diff options
Diffstat (limited to 'netwerk/sctp/src/netinet/sctp_os_userspace.h')
-rwxr-xr-x | netwerk/sctp/src/netinet/sctp_os_userspace.h | 1171 |
1 files changed, 1171 insertions, 0 deletions
diff --git a/netwerk/sctp/src/netinet/sctp_os_userspace.h b/netwerk/sctp/src/netinet/sctp_os_userspace.h new file mode 100755 index 000000000..bf4a3b519 --- /dev/null +++ b/netwerk/sctp/src/netinet/sctp_os_userspace.h @@ -0,0 +1,1171 @@ +/*- + * Copyright (c) 2006-2007, by Cisco Systems, Inc. All rights reserved. + * Copyright (c) 2008-2011, by Randall Stewart. All rights reserved. + * Copyright (c) 2008-2011, by Michael Tuexen. All rights reserved. + * Copyright (c) 2008-2011, by Brad Penoff. All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions are met: + * + * a) Redistributions of source code must retain the above copyright notice, + * this list of conditions and the following disclaimer. + * + * b) 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. + * + * c) Neither the name of Cisco Systems, Inc. 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 COPYRIGHT HOLDERS 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 COPYRIGHT OWNER 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. + */ + +#ifndef __sctp_os_userspace_h__ +#define __sctp_os_userspace_h__ +/* + * Userspace includes + * All the opt_xxx.h files are placed in the kernel build directory. + * We will place them in userspace stack build directory. + */ + +#include <errno.h> + +#if defined(__Userspace_os_Windows) +#include <winsock2.h> +#include <ws2tcpip.h> +#include <iphlpapi.h> +#include <Mswsock.h> +#include <Windows.h> +#include "user_environment.h" +typedef CRITICAL_SECTION userland_mutex_t; +#if WINVER < 0x0600 +enum { + C_SIGNAL = 0, + C_BROADCAST = 1, + C_MAX_EVENTS = 2 +}; +typedef struct +{ + u_int waiters_count; + CRITICAL_SECTION waiters_count_lock; + HANDLE events_[C_MAX_EVENTS]; +} userland_cond_t; +void InitializeXPConditionVariable(userland_cond_t *); +void DeleteXPConditionVariable(userland_cond_t *); +int SleepXPConditionVariable(userland_cond_t *, userland_mutex_t *); +void WakeAllXPConditionVariable(userland_cond_t *); +#define InitializeConditionVariable(cond) InitializeXPConditionVariable(cond) +#define DeleteConditionVariable(cond) DeleteXPConditionVariable(cond) +#define SleepConditionVariableCS(cond, mtx, time) SleepXPConditionVariable(cond, mtx) +#define WakeAllConditionVariable(cond) WakeAllXPConditionVariable(cond) +#else +#define DeleteConditionVariable(cond) +typedef CONDITION_VARIABLE userland_cond_t; +#endif +typedef HANDLE userland_thread_t; +#define ADDRESS_FAMILY unsigned __int8 +#define IPVERSION 4 +#define MAXTTL 255 +/* VS2010 comes with stdint.h */ +#if _MSC_VER >= 1600 +#include <stdint.h> +#else +#define uint64_t unsigned __int64 +#define uint32_t unsigned __int32 +#define int32_t __int32 +#define uint16_t unsigned __int16 +#define int16_t __int16 +#define uint8_t unsigned __int8 +#define int8_t __int8 +#endif +#ifndef _SIZE_T_DEFINED +#define size_t __int32 +#endif +#define u_long unsigned __int64 +#define u_int unsigned __int32 +#define u_int32_t unsigned __int32 +#define u_int16_t unsigned __int16 +#define u_int8_t unsigned __int8 +#define u_char unsigned char +#define n_short unsigned __int16 +#define u_short unsigned __int16 +#define n_time unsigned __int32 +#define sa_family_t unsigned __int8 +#define ssize_t __int64 +#define __func__ __FUNCTION__ + +#ifndef EWOULDBLOCK +#define EWOULDBLOCK WSAEWOULDBLOCK +#endif +#ifndef EINPROGRESS +#define EINPROGRESS WSAEINPROGRESS +#endif +#ifndef EALREADY +#define EALREADY WSAEALREADY +#endif +#ifndef ENOTSOCK +#define ENOTSOCK WSAENOTSOCK +#endif +#ifndef EDESTADDRREQ +#define EDESTADDRREQ WSAEDESTADDRREQ +#endif +#ifndef EMSGSIZE +#define EMSGSIZE WSAEMSGSIZE +#endif +#ifndef EPROTOTYPE +#define EPROTOTYPE WSAEPROTOTYPE +#endif +#ifndef ENOPROTOOPT +#define ENOPROTOOPT WSAENOPROTOOPT +#endif +#ifndef EPROTONOSUPPORT +#define EPROTONOSUPPORT WSAEPROTONOSUPPORT +#endif +#ifndef ESOCKTNOSUPPORT +#define ESOCKTNOSUPPORT WSAESOCKTNOSUPPORT +#endif +#ifndef EOPNOTSUPP +#define EOPNOTSUPP WSAEOPNOTSUPP +#endif +#ifndef ENOTSUP +#define