solaris.c

/*
 * Copyright (C) 1993-2011 by Darren Reed.
 *
 * See the IPFILTER.LICENCE file for details on licencing.
 */
/* #pragma ident   "@(#)solaris.c	1.12 6/5/96 (C) 1995 Darren Reed"*/
#pragma ident "@(#)$Id$"

#include <sys/systm.h>
#include <sys/types.h>
#include <sys/param.h>
#include <sys/errno.h>
#include <sys/uio.h>
#include <sys/buf.h>
#include <sys/modctl.h>
#include <sys/open.h>
#include <sys/kmem.h>
#include <sys/conf.h>
#include <sys/cmn_err.h>
#include <sys/stat.h>
#include <sys/cred.h>
#include <sys/dditypes.h>
#include <sys/stream.h>
#include <sys/poll.h>
#include <sys/autoconf.h>
#include <sys/byteorder.h>
#include <sys/socket.h>
#include <sys/dlpi.h>
#include <sys/stropts.h>
#include <sys/sockio.h>
#include <net/if.h>
#if SOLARIS2 >= 6
# include <net/if_types.h>
#endif
#include <net/af.h>
#include <net/route.h>
#include <netinet/in.h>
#include <netinet/in_systm.h>
#include <netinet/if_ether.h>
#include <netinet/ip.h>
#include <netinet/ip_var.h>
#include <netinet/tcp.h>
#include <netinet/udp.h>
#include <netinet/tcpip.h>
#include <netinet/ip_icmp.h>
#include <sys/ddi.h>
#include <sys/sunddi.h>
#include "netinet/ip_compat.h"
#include "netinet/ipl.h"
#include "netinet/ip_fil.h"
#include "netinet/ip_nat.h"
#include "netinet/ip_frag.h"
#include "netinet/ip_auth.h"
#include "netinet/ip_state.h"

struct pollhead iplpollhead[IPL_LOGSIZE];

extern	struct	filterstats	frstats[];
extern	int	fr_running;
extern	int	fr_flags;
extern	int	iplwrite __P((dev_t, struct uio *, cred_t *));

extern	ipnat_t	*nat_list;
extern	void	ipf_rand_push __P((void *, int));

static	int	ipf_getinfo __P((dev_info_t *, ddi_info_cmd_t,
				 void *, void **));
#if SOLARIS2 < 10
static	int	ipf_identify __P((dev_info_t *));
#endif
static	int	ipf_attach __P((dev_info_t *, ddi_attach_cmd_t));
static	int	ipf_detach __P((dev_info_t *, ddi_detach_cmd_t));
static	int	fr_qifsync __P((ip_t *, int, void *, int, void *, mblk_t **));
static	int	ipf_property_update __P((dev_info_t *));
static 	int	iplpoll __P((dev_t, short, int, short *, struct pollhead **));
static	char	*ipf_devfiles[] = { IPL_NAME, IPNAT_NAME, IPSTATE_NAME,
				    IPAUTH_NAME, IPSYNC_NAME, IPSCAN_NAME,
				    IPLOOKUP_NAME, NULL };


#if SOLARIS2 >= 7
extern	timeout_id_t	fr_timer_id;
#else
extern	int		fr_timer_id;
#endif


static struct cb_ops ipf_cb_ops = {
	iplopen,
	iplclose,
	nodev,		/* strategy */
	nodev,		/* print */
	nodev,		/* dump */
	iplread,
	iplwrite,	/* write */
	iplioctl,	/* ioctl */
	nodev,		/* devmap */
	nodev,		/* mmap */
	nodev,		/* segmap */
	iplpoll,	/* poll */
	ddi_prop_op,
	NULL,
	D_MTSAFE,
#if SOLARIS2 > 4
	CB_REV,
	nodev,		/* aread */
	nodev,		/* awrite */
#endif
};

static struct dev_ops ipf_ops = {
	DEVO_REV,
	0,
	ipf_getinfo,
#if SOLARIS2 >= 10
	nulldev,
#else
	ipf_identify,
#endif
	nulldev,
	ipf_attach,
	ipf_detach,
	nodev,		/* reset */
	&ipf_cb_ops,
	(struct bus_ops *)0
};

extern struct mod_ops mod_driverops;
static struct modldrv iplmod = {
	&mod_driverops, IPL_VERSION, &ipf_ops };
static struct modlinkage modlink1 = { MODREV_1, { &iplmod, NULL } };
static int ipf_pkts = 0;

