#include <stdint.h>
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/filedesc.h>
#include <sys/kernel.h>
#include <sys/proc_internal.h>
#include <sys/kauth.h>
#include <sys/malloc.h>
#include <sys/unistd.h>
#include <sys/file_internal.h>
#include <sys/fcntl.h>
#include <sys/select.h>
#include <sys/queue.h>
#include <sys/event.h>
#include <sys/eventvar.h>
#include <sys/protosw.h>
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/stat.h>
#include <sys/sysctl.h>
#include <sys/uio.h>
#include <sys/sysproto.h>
#include <sys/user.h>
#include <sys/vnode_internal.h>
#include <string.h>
#include <sys/proc_info.h>
#include <sys/codesign.h>
#include <kern/locks.h>
#include <kern/clock.h>
#include <kern/thread_call.h>
#include <kern/sched_prim.h>
#include <kern/zalloc.h>
#include <kern/assert.h>
#include <libkern/libkern.h>
#include "net/net_str_id.h"
#include <mach/task.h>
#if VM_PRESSURE_EVENTS
#include <kern/vm_pressure.h>
#endif
#if CONFIG_MEMORYSTATUS
#include <sys/kern_memorystatus.h>
#endif
MALLOC_DEFINE(M_KQUEUE, "kqueue", "memory for kqueue system");
#define KQ_EVENT NULL
static inline void kqlock(struct kqueue *kq);
static inline void kqunlock(struct kqueue *kq);
static int kqlock2knoteuse(struct kqueue *kq, struct knote *kn);
static int kqlock2knoteusewait(struct kqueue *kq, struct knote *kn);
static int kqlock2knotedrop(struct kqueue *kq, struct knote *kn);
static int knoteuse2kqlock(struct kqueue *kq, struct knote *kn);
static void kqueue_wakeup(struct kqueue *kq, int closed);
static int kqueue_read(struct fileproc *fp, struct uio *uio,
int flags, vfs_context_t ctx);
static int kqueue_write(struct fileproc *fp, struct uio *uio,
int flags, vfs_context_t ctx);
static int kqueue_ioctl(struct fileproc *fp, u_long com, caddr_t data,
vfs_context_t ctx);
static int kqueue_select(struct fileproc *fp, int which, void *wql,
vfs_context_t ctx);
static int kqueue_close(struct fileglob *fg, vfs_context_t ctx);
static int kqueue_kqfilter(struct fileproc *fp, struct knote *kn,
vfs_context_t ctx);
static int kqueue_drain(struct fileproc *fp, vfs_context_t ctx);
static const struct fileops kqueueops = {
.fo_type = DTYPE_KQUEUE,
.fo_read = kqueue_read,
.fo_write = kqueue_write,
.fo_ioctl = kqueue_ioctl,
.fo_select = kqueue_select,
.fo_close = kqueue_close,
.fo_kqfilter = kqueue_kqfilter,
.fo_drain = kqueue_drain,
};
static int kevent_internal(struct proc *p, int iskev64, user_addr_t changelist,
int nchanges, user_addr_t eventlist, int nevents, int fd,
user_addr_t utimeout, unsigned int flags, int32_t *retval);
static int kevent_copyin(user_addr_t *addrp, struct kevent64_s *kevp,
struct proc *p, int iskev64);
static int kevent_copyout(struct kevent64_s *kevp, user_addr_t *addrp,
struct proc *p, int iskev64);
char * kevent_description(struct kevent64_s *kevp, char *s, size_t n);
static int kevent_callback(struct kqueue *kq, struct kevent64_s *kevp,
void *data);
static void kevent_continue(struct kqueue *kq, void *data, int error);
static void kqueue_scan_continue(void *contp, wait_result_t wait_result);
static int kqueue_process(struct kqueue *kq, kevent_callback_t callback,
void *data, int *countp, struct proc *p);
static int kqueue_begin_processing(struct kqueue *kq);
static void kqueue_end_processing(struct kqueue *kq);
static int knote_process(struct knote *kn, kevent_callback_t callback,
void *data, struct kqtailq *inprocessp, struct proc *p);
static void knote_put(struct knote *kn);
static int knote_fdpattach(struct knote *kn, struct filedesc *fdp,
struct proc *p);
static void knote_drop(struct knote *kn, struct proc *p);
static void knote_activate(struct knote *kn, int);
static void knote_deactivate(struct knote *kn);
static void knote_enqueue(struct knote *kn);
static void knote_dequeue(struct knote *kn);
static struct knote *knote_alloc(void);
static void knote_free(struct knote *kn);
static int filt_fileattach(struct knote *kn);
static struct filterops file_filtops = {
.f_isfd = 1,
.f_attach = filt_fileattach,
};
static void filt_kqdetach(struct knote *kn);
static int filt_kqueue(struct knote *kn, long hint);
static struct filterops kqread_filtops = {
.f_isfd = 1,
.f_detach = filt_kqdetach,
.f_event = filt_kqueue,
};
static int filt_badattach(struct knote *kn);
static struct filterops bad_filtops = {
.f_attach = filt_badattach,
};
static int filt_procattach(struct knote *kn);
static void filt_procdetach(struct knote *kn);
static int filt_proc(struct knote *kn, long hint);
static struct filterops proc_filtops = {
.f_attach = filt_procattach,
.f_detach = filt_procdetach,
.f_event = filt_proc,
};
#if VM_PRESSURE_EVENTS
static int filt_vmattach(struct knote *kn);
static void filt_vmdetach(struct knote *kn);
static int filt_vm(struct knote *kn, long hint);
static struct filterops vm_filtops = {
.f_attach = filt_vmattach,
.f_detach = filt_vmdetach,
.f_event = filt_vm,
};
#endif
#if CONFIG_MEMORYSTATUS
extern struct filterops memorystatus_filtops;
#endif
extern struct filterops fs_filtops;
extern struct filterops sig_filtops;
static int filt_timerattach(struct knote *kn);
static void filt_timerdetach(struct knote *kn);