ENOTSUP WSAEOPNOTSUPP +#endif +#ifndef EPFNOSUPPORT +#define EPFNOSUPPORT WSAEPFNOSUPPORT +#endif +#ifndef EAFNOSUPPORT +#define EAFNOSUPPORT WSAEAFNOSUPPORT +#endif +#ifndef EADDRINUSE +#define EADDRINUSE WSAEADDRINUSE +#endif +#ifndef EADDRNOTAVAIL +#define EADDRNOTAVAIL WSAEADDRNOTAVAIL +#endif +#ifndef ENETDOWN +#define ENETDOWN WSAENETDOWN +#endif +#ifndef ENETUNREACH +#define ENETUNREACH WSAENETUNREACH +#endif +#ifndef ENETRESET +#define ENETRESET WSAENETRESET +#endif +#ifndef ECONNABORTED +#define ECONNABORTED WSAECONNABORTED +#endif +#ifndef ECONNRESET +#define ECONNRESET WSAECONNRESET +#endif +#ifndef ENOBUFS +#define ENOBUFS WSAENOBUFS +#endif +#ifndef EISCONN +#define EISCONN WSAEISCONN +#endif +#ifndef ENOTCONN +#define ENOTCONN WSAENOTCONN +#endif +#ifndef ESHUTDOWN +#define ESHUTDOWN WSAESHUTDOWN +#endif +#ifndef ETOOMANYREFS +#define ETOOMANYREFS WSAETOOMANYREFS +#endif +#ifndef ETIMEDOUT +#define ETIMEDOUT WSAETIMEDOUT +#endif +#ifndef ECONNREFUSED +#define ECONNREFUSED WSAECONNREFUSED +#endif +#ifndef ELOOP +#define ELOOP WSAELOOP +#endif +#ifndef EHOSTDOWN +#define EHOSTDOWN WSAEHOSTDOWN +#endif +#ifndef EHOSTUNREACH +#define EHOSTUNREACH WSAEHOSTUNREACH +#endif +#ifndef EPROCLIM +#define EPROCLIM WSAEPROCLIM +#endif +#ifndef EUSERS +#define EUSERS WSAEUSERS +#endif +#ifndef EDQUOT +#define EDQUOT WSAEDQUOT +#endif +#ifndef ESTALE +#define ESTALE WSAESTALE +#endif +#ifndef EREMOTE +#define EREMOTE WSAEREMOTE +#endif + +typedef char* caddr_t; + +#define bzero(buf, len) memset(buf, 0, len) +#define bcopy(srcKey, dstKey, len) memcpy(dstKey, srcKey, len) +#if _MSC_VER < 1900 +#define snprintf(data, size, format, ...) _snprintf_s(data, size, _TRUNCATE, format, __VA_ARGS__) +#endif +#define inline __inline +#define __inline__ __inline +#define MSG_EOR 0x8 /* data completes record */ +#define MSG_DONTWAIT 0x80 /* this message should be nonblocking */ + +#ifdef CMSG_DATA +#undef CMSG_DATA +#endif +/* + * The following definitions should apply iff WINVER < 0x0600 + * but that check doesn't work in all cases. So be more pedantic... + */ +#define CMSG_DATA(x) WSA_CMSG_DATA(x) +#define CMSG_ALIGN(x) WSA_CMSGDATA_ALIGN(x) +#ifndef CMSG_FIRSTHDR +#define CMSG_FIRSTHDR(x) WSA_CMSG_FIRSTHDR(x) +#endif +#ifndef CMSG_NXTHDR +#define CMSG_NXTHDR(x, y) WSA_CMSG_NXTHDR(x, y) +#endif +#ifndef CMSG_SPACE +#define CMSG_SPACE(x) WSA_CMSG_SPACE(x) +#endif +#ifndef CMSG_LEN +#define CMSG_LEN(x) WSA_CMSG_LEN(x) +#endif + +/**** from sctp_os_windows.h ***************/ +#define SCTP_IFN_IS_IFT_LOOP(ifn) ((ifn)->ifn_type == IFT_LOOP) +#define SCTP_ROUTE_IS_REAL_LOOP(ro) ((ro)->ro_rt && (ro)->ro_rt->rt_ifa && (ro)->ro_rt->rt_ifa->ifa_ifp && (ro)->ro_rt->rt_ifa->ifa_ifp->if_type == IFT_LOOP) + +/* + * Access to IFN's to help with src-addr-selection + */ +/* This could return VOID if the index works but for BSD we provide both. */ +#define SCTP_GET_IFN_VOID_FROM_ROUTE(ro) \ + ((ro)->ro_rt != NULL ? (ro)->ro_rt->rt_ifp : NULL) +#define SCTP_ROUTE_HAS_VALID_IFN(ro) \ + ((ro)->ro_rt && (ro)->ro_rt->rt_ifp) +/******************************************/ + +#define SCTP_GET_IF_INDEX_FROM_ROUTE(ro) 1 /* compiles... TODO use routing socket to determine */ + +#define BIG_ENDIAN 1 +#define LITTLE_ENDIAN 0 +#ifdef WORDS_BIGENDIAN +#define BYTE_ORDER BIG_ENDIAN +#else +#define BYTE_ORDER LITTLE_ENDIAN +#endif + +#else /* !defined(Userspace_os_Windows) */ +#include <sys/socket.h> +#if defined(__Userspace_os_DragonFly) || defined(__Userspace_os_FreeBSD) || defined(__Userspace_os_Linux) || defined(__Userspace_os_NetBSD) || defined(__Userspace_os_OpenBSD) || defined(__Userspace_os_NaCl) +#include <pthread.h> +#endif +typedef pthread_mutex_t userland_mutex_t; +typedef pthread_cond_t userland_cond_t; +typedef pthread_t userland_thread_t; +#endif + +#if defined(__Userspace_os_Windows) || defined(__Userspace_os_NaCl) + +#define IFNAMSIZ 64 + +#define random() rand() +#define srandom(s) srand(s) + +#define timeradd(tvp, uvp, vvp) \ + do { \ + (vvp)->tv_sec = (tvp)->tv_sec + (uvp)->tv_sec; \ + (vvp)->tv_usec = (tvp)->tv_usec + (uvp)->tv_usec; \ + if ((vvp)->tv_usec >= 1000000) { \ + (vvp)->tv_sec++; \ + (vvp)->tv_usec -= 1000000; \ + } \ + } while (0) + +#define timersub(tvp, uvp, vvp) \ + do { \ + (vvp)->tv_sec = (tvp)->tv_sec - (uvp)->tv_sec; \ + (vvp)->tv_usec = (tvp)->tv_usec - (uvp)->tv_usec; \ + if ((vvp)->tv_usec < 0) { \ + (vvp)->tv_sec--; \ + (vvp)->tv_usec += 1000000; \ + } \ + } while (0) + +/*#include <packon.h> +#pragma pack(push, 1)*/ +struct ip { + u_char ip_hl:4, ip_v:4; + u_char ip_tos; + u_short ip_len; + u_short ip_id; + u_short ip_off; +#define IP_RP 0x8000 +#define IP_DF 0x4000 +#define IP_MF 0x2000 +#define IP_OFFMASK 0x1fff + u_char ip_ttl; + u_char ip_p; + u_short ip_sum; + struct in_addr ip_src, ip_dst; +}; + +struct ifaddrs { + struct ifaddrs *ifa_next; + char *ifa_name; + unsigned int ifa_flags; + struct sockaddr *ifa_addr; + struct sockaddr *ifa_netmask; + struct sockaddr *ifa_dstaddr; + void *ifa_data; +}; + +struct udphdr { + uint16_t uh_sport; + uint16_t uh_dport; + uint16_t uh_ulen; + uint16_t uh_sum; +}; + +struct iovec { + unsigned long len; + char *buf; +}; + +#define iov_base buf +#define iov_len len + +struct ifa_msghdr { + uint16_t ifam_msglen; + unsigned char ifam_version; + unsigned char ifam_type; + uint32_t ifam_addrs; + uint32_t ifam_flags; + uint16_t ifam_index; + uint32_t ifam_metric; +}; + +struct ifdevmtu { + int ifdm_current; + int ifdm_min; + int ifdm_max; +}; + +struct ifkpi { + unsigned int ifk_module_id; + unsigned int ifk_type; + union { + void *ifk_ptr; + int ifk_value; + } ifk_data; +}; + +struct ifreq { + char ifr_name[16]; + union { + struct sockaddr ifru_addr; + struct sockaddr ifru_dstaddr; + struct sockaddr ifru_broadaddr; + short ifru_flags; + int ifru_metric; + int ifru_mtu; + int ifru_phys; + int ifru_media; + int ifru_intval; + char* ifru_data; + struct ifdevmtu ifru_devmtu; + struct ifkpi ifru_kpi; + uint32_t ifru_wake_flags; + } ifr_ifru; +#define ifr_addr ifr_ifru.ifru_addr +#define ifr_dstaddr ifr_ifru.ifru_dstaddr +#define ifr_broadaddr ifr_ifru.ifru_broadaddr +#define ifr_flags ifr_ifru.ifru_flags[0] +#define ifr_prevflags ifr_ifru.ifru_flags[1] +#define ifr_metric ifr_ifru.ifru_metric +#define ifr_mtu ifr_ifru.ifru_mtu +#define ifr_phys ifr_ifru.ifru_phys +#define ifr_media ifr_ifru.ifru_media +#define ifr_data ifr_ifru.ifru_data +#define ifr_devmtu ifr_ifru.ifru_devmtu +#define ifr_intval ifr_ifru.ifru_intval +#define ifr_kpi ifr_ifru.ifru_kpi +#define ifr_wake_flags ifr_ifru.ifru_wake_flags +}; + +#endif + +#if defined(__Userspace_os_Windows) +int Win_getifaddrs(struct ifaddrs**); +#define getifaddrs(interfaces) (int)Win_getifaddrs(interfaces) +int win_if_nametoindex(const char *); +#define if_nametoindex(x) win_if_nametoindex(x) +#endif + +#define mtx_lock(arg1) +#define mtx_unlock(arg1) +#define mtx_assert(arg1,arg2) +#define MA_OWNED 7 /* sys/mutex.h typically on FreeBSD */ +#if !defined(__Userspace_os_FreeBSD) +struct mtx {int dummy;}; +#if !defined(__Userspace_os_NetBSD) +struct selinfo {int dummy;}; +#endif +struct sx {int dummy;}; +#endif + +#include <stdio.h> +#include <string.h> +/* #include <sys/param.h> in FreeBSD defines MSIZE */ +/* #include <sys/ktr.h> */ +/* #include <sys/systm.h> */ +#if defined(HAVE_SYS_QUEUE_H) +#include <sys/queue.h> +#else +#include <user_queue.h> +#endif +#include <user_malloc.h> +/* #include <sys/kernel.h> */ +/* #include <sys/sysctl.h> */ +/* #include <sys/protosw.h> */ +/* on FreeBSD, this results in a redefintion of SOCK(BUF)_(UN)LOCK and + * uknown type of struct mtx for sb_mtx in struct sockbuf */ +#include "user_socketvar.h" /* MALLOC_DECLARE's M_PCB. Replacement for sys/socketvar.h */ +/* #include <sys/jail.h> */ +/* #include <sys/sysctl.h> */ +#include <user_environment.h> +#include <user_atomic.h> +#include <user_mbuf.h> +/* #include <sys/uio.h> */ +/* #include <sys/lock.h> */ +#if defined(__FreeBSD__) && __FreeBSD_version > 602000 +#include <sys/rwlock.h> +#endif +/* #include <sys/kthread.h> */ +#if defined(__FreeBSD__) && __FreeBSD_version > 602000 +#include <sys/priv.h> +#endif +/* #include <sys/random.h> */ +/* #include <sys/limits.h> */ +/* #include <machine/cpu.h> */ + +#if defined(__Userspace_os_Darwin) +/* was a 0 byte file. needed for structs if_data(64) and net_event_data */ +#include <net/if_var.h> +#endif +#if defined(__Userspace_os_FreeBSD) +#include <net/if_types.h> +/* #include <net/if_var.h> was a 0 byte file. causes struct mtx redefinition */ +#endif +/* OOTB only - dummy route used at the moment. should we port route to + * userspace as well? */ +/* on FreeBSD, this results in a redefintion of struct route */ +/* #include <net/route.h> */ +#if !defined(__Userspace_os_Windows) && !defined(__Userspace_os_NaCl) +#include <net/if.h> +#include <netinet/in.h> +#include <netinet/in_systm.h> +#include <netinet/ip.h> +#endif +#if defined(HAVE_NETINET_IP_ICMP_H) +#include <netinet/ip_icmp.h> +#else +#include <user_ip_icmp.h> +#endif +/* #include <netinet/in_pcb.h> ported to userspace */ +#include <user_inpcb.h> + +/* for getifaddrs */ +#include <sys/types.h> +#if !defined(__Userspace_os_Windows) +#if !defined(ANDROID) && (defined(INET) || defined(INET6)) +#include <ifaddrs.h> +#endif + +/* for ioctl */ +#include <sys/ioctl.h> + +/* for close, etc. */ +#include <unistd.h> +#endif + +/* lots of errno's used and needed in userspace */ + +/* for offsetof */ +#include <stddef.h> + +#if defined(SCTP_PROCESS_LEVEL_LOCKS) && !defined(__Userspace_os_Windows) +/* for pthread_mutex_lock, pthread_mutex_unlock, etc. */ +#include <pthread.h> +#endif + +#ifdef IPSEC +#include <netipsec/ipsec.h> +#include <netipsec/key.h> +#endif /* IPSEC */ + +#ifdef INET6 +#if defined(__Userspace_os_FreeBSD) +#include <sys/domain.h> +#endif +#ifdef IPSEC +#include <netipsec/ipsec6.h> +#endif +#if !defined(__Userspace_os_Windows) +#include <netinet/ip6.h> +#include <netinet/icmp6.h> +#endif +#if defined(__Userspace_os_Darwin) || defined(__Userspace_os_FreeBSD) || defined(__Userspace_os_Linux) || defined(__Userspace_os_NetBSD) || defined(__Userspace_os_OpenBSD) || defined(__Userspace_os_Windows) +#include "user_ip6_var.h" +#else +#include <netinet6/ip6_var.h> +#endif +#if defined(__Userspace_os_FreeBSD) +#include <netinet6/in6_pcb.h> +#include <netinet6/ip6protosw.h> +/* #include <netinet6/nd6.h> was a 0 byte file */ +#include <netinet6/scope6_var.h> +#endif +#endif /* INET6 */ + +#if defined(HAVE_SCTP_PEELOFF_SOCKOPT) +#include <sys/file.h> +#include <sys/filedesc.h> +#endif + +#include "netinet/sctp_sha1.h" + +#if __FreeBSD_version >= 700000 +#include <netinet/ip_options.h> +#endif + +#define SCTP_PRINTF(...) \ + if (SCTP_BASE_VAR(debug_printf)) { \ + SCTP_BASE_VAR(debug_printf)(__VA_ARGS__); \ + } + +#if defined(__FreeBSD__) +#ifndef in6pcb +#define in6pcb inpcb +#endif +#endif +/* Declare all the malloc names for all the various mallocs */ +MALLOC_DECLARE(SCTP_M_MAP); +MALLOC_DECLARE(SCTP_M_STRMI); +MALLOC_DECLARE(SCTP_M_STRMO); +MALLOC_DECLARE(SCTP_M_ASC_ADDR); +MALLOC_DECLARE(SCTP_M_ASC_IT); +MALLOC_DECLARE(SCTP_M_AUTH_CL); +MALLOC_DECLARE(SCTP_M_AUTH_KY); +MALLOC_DECLARE(SCTP_M_AUTH_HL); +MALLOC_DECLARE(SCTP_M_AUTH_IF); +MALLOC_DECLARE(SCTP_M_STRESET); +MALLOC_DECLARE(SCTP_M_CMSG); +MALLOC_DECLARE(SCTP_M_COPYAL); +MALLOC_DECLARE(SCTP_M_VRF); +MALLOC_DECLARE(SCTP_M_IFA); +MALLOC_DECLARE(SCTP_M_IFN); +MALLOC_DECLARE(SCTP_M_TIMW); +MALLOC_DECLARE(SCTP_M_MVRF); +MALLOC_DECLARE(SCTP_M_ITER); +MALLOC_DECLARE(SCTP_M_SOCKOPT); + +#if defined(SCTP_LOCAL_TRACE_BUF) + +#define SCTP_GET_CYCLECOUNT get_cyclecount() +#define SCTP_CTR6 sctp_log_trace + +#else +#define SCTP_CTR6 CTR6 +#endif + +/* Empty ktr statement for _Userspace__ (similar to what is done for mac) */ +#define CTR6(m, d, p1, p2, p3, p4, p5, p6) + + + +#define SCTP_BASE_INFO(__m) system_base_info.sctppcbinfo.__m +#define SCTP_BASE_STATS system_base_info.sctpstat +#define SCTP_BASE_STAT(__m) system_base_info.sctpstat.__m +#define SCTP_BASE_SYSCTL(__m) system_base_info.sctpsysctl.__m +#define SCTP_BASE_VAR(__m) system_base_info.