#if SOLARIS2 >= 6
static	size_t	hdrsizes[57][2] = {
	{ 0, 0 },
	{ IFT_OTHER, 0 },
	{ IFT_1822, 0 },
	{ IFT_HDH1822, 0 },
	{ IFT_X25DDN, 0 },
	{ IFT_X25, 0 },
	{ IFT_ETHER, 14 },
	{ IFT_ISO88023, 0 },
	{ IFT_ISO88024, 0 },
	{ IFT_ISO88025, 0 },
	{ IFT_ISO88026, 0 },
	{ IFT_STARLAN, 0 },
	{ IFT_P10, 0 },
	{ IFT_P80, 0 },
	{ IFT_HY, 0 },
	{ IFT_FDDI, 24 },
	{ IFT_LAPB, 0 },
	{ IFT_SDLC, 0 },
	{ IFT_T1, 0 },
	{ IFT_CEPT, 0 },
	{ IFT_ISDNBASIC, 0 },
	{ IFT_ISDNPRIMARY, 0 },
	{ IFT_PTPSERIAL, 0 },
	{ IFT_PPP, 0 },
	{ IFT_LOOP, 0 },
	{ IFT_EON, 0 },
	{ IFT_XETHER, 0 },
	{ IFT_NSIP, 0 },
	{ IFT_SLIP, 0 },
	{ IFT_ULTRA, 0 },
	{ IFT_DS3, 0 },
	{ IFT_SIP, 0 },
	{ IFT_FRELAY, 0 },
	{ IFT_RS232, 0 },
	{ IFT_PARA, 0 },
	{ IFT_ARCNET, 0 },
	{ IFT_ARCNETPLUS, 0 },
	{ IFT_ATM, 0 },
	{ IFT_MIOX25, 0 },
	{ IFT_SONET, 0 },
	{ IFT_X25PLE, 0 },
	{ IFT_ISO88022LLC, 0 },
	{ IFT_LOCALTALK, 0 },
	{ IFT_SMDSDXI, 0 },
	{ IFT_FRELAYDCE, 0 },
	{ IFT_V35, 0 },
	{ IFT_HSSI, 0 },
	{ IFT_HIPPI, 0 },
	{ IFT_MODEM, 0 },
	{ IFT_AAL5, 0 },
	{ IFT_SONETPATH, 0 },
	{ IFT_SONETVT, 0 },
	{ IFT_SMDSICIP, 0 },
	{ IFT_PROPVIRTUAL, 0 },
	{ IFT_PROPMUX, 0 },
};
#endif /* SOLARIS2 >= 6 */

static dev_info_t *ipf_dev_info = NULL;

#if defined(_INET_IP_STACK_H)
static hook_t *ipfhook;
static int ipf_hook(hook_event_token_t, hook_data_t, void *);
net_handle_t ipfipv4, ipfipv6;
#endif


int _init()
{
	int ipfinst;

	ipfinst = mod_install(&modlink1);
#ifdef	IPFDEBUG
	cmn_err(CE_NOTE, "IP Filter: _init() = %d", ipfinst);
#endif
	return ipfinst;
}


int _fini(void)
{
	int ipfinst;

	ipfinst = mod_remove(&modlink1);
#ifdef	IPFDEBUG
	cmn_err(CE_NOTE, "IP Filter: _fini() = %d", ipfinst);
#endif
	return ipfinst;
}


int _info(modinfop)
struct modinfo *modinfop;
{
	int ipfinst;

	ipfinst = mod_info(&modlink1, modinfop);
#ifdef	IPFDEBUG
	cmn_err(CE_NOTE, "IP Filter: _info(%x) = %x", modinfop, ipfinst);
#endif
	return ipfinst;
}