static int filt_timer(struct knote *kn, long hint);
static void filt_timertouch(struct knote *kn, struct kevent64_s *kev,
long type);
static struct filterops timer_filtops = {
.f_attach = filt_timerattach,
.f_detach = filt_timerdetach,
.f_event = filt_timer,
.f_touch = filt_timertouch,
};
static void filt_timerexpire(void *knx, void *param1);
static int filt_timervalidate(struct knote *kn);
static void filt_timerupdate(struct knote *kn);
static void filt_timercancel(struct knote *kn);
#define TIMER_RUNNING 0x1
#define TIMER_CANCELWAIT 0x2
static lck_mtx_t _filt_timerlock;
static void filt_timerlock(void);
static void filt_timerunlock(void);
static zone_t knote_zone;
#define KN_HASH(val, mask) (((val) ^ (val >> 8)) & (mask))
#if 0
extern struct filterops aio_filtops;
#endif
extern struct filterops machport_filtops;
static int filt_userattach(struct knote *kn);
static void filt_userdetach(struct knote *kn);
static int filt_user(struct knote *kn, long hint);
static void filt_usertouch(struct knote *kn, struct kevent64_s *kev,
long type);
static struct filterops user_filtops = {
.f_attach = filt_userattach,
.f_detach = filt_userdetach,
.f_event = filt_user,
.f_touch = filt_usertouch,
};
static struct filterops *sysfilt_ops[] = {
&file_filtops,
&file_filtops,
#if 0
&aio_filtops,
#else
&bad_filtops,
#endif
&file_filtops,
&proc_filtops,
&sig_filtops,
&timer_filtops,
&machport_filtops,
&fs_filtops,
&user_filtops,
&bad_filtops,
#if VM_PRESSURE_EVENTS
&vm_filtops,
#else
&bad_filtops,
#endif
&file_filtops,
#if CONFIG_MEMORYSTATUS
&memorystatus_filtops,
#else
&bad_filtops,
#endif
};
lck_grp_attr_t * kq_lck_grp_attr;
lck_grp_t * kq_lck_grp;
lck_attr_t * kq_lck_attr;
static inline void
kqlock(struct kqueue *kq)
{
lck_spin_lock(&kq->kq_lock);
}
static inline void
kqunlock(struct kqueue *kq)
{
lck_spin_unlock(&kq->kq_lock);
}
static int
kqlock2knoteuse(struct kqueue *kq, struct knote *kn)
{
if (kn->kn_status & KN_DROPPING)
return (0);
kn->kn_inuse++;
kqunlock(kq);
return (1);
}
static int
kqlock2knoteusewait(struct kqueue *kq, struct knote *kn)
{
if ((kn->kn_status & (KN_DROPPING | KN_ATTACHING)) != 0) {
kn->kn_status |= KN_USEWAIT;
wait_queue_assert_wait((wait_queue_t)kq->kq_wqs,
&kn->kn_status, THREAD_UNINT, 0);
kqunlock(kq);
thread_block(THREAD_CONTINUE_NULL);
return (0);
}
kn->kn_inuse++;
kqunlock(kq);
return (1);
}
static int
knoteuse2kqlock(struct kqueue *kq, struct knote *kn)
{
kqlock(kq);
if (--kn->kn_inuse == 0) {
if ((kn->kn_status & KN_ATTACHING) != 0) {
kn->kn_status &= ~KN_ATTACHING;
}
if ((kn->kn_status & KN_USEWAIT) != 0) {
kn->kn_status &= ~KN_USEWAIT;
wait_queue_wakeup_all((wait_queue_t)kq->kq_wqs,
&kn->kn_status, THREAD_AWAKENED);
}
}
return ((kn->kn_status & KN_DROPPING) == 0);
}
static int
kqlock2knotedrop(struct kqueue *kq, struct knote *kn)
{
int oktodrop;
oktodrop = ((kn->kn_status & (KN_DROPPING | KN_ATTACHING)) == 0);
kn->kn_status |= KN_DROPPING;
if (oktodrop) {
if (kn->kn_inuse == 0) {
kqunlock(kq);
return (oktodrop);
}
}
kn->kn_status |= KN_USEWAIT;
wait_queue_assert_wait((wait_queue_t)kq->kq_wqs, &kn->kn_status,
THREAD_UNINT, 0);
kqunlock(kq);
thread_block(THREAD_CONTINUE_NULL);
return (oktodrop);
}
static void
knote_put(struct knote *kn)
{
struct kqueue *kq = kn->kn_kq;
kqlock(kq);
if (--kn->kn_inuse == 0) {
if ((kn->kn_status & KN_USEWAIT) != 0) {
kn->kn_status &= ~KN_USEWAIT;
wait_queue_wakeup_all((wait_queue_t)kq->kq_wqs,
&kn->kn_status, THREAD_AWAKENED);
}
}
kqunlock(kq);
}
static int
filt_fileattach(struct knote *kn)
{
return (fo_kqfilter(kn->kn_fp, kn, vfs_context_current()));
}
#define f_flag f_fglob->fg_flag
#define f_msgcount f_fglob->fg_msgcount
#define f_cred f_fglob->fg_cred
#define f_ops f_fglob->fg_ops
#define f_offset f_fglob->fg_offset
#define f_data f_fglob->fg_data
static void
filt_kqdetach(struct knote *kn)
{
struct kqueue *kq = (struct kqueue *)kn->kn_fp->f_data;
kqlock(kq);
KNOTE_DETACH(&kq->kq_sel.si_note, kn);
kqunlock(kq);
}
static int
filt_kqueue(struct knote *kn, __unused long hint)
{
struct kqueue *kq = (struct kqueue *)kn->kn_fp->f_data;
kn->kn_data = kq->kq_count;
return (kn->kn_data > 0);
}
static int
filt_procattach(struct knote *kn)
{
struct proc *p;
assert(PID_MAX < NOTE_PDATAMASK);
if ((kn->kn_sfflags & (NOTE_TRACK | NOTE_TRACKERR | NOTE_CHILD)) != 0)
return (ENOTSUP);
p = proc_find(kn->kn_id);
if (p == NULL) {
return (ESRCH);
}
const int NoteExitStatusBits = NOTE_EXIT | NOTE_EXITSTATUS;
if ((kn->kn_sfflags & NoteExitStatusBits) == NoteExitStatusBits)
do {
pid_t selfpid = proc_selfpid();
if (p->p_ppid == selfpid)
break;
if ((p->p_lflag & P_LTRACED) != 0 &&
(p->p_oppid == selfpid))
break;
proc_rele(p);
return (EACCES);
} while (0);
proc_klist_lock();
kn->kn_flags |= EV_CLEAR;
kn->kn_ptr.p_proc = p;
KNOTE_ATTACH(&p->p_klist, kn);
proc_klist_unlock();
proc_rele(p);
return (0);
}
static void
filt_procdetach(struct knote *kn)
{
struct proc *p;
proc_klist_lock();
p = kn->kn_ptr.p_proc;
if (p != PROC_NULL) {
kn->kn_ptr.p_proc = PROC_NULL;
KNOTE_DETACH(&p->p_klist, kn);
}
proc_klist_unlock();
}
static int
filt_proc(struct knote *kn, long hint)
{
