t tm_pktlen; /* LRO - max segment size encountered */
|
|
u_int16_t tm_npkts; /* LRO - number of coalesced TCP pkts */
|
|
};
|
|
|
|
/* record/packet header in first mbuf of chain; valid if M_PKTHDR set */
|
|
struct pkthdr {
|
|
int len; /* total packet length */
|
|
struct ifnet *rcvif; /* rcv interface */
|
|
|
|
/* variables for ip and tcp reassembly */
|
|
void *header; /* pointer to packet header */
|
|
/* variables for hardware checksum */
|
|
/* Note: csum_flags is used for hardware checksum and VLAN */
|
|
int csum_flags; /* flags regarding checksum */
|
|
int csum_data; /* data field used by csum routines */
|
|
u_short vlan_tag; /* VLAN tag, host byte order */
|
|
u_short socket_id; /* socket id */
|
|
SLIST_HEAD(packet_tags, m_tag) tags; /* list of packet tags */
|
|
struct pf_mtag pf_mtag; /* built-in PF tag */
|
|
#define m_flowhash pf_mtag.pftag_flowhash
|
|
#define m_fhflags pf_mtag.pftag_flags
|
|
u_int32_t svc; /* MBUF_SVC value */
|
|
u_int16_t vt_nrecs; /* # of IGMPv3/MLDv2 records */
|
|
u_int16_t aux_flags; /* auxiliary packet flags */
|
|
struct tcp_mtag tcp_mtag; /* tcp related data */
|
|
#define tso_segsz tcp_mtag.tm_tso_segz
|
|
#define lro_pktlen tcp_mtag.tm_pktlen
|
|
#define lro_npkts tcp_mtag.tm_npkts
|
|
};
|
|
|
|
/* description of external storage mapped into mbuf, valid if M_EXT set */
|
|
struct m_ext {
|
|
caddr_t ext_buf; /* start of buffer */
|
|
void (*ext_free)(caddr_t, u_int, caddr_t); /* free routine if not the usual */
|
|
u_int ext_size; /* size of buffer, for ext_free */
|
|
caddr_t ext_arg; /* additional ext_free argument */
|
|
struct ext_refsq { /* references held */
|
|
struct ext_refsq *forward, *backward;
|
|
} ext_refs;
|
|
struct ext_ref {
|
|
u_int32_t refcnt;
|
|
u_int32_t flags;
|
|
} *ext_refflags;
|
|
};
|
|
|
|
/* define m_ext to a type since it gets redefined below */
|
|
typedef struct m_ext _m_ext_t;
|
|
|
|
struct mbuf {
|
|
struct m_hdr m_hdr;
|
|
union {
|
|
struct {
|
|
struct pkthdr MH_pkthdr; /* M_PKTHDR set */
|
|
union {
|
|
struct m_ext MH_ext; /* M_EXT set */
|
|
char MH_databuf[_MHLEN];
|
|
} MH_dat;
|
|
} MH;
|
|
char M_databuf[_MLEN]; /* !M_PKTHDR, !M_EXT */
|
|
} M_dat;
|
|
};
|
|
|
|
#define m_next m_hdr.mh_next
|
|
#define m_len m_hdr.mh_len
|
|
#define m_data m_hdr.mh_data
|
|
#define m_type m_hdr.mh_type
|
|
#define m_flags m_hdr.mh_flags
|
|
#define m_nextpkt m_hdr.mh_nextpkt
|
|
#define m_act m_nextpkt
|
|
#define m_pkthdr M_dat.MH.MH_pkthdr
|
|
#define m_ext M_dat.MH.MH_dat.MH_ext
|
|
#define m_pktdat M_dat.MH.MH_dat.MH_databuf
|
|
#define m_dat M_dat.M_databuf
|
|
#define m_pktlen(_m) ((_m)->m_pkthdr.len)
|
|
#define m_pftag(_m) (&(_m)->m_pkthdr.pf_mtag)
|
|
|
|
/* mbuf flags (private) */
|
|
#define M_EXT 0x0001 /* has associated external storage */
|
|
#define M_PKTHDR 0x0002 /* start of record */
|
|
#define M_EOR 0x0004 /* end of record */
|
|
#define M_PROTO1 0x0008 /* protocol-specific */
|
|
#define M_PROTO2 0x0010 /* protocol-specific */
|
|
#define M_PROTO3 0x0020 /* protocol-specific */
|
|
#define M_LOOP 0x0040 /* packet is looped back */
|
|
#define M_PROTO5 0x0080 /* protocol-specific */
|
|
|
|
/* mbuf pkthdr flags, also in m_flags (private) */
|
|
#define M_BCAST 0x0100 /* send/received as link-level broadcast */
|
|
#define M_MCAST 0x0200 /* send/received as link-level multicast */
|
|
#define M_FRAG 0x0400 /* packet is a fragment of a larger packet */
|
|
#define M_FIRSTFRAG 0x0800 /* packet is first fragment */
|
|
#define M_LASTFRAG 0x1000 /* packet is last fragment */
|
|
#define M_PROMISC 0x2000 /* packet is promiscuous (shouldn't go to stack) */
|
|
#define M_HASFCS 0x4000 /* packet has FCS */
|
|
#define M_TAGHDR 0x8000 /* m_tag hdr structure at top of mbuf data */
|
|
|
|
/*
|
|
* Flags to purge when crossing layers.