#if SOLARIS2 < 10
static int ipf_identify(dip)
dev_info_t *dip;
{
# ifdef	IPFDEBUG
	cmn_err(CE_NOTE, "IP Filter: ipf_identify(%x)", dip);
# endif
	if (strcmp(ddi_get_name(dip), "ipf") == 0)
		return (DDI_IDENTIFIED);
	return (DDI_NOT_IDENTIFIED);
}
#endif


static int ipf_attach(dip, cmd)
dev_info_t *dip;
ddi_attach_cmd_t cmd;
{
	char *s;
	int i;
#ifdef	IPFDEBUG
	int instance;

	cmn_err(CE_NOTE, "IP Filter: ipf_attach(%x,%x)", dip, cmd);
#endif

	ipf_rand_push(dip, sizeof(*dip));

#if !defined(_INET_IP_STACK_H)
	if ((pfilinterface != PFIL_INTERFACE) || (PFIL_INTERFACE < 2000000)) {
		cmn_err(CE_NOTE, "pfilinterface(%d) != %d\n", pfilinterface,
			PFIL_INTERFACE);
		return EINVAL;
	}
#else
#endif

	switch (cmd)
	{
	case DDI_ATTACH:
		if (fr_running != 0)
			break;
#ifdef	IPFDEBUG
		instance = ddi_get_instance(dip);

		cmn_err(CE_NOTE, "IP Filter: attach ipf instance %d", instance);
#endif

		(void) ipf_property_update(dip);

		for (i = 0; ((s = ipf_devfiles[i]) != NULL); i++) {
			s = strrchr(s, '/');
			if (s == NULL)
				continue;
			s++;
			if (ddi_create_minor_node(dip, s, S_IFCHR, i,
						  DDI_PSEUDO, 0) ==
			    DDI_FAILURE) {
				ddi_remove_minor_node(dip, NULL);
				goto attach_failed;
			}
		}

		ipf_dev_info = dip;
		/*
		 * Initialize mutex's
		 */
		RWLOCK_INIT(&ipf_global, "ipf filter load/unload mutex");
		RWLOCK_INIT(&ipf_mutex, "ipf filter rwlock");
		RWLOCK_INIT(&ipf_frcache, "ipf cache rwlock");

		/*
		 * Lock people out while we set things up.
		 */
		WRITE_ENTER(&ipf_global);
		if ((fr_running != 0) || (ipfattach() == -1)) {
			RWLOCK_EXIT(&ipf_global);
			goto attach_failed;
		}

#if !defined(_INET_IP_STACK_H)
		if (pfil_add_hook(fr_check, PFIL_IN|PFIL_OUT, &pfh_inet4))
			cmn_err(CE_WARN, "IP Filter: %s(pfh_inet4) failed",
				"pfil_add_hook");
# ifdef USE_INET6
		if (pfil_add_hook(fr_check, PFIL_IN|PFIL_OUT, &pfh_inet6))
			cmn_err(CE_WARN, "IP Filter: %s(pfh_inet6) failed",
				"pfil_add_hook");
# endif
		if (pfil_add_hook(fr_qifsync, PFIL_IN|PFIL_OUT, &pfh_sync))
			cmn_err(CE_WARN, "IP Filter: %s(pfh_sync) failed",
				"pfil_add_hook");
#else
		HOOK_INIT(ipfhook, ipf_hook, "ipf_v4", NULL);