if (hint != 0) {
u_int event;
event = (u_int)hint & NOTE_PCTRLMASK;
if (event & NOTE_EXIT) {
if ((kn->kn_ptr.p_proc->p_oppid != 0)
&& (kn->kn_kq->kq_p->p_pid != kn->kn_ptr.p_proc->p_ppid)) {
return 0;
}
}
if (kn->kn_sfflags & event)
kn->kn_fflags |= event;
#pragma clang diagnostic push
#pragma clang diagnostic ignored "-Wdeprecated-declarations"
if ((event == NOTE_REAP) || ((event == NOTE_EXIT) && !(kn->kn_sfflags & NOTE_REAP))) {
kn->kn_flags |= (EV_EOF | EV_ONESHOT);
}
#pragma clang diagnostic pop
if (event == NOTE_EXIT) {
kn->kn_data = 0;
if ((kn->kn_sfflags & NOTE_EXITSTATUS) != 0) {
kn->kn_fflags |= NOTE_EXITSTATUS;
kn->kn_data |= (hint & NOTE_PDATAMASK);
}
if ((kn->kn_sfflags & NOTE_EXIT_DETAIL) != 0) {
kn->kn_fflags |= NOTE_EXIT_DETAIL;
if ((kn->kn_ptr.p_proc->p_lflag &
P_LTERM_DECRYPTFAIL) != 0) {
kn->kn_data |= NOTE_EXIT_DECRYPTFAIL;
}
if ((kn->kn_ptr.p_proc->p_lflag &
P_LTERM_JETSAM) != 0) {
kn->kn_data |= NOTE_EXIT_MEMORY;
switch (kn->kn_ptr.p_proc->p_lflag &
P_JETSAM_MASK) {
case P_JETSAM_VMPAGESHORTAGE:
kn->kn_data |= NOTE_EXIT_MEMORY_VMPAGESHORTAGE;
break;
case P_JETSAM_VMTHRASHING:
kn->kn_data |= NOTE_EXIT_MEMORY_VMTHRASHING;
break;
case P_JETSAM_FCTHRASHING:
kn->kn_data |= NOTE_EXIT_MEMORY_FCTHRASHING;
break;
case P_JETSAM_VNODE:
kn->kn_data |= NOTE_EXIT_MEMORY_VNODE;
break;
case P_JETSAM_HIWAT:
kn->kn_data |= NOTE_EXIT_MEMORY_HIWAT;
break;
case P_JETSAM_PID:
kn->kn_data |= NOTE_EXIT_MEMORY_PID;
break;
case P_JETSAM_IDLEEXIT:
kn->kn_data |= NOTE_EXIT_MEMORY_IDLE;
break;
}
}
if ((kn->kn_ptr.p_proc->p_csflags &
CS_KILLED) != 0) {
kn->kn_data |= NOTE_EXIT_CSERROR;
}
}
}
}
return (kn->kn_fflags != 0);
}
#if VM_PRESSURE_EVENTS
static int
filt_vmattach(struct knote *kn)
{
kn->kn_flags |= EV_CLEAR;
return (vm_knote_register(kn));
}
static void
filt_vmdetach(struct knote *kn)
{
vm_knote_unregister(kn);
}
static int
filt_vm(struct knote *kn, long hint)
{
if (hint != 0) {
const pid_t pid = (pid_t)hint;
if ((kn->kn_sfflags & NOTE_VM_PRESSURE) &&
(kn->kn_kq->kq_p->p_pid == pid)) {
kn->kn_fflags |= NOTE_VM_PRESSURE;
}
}
return (kn->kn_fflags != 0);
}
#endif
static int
filt_timervalidate(struct knote *kn)
{
uint64_t multiplier;
uint64_t raw = 0;
switch (kn->kn_sfflags & (NOTE_SECONDS|NOTE_USECONDS|NOTE_NSECONDS)) {
case NOTE_SECONDS:
multiplier = NSEC_PER_SEC;
break;
case NOTE_USECONDS:
multiplier = NSEC_PER_USEC;
break;
case NOTE_NSECONDS:
multiplier = 1;
break;
case 0:
multiplier = NSEC_PER_SEC / 1000;
break;
default:
return (EINVAL);
}
if(kn->kn_sfflags & NOTE_LEEWAY){
nanoseconds_to_absolutetime((uint64_t)kn->kn_ext[1] * multiplier, &raw);
kn->kn_ext[1] = raw;
}
nanoseconds_to_absolutetime((uint64_t)kn->kn_sdata * multiplier, &raw);
kn->kn_ext[0] = 0;
kn->kn_sdata = 0;
if (kn->kn_sfflags & NOTE_ABSOLUTE) {
clock_sec_t seconds;
clock_nsec_t nanoseconds;
uint64_t now;
clock_get_calendar_nanotime(&seconds, &nanoseconds);
nanoseconds_to_absolutetime((uint64_t)seconds * NSEC_PER_SEC +
nanoseconds, &now);
if (raw < now) {
kn->kn_ext[0] = 0;
} else {
raw -= now;
clock_absolutetime_interval_to_deadline(raw,
&kn->kn_ext[0]);
}
} else {
kn->kn_sdata = raw;
}
return (0);
}
static void
filt_timerupdate(struct knote *kn)
{
if (kn->kn_sdata == 0)
return;
if (kn->kn_ext[0] == 0) {
clock_absolutetime_interval_to_deadline(kn->kn_sdata,
&kn->kn_ext[0]);
} else {
kn->kn_ext[0] += kn->kn_sdata;
}
}
static void
filt_timerexpire(void *knx, __unused void *spare)
{
struct klist timer_list;
struct knote *kn = knx;
filt_timerlock();
kn->kn_hookid &= ~TIMER_RUNNING;
SLIST_INIT(&timer_list);
SLIST_INSERT_HEAD(&timer_list, kn, kn_selnext);
KNOTE(&timer_list, 1);
if (kn->kn_hookid & TIMER_CANCELWAIT) {
struct kqueue *kq = kn->kn_kq;
wait_queue_wakeup_all((wait_queue_t)kq->kq_wqs, &kn->kn_hook,
THREAD_AWAKENED);
}
filt_timerunlock();
}
static void
filt_timercancel(struct knote *kn)
{
struct kqueue *kq = kn->kn_kq;
thread_call_t callout = kn->kn_hook;
boolean_t cancelled;
if (kn->kn_hookid & TIMER_RUNNING) {
cancelled = thread_call_cancel(callout);
if (cancelled) {
kn->kn_hookid &= ~TIMER_RUNNING;
} else {
kn->kn_hookid |= TIMER_CANCELWAIT;
wait_queue_assert_wait((wait_queue_t)kq->kq_wqs,
&kn->kn_hook, THREAD_UNINT, 0);
filt_timerunlock();
thread_block(THREAD_CONTINUE_NULL);
filt_timerlock();
assert((kn->kn_hookid & TIMER_RUNNING) == 0);
}
}
}
static int
filt_timerattach(struct knote *kn)
{
thread_call_t callout;
int error;
callout = thread_call_allocate(filt_timerexpire, kn);
if (NULL == callout)
return (ENOMEM);
filt_timerlock();
error = filt_timervalidate(kn);
if (error != 0) {
filt_timerunlock();
return (error);
}
kn->kn_hook = (void*)callout;
kn->kn_hookid = 0;
if (kn->kn_sfflags & NOTE_ABSOLUTE)
kn->kn_flags |= EV_ONESHOT;
filt_timerupdate(kn);
if (kn->kn_ext[0]) {
kn->kn_flags |= EV_CLEAR;
unsigned int timer_flags = 0;
if (kn->kn_sfflags & NOTE_CRITICAL)
timer_flags |= THREAD_CALL_DELAY_USER_CRITICAL;
else if (kn->kn_sfflags & NOTE_BACKGROUND)
timer_flags |= THREAD_CALL_DELAY_USER_BACKGROUND;
else
timer_flags |= THREAD_CALL_DELAY_USER_NORMAL;
if (kn->kn_sfflags & NOTE_LEEWAY)
timer_flags |= THREAD_CALL_DELAY_LEEWAY;