|
|
*/
|
|
#define M_PROTOFLAGS \
|
|
(M_PROTO1|M_PROTO2|M_PROTO3|M_PROTO5)
|
|
|
|
/* flags copied when copying m_pkthdr */
|
|
#define M_COPYFLAGS \
|
|
(M_PKTHDR|M_EOR|M_PROTO1|M_PROTO2|M_PROTO3 | \
|
|
M_LOOP|M_PROTO5|M_BCAST|M_MCAST|M_FRAG | \
|
|
M_FIRSTFRAG|M_LASTFRAG|M_PROMISC|M_HASFCS)
|
|
|
|
/* flags indicating hw checksum support and sw checksum requirements */
|
|
#define CSUM_IP 0x0001 /* will csum IP */
|
|
#define CSUM_TCP 0x0002 /* will csum TCP */
|
|
#define CSUM_UDP 0x0004 /* will csum UDP */
|
|
#define CSUM_IP_FRAGS 0x0008 /* will csum IP fragments */
|
|
#define CSUM_FRAGMENT 0x0010 /* will do IP fragmentation */
|
|
#define CSUM_TCPIPV6 0x0020 /* will csum TCP for IPv6 */
|
|
#define CSUM_UDPIPV6 0x0040 /* will csum UDP for IPv6 */
|
|
#define CSUM_FRAGMENT_IPV6 0x0080 /* will do IPv6 fragmentation */
|
|
|
|
#define CSUM_IP_CHECKED 0x0100 /* did csum IP */
|
|
#define CSUM_IP_VALID 0x0200 /* ... the csum is valid */
|
|
#define CSUM_DATA_VALID 0x0400 /* csum_data field is valid */
|
|
#define CSUM_PSEUDO_HDR 0x0800 /* csum_data has pseudo hdr */
|
|
#define CSUM_TCP_SUM16 0x1000 /* simple TCP Sum16 computation */
|
|
|
|
#define CSUM_DELAY_DATA (CSUM_TCP | CSUM_UDP)
|
|
#define CSUM_DELAY_IP (CSUM_IP) /* IPv4 only: no IPv6 IP cksum */
|
|
#define CSUM_DELAY_IPV6_DATA (CSUM_TCPIPV6 | CSUM_UDPIPV6)
|
|
#define CSUM_DATA_IPV6_VALID CSUM_DATA_VALID /* csum_data field is valid */
|
|
/*
|
|
* Note: see also IF_HWASSIST_CSUM defined in <net/if_var.h>
|
|
*/
|
|
/* bottom 16 bits reserved for hardware checksum */
|
|
#define CSUM_CHECKSUM_MASK 0xffff
|
|
|
|
/* VLAN tag present */
|
|
#define CSUM_VLAN_TAG_VALID 0x10000 /* vlan_tag field is valid */
|
|
|
|
/* TCP Segment Offloading requested on this mbuf */
|
|
#define CSUM_TSO_IPV4 0x100000 /* This mbuf needs to be segmented by the NIC */
|
|
#define CSUM_TSO_IPV6 0x200000 /* This mbuf needs to be segmented by the NIC */
|
|
|
|
/*
|
|
* Auxiliary packet flags. Unlike m_flags, all auxiliary flags are copied
|
|
* along when copying m_pkthdr, i.e. no equivalent of M_COPYFLAGS here.
|
|
* Note that this flag is 16-bit wide.
|
|
*/
|
|
#define MAUXF_PRIO_PRIVILEGED 0x0001 /* packet priority is privileged */
|
|
#define MAUXF_PROXY_DST 0x0002 /* processed but not locally destined */
|
|
#define MAUXF_INET_RESOLVE_RTR 0x0004 /* pkt is for resolving IPv4 router */
|
|
#define MAUXF_INET6_RESOLVE_RTR 0x0008 /* pkt is for resolving IPv6 router */
|
|
#define MAUXF_SW_LRO_PKT 0x0010 /* pkt is a large coalesced pkt */
|
|
#define MAUXF_SW_LRO_DID_CSUM 0x0020 /* IP and TCP checksums done by LRO*/
|
|
#endif /* XNU_KERNEL_PRIVATE */
|
|
|
|
/* mbuf types */
|
|
#define MT_FREE 0 /* should be on free list */
|
|
#define MT_DATA 1 /* dynamic (data) allocation */
|
|
#define MT_HEADER 2 /* packet header */
|
|
#define MT_SOCKET 3 /* socket structure */
|
|
#define MT_PCB 4 /* protocol control block */
|
|
#define MT_RTABLE 5 /* routing tables */
|
|
#define MT_HTABLE 6 /* IMP host tables */
|
|
#define MT_ATABLE 7 /* address resolution tables */
|
|
#define MT_SONAME 8 /* socket name */
|
|
#define MT_SOOPTS 10 /* socket options */
|
|
#define MT_FTABLE 11 /* fragment reassembly header */
|
|
#define MT_RIGHTS 12 /* access rights */
|
|
#define MT_IFADDR 13 /* interface address */
|
|
#define MT_CONTROL 14 /* extra-data protocol message */
|
|
#define MT_OOBDATA 15 /* expedited data */
|
|
#define MT_TAG 16 /* volatile metadata associated to pkts */
|
|
#define MT_MAX 32 /* enough? */
|
|
|
|
#ifdef XNU_KERNEL_PRIVATE
|
|
/*
|
|
* mbuf allocation/deallocation macros:
|
|
*
|
|
* MGET(struct mbuf *m, int how, int type)
|
|
* allocates an mbuf and initializes it to contain internal data.
|
|
*
|
|
* MGETHDR(struct mbuf *m, int how, int type)
|
|
* allocates an mbuf and initializes it to contain a packet header
|
|
* and internal data.
|
|
*/
|
|
|
|
#if 1
|
|
#define MCHECK(m) m_mcheck(m)
|
|
#else
|
|
#define MCHECK(m)
|
|
#endif
|
|
|
|
#define MGET(m, how, type) ((m) = m_get((how), (type)))
|
|
|
|
#define MGETHDR(m, how, type) ((m) = m_gethdr((how), (type)))
|
|
|
|
/*
|
|
* Mbuf cluster macros.
|
|
* MCLALLOC(caddr_t p, int how) allocates an mbuf cluster.
|
|
* MCLGET adds such clusters to a normal mbuf;
|
|
* the flag M_EXT is set upon success.
|
|
* MCLFREE releases a reference to a cluster allocated by MCLALLOC,
|
|
* freeing the cluster if the reference count has reached 0.
|
|
*
|
|
* Normal mbuf clusters are normally treated as character arrays
|
|
* after allocation, but use the first word of the buffer as a free list
|
|
* pointer while on the free list.
|
|
*/
|
|
union mcluster {
|
|
union mcluster *mcl_next;
|
|
char mcl_buf[MCLBYTES];
|
|
};
|
|
|
|
#define MCLALLOC(p, how) ((p) = m_mclalloc(how))
|
|
|
|
#define MCLFREE(p) m_mclfree(p)
|
|
|
|
#define MCLGET(m, how) ((m) = m_mclget(m, how))
|
|
|
|
/*
|
|
* Mbuf big cluster
|
|
*/
|
|
union mbigcluster {
|
|
union mbigcluster *mbc_next;
|
|
char mbc_buf[MBIGCLBYTES];
|
|
};
|
|
|
|
/*
|
|
* Mbuf jumbo cluster
|
|
*/
|
|
union m16kcluster {
|
|
union m16kcluster *m16kcl_next;
|
|
char m16kcl_buf[M16KCLBYTES];
|
|
};
|
|
|
|
#define MCLHASREFERENCE(m) m_mclhasreference(m)
|
|
|
|
/*
|
|
* MFREE(struct mbuf *m, struct mbuf *n)
|
|
* Free a single mbuf and associated external storage.