		ipfipv4 = net_protocol_lookup(0, NHF_INET);
		if (ipfipv4 == NULL)
			goto attach_failed;
		if (net_hook_register(ipfipv4, NH_PHYSICAL_IN, ipfhook))
			goto attach_failed;
		if (net_hook_register(ipfipv4, NH_PHYSICAL_OUT, ipfhook))
			goto attach_failed;
# ifdef USE_INET6
		ipfipv6 = net_protocol_lookup(0, NHF_INET6);
		if (ipfipv6 == NULL)
			goto attach_failed;
		if (net_hook_register(ipfipv6, NH_PHYSICAL_IN, ipfhook))
			goto attach_failed;
		if (net_hook_register(ipfipv6, NH_PHYSICAL_OUT, ipfhook))
			goto attach_failed;
# endif
#endif

		fr_timer_id = timeout(fr_slowtimer, NULL,
				      drv_usectohz(1000000 * IPF_HZ_MULT /
						   IPF_HZ_DIVIDE));


		fr_running = 1;

		RWLOCK_EXIT(&ipf_global);

		cmn_err(CE_CONT, "!%s, running.\n", ipfilter_version);

		return DDI_SUCCESS;
		/* NOTREACHED */
	default:
		break;
	}

attach_failed:
#ifdef	IPFDEBUG
	cmn_err(CE_NOTE, "IP Filter: failed to attach\n");
#endif
	/*
	 * Use our own detach routine to toss
	 * away any stuff we allocated above.
	 */
	(void) ipf_detach(dip, DDI_DETACH);
	return DDI_FAILURE;
}


static int ipf_detach(dip, cmd)
dev_info_t *dip;
ddi_detach_cmd_t cmd;
{
	int i;

#ifdef	IPFDEBUG
	cmn_err(CE_NOTE, "IP Filter: ipf_detach(%x,%x)", dip, cmd);
#endif
	switch (cmd) {
	case DDI_DETACH:
		if (fr_refcnt != 0)
			return DDI_FAILURE;

		if (fr_running == -2 || fr_running == 0)
			break;
		/*
		 * Make sure we're the only one's modifying things.  With
		 * this lock others should just fall out of the loop.
		 */
		WRITE_ENTER(&ipf_global);
		if (fr_running == -2) {
			RWLOCK_EXIT(&ipf_global);
			return DDI_FAILURE;
		}
		fr_running = -2;

#if !defined(_INET_IP_STACK_H)
		if (pfil_remove_hook(fr_check, PFIL_IN|PFIL_OUT, &pfh_inet4))
			cmn_err(CE_WARN, "IP Filter: %s(pfh_inet4) failed",
				"pfil_remove_hook");
# ifdef USE_INET6
		if (pfil_remove_hook(fr_check, PFIL_IN|PFIL_OUT, &pfh_inet6))
			cmn_err(CE_WARN, "IP Filter: %s(pfh_inet6) failed",
				"pfil_add_hook");
# endif
		if (pfil_remove_hook(fr_qifsync, PFIL_IN|PFIL_OUT, &pfh_sync))
			cmn_err(CE_WARN, "IP Filter: %s(pfh_sync) failed",
				"pfil_remove_hook");
#else
		if (net_hook_unregister(ipfipv4, NH_PHYSICAL_IN, ipfhook))
			cmn_err(CE_WARN, "IP Filter: v4-IN unregister failed");
		if (net_hook_unregister(ipfipv4, NH_PHYSICAL_OUT, ipfhook))
			cmn_err(CE_WARN, "IP Filter: v4-OUT unregister failed");
		if (net_protocol_release(ipfipv4))
			cmn_err(CE_WARN,
				"IP Filter: v4 net_protocol_release failed");
# ifdef USE_INET6
		if (net_hook_unregister(ipfipv6, NH_PHYSICAL_IN, ipfhook))
			cmn_err(CE_WARN, "IP Filter: v6-IN unregister failed");
		if (net_hook_unregister(ipfipv6, NH_PHYSICAL_OUT, ipfhook))
			cmn_err(CE_WARN, "IP Filter: v6-OUT unregister failed");
		if (net_protocol_release(ipfipv6))
			cmn_err(CE_WARN,
				"IP Filter: v6 net_protocol_release failed");
# endif
#endif