thread_call_enter_delayed_with_leeway(callout, NULL,
kn->kn_ext[0], kn->kn_ext[1], timer_flags);
kn->kn_hookid |= TIMER_RUNNING;
} else {
kn->kn_data = 1;
}
filt_timerunlock();
return (0);
}
static void
filt_timerdetach(struct knote *kn)
{
thread_call_t callout;
filt_timerlock();
callout = (thread_call_t)kn->kn_hook;
filt_timercancel(kn);
filt_timerunlock();
thread_call_free(callout);
}
static int
filt_timer(struct knote *kn, long hint)
{
int result;
if (hint) {
kn->kn_data++;
if (((kn->kn_hookid & TIMER_CANCELWAIT) == 0) &&
((kn->kn_flags & EV_ONESHOT) == 0)) {
filt_timerupdate(kn);
if (kn->kn_ext[0]) {
unsigned int timer_flags = 0;
if (kn->kn_sfflags & NOTE_CRITICAL)
timer_flags |= THREAD_CALL_DELAY_USER_CRITICAL;
else if (kn->kn_sfflags & NOTE_BACKGROUND)
timer_flags |= THREAD_CALL_DELAY_USER_BACKGROUND;
else
timer_flags |= THREAD_CALL_DELAY_USER_NORMAL;
if (kn->kn_sfflags & NOTE_LEEWAY)
timer_flags |= THREAD_CALL_DELAY_LEEWAY;
thread_call_enter_delayed_with_leeway(kn->kn_hook, NULL,
kn->kn_ext[0], kn->kn_ext[1], timer_flags);
kn->kn_hookid |= TIMER_RUNNING;
}
}
return (1);
}
filt_timerlock();
result = (kn->kn_data != 0);
filt_timerunlock();
return (result);
}
static void
filt_timertouch(struct knote *kn, struct kevent64_s *kev, long type)
{
int error;
filt_timerlock();
switch (type) {
case EVENT_REGISTER:
filt_timercancel(kn);
kn->kn_sdata = kev->data;
kn->kn_sfflags = kev->fflags;
kn->kn_ext[0] = kev->ext[0];
kn->kn_ext[1] = kev->ext[1];
error = filt_timervalidate(kn);
if (error) {
kn->kn_flags |= EV_ERROR;
kn->kn_data = error;
break;
}
filt_timerupdate(kn);
if (kn->kn_ext[0]) {
unsigned int timer_flags = 0;
if (kn->kn_sfflags & NOTE_CRITICAL)
timer_flags |= THREAD_CALL_DELAY_USER_CRITICAL;
else if (kn->kn_sfflags & NOTE_BACKGROUND)
timer_flags |= THREAD_CALL_DELAY_USER_BACKGROUND;
else
timer_flags |= THREAD_CALL_DELAY_USER_NORMAL;
if (kn->kn_sfflags & NOTE_LEEWAY)
timer_flags |= THREAD_CALL_DELAY_LEEWAY;
thread_call_enter_delayed_with_leeway(kn->kn_hook, NULL,
kn->kn_ext[0], kn->kn_ext[1], timer_flags);
kn->kn_hookid |= TIMER_RUNNING;
} else {
kn->kn_data = 1;
}
break;
case EVENT_PROCESS:
*kev = kn->kn_kevent;
kev->ext[0] = 0;
kn->kn_data = 0;
if (kn->kn_flags & EV_CLEAR)
kn->kn_fflags = 0;
break;
default:
panic("%s: - invalid type (%ld)", __func__, type);
break;
}
filt_timerunlock();
}
static void
filt_timerlock(void)
{
lck_mtx_lock(&_filt_timerlock);
}
static void
filt_timerunlock(void)
{
lck_mtx_unlock(&_filt_timerlock);
}
static int
filt_userattach(struct knote *kn)
{
kn->kn_hook = NULL;
if (kn->kn_fflags & NOTE_TRIGGER) {
kn->kn_hookid = 1;
} else {
kn->kn_hookid = 0;
}
return (0);
}
static void
filt_userdetach(__unused struct knote *kn)
{
}
static int
filt_user(struct knote *kn, __unused long hint)
{
return (kn->kn_hookid);
}
static void
filt_usertouch(struct knote *kn, struct kevent64_s *kev, long type)
{
uint32_t ffctrl;
switch (type) {
case EVENT_REGISTER:
if (kev->fflags & NOTE_TRIGGER) {
kn->kn_hookid = 1;
}
ffctrl = kev->fflags & NOTE_FFCTRLMASK;
kev->fflags &= NOTE_FFLAGSMASK;
switch (ffctrl) {
case NOTE_FFNOP:
break;
case NOTE_FFAND:
OSBitAndAtomic(kev->fflags, &kn->kn_sfflags);
break;
case NOTE_FFOR:
OSBitOrAtomic(kev->fflags, &kn->kn_sfflags);
break;
case NOTE_FFCOPY:
kn->kn_sfflags = kev->fflags;
break;
}
kn->kn_sdata = kev->data;
break;
case EVENT_PROCESS:
*kev = kn->kn_kevent;
kev->fflags = (volatile UInt32)kn->kn_sfflags;
kev->data = kn->kn_sdata;
if (kn->kn_flags & EV_CLEAR) {
kn->kn_hookid = 0;
kn->kn_data = 0;
kn->kn_fflags = 0;
}
break;
default:
panic("%s: - invalid type (%ld)", __func__, type);
break;
}
}
static int
filt_badattach(__unused struct knote *kn)
{
return (ENOTSUP);
}
struct kqueue *
kqueue_alloc(struct proc *p)
{
struct filedesc *fdp = p->p_fd;
struct kqueue *kq;
MALLOC_ZONE(kq, struct kqueue *, sizeof (struct kqueue), M_KQUEUE,
M_WAITOK);
if (kq != NULL) {
wait_queue_set_t wqs;
wqs = wait_queue_set_alloc(SYNC_POLICY_FIFO |
SYNC_POLICY_PREPOST);
if (wqs != NULL) {
bzero(kq, sizeof (struct kqueue));
lck_spin_init(&kq->kq_lock, kq_lck_grp, kq_lck_attr);
TAILQ_INIT(&kq->kq_head);
kq->kq_wqs = wqs;
kq->kq_p = p;
} else {
FREE_ZONE(kq, sizeof (struct kqueue), M_KQUEUE);
}
}
if (fdp->fd_knlistsize < 0) {
proc_fdlock(p);
if (fdp->fd_knlistsize < 0)
fdp->fd_knlistsize = 0;
proc_fdunlock(p);
}
return (kq);
}
void
kqueue_dealloc(struct kqueue *kq)
{
struct proc *p = kq->kq_p;
struct filedesc *fdp = p->p_fd;
struct knote *kn;
int i;
proc_fdlock(p);
for (i = 0; i < fdp->fd_knlistsize; i++) {
kn = SLIST_FIRST(&fdp->fd_knlist[i]);
while (kn != NULL) {
if (kq == kn->kn_kq) {
kqlock(kq);
proc_fdunlock(p);
if (kqlock2knotedrop(kq, kn)) {
kn->kn_fop->f_detach(kn);
knote_drop(kn, p);
}
proc_fdlock(p);
kn = SLIST_FIRST(&fdp->fd_knlist[i]);
continue;
}
kn = SLIST_NEXT(kn, kn_link);
}
}
if (fdp->fd_knhashmask != 0) {
for (i = 0; i < (int)fdp->fd_knhashmask + 1; i++) {
kn = SLIST_FIRST(&fdp->fd_knhash[i]);
while (kn != NULL) {
if (kq == kn->kn_kq) {
kqlock(kq);
proc_fdunlock(p);
if (kqlock2knotedrop(kq, kn)) {
kn->kn_fop->f_detach(kn);
knote_drop(kn, p);
}
proc_fdlock(p);
kn = SLIST_FIRST(&fdp->fd_knhash[i]);