__m + +/* + * + */ +#if !defined(__Userspace_os_Darwin) +#define USER_ADDR_NULL (NULL) /* FIX ME: temp */ +#endif + +#if defined(SCTP_DEBUG) +#include <netinet/sctp_constants.h> +#define SCTPDBG(level, ...) \ +{ \ + do { \ + if (SCTP_BASE_SYSCTL(sctp_debug_on) & level) { \ + SCTP_PRINTF(__VA_ARGS__); \ + } \ + } while (0); \ +} +#define SCTPDBG_ADDR(level, addr) \ +{ \ + do { \ + if (SCTP_BASE_SYSCTL(sctp_debug_on) & level ) { \ + sctp_print_address(addr); \ + } \ + } while (0); \ +} +#else +#define SCTPDBG(level, ...) +#define SCTPDBG_ADDR(level, addr) +#endif + +#ifdef SCTP_LTRACE_CHUNKS +#define SCTP_LTRACE_CHK(a, b, c, d) if(sctp_logging_level & SCTP_LTRACE_CHUNK_ENABLE) CTR6(KTR_SUBSYS, "SCTP:%d[%d]:%x-%x-%x-%x", SCTP_LOG_CHUNK_PROC, 0, a, b, c, d) +#else +#define SCTP_LTRACE_CHK(a, b, c, d) +#endif + +#ifdef SCTP_LTRACE_ERRORS +#define SCTP_LTRACE_ERR_RET_PKT(m, inp, stcb, net, file, err) \ + if (sctp_logging_level & SCTP_LTRACE_ERROR_ENABLE) \ + SCTP_PRINTF("mbuf:%p inp:%p stcb:%p net:%p file:%x line:%d error:%d\n", \ + (void *)m, (void *)inp, (void *)stcb, (void *)net, file, __LINE__, err); +#define SCTP_LTRACE_ERR_RET(inp, stcb, net, file, err) \ + if (sctp_logging_level & SCTP_LTRACE_ERROR_ENABLE) \ + SCTP_PRINTF("inp:%p stcb:%p net:%p file:%x line:%d error:%d\n", \ + (void *)inp, (void *)stcb, (void *)net, file, __LINE__, err); +#else +#define SCTP_LTRACE_ERR_RET_PKT(m, inp, stcb, net, file, err) +#define SCTP_LTRACE_ERR_RET(inp, stcb, net, file, err) +#endif + + +/* + * Local address and interface list handling + */ +#define SCTP_MAX_VRF_ID 0 +#define SCTP_SIZE_OF_VRF_HASH 3 +#define SCTP_IFNAMSIZ IFNAMSIZ +#define SCTP_DEFAULT_VRFID 0 +#define SCTP_VRF_ADDR_HASH_SIZE 16 +#define SCTP_VRF_IFN_HASH_SIZE 3 +#define SCTP_INIT_VRF_TABLEID(vrf) + +#if !defined(__Userspace_os_Windows) +#define SCTP_IFN_IS_IFT_LOOP(ifn) (strncmp((ifn)->ifn_name, "lo", 2) == 0) +/* BSD definition */ +/* #define SCTP_ROUTE_IS_REAL_LOOP(ro) ((ro)->ro_rt && (ro)->ro_rt->rt_ifa && (ro)->ro_rt->rt_ifa->ifa_ifp && (ro)->ro_rt->rt_ifa->ifa_ifp->if_type == IFT_LOOP) */ +/* only used in IPv6 scenario, which isn't supported yet */ +#define SCTP_ROUTE_IS_REAL_LOOP(ro) 0 + +/* + * Access to IFN's to help with src-addr-selection + */ +/* This could return VOID if the index works but for BSD we provide both. */ +#define SCTP_GET_IFN_VOID_FROM_ROUTE(ro) (void *)ro->ro_rt->rt_ifp +#define SCTP_GET_IF_INDEX_FROM_ROUTE(ro) 1 /* compiles... TODO use routing socket to determine */ +#define SCTP_ROUTE_HAS_VALID_IFN(ro) ((ro)->ro_rt && (ro)->ro_rt->rt_ifp) +#endif + +/* + * general memory allocation + */ +#define SCTP_MALLOC(var, type, size, name) \ + do { \ + MALLOC(var, type, size, name, M_NOWAIT); \ + } while (0) + +#define SCTP_FREE(var, type) FREE(var, type) + +#define SCTP_MALLOC_SONAME(var, type, size) \ + do { \ + MALLOC(var, type, size, M_SONAME, (M_WAITOK | M_ZERO)); \ + } while (0) + +#define SCTP_FREE_SONAME(var) FREE(var, M_SONAME) + +#define SCTP_PROCESS_STRUCT struct proc * + +/* + * zone allocation functions + */ + + +#if defined(SCTP_SIMPLE_ALLOCATOR) +/*typedef size_t sctp_zone_t;*/ +#define SCTP_ZONE_INIT(zone, name, size, number) { \ + zone = size; \ +} + +/* __Userspace__ SCTP_ZONE_GET: allocate element from the zone */ +#define SCTP_ZONE_GET(zone, type) \ + (type *)malloc(zone); + + +/* __Userspace__ SCTP_ZONE_FREE: free element from the zone */ +#define SCTP_ZONE_FREE(zone, element) { \ + free(element); \ +} + +#define SCTP_ZONE_DESTROY(zone) +#else +/*__Userspace__ + Compiling & linking notes: Needs libumem, which has been placed in ./user_lib + All userspace header files are in ./user_include. Makefile will need the + following. + CFLAGS = -I./ -Wall + LDFLAGS = -L./user_lib -R./user_lib -lumem +*/ +#include "user_include/umem.h" + +/* __Userspace__ SCTP_ZONE_INIT: initialize the zone */ +/* + __Userspace__ + No equivalent function to uma_zone_set_max added yet. (See SCTP_ZONE_INIT in sctp_os_bsd.h + for reference). It may not be required as mentioned in + http://nixdoc.net/man-pages/FreeBSD/uma_zalloc.9.html that + max limits may not enforced on systems with more than one CPU. +*/ +#define SCTP_ZONE_INIT(zone, name, size, number) { \ + zone = umem_cache_create(name, size, 0, NULL, NULL, NULL, NULL, NULL, 0); \ + } + +/* __Userspace__ SCTP_ZONE_GET: allocate element from the zone */ +#define SCTP_ZONE_GET(zone, type) \ + (type *)umem_cache_alloc(zone, UMEM_DEFAULT); + + +/* __Userspace__ SCTP_ZONE_FREE: free element from the zone */ +#define SCTP_ZONE_FREE(zone, element) \ + umem_cache_free(zone, element); + + +/* __Userspace__ SCTP_ZONE_DESTROY: destroy the zone */ +#define SCTP_ZONE_DESTROY(zone) \ + umem_cache_destroy(zone); +#endif + +/* global struct ifaddrs used in sctp_init_ifns_for_vrf getifaddrs call + * but references to fields are needed to persist as the vrf is queried. + * getifaddrs allocates memory that needs to be freed with a freeifaddrs + * call; this global is used to call freeifaddrs upon in sctp_pcb_finish + */ +extern struct ifaddrs *g_interfaces; + + +/* + * __Userspace__ Defining sctp_hashinit_flags() and sctp_hashdestroy() for userland. + */ +void *sctp_hashinit_flags(int elements, struct malloc_type *type, + u_long *hashmask, int flags); +void +sctp_hashdestroy(void *vhashtbl, struct malloc_type *type, u_long hashmask); + +void +sctp_hashfreedestroy(void *vhashtbl, struct malloc_type *type, u_long hashmask); + + +#define HASH_NOWAIT 0x00000001 +#define HASH_WAITOK 0x00000002 + +/* M_PCB is MALLOC_DECLARE'd in sys/socketvar.h */ +#define SCTP_HASH_INIT(size, hashmark) sctp_hashinit_flags(size, M_PCB, hashmark, HASH_NOWAIT) + +#define SCTP_HASH_FREE(table, hashmark) sctp_hashdestroy(table, M_PCB, hashmark) + +#define SCTP_HASH_FREE_DESTROY(table, hashmark) sctp_hashfreedestroy(table, M_PCB, hashmark) +#define SCTP_M_COPYM m_copym + +/* + * timers + */ +/* __Userspace__ + * user_sctp_callout.h has typedef struct sctp_callout sctp_os_timer_t; + * which is used in the timer related functions such as + * SCTP_OS_TIMER_INIT etc. +*/ +#include <netinet/sctp_callout.h> + +/* __Userspace__ Creating a receive thread */ +#include <user_recv_thread.h> + +/*__Userspace__ defining KTR_SUBSYS 1 as done in sctp_os_macosx.h */ +#define KTR_SUBSYS 1 + +#define sctp_get_tick_count() (ticks) + +/* The packed define for 64 bit platforms */ +#if !defined(__Userspace_os_Windows) +#define SCTP_PACKED __attribute__((packed)) +#define SCTP_UNUSED __attribute__((unused)) +#else +#define SCTP_PACKED +#define SCTP_UNUSED +#endif + +/* + * Functions + */ +/* Mbuf manipulation and access macros */ +#define SCTP_BUF_LEN(m) (m->m_len) +#define SCTP_BUF_NEXT(m) (m->m_next) +#define SCTP_BUF_NEXT_PKT(m) (m->m_nextpkt) +#define SCTP_BUF_RESV_UF(m, size) m->m_data += size +#define SCTP_BUF_AT(m, size) m->m_data + size +#define SCTP_BUF_IS_EXTENDED(m) (m->m_flags & M_EXT) +#define SCTP_BUF_EXTEND_SIZE(m) (m->m_ext.ext_size) +#define SCTP_BUF_TYPE(m) (m->m_type) +#define SCTP_BUF_RECVIF(m) (m->m_pkthdr.rcvif) +#define SCTP_BUF_PREPEND M_PREPEND + +#define SCTP_ALIGN_TO_END(m, len) if(m->m_flags & M_PKTHDR) { \ + MH_ALIGN(m, len); \ + } else if ((m->m_flags & M_EXT) == 0) { \ + M_ALIGN(m, len); \ + } + +/* We make it so if you have up to 4 threads + * writting based on the default size of + * the packet log 65 k, that would be + * 4 16k packets before we would hit + * a problem. + */ +#define SCTP_PKTLOG_WRITERS_NEED_LOCK 3 + + +/* + * routes, output, etc. + */ + +typedef struct sctp_route sctp_route_t; +typedef struct sctp_rtentry sctp_rtentry_t; + +static inline void sctp_userspace_rtalloc(sctp_route_t *ro) +{ + if (ro->ro_rt != NULL) { + ro->ro_rt->rt_refcnt++; + return; + } + + ro->ro_rt = (sctp_rtentry_t *) malloc(sizeof(sctp_rtentry_t)); + if (ro->ro_rt == NULL) + return; + + /* initialize */ + memset(ro->ro_rt, 0, sizeof(sctp_rtentry_t)); + ro->ro_rt->rt_refcnt = 1; + + /* set MTU */ + /* TODO set this based on the ro->ro_dst, looking up MTU with routing socket */ +#if 0 + if (userspace_rawroute == -1) { + userspace_rawroute = socket(AF_ROUTE, SOCK_RAW, 0); + if (userspace_rawroute == -1) + return; + } +#endif + ro->ro_rt->rt_rmx.rmx_mtu = 1500; /* FIXME temporary solution */ + + /* TODO enable the ability to obtain interface index of route for + * SCTP_GET_IF_INDEX_FROM_ROUTE macro. + */ +} +#define SCTP_RTALLOC(ro, vrf_id) sctp_userspace_rtalloc((sctp_route_t *)ro) + +/* dummy rtfree needed once user_route.h is included */ +static inline void sctp_userspace_rtfree(sctp_rtentry_t *rt) +{ + if(rt == NULL) { + return; + } + if(--rt->rt_refcnt > 0) { + return; + } + free(rt); + rt = NULL; +} +#define rtfree(arg1) sctp_userspace_rtfree(arg1) + + +/*************************/ +/* MTU */ +/*************************/ +int