|
|
* Place the successor, if any, in n.
|
|
*/
|
|
|
|
#define MFREE(m, n) ((n) = m_free(m))
|
|
|
|
/*
|
|
* Copy mbuf pkthdr from from to to.
|
|
* from must have M_PKTHDR set, and to must be empty.
|
|
* aux pointer will be moved to `to'.
|
|
*/
|
|
#define M_COPY_PKTHDR(to, from) m_copy_pkthdr(to, from)
|
|
|
|
#define M_COPY_PFTAG(to, from) m_copy_pftag(to, from)
|
|
|
|
/*
|
|
* Set the m_data pointer of a newly-allocated mbuf (m_get/MGET) to place
|
|
* an object of the specified size at the end of the mbuf, longword aligned.
|
|
*/
|
|
#define M_ALIGN(m, len) \
|
|
do { \
|
|
(m)->m_data += (MLEN - (len)) &~ (sizeof (long) - 1); \
|
|
} while (0)
|
|
|
|
/*
|
|
* As above, for mbufs allocated with m_gethdr/MGETHDR
|
|
* or initialized by M_COPY_PKTHDR.
|
|
*/
|
|
#define MH_ALIGN(m, len) \
|
|
do { \
|
|
(m)->m_data += (MHLEN - (len)) &~ (sizeof (long) - 1); \
|
|
} while (0)
|
|
|
|
/*
|
|
* Compute the amount of space available
|
|
* before the current start of data in an mbuf.
|
|
* Subroutine - data not available if certain references.
|
|
*/
|
|
#define M_LEADINGSPACE(m) m_leadingspace(m)
|
|
|
|
/*
|
|
* Compute the amount of space available
|
|
* after the end of data in an mbuf.
|
|
* Subroutine - data not available if certain references.
|
|
*/
|
|
#define M_TRAILINGSPACE(m) m_trailingspace(m)
|
|
|
|
/*
|
|
* Arrange to prepend space of size plen to mbuf m.
|
|
* If a new mbuf must be allocated, how specifies whether to wait.
|
|
* If how is M_DONTWAIT and allocation fails, the original mbuf chain
|
|
* is freed and m is set to NULL.
|
|
*/
|
|
#define M_PREPEND(m, plen, how) ((m) = m_prepend_2((m), (plen), (how)))
|
|
|
|
/* change mbuf to new type */
|
|
#define MCHTYPE(m, t) m_mchtype(m, t)
|
|
|
|
/* compatiblity with 4.3 */
|
|
#define m_copy(m, o, l) m_copym((m), (o), (l), M_DONTWAIT)
|
|
|
|
#define MBSHIFT 20 /* 1MB */
|
|
#define MBSIZE (1 << MBSHIFT)
|
|
#define GBSHIFT 30 /* 1GB */
|
|
#define GBSIZE (1 << GBSHIFT)
|
|
|
|
/*
|
|
* M_STRUCT_GET ensures that intermediate protocol header (from "off" to
|
|
* "off+len") is located in single mbuf, on contiguous memory region.
|
|
* The pointer to the region will be returned to pointer variable "val",
|
|
* with type "typ".
|
|
*
|
|
* M_STRUCT_GET0 does the same, except that it aligns the structure at
|
|
* very top of mbuf. GET0 is likely to make memory copy than GET.
|
|
*/
|
|
#define M_STRUCT_GET(val, typ, m, off, len) \
|
|
do { \
|
|
struct mbuf *t; \
|
|
int tmp; \
|
|
\
|
|
if ((m)->m_len >= (off) + (len)) { \
|
|
(val) = (typ)(mtod((m), caddr_t) + (off)); \
|
|
} else { \
|
|
t = m_pulldown((m), (off), (len), &tmp); \
|
|
if (t != NULL) { \
|
|
if (t->m_len < tmp + (len)) \
|
|
panic("m_pulldown malfunction"); \
|
|
(val) = (typ)(mtod(t, caddr_t) + tmp); \
|
|
} else { \
|
|
(val) = (typ)NULL; \
|
|
(m) = NULL; \
|
|
} \
|
|
} \
|
|
} while (0)
|
|
|
|
#define M_STRUCT_GET0(val, typ, m, off, len) \
|
|
do { \
|
|
struct mbuf *t; \
|
|
\
|
|
if ((off) == 0 && ((m)->m_len >= (len))) { \
|
|
(val) = (typ)(void *)mtod(m, caddr_t); \
|
|
} else { \
|
|
t = m_pulldown((m), (off), (len), NULL); \
|
|
if (t != NULL) { \
|
|
if (t->m_len < (len)) \
|
|
panic("m_pulldown malfunction"); \
|
|
(val) = (typ)(void *)mtod(t, caddr_t); \
|
|
} else { \
|
|
(val) = (typ)NULL; \
|
|
(m) = NULL; \
|
|
} \
|
|
} \
|
|
} while (0)
|
|
|
|
#define MBUF_INPUT_CHECK(m, rcvif) \
|
|
do { \
|
|
if (!(m->m_flags & MBUF_PKTHDR) || \
|
|
m->m_len < 0 || \
|
|
m->m_len > ((njcl > 0) ? njclbytes : MBIGCLBYTES) || \
|
|
m->m_type == MT_FREE || \
|
|
((m->m_flags & M_EXT) != 0 && m->m_ext.ext_buf == NULL)) { \
|
|
panic_plain("Failed mbuf validity check: mbuf %p len %d " \
|
|
"type %d flags 0x%x data %p rcvif %s%d ifflags 0x%x", \
|
|
m, m->m_len, m->m_type, m->m_flags, \
|
|
((m->m_flags & M_EXT) ? m->m_ext.ext_buf : m->m_data), \
|
|
rcvif->if_name, rcvif->if_unit, \
|
|
(rcvif->if_flags & 0xffff)); \
|
|
} \
|
|
} while (0)
|
|
|
|
/*
|
|
* Simple mbuf queueing system
|
|
*
|
|
* This is basically a SIMPLEQ adapted to mbuf use (i.e. using
|
|
* m_nextpkt instead of field.sqe_next).