continue;
}
kn = SLIST_NEXT(kn, kn_link);
}
}
}
proc_fdunlock(p);
wait_queue_unlink_all((wait_queue_t)kq->kq_wqs);
wait_queue_set_free(kq->kq_wqs);
lck_spin_destroy(&kq->kq_lock, kq_lck_grp);
FREE_ZONE(kq, sizeof (struct kqueue), M_KQUEUE);
}
int
kqueue_body(struct proc *p, fp_allocfn_t fp_zalloc, void *cra, int32_t *retval)
{
struct kqueue *kq;
struct fileproc *fp;
int fd, error;
error = falloc_withalloc(p,
&fp, &fd, vfs_context_current(), fp_zalloc, cra);
if (error) {
return (error);
}
kq = kqueue_alloc(p);
if (kq == NULL) {
fp_free(p, fd, fp);
return (ENOMEM);
}
fp->f_flag = FREAD | FWRITE;
fp->f_ops = &kqueueops;
fp->f_data = kq;
proc_fdlock(p);
*fdflags(p, fd) |= UF_EXCLOSE;
procfdtbl_releasefd(p, fd, NULL);
fp_drop(p, fd, fp, 1);
proc_fdunlock(p);
*retval = fd;
return (error);
}
int
kqueue(struct proc *p, __unused struct kqueue_args *uap, int32_t *retval)
{
return (kqueue_body(p, fileproc_alloc_init, NULL, retval));
}
static int
kevent_copyin(user_addr_t *addrp, struct kevent64_s *kevp, struct proc *p,
int iskev64)
{
int advance;
int error;
if (iskev64) {
advance = sizeof (struct kevent64_s);
error = copyin(*addrp, (caddr_t)kevp, advance);
} else if (IS_64BIT_PROCESS(p)) {
struct user64_kevent kev64;
bzero(kevp, sizeof (struct kevent64_s));
advance = sizeof (kev64);
error = copyin(*addrp, (caddr_t)&kev64, advance);
if (error)
return (error);
kevp->ident = kev64.ident;
kevp->filter = kev64.filter;
kevp->flags = kev64.flags;
kevp->fflags = kev64.fflags;
kevp->data = kev64.data;
kevp->udata = kev64.udata;
} else {
struct user32_kevent kev32;
bzero(kevp, sizeof (struct kevent64_s));
advance = sizeof (kev32);
error = copyin(*addrp, (caddr_t)&kev32, advance);
if (error)
return (error);
kevp->ident = (uintptr_t)kev32.ident;
kevp->filter = kev32.filter;
kevp->flags = kev32.flags;
kevp->fflags = kev32.fflags;
kevp->data = (intptr_t)kev32.data;
kevp->udata = CAST_USER_ADDR_T(kev32.udata);
}
if (!error)
*addrp += advance;
return (error);
}
static int
kevent_copyout(struct kevent64_s *kevp, user_addr_t *addrp, struct proc *p,
int iskev64)
{
int advance;
int error;
if (iskev64) {
advance = sizeof (struct kevent64_s);
error = copyout((caddr_t)kevp, *addrp, advance);
} else if (IS_64BIT_PROCESS(p)) {
struct user64_kevent kev64;
kev64.ident = (kevp->ident == (uintptr_t)-1) ?
(uint64_t)-1LL : (uint64_t)kevp->ident;
kev64.filter = kevp->filter;
kev64.flags = kevp->flags;
kev64.fflags = kevp->fflags;
kev64.data = (int64_t) kevp->data;
kev64.udata = kevp->udata;
advance = sizeof (kev64);
error = copyout((caddr_t)&kev64, *addrp, advance);
} else {
struct user32_kevent kev32;
kev32.ident = (uint32_t)kevp->ident;
kev32.filter = kevp->filter;
kev32.flags = kevp->flags;
kev32.fflags = kevp->fflags;
kev32.data = (int32_t)kevp->data;
kev32.udata = kevp->udata;
advance = sizeof (kev32);
error = copyout((caddr_t)&kev32, *addrp, advance);
}
if (!error)
*addrp += advance;
return (error);
}
static void
kevent_continue(__unused struct kqueue *kq, void *data, int error)
{
struct _kevent *cont_args;
struct fileproc *fp;
int32_t *retval;
int noutputs;
int fd;
struct proc *p = current_proc();
cont_args = (struct _kevent *)data;
noutputs = cont_args->eventout;
retval = cont_args->retval;
fd = cont_args->fd;
fp = cont_args->fp;
fp_drop(p, fd, fp, 0);
if (error == ERESTART)
error = EINTR;
else if (error == EWOULDBLOCK)
error = 0;
if (error == 0)
*retval = noutputs;
unix_syscall_return(error);
}
int
kevent(struct proc *p, struct kevent_args *uap, int32_t *retval)
{
return (kevent_internal(p,
0,
uap->changelist,
uap->nchanges,
uap->eventlist,
uap->nevents,
uap->fd,
uap->timeout,
0,
retval));
}
int
kevent64(struct proc *p, struct kevent64_args *uap, int32_t *retval)
{
return (kevent_internal(p,
1,
uap->changelist,
uap->nchanges,
uap->eventlist,
uap->nevents,
uap->fd,
uap->timeout,
uap->flags,
retval));
}
static int
kevent_internal(struct proc *p, int iskev64, user_addr_t changelist,
int nchanges, user_addr_t ueventlist, int nevents, int fd,
user_addr_t utimeout, __unused unsigned int flags,
int32_t *retval)
{
struct _kevent *cont_args;
uthread_t ut;
struct kqueue *kq;
struct fileproc *fp;
struct kevent64_s kev;
int error, noutputs;
struct timeval atv;
if (utimeout != USER_ADDR_NULL) {
struct timeval rtv;
if (IS_64BIT_PROCESS(p)) {
struct user64_timespec ts;
error = copyin(utimeout, &ts, sizeof(ts));
if ((ts.tv_sec & 0xFFFFFFFF00000000ull) != 0)
error = EINVAL;
else
TIMESPEC_TO_TIMEVAL(&rtv, &ts);
} else {
struct user32_timespec ts;
error = copyin(utimeout, &ts, sizeof(ts));
TIMESPEC_TO_TIMEVAL(&rtv, &ts);
}
if (error)
return (error);
if (itimerfix(&rtv))
return (EINVAL);
getmicrouptime(&atv);
timevaladd(&atv, &rtv);
} else {
atv.tv_sec = 0;
atv.tv_usec = 0;
}
if ((error = fp_getfkq(p, fd, &fp, &kq)) != 0)
return (error);
kqlock(kq);
if (kq->kq_state & (KQ_KEV32 | KQ_KEV64)) {
if (((iskev64 && (kq->kq_state & KQ_KEV32)) ||
(!iskev64 && (kq->kq_state & KQ_KEV64)))) {
error = EINVAL;
kqunlock(kq);
goto errorout;
}
} else {
kq->kq_state |= (iskev64 ? KQ_KEV64 : KQ_KEV32);
}
kqunlock(kq);
noutputs = 0;
while (nchanges > 0 && error == 0) {