sctp_userspace_get_mtu_from_ifn(uint32_t if_index, int af); + +#define SCTP_GATHER_MTU_FROM_IFN_INFO(ifn, ifn_index, af) sctp_userspace_get_mtu_from_ifn(ifn_index, af) + +#define SCTP_GATHER_MTU_FROM_ROUTE(sctp_ifa, sa, rt) ((rt != NULL) ? rt->rt_rmx.rmx_mtu : 0) + +#define SCTP_GATHER_MTU_FROM_INTFC(sctp_ifn) sctp_userspace_get_mtu_from_ifn(if_nametoindex(((struct ifaddrs *) (sctp_ifn))->ifa_name), AF_INET) + +#define SCTP_SET_MTU_OF_ROUTE(sa, rt, mtu) do { \ + if (rt != NULL) \ + rt->rt_rmx.rmx_mtu = mtu; \ + } while(0) + +/* (de-)register interface event notifications */ +#define SCTP_REGISTER_INTERFACE(ifhandle, af) +#define SCTP_DEREGISTER_INTERFACE(ifhandle, af) + + +/*************************/ +/* These are for logging */ +/*************************/ +/* return the base ext data pointer */ +#define SCTP_BUF_EXTEND_BASE(m) (m->m_ext.ext_buf) + /* return the refcnt of the data pointer */ +#define SCTP_BUF_EXTEND_REFCNT(m) (*m->m_ext.ref_cnt) +/* return any buffer related flags, this is + * used beyond logging for apple only. + */ +#define SCTP_BUF_GET_FLAGS(m) (m->m_flags) + +/* For BSD this just accesses the M_PKTHDR length + * so it operates on an mbuf with hdr flag. Other + * O/S's may have seperate packet header and mbuf + * chain pointers.. thus the macro. + */ +#define SCTP_HEADER_TO_CHAIN(m) (m) +#define SCTP_DETACH_HEADER_FROM_CHAIN(m) +#define SCTP_HEADER_LEN(m) ((m)->m_pkthdr.len) +#define SCTP_GET_HEADER_FOR_OUTPUT(o_pak) 0 +#define SCTP_RELEASE_HEADER(m) +#define SCTP_RELEASE_PKT(m) sctp_m_freem(m) +/* UDP __Userspace__ - dummy definition */ +#define SCTP_ENABLE_UDP_CSUM(m) m=m +/* BSD definition */ +/* #define SCTP_ENABLE_UDP_CSUM(m) do { \ */ +/* m->m_pkthdr.csum_flags = CSUM_UDP; \ */ +/* m->m_pkthdr.csum_data = offsetof(struct udphdr, uh_sum); \ */ +/* } while (0) */ + +#define SCTP_GET_PKT_VRFID(m, vrf_id) ((vrf_id = SCTP_DEFAULT_VRFID) != SCTP_DEFAULT_VRFID) + + + +/* Attach the chain of data into the sendable packet. */ +#define SCTP_ATTACH_CHAIN(pak, m, packet_length) do { \ + pak = m; \ + pak->m_pkthdr.len = packet_length; \ + } while(0) + +/* Other m_pkthdr type things */ +/* FIXME need real definitions */ +#define SCTP_IS_IT_BROADCAST(dst, m) 0 +/* OOTB only #define SCTP_IS_IT_BROADCAST(dst, m) ((m->m_flags & M_PKTHDR) ? in_broadcast(dst, m->m_pkthdr.rcvif) : 0) BSD def */ +#define SCTP_IS_IT_LOOPBACK(m) 0 +/* OOTB ONLY #define SCTP_IS_IT_LOOPBACK(m) ((m->m_flags & M_PKTHDR) && ((m->m_pkthdr.rcvif == NULL) || (m->m_pkthdr.rcvif->if_type == IFT_LOOP))) BSD def */ + + +/* This converts any input packet header + * into the chain of data holders, for BSD + * its a NOP. + */ + +/* get the v6 hop limit */ +#define SCTP_GET_HLIM(inp, ro) 128 /* As done for __Windows__ */ +#define IPv6_HOP_LIMIT 128 + +/* is the endpoint v6only? */ +#define SCTP_IPV6_V6ONLY(inp) (((struct inpcb *)inp)->inp_flags & IN6P_IPV6_V6ONLY) +/* is the socket non-blocking? */ +#define SCTP_SO_IS_NBIO(so) ((so)->so_state & SS_NBIO) +#define SCTP_SET_SO_NBIO(so) ((so)->so_state |= SS_NBIO) +#define SCTP_CLEAR_SO_NBIO(so) ((so)->so_state &= ~SS_NBIO) +/* get the socket type */ +#define SCTP_SO_TYPE(so) ((so)->so_type) + +/* reserve sb space for a socket */ +#define SCTP_SORESERVE(so, send, recv) soreserve(so, send, recv) + +/* wakeup a socket */ +#define SCTP_SOWAKEUP(so) wakeup(&(so)->so_timeo, so) +/* clear the socket buffer state */ +#define SCTP_SB_CLEAR(sb) \ + (sb).sb_cc = 0; \ + (sb).sb_mb = NULL; \ + (sb).sb_mbcnt = 0; + +#define SCTP_SB_LIMIT_RCV(so) so->so_rcv.sb_hiwat +#define SCTP_SB_LIMIT_SND(so) so->so_snd.sb_hiwat + +/* Future zero copy wakeup/send function */ +#define SCTP_ZERO_COPY_EVENT(inp, so) +/* This is re-pulse ourselves for sendbuf */ +#define SCTP_ZERO_COPY_SENDQ_EVENT(inp, so) + +#define SCTP_READ_RANDOM(buf, len) read_random(buf, len) + +#define SCTP_SHA1_CTX struct sctp_sha1_context +#define SCTP_SHA1_INIT sctp_sha1_init +#define SCTP_SHA1_UPDATE sctp_sha1_update +#define SCTP_SHA1_FINAL(x,y) sctp_sha1_final((unsigned char *)x, y) + +/* start OOTB only stuff */ +/* TODO IFT_LOOP is in net/if_types.h on Linux */ +#define IFT_LOOP 0x18 + +/* sctp_pcb.h */ + +#if defined(__Userspace_os_Windows) +#define