|
|
*
|
|
* m_next is ignored, so queueing chains of mbufs is possible
|
|
*/
|
|
#define MBUFQ_HEAD(name) \
|
|
struct name { \
|
|
struct mbuf *mq_first; /* first packet */ \
|
|
struct mbuf **mq_last; /* addr of last next packet */ \
|
|
}
|
|
|
|
#define MBUFQ_INIT(q) do { \
|
|
MBUFQ_FIRST(q) = NULL; \
|
|
(q)->mq_last = &MBUFQ_FIRST(q); \
|
|
} while (0)
|
|
|
|
#define MBUFQ_PREPEND(q, m) do { \
|
|
if ((MBUFQ_NEXT(m) = MBUFQ_FIRST(q)) == NULL) \
|
|
(q)->mq_last = &MBUFQ_NEXT(m); \
|
|
MBUFQ_FIRST(q) = (m); \
|
|
} while (0)
|
|
|
|
#define MBUFQ_ENQUEUE(q, m) do { \
|
|
MBUFQ_NEXT(m) = NULL; \
|
|
*(q)->mq_last = (m); \
|
|
(q)->mq_last = &MBUFQ_NEXT(m); \
|
|
} while (0)
|
|
|
|
#define MBUFQ_ENQUEUE_MULTI(q, m, n) do { \
|
|
MBUFQ_NEXT(n) = NULL; \
|
|
*(q)->mq_last = (m); \
|
|
(q)->mq_last = &MBUFQ_NEXT(n); \
|
|
} while (0)
|
|
|
|
#define MBUFQ_DEQUEUE(q, m) do { \
|
|
if (((m) = MBUFQ_FIRST(q)) != NULL) { \
|
|
if ((MBUFQ_FIRST(q) = MBUFQ_NEXT(m)) == NULL) \
|
|
(q)->mq_last = &MBUFQ_FIRST(q); \
|
|
else \
|
|
MBUFQ_NEXT(m) = NULL; \
|
|
} \
|
|
} while (0)
|
|
|
|
#define MBUFQ_REMOVE(q, m) do { \
|
|
if (MBUFQ_FIRST(q) == (m)) { \
|
|
MBUFQ_DEQUEUE(q, m); \
|
|
} else { \
|
|
struct mbuf *_m = MBUFQ_FIRST(q); \
|
|
while (MBUFQ_NEXT(_m) != (m)) \
|
|
_m = MBUFQ_NEXT(_m); \
|
|
if ((MBUFQ_NEXT(_m) = \
|
|
MBUFQ_NEXT(MBUFQ_NEXT(_m))) == NULL) \
|
|
(q)->mq_last = &MBUFQ_NEXT(_m); \
|
|
} \
|
|
} while (0)
|
|
|
|
#define MBUFQ_DRAIN(q) do { \
|
|
struct mbuf *__m0; \
|
|
while ((__m0 = MBUFQ_FIRST(q)) != NULL) { \
|
|
MBUFQ_FIRST(q) = MBUFQ_NEXT(__m0); \
|
|
MBUFQ_NEXT(__m0) = NULL; \
|
|
m_freem(__m0); \
|
|
} \
|
|
(q)->mq_last = &MBUFQ_FIRST(q); \
|
|
} while (0)
|
|
|
|
#define MBUFQ_FOREACH(m, q) \
|
|
for ((m) = MBUFQ_FIRST(q); \
|
|
(m); \
|
|
(m) = MBUFQ_NEXT(m))
|
|
|
|
#define MBUFQ_FOREACH_SAFE(m, q, tvar) \
|
|
for ((m) = MBUFQ_FIRST(q); \
|
|
(m) && ((tvar) = MBUFQ_NEXT(m), 1); \
|
|
(m) = (tvar))
|
|
|
|
#define MBUFQ_EMPTY(q) ((q)->mq_first == NULL)
|
|
#define MBUFQ_FIRST(q) ((q)->mq_first)
|
|
#define MBUFQ_NEXT(m) ((m)->m_nextpkt)
|
|
#define MBUFQ_LAST(q) (*(q)->mq_last)
|
|
|
|
#define max_linkhdr P2ROUNDUP(_max_linkhdr, sizeof (u_int32_t))
|
|
#define max_protohdr P2ROUNDUP(_max_protohdr, sizeof (u_int32_t))
|
|
#endif /* XNU_KERNEL_PRIVATE */
|
|
|
|
/*
|
|
* Mbuf statistics (legacy).
|
|
*/
|
|
struct mbstat {
|
|
u_int32_t m_mbufs; /* mbufs obtained from page pool */
|
|
u_int32_t m_clusters; /* clusters obtained from page pool */
|
|
u_int32_t m_spare; /* spare field */
|
|
u_int32_t m_clfree; /* free clusters */
|
|
u_int32_t m_drops; /* times failed to find space */
|
|
u_int32_t m_wait; /* times waited for space */
|
|
u_int32_t m_drain; /* times drained protocols for space */
|
|
u_short m_mtypes[256]; /* type specific mbuf allocations */
|
|
u_int32_t m_mcfail; /* times m_copym failed */
|
|
u_int32_t m_mpfail; /* times m_pullup failed */
|
|
u_int32_t m_msize; /* length of an mbuf */
|
|
u_int32_t m_mclbytes; /* length of an mbuf cluster */
|
|
u_int32_t m_minclsize; /* min length of data to allocate a cluster */
|
|
u_int32_t m_mlen; /* length of data in an mbuf */
|
|
u_int32_t m_mhlen; /* length of data in a header mbuf */
|
|
u_int32_t m_bigclusters; /* clusters obtained from page pool */
|
|
u_int32_t m_bigclfree; /* free clusters */
|
|
u_int32_t m_bigmclbytes; /* length of an mbuf cluster */
|
|
};
|
|
|
|
/* Compatibillity with 10.3 */
|
|
struct ombstat {
|
|
u_int32_t m_mbufs; /* mbufs obtained from page pool */
|
|
u_int32_t m_clusters; /* clusters obtained from page pool */
|
|
u_int32_t m_spare; /* spare field */
|
|
u_int32_t m_clfree; /* free clusters */
|
|
u_int32_t m_drops; /* times failed to find space */
|
|
u_int32_t m_wait; /* times waited for space */
|
|
u_int32_t m_drain; /* times drained protocols for space */
|
|
u_short m_mtypes[256]; /* type specific mbuf allocations */
|
|
u_int32_t m_mcfail; /* times m_copym failed */
|
|
u_int32_t m_mpfail; /* times m_pullup failed */
|
|
u_int32_t m_msize; /* length of an mbuf */
|
|
u_int32_t m_mclbytes; /* length of an mbuf cluster */
|
|
u_int32_t m_minclsize; /* min length of data to allocate a cluster */
|
|
u_int32_t m_mlen; /* length of data in an mbuf */
|
|
u_int32_t m_mhlen; /* length of data in a header mbuf */
|
|
};
|
|
|
|
/*
|
|
* mbuf class statistics.