error = kevent_copyin(&changelist, &kev, p, iskev64);
if (error)
break;
kev.flags &= ~EV_SYSFLAGS;
error = kevent_register(kq, &kev, p);
if ((error || (kev.flags & EV_RECEIPT)) && nevents > 0) {
kev.flags = EV_ERROR;
kev.data = error;
error = kevent_copyout(&kev, &ueventlist, p, iskev64);
if (error == 0) {
nevents--;
noutputs++;
}
}
nchanges--;
}
ut = (uthread_t)get_bsdthread_info(current_thread());
cont_args = &ut->uu_kevent.ss_kevent;
cont_args->fp = fp;
cont_args->fd = fd;
cont_args->retval = retval;
cont_args->eventlist = ueventlist;
cont_args->eventcount = nevents;
cont_args->eventout = noutputs;
cont_args->eventsize = iskev64;
if (nevents > 0 && noutputs == 0 && error == 0)
error = kqueue_scan(kq, kevent_callback,
kevent_continue, cont_args,
&atv, p);
kevent_continue(kq, cont_args, error);
errorout:
fp_drop(p, fd, fp, 0);
return (error);
}
static int
kevent_callback(__unused struct kqueue *kq, struct kevent64_s *kevp,
void *data)
{
struct _kevent *cont_args;
int error;
int iskev64;
cont_args = (struct _kevent *)data;
assert(cont_args->eventout < cont_args->eventcount);
iskev64 = cont_args->eventsize;
error = kevent_copyout(kevp, &cont_args->eventlist, current_proc(),
iskev64);
if (error == 0 && ++cont_args->eventout == cont_args->eventcount)
error = EWOULDBLOCK;
return (error);
}
char *
kevent_description(struct kevent64_s *kevp, char *s, size_t n)
{
snprintf(s, n,
"kevent="
"{.ident=%#llx, .filter=%d, .flags=%#x, .fflags=%#x, .data=%#llx, .udata=%#llx, .ext[0]=%#llx, .ext[1]=%#llx}",
kevp->ident,
kevp->filter,
kevp->flags,
kevp->fflags,
kevp->data,
kevp->udata,
kevp->ext[0],
kevp->ext[1]);
return (s);
}
int
kevent_register(struct kqueue *kq, struct kevent64_s *kev,
__unused struct proc *ctxp)
{
struct proc *p = kq->kq_p;
struct filedesc *fdp = p->p_fd;
struct filterops *fops;
struct fileproc *fp = NULL;
struct knote *kn = NULL;
int error = 0;
if (kev->filter < 0) {
if (kev->filter + EVFILT_SYSCOUNT < 0)
return (EINVAL);
fops = sysfilt_ops[~kev->filter];
} else {
printf("unknown filter: %d\n", kev->filter);
return (EINVAL);
}
restart:
proc_fdlock(p);
if (fops->f_isfd && (error = fp_lookup(p, kev->ident, &fp, 1)) != 0) {
proc_fdunlock(p);
return (error);
}
if (fops->f_isfd) {
if (kev->ident < (u_int)fdp->fd_knlistsize) {
SLIST_FOREACH(kn, &fdp->fd_knlist[kev->ident], kn_link)
if (kq == kn->kn_kq &&
kev->filter == kn->kn_filter)
break;
}
} else {
if (fdp->fd_knhashmask != 0) {
struct klist *list;
list = &fdp->fd_knhash[
KN_HASH((u_long)kev->ident, fdp->fd_knhashmask)];
SLIST_FOREACH(kn, list, kn_link)
if (kev->ident == kn->kn_id &&
kq == kn->kn_kq &&
kev->filter == kn->kn_filter)
break;
}
}
if (kn == NULL) {
if ((kev->flags & (EV_ADD|EV_DELETE)) == EV_ADD) {
kn = knote_alloc();
if (kn == NULL) {
proc_fdunlock(p);
error = ENOMEM;
goto done;
}
kn->kn_fp = fp;
kn->kn_kq = kq;
kn->kn_tq = &kq->kq_head;
kn->kn_fop = fops;
kn->kn_sfflags = kev->fflags;
kn->kn_sdata = kev->data;
kev->fflags = 0;
kev->data = 0;
kn->kn_kevent = *kev;
kn->kn_inuse = 1;
kn->kn_status = KN_ATTACHING;
if (kev->flags & EV_DISABLE)
kn->kn_status |= KN_DISABLED;
error = knote_fdpattach(kn, fdp, p);
proc_fdunlock(p);
if (error) {
knote_free(kn);
goto done;
}
fp = NULL;
error = fops->f_attach(kn);
kqlock(kq);
if (error != 0) {
kn->kn_status |= KN_DROPPING;
kqunlock(kq);
knote_drop(kn, p);
goto done;
} else if (kn->kn_status & KN_DROPPING) {
kqunlock(kq);
kn->kn_fop->f_detach(kn);
knote_drop(kn, p);
goto done;
}
kn->kn_status &= ~KN_ATTACHING;
kqunlock(kq);
} else {
proc_fdunlock(p);
error = ENOENT;
goto done;
}
} else {
kqlock(kq);
proc_fdunlock(p);
if (kev->flags & EV_DELETE) {
knote_dequeue(kn);
kn->kn_status |= KN_DISABLED;
if (kqlock2knotedrop(kq, kn)) {
kn->kn_fop->f_detach(kn);
knote_drop(kn, p);
}
goto done;
}
if (kev->flags & EV_DISABLE) {
knote_dequeue(kn);
kn->kn_status |= KN_DISABLED;
} else if (kev->flags & EV_ENABLE) {
kn->kn_status &= ~KN_DISABLED;
if (kn->kn_status & KN_ACTIVE)
knote_enqueue(kn);
}
kn->kn_kevent.udata = kev->udata;
if (fops->f_isfd || fops->f_touch == NULL) {
kn->kn_sfflags = kev->fflags;
kn->kn_sdata = kev->data;
}
if (!kqlock2knoteusewait(kq, kn)) {
goto restart;
}
if (!fops->f_isfd && fops->f_touch != NULL)
fops->f_touch(kn, kev, EVENT_REGISTER);
}
if ((kn->kn_status & KN_STAYQUEUED) == 0 && kn->kn_fop->f_event(kn, 0)) {
if (knoteuse2kqlock(kq, kn))
knote_activate(kn, 1);
kqunlock(kq);
} else {
knote_put(kn);
}
done:
if (fp != NULL)
fp_drop(p, kev->ident, fp, 0);
return (error);
}
static int
knote_process(struct knote *kn,
kevent_callback_t callback,
void *data,
struct kqtailq *inprocessp,
struct proc *p)
{
struct kqueue *kq = kn->kn_kq;
struct kevent64_s kev;
int touch;
int result;
int error;
if ((kn->kn_status & KN_DISABLED) != 0) {
result = 0;
touch = 0;
} else {
int revalidate;
result = 1;
revalidate = ((kn->kn_status & KN_STAYQUEUED) != 0 ||
(kn->kn_flags & EV_ONESHOT) == 0);
touch = (!kn->kn_fop->f_isfd && kn->kn_fop->f_touch != NULL);
if (revalidate || touch) {
if (revalidate)
knote_deactivate(kn);
if (kqlock2knoteuse(kq, kn)) {
if (revalidate) {
result = kn->kn_fop->f_event(kn, 0);