SHUT_RD 1 +#define SHUT_WR 2 +#define SHUT_RDWR 3 +#endif +#define PRU_FLUSH_RD SHUT_RD +#define PRU_FLUSH_WR SHUT_WR +#define PRU_FLUSH_RDWR SHUT_RDWR + +/* netinet/ip_var.h defintions are behind an if defined for _KERNEL on FreeBSD */ +#define IP_RAWOUTPUT 0x2 + + +/* end OOTB only stuff */ + +#define AF_CONN 123 +struct sockaddr_conn { +#ifdef HAVE_SCONN_LEN + uint8_t sconn_len; +#endif + uint8_t sconn_family; + uint16_t sconn_port; + void *sconn_addr; +}; + +/* + * SCTP protocol specific mbuf flags. + */ +#define M_NOTIFICATION M_PROTO5 /* SCTP notification */ + +/* + * IP output routines + */ + +/* Defining SCTP_IP_ID macro. + In netinet/ip_output.c, we have u_short ip_id; + In netinet/ip_var.h, we have extern u_short ip_id; (enclosed within _KERNEL_) + See static __inline uint16_t ip_newid(void) in netinet/ip_var.h + */ +#define SCTP_IP_ID(inp) (ip_id) + +/* need sctphdr to get port in SCTP_IP_OUTPUT. sctphdr defined in sctp.h */ +#include <netinet/sctp.h> +extern void sctp_userspace_ip_output(int *result, struct mbuf *o_pak, + sctp_route_t *ro, void *stcb, + uint32_t vrf_id); + +#define SCTP_IP_OUTPUT(result, o_pak, ro, stcb, vrf_id) sctp_userspace_ip_output(&result, o_pak, ro, stcb, vrf_id); + +#if defined(INET6) +extern void sctp_userspace_ip6_output(int *result, struct mbuf *o_pak, + struct route_in6 *ro, void *stcb, + uint32_t vrf_id); +#define SCTP_IP6_OUTPUT(result, o_pak, ro, ifp, stcb, vrf_id) sctp_userspace_ip6_output(&result, o_pak, ro, stcb, vrf_id); +#endif + + + +#if 0 +#define SCTP_IP6_OUTPUT(result, o_pak, ro, ifp, stcb, vrf_id) \ +{ \ + if (stcb && stcb->sctp_ep) \ + result = ip6_output(o_pak, \ + ((struct in6pcb *)(stcb->sctp_ep))->in6p_outputopts, \ + (ro), 0, 0, ifp, NULL); \ + else \ + result = ip6_output(o_pak, NULL, (ro), 0, 0, ifp, NULL); \ +} +#endif + +struct mbuf * +sctp_get_mbuf_for_msg(unsigned int space_needed, int want_header, int how, int allonebuf, int type); + + +/* with the current included files, this is defined in Linux but + * in FreeBSD, it is behind a _KERNEL in sys/socket.h ... + */ +#if defined(__Userspace_os_DragonFly) || defined(__Userspace_os_FreeBSD) || defined(__Userspace_os_OpenBSD) || defined(__Userspace_os_NaCl) +/* stolen from /usr/include/sys/socket.h */ +#define CMSG_ALIGN(n) _ALIGN(n) +#elif defined(__Userspace_os_NetBSD) +#define CMSG_ALIGN(n) (((n) + __ALIGNBYTES) & ~__ALIGNBYTES) +#elif defined(__Userspace_os_Darwin) +#if !defined(__DARWIN_ALIGNBYTES) +#define __DARWIN_ALIGNBYTES (sizeof(__darwin_size_t) - 1) +#endif + +#if !defined(__DARWIN_ALIGN) +#define __DARWIN_ALIGN(p) ((__darwin_size_t)((char *)(uintptr_t)(p) + __DARWIN_ALIGNBYTES) &~ __DARWIN_ALIGNBYTES) +#endif + +#if !defined(__DARWIN_ALIGNBYTES32) +#define __DARWIN_ALIGNBYTES32 (sizeof(__uint32_t) - 1) +#endif + +#if !defined(__DARWIN_ALIGN32) +#define __DARWIN_ALIGN32(p) ((__darwin_size_t)((char *)(uintptr_t)(p) + __DARWIN_ALIGNBYTES32) &~ __DARWIN_ALIGNBYTES32) +#endif +#define CMSG_ALIGN(n) __DARWIN_ALIGN32(n) +#endif +#define I_AM_HERE \ + do { \ + SCTP_PRINTF("%s:%d at %s\n", __FILE__, __LINE__ , __FUNCTION__); \ + } while (0) + +#ifndef timevalsub +#define timevalsub(tp1, tp2) \ + do { \ + (tp1)->tv_sec -= (tp2)->tv_sec; \ + (tp1)->tv_usec -= (tp2)->tv_usec; \ + if ((tp1)->tv_usec < 0) { \ + (tp1)->tv_sec--; \ + (tp1)->tv_usec += 1000000; \ + } \ + } while (0) +#endif + +#if defined(__Userspace_os_Linux) +#if !defined(TAILQ_FOREACH_SAFE) +#define TAILQ_FOREACH_SAFE(var, head, field, tvar) \ + for ((var) = ((head)->tqh_first); \ + (var) && ((tvar) = TAILQ_NEXT((var), field), 1); \ + (var) = (tvar)) +#endif +#if !defined(LIST_FOREACH_SAFE) +#define LIST_FOREACH_SAFE(var, head, field, tvar) \ + for ((var) = ((head)->lh_first); \ + (var) && ((tvar) = LIST_NEXT((var), field), 1); \ + (var) = (tvar)) +#endif +#endif +#if defined(__Userspace_os_DragonFly) +#define TAILQ_FOREACH_SAFE TAILQ_FOREACH_MUTABLE +#define LIST_FOREACH_SAFE LIST_FOREACH_MUTABLE +#endif + +#if defined(__Userspace_os_NaCl) +#define timercmp(tvp, uvp, cmp) \ + (((tvp)->tv_sec == (uvp)->tv_sec) ? \ + ((tvp)->tv_usec cmp (uvp)->tv_usec) : \ + ((tvp)->tv_sec cmp (uvp)->tv_sec)) +#endif + +#endif |