|
|
*/
|
|
#define MAX_MBUF_CNAME 15
|
|
|
|
#if defined(XNU_KERNEL_PRIVATE)
|
|
/* For backwards compatibility with 32-bit userland process */
|
|
struct omb_class_stat {
|
|
char mbcl_cname[MAX_MBUF_CNAME + 1]; /* class name */
|
|
u_int32_t mbcl_size; /* buffer size */
|
|
u_int32_t mbcl_total; /* # of buffers created */
|
|
u_int32_t mbcl_active; /* # of active buffers */
|
|
u_int32_t mbcl_infree; /* # of available buffers */
|
|
u_int32_t mbcl_slab_cnt; /* # of available slabs */
|
|
u_int64_t mbcl_alloc_cnt; /* # of times alloc is called */
|
|
u_int64_t mbcl_free_cnt; /* # of times free is called */
|
|
u_int64_t mbcl_notified; /* # of notified wakeups */
|
|
u_int64_t mbcl_purge_cnt; /* # of purges so far */
|
|
u_int64_t mbcl_fail_cnt; /* # of allocation failures */
|
|
u_int32_t mbcl_ctotal; /* total only for this class */
|
|
/*
|
|
* Cache layer statistics
|
|
*/
|
|
u_int32_t mbcl_mc_state; /* cache state (see below) */
|
|
u_int32_t mbcl_mc_cached; /* # of cached buffers */
|
|
u_int32_t mbcl_mc_waiter_cnt; /* # waiters on the cache */
|
|
u_int32_t mbcl_mc_wretry_cnt; /* # of wait retries */
|
|
u_int32_t mbcl_mc_nwretry_cnt; /* # of no-wait retry attempts */
|
|
u_int64_t mbcl_reserved[4]; /* for future use */
|
|
} __attribute__((__packed__));
|
|
#endif /* XNU_KERNEL_PRIVATE */
|
|
|
|
typedef struct mb_class_stat {
|
|
char mbcl_cname[MAX_MBUF_CNAME + 1]; /* class name */
|
|
u_int32_t mbcl_size; /* buffer size */
|
|
u_int32_t mbcl_total; /* # of buffers created */
|
|
u_int32_t mbcl_active; /* # of active buffers */
|
|
u_int32_t mbcl_infree; /* # of available buffers */
|
|
u_int32_t mbcl_slab_cnt; /* # of available slabs */
|
|
#if defined(KERNEL) || defined(__LP64__)
|
|
u_int32_t mbcl_pad; /* padding */
|
|
#endif /* KERNEL || __LP64__ */
|
|
u_int64_t mbcl_alloc_cnt; /* # of times alloc is called */
|
|
u_int64_t mbcl_free_cnt; /* # of times free is called */
|
|
u_int64_t mbcl_notified; /* # of notified wakeups */
|
|
u_int64_t mbcl_purge_cnt; /* # of purges so far */
|
|
u_int64_t mbcl_fail_cnt; /* # of allocation failures */
|
|
u_int32_t mbcl_ctotal; /* total only for this class */
|
|
/*
|
|
* Cache layer statistics
|
|
*/
|
|
u_int32_t mbcl_mc_state; /* cache state (see below) */
|
|
u_int32_t mbcl_mc_cached; /* # of cached buffers */
|
|
u_int32_t mbcl_mc_waiter_cnt; /* # waiters on the cache */
|
|
u_int32_t mbcl_mc_wretry_cnt; /* # of wait retries */
|
|
u_int32_t mbcl_mc_nwretry_cnt; /* # of no-wait retry attempts */
|
|
u_int64_t mbcl_reserved[4]; /* for future use */
|
|
} mb_class_stat_t;
|
|
|
|
#define MCS_DISABLED 0 /* cache is permanently disabled */
|
|
#define MCS_ONLINE 1 /* cache is online */
|
|
#define MCS_PURGING 2 /* cache is being purged */
|
|
#define MCS_OFFLINE 3 /* cache is offline (resizing) */
|
|
|
|
#if defined(XNU_KERNEL_PRIVATE)
|
|
/* For backwards compatibility with 32-bit userland process */
|
|
struct omb_stat {
|
|
u_int32_t mbs_cnt; /* number of classes */
|
|
struct omb_class_stat mbs_class[1]; /* class array */
|
|
} __attribute__((__packed__));
|
|
#endif /* XNU_KERNEL_PRIVATE */
|
|
|
|
typedef struct mb_stat {
|
|
u_int32_t mbs_cnt; /* number of classes */
|
|
#if defined(KERNEL) || defined(__LP64__)
|
|
u_int32_t mbs_pad; /* padding */
|
|
#endif /* KERNEL || __LP64__ */
|
|
mb_class_stat_t mbs_class[1]; /* class array */
|
|
} mb_stat_t;
|
|
|
|
#ifdef PRIVATE
|
|
#define MLEAK_STACK_DEPTH 16 /* Max PC stack depth */
|
|
|
|
typedef struct mleak_trace_stat {
|
|
u_int64_t mltr_collisions;
|
|
u_int64_t mltr_hitcount;
|
|
u_int64_t mltr_allocs;
|
|
u_int64_t mltr_depth;
|
|
u_int64_t mltr_addr[MLEAK_STACK_DEPTH];
|
|
} mleak_trace_stat_t;
|
|
|
|
typedef struct mleak_stat {
|
|
u_int32_t ml_isaddr64; /* 64-bit KVA? */
|
|
u_int32_t ml_cnt; /* number of traces */
|
|
mleak_trace_stat_t ml_trace[1]; /* trace array */
|
|
} mleak_stat_t;
|
|
|
|
struct mleak_table {
|
|
u_int32_t mleak_capture; /* sampling capture counter */
|
|
u_int32_t mleak_sample_factor; /* sample factor */
|
|
|
|
/* Times two active records want to occupy the same spot */
|
|
u_int64_t alloc_collisions;
|
|
u_int64_t trace_collisions;
|
|
|
|
/* Times new record lands on spot previously occupied by freed alloc */
|
|
u_int64_t alloc_overwrites;
|
|
u_int64_t trace_overwrites;
|
|
|
|
/* Times a new alloc or trace is put into the hash table */
|
|
u_int64_t alloc_recorded;
|
|
u_int64_t trace_recorded;
|
|
|
|
/* Total number of outstanding allocs */
|
|
u_int64_t outstanding_allocs;
|
|
|
|
/* Times mleak_log returned false because couldn't acquire the lock */
|
|
u_int64_t total_conflicts;
|
|
};
|
|
#endif /* PRIVATE */
|
|
|
|
#ifdef KERNEL_PRIVATE
|
|
__BEGIN_DECLS
|
|
|
|
/*
|
|
* Exported (private)
|
|
*/
|
|
|
|
extern struct mbstat mbstat; /* statistics */
|
|
|
|
__END_DECLS
|
|
#endif /* KERNEL_PRIVATE */
|
|
|
|
#ifdef XNU_KERNEL_PRIVATE
|
|
__BEGIN_DECLS
|
|
|
|
/*
|
|
* Not exported (xnu private)
|
|
*/
|
|
|
|
/* flags to m_get/MGET */
|
|
/* Need to include malloc.h to get right options for malloc */
|
|
#include <sys/malloc.h>
|
|
|
|
struct mbuf;
|
|
|
|
/* length to m_copy to copy all */
|
|
#define M_COPYALL 1000000000
|
|
|
|
#define M_DONTWAIT M_NOWAIT
|
|
#define M_WAIT M_WAITOK
|
|
|
|
/*
|
|
* These macros are mapped to the appropriate KPIs, so that private code
|
|
* can be simply recompiled in order to be forward-compatible with future
|
|
* changes toward the struture sizes.