		RWLOCK_EXIT(&ipf_global);

		if (fr_timer_id != 0) {
			(void) untimeout(fr_timer_id);
			fr_timer_id = 0;
		}

		/*
		 * Undo what we did in ipf_attach, freeing resources
		 * and removing things we installed.  The system
		 * framework guarantees we are not active with this devinfo
		 * node in any other entry points at this time.
		 */
		ddi_prop_remove_all(dip);
		i = ddi_get_instance(dip);
		ddi_remove_minor_node(dip, NULL);
		if (i > 0) {
			cmn_err(CE_CONT, "IP Filter: still attached (%d)\n", i);
			return DDI_FAILURE;
		}

		WRITE_ENTER(&ipf_global);
		if (!ipfdetach()) {
			RWLOCK_EXIT(&ipf_global);
			RW_DESTROY(&ipf_mutex);
			RW_DESTROY(&ipf_frcache);
			RW_DESTROY(&ipf_global);
			cmn_err(CE_CONT, "!%s detached.\n", ipfilter_version);
			return (DDI_SUCCESS);
		}
		RWLOCK_EXIT(&ipf_global);
		break;
	default:
		break;
	}
	cmn_err(CE_NOTE, "IP Filter: failed to detach\n");
	return DDI_FAILURE;
}


/*ARGSUSED*/
static int ipf_getinfo(dip, infocmd, arg, result)
dev_info_t *dip;
ddi_info_cmd_t infocmd;
void *arg, **result;
{
	int error;

	if (fr_running <= 0)
		return DDI_FAILURE;
	error = DDI_FAILURE;
#ifdef	IPFDEBUG
	cmn_err(CE_NOTE, "IP Filter: ipf_getinfo(%x,%x,%x)", dip, infocmd, arg);
#endif
	switch (infocmd) {
	case DDI_INFO_DEVT2DEVINFO:
		*result = ipf_dev_info;
		error = DDI_SUCCESS;
		break;
	case DDI_INFO_DEVT2INSTANCE:
		*result = (void *)0;
		error = DDI_SUCCESS;
		break;
	default:
		break;
	}
	return (error);
}


/*
 * look for bad consistancies between the list of interfaces the filter knows
 * about and those which are currently configured.
 */
/*ARGSUSED*/
static int fr_qifsync(ip, hlen, il, out, qif, mp)
ip_t *ip;
int hlen;
void *il;
int out;
void *qif;
mblk_t **mp;
{

	frsync(qif);
	/*
	 * Resync. any NAT `connections' using this interface and its IP #.
	 */
	fr_natsync(qif);
	fr_statesync(qif);
	return 0;
}


/*
 * look for bad consistancies between the list of interfaces the filter knows
 * about and those which are currently configured.
 */
int ipfsync()
{
	frsync(NULL);
	return 0;
}


/*
 * Fetch configuration file values that have been entered into the ipf.conf
 * driver file.
 */
static int ipf_property_update(dip)
dev_info_t *dip;
{
	ipftuneable_t *ipft;
	u_long i64;
	char *name;
	u_int one;
	int *i32p;
	int err;

#ifdef DDI_NO_AUTODETACH
	if (ddi_prop_update_int(DDI_DEV_T_NONE, dip,
				DDI_NO_AUTODETACH, 1) != DDI_PROP_SUCCESS) {
		cmn_err(CE_WARN, "!updating DDI_NO_AUTODETACH failed");
		return DDI_FAILURE;
	}
#else
	if (ddi_prop_update_int(DDI_DEV_T_NONE, dip,
				"ddi-no-autodetach", 1) != DDI_PROP_SUCCESS) {
		cmn_err(CE_WARN, "!updating ddi-no-autodetach failed");
		return DDI_FAILURE;
	}
#endif