}
if (result && touch) {
kn->kn_fop->f_touch(kn, &kev,
EVENT_PROCESS);
}
if (!knoteuse2kqlock(kq, kn)) {
return (EJUSTRETURN);
} else if (result) {
if (!(kn->kn_status & KN_ACTIVE)) {
knote_activate(kn, 0);
}
if (!touch) {
kev = kn->kn_kevent;
}
} else if ((kn->kn_status & KN_STAYQUEUED) == 0) {
return (EJUSTRETURN);
}
} else {
return (EJUSTRETURN);
}
} else {
kev = kn->kn_kevent;
}
}
assert(kn->kn_tq == &kq->kq_head);
TAILQ_REMOVE(&kq->kq_head, kn, kn_tqe);
kn->kn_tq = inprocessp;
TAILQ_INSERT_TAIL(inprocessp, kn, kn_tqe);
if (result == 0) {
return (EJUSTRETURN);
} else if ((kn->kn_flags & EV_ONESHOT) != 0) {
knote_deactivate(kn);
if (kqlock2knotedrop(kq, kn)) {
kn->kn_fop->f_detach(kn);
knote_drop(kn, p);
}
} else if ((kn->kn_flags & (EV_CLEAR | EV_DISPATCH)) != 0) {
if ((kn->kn_flags & EV_DISPATCH) != 0) {
knote_deactivate(kn);
kn->kn_status |= KN_DISABLED;
} else if (!touch || kn->kn_fflags == 0) {
knote_deactivate(kn);
}
if (!touch && (kn->kn_flags & EV_CLEAR) != 0) {
kn->kn_data = 0;
kn->kn_fflags = 0;
}
kqunlock(kq);
} else {
kqunlock(kq);
}
error = (callback)(kq, &kev, data);
kqlock(kq);
return (error);
}
static int
kqueue_begin_processing(struct kqueue *kq)
{
for (;;) {
if (kq->kq_count == 0) {
return (-1);
}
if (kq->kq_nprocess != 0) {
wait_queue_assert_wait((wait_queue_t)kq->kq_wqs,
&kq->kq_nprocess, THREAD_UNINT, 0);
kq->kq_state |= KQ_PROCWAIT;
kqunlock(kq);
thread_block(THREAD_CONTINUE_NULL);
kqlock(kq);
} else {
kq->kq_nprocess = 1;
return (0);
}
}
}
static void
kqueue_end_processing(struct kqueue *kq)
{
kq->kq_nprocess = 0;
if (kq->kq_state & KQ_PROCWAIT) {
kq->kq_state &= ~KQ_PROCWAIT;
wait_queue_wakeup_all((wait_queue_t)kq->kq_wqs,
&kq->kq_nprocess, THREAD_AWAKENED);
}
}
static int
kqueue_process(struct kqueue *kq,
kevent_callback_t callback,
void *data,
int *countp,
struct proc *p)
{
struct kqtailq inprocess;
struct knote *kn;
int nevents;
int error;
TAILQ_INIT(&inprocess);
if (kqueue_begin_processing(kq) == -1) {
*countp = 0;
return (0);
}
wait_queue_sub_clearrefs(kq->kq_wqs);
error = 0;
nevents = 0;
while (error == 0 &&
(kn = TAILQ_FIRST(&kq->kq_head)) != NULL) {
error = knote_process(kn, callback, data, &inprocess, p);
if (error == EJUSTRETURN)
error = 0;
else
nevents++;
}
while ((kn = TAILQ_FIRST(&inprocess)) != NULL) {
assert(kn->kn_tq == &inprocess);
TAILQ_REMOVE(&inprocess, kn, kn_tqe);
kn->kn_tq = &kq->kq_head;
TAILQ_INSERT_TAIL(&kq->kq_head, kn, kn_tqe);
}
kqueue_end_processing(kq);
*countp = nevents;
return (error);
}
static void
kqueue_scan_continue(void *data, wait_result_t wait_result)
{
thread_t self = current_thread();
uthread_t ut = (uthread_t)get_bsdthread_info(self);
struct _kqueue_scan * cont_args = &ut->uu_kevent.ss_kqueue_scan;
struct kqueue *kq = (struct kqueue *)data;
int error;
int count;
switch (wait_result) {
case THREAD_AWAKENED:
kqlock(kq);
error = kqueue_process(kq, cont_args->call, cont_args, &count,
current_proc());
if (error == 0 && count == 0) {
wait_queue_assert_wait((wait_queue_t)kq->kq_wqs,
KQ_EVENT, THREAD_ABORTSAFE, cont_args->deadline);
kq->kq_state |= KQ_SLEEP;
kqunlock(kq);
thread_block_parameter(kqueue_scan_continue, kq);
}
kqunlock(kq);
break;
case THREAD_TIMED_OUT:
error = EWOULDBLOCK;
break;
case THREAD_INTERRUPTED:
error = EINTR;
break;
default:
panic("%s: - invalid wait_result (%d)", __func__,
wait_result);
error = 0;
}
assert(cont_args->cont != NULL);
(cont_args->cont)(kq, cont_args->data, error);
}
int
kqueue_scan(struct kqueue *kq,
kevent_callback_t callback,
kqueue_continue_t continuation,
void *data,
struct timeval *atvp,
struct proc *p)
{
thread_continue_t cont = THREAD_CONTINUE_NULL;
uint64_t deadline;
int error;
int first;
assert(callback != NULL);
first = 1;
for (;;) {
wait_result_t wait_result;
int count;
kqlock(kq);
error = kqueue_process(kq, callback, data, &count, p);
if (error || count)
break;
if (first) {
first = 0;
if (atvp->tv_sec || atvp->tv_usec) {
uint64_t now;
clock_get_uptime(&now);
nanoseconds_to_absolutetime((uint64_t)atvp->tv_sec * NSEC_PER_SEC +
atvp->tv_usec * (long)NSEC_PER_USEC,
&deadline);
if (now >= deadline) {
error = EWOULDBLOCK;
break;
}
deadline -= now;
clock_absolutetime_interval_to_deadline(deadline, &deadline);
} else {
deadline = 0;
}
if (continuation) {
uthread_t ut = (uthread_t)get_bsdthread_info(current_thread());
struct _kqueue_scan *cont_args = &ut->uu_kevent.ss_kqueue_scan;
cont_args->call = callback;
cont_args->cont = continuation;
cont_args->deadline = deadline;
cont_args->data = data;
cont = kqueue_scan_continue;
}
}
wait_queue_assert_wait_with_leeway((wait_queue_t)kq->kq_wqs,
KQ_EVENT, THREAD_ABORTSAFE, TIMEOUT_URGENCY_USER_NORMAL,
deadline, 0);
kq->kq_state |= KQ_SLEEP;
kqunlock(kq);
wait_result = thread_block_parameter(cont, kq);
switch (wait_result) {
case THREAD_AWAKENED:
continue;
case THREAD_TIMED_OUT:
return (EWOULDBLOCK);
case THREAD_INTERRUPTED:
return (EINTR);
default:
panic("%s: - bad wait_result (%d)", __func__,
wait_result);
error = 0;
}
}
kqunlock(kq);
return (error);
}
static int
kqueue_read(__unused struct fileproc *fp,
__unused struct uio *uio,
__unused int flags,
__unused vfs_context_t ctx)
{