|
|
*/
|
|
#define MLEN mbuf_get_mlen() /* normal data len */
|
|
#define MHLEN mbuf_get_mhlen() /* data len w/pkthdr */
|
|
|
|
#define MINCLSIZE mbuf_get_minclsize() /* cluster usage threshold */
|
|
|
|
extern void m_freem(struct mbuf *);
|
|
extern char *mcl_to_paddr(char *);
|
|
extern void m_adj(struct mbuf *, int);
|
|
extern void m_cat(struct mbuf *, struct mbuf *);
|
|
extern void m_copydata(struct mbuf *, int, int, void *);
|
|
extern struct mbuf *m_copym(struct mbuf *, int, int, int);
|
|
extern struct mbuf *m_get(int, int);
|
|
extern struct mbuf *m_gethdr(int, int);
|
|
extern struct mbuf *m_getpacket(void);
|
|
extern struct mbuf *m_getpackets(int, int, int);
|
|
extern struct mbuf *m_mclget(struct mbuf *, int);
|
|
extern void *m_mtod(struct mbuf *);
|
|
extern struct mbuf *m_prepend_2(struct mbuf *, int, int);
|
|
extern struct mbuf *m_pullup(struct mbuf *, int);
|
|
extern struct mbuf *m_split(struct mbuf *, int, int);
|
|
extern void m_mclfree(caddr_t p);
|
|
|
|
/*
|
|
* On platforms which require strict alignment (currently for anything but
|
|
* i386 or x86_64), this macro checks whether the data pointer of an mbuf
|
|
* is 32-bit aligned (this is the expected minimum alignment for protocol
|
|
* headers), and assert otherwise.
|
|
*/
|
|
#if defined(__i386__) || defined(__x86_64__)
|
|
#define MBUF_STRICT_DATA_ALIGNMENT_CHECK_32(_m)
|
|
#else /* !__i386__ && !__x86_64__ */
|
|
#define MBUF_STRICT_DATA_ALIGNMENT_CHECK_32(_m) do { \
|
|
if (!IS_P2ALIGNED((_m)->m_data, sizeof (u_int32_t))) { \
|
|
if (((_m)->m_flags & M_PKTHDR) && \
|
|
(_m)->m_pkthdr.rcvif != NULL) { \
|
|
panic_plain("\n%s: mbuf %p data ptr %p is not " \
|
|
"32-bit aligned [%s%d: alignerrs=%lld]\n", \
|
|
__func__, (_m), (_m)->m_data, \
|
|
(_m)->m_pkthdr.rcvif->if_name, \
|
|
(_m)->m_pkthdr.rcvif->if_unit, \
|
|
(_m)->m_pkthdr.rcvif->if_alignerrs); \
|
|
} else { \
|
|
panic_plain("\n%s: mbuf %p data ptr %p is not " \
|
|
"32-bit aligned\n", \
|
|
__func__, (_m), (_m)->m_data); \
|
|
} \
|
|
} \
|
|
} while (0)
|
|
#endif /* !__i386__ && !__x86_64__ */
|
|
|
|
/* Maximum number of MBUF_SC values (excluding MBUF_SC_UNSPEC) */
|
|
#define MBUF_SC_MAX_CLASSES 10
|
|
|
|
/*
|
|
* These conversion macros rely on the corresponding MBUF_SC and
|
|
* MBUF_TC values in order to establish the following mapping:
|
|
*
|
|
* MBUF_SC_BK_SYS ] ==> MBUF_TC_BK
|
|
* MBUF_SC_BK ]
|
|
*
|
|
* MBUF_SC_BE ] ==> MBUF_TC_BE
|
|
* MBUF_SC_RD ]
|
|
* MBUF_SC_OAM ]
|
|
*
|
|
* MBUF_SC_AV ] ==> MBUF_TC_VI
|
|
* MBUF_SC_RV ]
|
|
* MBUF_SC_VI ]
|
|
*
|
|
* MBUF_SC_VO ] ==> MBUF_TC_VO
|
|
* MBUF_SC_CTL ]
|
|
*
|
|
* The values assigned to each service class allows for a fast mapping to
|
|
* the corresponding MBUF_TC traffic class values, as well as to retrieve the
|
|
* assigned index; therefore care must be taken when comparing against these
|
|
* values. Use the corresponding class and index macros to retrieve the
|
|
* corresponding portion, and never assume that a higher class corresponds
|
|
* to a higher index.