	err = DDI_SUCCESS;
	ipft = ipf_tuneables;
	for (ipft = ipf_tuneables; (name = (char *)ipft->ipft_name) != NULL;
	     ipft++) {
		one = 1;
		i32p = NULL;
		err = ddi_prop_lookup_int_array(DDI_DEV_T_ANY, dip,
						0, name, &i32p, &one);
		if (err == DDI_PROP_NOT_FOUND)
			continue;
		if (*i32p >= ipft->ipft_min && *i32p <= ipft->ipft_max) {
			if (ipft->ipft_sz == sizeof(int)) {
				*ipft->ipft_pint = *i32p;
			} else {
				i64 = *(u_int *)i32p;
				*(u_long *)ipft->ipft_pint = i64;
			}
		}
	}

	return err;
}


static int iplpoll(dev, events, anyyet, reventsp, phpp)
dev_t dev;
short events;
int anyyet;
short *reventsp;
struct pollhead **phpp;
{
	u_int xmin = getminor(dev);
	int revents = 0;

	if (xmin > IPL_LOGMAX)
		return ENXIO;

	switch (xmin)
	{
	case IPL_LOGIPF :
	case IPL_LOGNAT :
	case IPL_LOGSTATE :
#ifdef IPFILTER_LOG
		if ((events & (POLLIN | POLLRDNORM)) && ipflog_canread(xmin))
			revents |= events & (POLLIN | POLLRDNORM);
#endif
		break;
	case IPL_LOGAUTH :
		if ((events & (POLLIN | POLLRDNORM)) && fr_auth_waiting())
			revents |= events & (POLLIN | POLLRDNORM);
		break;
	case IPL_LOGSYNC :
#ifdef IPFILTER_SYNC
		if ((events & (POLLIN | POLLRDNORM)) && ipfsync_canread())
			revents |= events & (POLLIN | POLLRDNORM);
		if ((events & (POLLOUT | POLLWRNORM)) && ipfsync_canwrite())
			revents |= events & (POLLOUT | POLLOUTNORM);
#endif
		break;
	case IPL_LOGSCAN :
	case IPL_LOGLOOKUP :
	default :
		break;
	}

	if (revents) {
		*reventsp = revents;
	} else {
		*reventsp = 0;
		if (!anyyet)
			*phpp = &iplpollhead[xmin];
	}
	return 0;
}


#if defined(_INET_IP_STACK_H)
static int
ipf_hook(hook_event_token_t event, hook_data_t data, void *stp)
{
	hook_pkt_event_t *hpe;
	qpktinfo_t qpi;
	int out, hlen;
	ip_t *ip;

	hpe = (hook_pkt_event_t *)data;
	qpi.qpi_data = hpe->hpe_hdr;
	qpi.qpi_mp = hpe->hpe_mp;
	qpi.qpi_m = hpe->hpe_mb;

	if (hpe->hpe_ifp > 0) {
		qpi.qpi_real = hpe->hpe_ifp;
		out = 0;
	} else {
		qpi.qpi_real = hpe->hpe_ofp;
		out = 1;
	}
	qpi.qpi_flags = 0;
	qpi.qpi_num = hpe->hpe_ifp;
	qpi.qpi_off = 0;

	if ((++ipf_pkts & 0xffff) == 0) {
		ipf_rand_push(MTOD(qpi.qpi_m, void *), M_LEN(qpi.qpi_m));
	}

	ip = hpe->hpe_hdr;
	if (ip->ip_v == 4) {
		hlen = ip->ip_hl << 2;
		ip->ip_off = ntohs(ip->ip_off);
		ip->ip_len = ntohs(ip->ip_len);
	} else {
		hlen = sizeof(ip6_t);
	}

	int r = fr_check(hpe->hpe_hdr, hlen, (void *)qpi.qpi_real, out, &qpi, qpi.qpi_mp);
	hpe->hpe_hdr = qpi.qpi_data;
	hpe->hpe_mb = qpi.qpi_m;
	ip = hpe->hpe_hdr;
	if(r == 0) {
		if(!*qpi.qpi_mp) r = 1;
		else if (ip->ip_v == 4) {
			ip->ip_off = htons(ip->ip_off);
			ip->ip_len = htons(ip->ip_len);
		}
	}
	return r;
}
#endif