|
|
*/
|
|
#define MBUF_SCVAL(x) ((x) & 0xffff)
|
|
#define MBUF_SCIDX(x) ((((x) >> 16) & 0xff) >> 3)
|
|
#define MBUF_SC2TC(_sc) (MBUF_SCVAL(_sc) >> 7)
|
|
#define MBUF_TC2SCVAL(_tc) ((_tc) << 7)
|
|
#define IS_MBUF_SC_BACKGROUND(_sc) (((_sc) == MBUF_SC_BK_SYS) || \
|
|
((_sc) == MBUF_SC_BK))
|
|
|
|
#define SCIDX_BK_SYS MBUF_SCIDX(MBUF_SC_BK_SYS)
|
|
#define SCIDX_BK MBUF_SCIDX(MBUF_SC_BK)
|
|
#define SCIDX_BE MBUF_SCIDX(MBUF_SC_BE)
|
|
#define SCIDX_RD MBUF_SCIDX(MBUF_SC_RD)
|
|
#define SCIDX_OAM MBUF_SCIDX(MBUF_SC_OAM)
|
|
#define SCIDX_AV MBUF_SCIDX(MBUF_SC_AV)
|
|
#define SCIDX_RV MBUF_SCIDX(MBUF_SC_RV)
|
|
#define SCIDX_VI MBUF_SCIDX(MBUF_SC_VI)
|
|
#define SCIDX_VO MBUF_SCIDX(MBUF_SC_VO)
|
|
#define SCIDX_CTL MBUF_SCIDX(MBUF_SC_CTL)
|
|
|
|
#define SCVAL_BK_SYS MBUF_SCVAL(MBUF_SC_BK_SYS)
|
|
#define SCVAL_BK MBUF_SCVAL(MBUF_SC_BK)
|
|
#define SCVAL_BE MBUF_SCVAL(MBUF_SC_BE)
|
|
#define SCVAL_RD MBUF_SCVAL(MBUF_SC_RD)
|
|
#define SCVAL_OAM MBUF_SCVAL(MBUF_SC_OAM)
|
|
#define SCVAL_AV MBUF_SCVAL(MBUF_SC_AV)
|
|
#define SCVAL_RV MBUF_SCVAL(MBUF_SC_RV)
|
|
#define SCVAL_VI MBUF_SCVAL(MBUF_SC_VI)
|
|
#define SCVAL_VO MBUF_SCVAL(MBUF_SC_VO)
|
|
#define SCVAL_CTL MBUF_SCVAL(MBUF_SC_CTL)
|
|
|
|
#define MBUF_VALID_SC(c) \
|
|
(c == MBUF_SC_BK_SYS || c == MBUF_SC_BK || c == MBUF_SC_BE || \
|
|
c == MBUF_SC_RD || c == MBUF_SC_OAM || c == MBUF_SC_AV || \
|
|
c == MBUF_SC_RV || c == MBUF_SC_VI || c == MBUF_SC_VO || \
|
|
c == MBUF_SC_CTL)
|
|
|
|
#define MBUF_VALID_SCIDX(c) \
|
|
(c == SCIDX_BK_SYS || c == SCIDX_BK || c == SCIDX_BE || \
|
|
c == SCIDX_RD || c == SCIDX_OAM || c == SCIDX_AV || \
|
|
c == SCIDX_RV || c == SCIDX_VI || c == SCIDX_VO || \
|
|
c == SCIDX_CTL)
|
|
|
|
#define MBUF_VALID_SCVAL(c) \
|
|
(c == SCVAL_BK_SYS || c == SCVAL_BK || c == SCVAL_BE || \
|
|
c == SCVAL_RD || c == SCVAL_OAM || c == SCVAL_AV || \
|
|
c == SCVAL_RV || c == SCVAL_VI || c == SCVAL_VO || \
|
|
c == SCVAL_CTL)
|
|
|
|
__private_extern__ union mbigcluster *mbutl; /* start VA of mbuf pool */
|
|
__private_extern__ union mbigcluster *embutl; /* end VA of mbuf pool */
|
|
__private_extern__ unsigned int nmbclusters; /* number of mapped clusters */
|
|
__private_extern__ int njcl; /* # of jumbo clusters */
|
|
__private_extern__ int njclbytes; /* size of a jumbo cluster */
|
|
__private_extern__ int max_hdr; /* largest link+protocol header */
|
|
__private_extern__ int max_datalen; /* MHLEN - max_hdr */
|
|
|
|
/* Use max_linkhdr instead of _max_linkhdr */
|
|
__private_extern__ int _max_linkhdr; /* largest link-level header */
|
|
|
|
/* Use max_protohdr instead of _max_protohdr */
|
|
__private_extern__ int _max_protohdr; /* largest protocol header */
|
|
|
|
__private_extern__ unsigned int mbuf_default_ncl(int, u_int64_t);
|
|
__private_extern__ void mbinit(void);
|
|
__private_extern__ struct mbuf *m_clattach(struct mbuf *, int, caddr_t,
|
|
void (*)(caddr_t, u_int, caddr_t), u_int, caddr_t, int);
|
|
__private_extern__ caddr_t m_bigalloc(int);
|
|
__private_extern__ void m_bigfree(caddr_t, u_int, caddr_t);
|
|
__private_extern__ struct mbuf *m_mbigget(struct mbuf *, int);
|
|
__private_extern__ caddr_t m_16kalloc(int);
|
|
__private_extern__ void m_16kfree(caddr_t, u_int, caddr_t);
|
|
__private_extern__ struct mbuf *m_m16kget(struct mbuf *, int);
|
|
|
|
__private_extern__ struct mbuf *m_free(struct mbuf *);
|
|
__private_extern__ struct mbuf *m_getclr(int, int);
|
|
__private_extern__ struct mbuf *m_getptr(struct mbuf *, int, int *);
|
|
__private_extern__ unsigned int m_length(struct mbuf *);
|
|
__private_extern__ unsigned int m_length2(struct mbuf *, struct mbuf **);
|
|
__private_extern__ unsigned int m_fixhdr(struct mbuf *);
|
|
__private_extern__ struct mbuf *m_defrag(struct mbuf *, int);
|
|
__private_extern__ struct mbuf *m_defrag_offset(struct mbuf *, u_int32_t, int);
|
|
__private_extern__ struct mbuf *m_prepend(struct mbuf *, int, int);
|
|
__private_extern__ struct mbuf *m_copyup(struct mbuf *, int, int);
|
|
__private_extern__ struct mbuf *m_retry(int, int);
|
|
__private_extern__ struct mbuf *m_retryhdr(int, int);
|
|
__private_extern__ int m_freem_list(struct mbuf *);
|
|
__private_extern__ int m_append(struct mbuf *, int, caddr_t);
|
|
__private_extern__ struct mbuf *m_last(struct mbuf *);
|
|
__private_extern__ struct mbuf *m_devget(char *, int, int, struct ifnet *,
|
|
void (*)(const void *, void *, size_t));
|
|
__private_extern__ struct mbuf *m_pulldown(struct mbuf *, int, int, int *);
|
|
|
|
__private_extern__ struct mbuf *m_getcl(int, int, int);
|
|
__private_extern__ caddr_t m_mclalloc(int);
|
|
__private_extern__ int m_mclhasreference(struct mbuf *);
|
|
__private_extern__ void m_copy_pkthdr(struct mbuf *, struct mbuf *);
|
|
__private_extern__ void m_copy_pftag(struct mbuf *, struct mbuf *);
|
|
|
|
__private_extern__ struct mbuf *m_dtom(void *);
|
|
__private_extern__ int m_mtocl(void *);
|
|
__private_extern__ union mcluster *m_cltom(int);
|
|
|
|
__private_extern__ int m_trailingspace(struct mbuf *);
|
|
__private_extern__ int m_leadingspace(struct mbuf *);
|
|
|
|
__private_extern__ struct mbuf *m_normalize(struct mbuf *m);
|
|
__private_extern__ void m_mchtype(struct mbuf *m, int t);
|
|
__private_extern__ void m_mcheck(struct mbuf *);
|
|
|
|
__private_extern__ void m_copyback(struct mbuf *, int, int, const void *);
|
|
__private_extern__ struct mbuf *m_copyback_cow(struct mbuf *, int, int,
|
|
const void *, int);
|
|
__private_extern__ int m_makewritable(struct mbuf **, int, int, int);
|
|
__private_extern__ struct mbuf *m_dup(struct mbuf *m, int how);
|
|
__private_extern__ struct mbuf *m_copym_with_hdrs(struct mbuf *, int, int, int,
|
|
struct mbuf **, int *);
|
|
__private_extern__ struct mbuf *m_getpackethdrs(int, int);
|
|
__private_extern__ struct mbuf *m_getpacket_how(int);
|
|
__private_extern__ struct mbuf *m_getpackets_internal(unsigned int *, int,
|
|
int, int, size_t);
|
|
__private_extern__ struct mbuf *m_allocpacket_internal(unsigned int *, size_t,
|
|
unsigned int *, int, int, size_t);
|
|
|
|
/*
|
|
* Packets may have annotations attached by affixing a list of "packet
|
|
* tags" to the pkthdr structure. Packet tags are dynamically allocated
|
|
* semi-opaque data structures that have a fixed header (struct m_tag)
|
|
* that specifies the size of the memory block and an <id,type> pair that
|
|
* identifies it. The id identifies the module and the type identifies the
|
|
* type of data for that module. The id of zero is reserved for the kernel.
|
|
*
|
|
* Note that the packet tag returned by m_tag_allocate has the default
|
|
* memory alignment implemented by malloc. To reference private data one
|
|
* can use a construct like:
|
|
*
|
|
* struct m_tag *mtag = m_tag_allocate(...);
|
|
* struct foo *p = (struct foo *)(mtag+1);
|
|
*
|
|
* if the alignment of struct m_tag is sufficient for referencing members
|
|
* of struct foo. Otherwise it is necessary to embed struct m_tag within
|
|
* the private data structure to insure proper alignment; e.g.
|
|
*
|
|
* struct foo {
|
|
* struct m_tag tag;
|
|
* ...
|
|
* };
|
|
* struct foo *p = (struct foo *) m_tag_allocate(...);
|
|
* struct m_tag *mtag = &p->tag;
|
|
*/
|
|
|
|
#define KERNEL_MODULE_TAG_ID 0
|
|
|
|
enum {
|
|
KERNEL_TAG_TYPE_NONE = 0,
|
|
KERNEL_TAG_TYPE_DUMMYNET = 1,
|
|
KERNEL_TAG_TYPE_DIVERT = 2,
|
|
KERNEL_TAG_TYPE_IPFORWARD = 3,
|
|
KERNEL_TAG_TYPE_IPFILT = 4,
|
|
KERNEL_TAG_TYPE_MACLABEL = 5,
|
|
KERNEL_TAG_TYPE_MAC_POLICY_LABEL = 6,
|
|
KERNEL_TAG_TYPE_ENCAP = 8,
|
|
KERNEL_TAG_TYPE_INET6 = 9,
|
|
KERNEL_TAG_TYPE_IPSEC = 10,
|
|
KERNEL_TAG_TYPE_PF = 11
|
|
};
|
|
|
|
/* Packet tag routines */
|
|
__private_extern__ struct m_tag *m_tag_alloc(u_int32_t, u_int16_t, int, int);
|
|
__private_extern__ struct m_tag *m_tag_create(u_int32_t, u_int16_t, int, int,
|
|
struct mbuf *);
|
|
__private_extern__ void m_tag_free(struct m_tag *);
|
|
__private_extern__ void m_tag_prepend(struct mbuf *, struct m_tag *);
|
|
__private_extern__ void m_tag_unlink(struct mbuf *, struct m_tag *);
|
|
__private_extern__ void m_tag_delete(struct mbuf *, struct m_tag *);
|
|
__private_extern__ void m_tag_delete_chain(struct mbuf *, struct m_tag *);
|
|
__private_extern__ struct m_tag *m_tag_locate(struct mbuf *, u_int32_t,
|
|
u_int16_t, struct m_tag *);
|
|
__private_extern__ struct m_tag *m_tag_copy(struct m_tag *, int);
|
|
__private_extern__ int m_tag_copy_chain(struct mbuf *, struct mbuf *, int);
|
|
__private_extern__ void m_tag_init(struct mbuf *);
|
|
__private_extern__ struct m_tag *m_tag_first(struct mbuf *);
|
|
__private_extern__ struct m_tag *m_tag_next(struct mbuf *, struct m_tag *);
|
|
|
|
__END_DECLS
|
|
#endif /* XNU_KERNEL_PRIVATE */
|
|
#ifdef KERNEL
|
|
#include <sys/kpi_mbuf.h>
|
|
#ifdef XNU_KERNEL_PRIVATE
|
|
__BEGIN_DECLS
|
|
|
|
__private_extern__ void m_service_class_init(struct mbuf *);
|
|
__private_extern__ int m_set_service_class(struct mbuf *, mbuf_svc_class_t);
|
|
__private_extern__ mbuf_svc_class_t m_get_service_class(struct mbuf *);
|
|
__private_extern__ mbuf_svc_class_t m_service_class_from_idx(u_int32_t);
|
|
__private_extern__ mbuf_svc_class_t m_service_class_from_val(u_int32_t);
|
|
__private_extern__ int m_set_traffic_class(struct mbuf *, mbuf_traffic_class_t);
|
|
__private_extern__ mbuf_traffic_class_t m_get_traffic_class(struct mbuf *);
|
|
|
|
__END_DECLS
|
|
#endif /* XNU_KERNEL_PRIVATE */
|
|
#endif /* KERNEL */
|
|
#endif /* !_SYS_MBUF_H_ */
|