return (ENXIO);
}
static int
kqueue_write(__unused struct fileproc *fp,
__unused struct uio *uio,
__unused int flags,
__unused vfs_context_t ctx)
{
return (ENXIO);
}
static int
kqueue_ioctl(__unused struct fileproc *fp,
__unused u_long com,
__unused caddr_t data,
__unused vfs_context_t ctx)
{
return (ENOTTY);
}
static int
kqueue_select(struct fileproc *fp, int which, void *wql,
__unused vfs_context_t ctx)
{
struct kqueue *kq = (struct kqueue *)fp->f_data;
struct knote *kn;
struct kqtailq inprocessq;
int retnum = 0;
if (which != FREAD)
return (0);
TAILQ_INIT(&inprocessq);
kqlock(kq);
if (wql != NULL) {
thread_t cur_act = current_thread();
struct uthread * ut = get_bsdthread_info(cur_act);
kq->kq_state |= KQ_SEL;
wait_queue_link_noalloc((wait_queue_t)kq->kq_wqs, ut->uu_wqset,
(wait_queue_link_t)wql);
}
if (kqueue_begin_processing(kq) == -1) {
kqunlock(kq);
return (0);
}
if (kq->kq_count != 0) {
while ((kn = (struct knote *)TAILQ_FIRST(&kq->kq_head)) != NULL) {
if ((kn->kn_status & KN_STAYQUEUED) == 0) {
retnum = 1;
goto out;
}
TAILQ_REMOVE(&kq->kq_head, kn, kn_tqe);
TAILQ_INSERT_TAIL(&inprocessq, kn, kn_tqe);
if (kqlock2knoteuse(kq, kn)) {
unsigned peek;
peek = kn->kn_fop->f_peek(kn);
if (knoteuse2kqlock(kq, kn)) {
if (peek > 0) {
retnum = 1;
goto out;
}
} else {
retnum = 0;
}
}
}
}
out:
while ((kn = TAILQ_FIRST(&inprocessq)) != NULL) {
TAILQ_REMOVE(&inprocessq, kn, kn_tqe);
kn->kn_tq = &kq->kq_head;
TAILQ_INSERT_TAIL(&kq->kq_head, kn, kn_tqe);
}
kqueue_end_processing(kq);
kqunlock(kq);
return (retnum);
}
static int
kqueue_close(struct fileglob *fg, __unused vfs_context_t ctx)
{
struct kqueue *kq = (struct kqueue *)fg->fg_data;
kqueue_dealloc(kq);
fg->fg_data = NULL;
return (0);
}
static int
kqueue_kqfilter(__unused struct fileproc *fp, struct knote *kn, __unused vfs_context_t ctx)
{
struct kqueue *kq = (struct kqueue *)kn->kn_fp->f_data;
struct kqueue *parentkq = kn->kn_kq;
if (parentkq == kq ||
kn->kn_filter != EVFILT_READ)
return (1);
kqlock(parentkq);
if (parentkq->kq_level > 0 &&
parentkq->kq_level < kq->kq_level)
{
kqunlock(parentkq);
return (1);
} else {
if (parentkq->kq_level == 0)
parentkq->kq_level = 2;
if (parentkq->kq_level < kq->kq_level + 1)
parentkq->kq_level = kq->kq_level + 1;
kqunlock(parentkq);
kn->kn_fop = &kqread_filtops;
kqlock(kq);
KNOTE_ATTACH(&kq->kq_sel.si_note, kn);
if (kq->kq_level == 0)
kq->kq_level = 1;
kqunlock(kq);
return (0);
}
}
static int
kqueue_drain(struct fileproc *fp, __unused vfs_context_t ctx)
{
struct kqueue *kq = (struct kqueue *)fp->f_fglob->fg_data;
kqlock(kq);
kqueue_wakeup(kq, 1);
kqunlock(kq);
return (0);
}
int
kqueue_stat(struct kqueue *kq, void *ub, int isstat64, proc_t p)
{
kqlock(kq);
if (isstat64 != 0) {
struct stat64 *sb64 = (struct stat64 *)ub;
bzero((void *)sb64, sizeof(*sb64));
sb64->st_size = kq->kq_count;
if (kq->kq_state & KQ_KEV64)
sb64->st_blksize = sizeof(struct kevent64_s);
else
sb64->st_blksize = IS_64BIT_PROCESS(p) ? sizeof(struct user64_kevent) : sizeof(struct user32_kevent);
sb64->st_mode = S_IFIFO;
} else {
struct stat *sb = (struct stat *)ub;
bzero((void *)sb, sizeof(*sb));
sb->st_size = kq->kq_count;
if (kq->kq_state & KQ_KEV64)
sb->st_blksize = sizeof(struct kevent64_s);
else
sb->st_blksize = IS_64BIT_PROCESS(p) ? sizeof(struct user64_kevent) : sizeof(struct user32_kevent);
sb->st_mode = S_IFIFO;
}
kqunlock(kq);
return (0);
}
static void
kqueue_wakeup(struct kqueue *kq, int closed)
{
if ((kq->kq_state & (KQ_SLEEP | KQ_SEL)) != 0 || kq->kq_nprocess > 0) {
kq->kq_state &= ~(KQ_SLEEP | KQ_SEL);
wait_queue_wakeup_all((wait_queue_t)kq->kq_wqs, KQ_EVENT,
(closed) ? THREAD_INTERRUPTED : THREAD_AWAKENED);
}
}
void
klist_init(struct klist *list)
{
SLIST_INIT(list);
}
void
knote(struct klist *list, long hint)
{
struct knote *kn;
SLIST_FOREACH(kn, list, kn_selnext) {
struct kqueue *kq = kn->kn_kq;
kqlock(kq);
if (kqlock2knoteuse(kq, kn)) {
int result;
result = kn->kn_fop->f_event(kn, hint);
if (knoteuse2kqlock(kq, kn) && result)
knote_activate(kn, 1);
}
kqunlock(kq);
}
}
int
knote_attach(struct klist *list, struct knote *kn)
{
int ret = SLIST_EMPTY(list);
SLIST_INSERT_HEAD(list, kn, kn_selnext);
return (ret);
}
int
knote_detach(struct klist *list, struct knote *kn)
{
SLIST_REMOVE(list, kn, knote, kn_selnext);
return (SLIST_EMPTY(list));
}
int
knote_link_wait_queue(struct knote *kn, struct wait_queue *wq, wait_queue_link_t wql)
{
struct kqueue *kq = kn->kn_kq;
kern_return_t kr;
kr = wait_queue_link_noalloc(wq, kq->kq_wqs, wql);
if (kr == KERN_SUCCESS) {
knote_markstayqueued(kn);
return (0);
} else {
return (EINVAL);
}
}
int
knote_unlink_wait_queue(struct knote *kn, struct wait_queue *wq, wait_queue_link_t *wqlp)
{
struct kqueue *kq = kn->kn_kq;
kern_return_t kr;
kr = wait_queue_unlink_nofree(wq, kq->kq_wqs, wqlp);
kqlock(kq);
kn->kn_status &= ~KN_STAYQUEUED;
knote_dequeue(kn);
kqunlock(kq);
return ((kr != KERN_SUCCESS) ? EINVAL : 0);
}
void
knote_fdclose(struct proc *p, int fd)
{
struct filedesc *fdp = p->p_fd;
struct klist *list;
struct knote *kn;
list = &fdp->fd_knlist[fd];
while ((kn = SLIST_FIRST(list)) != NULL) {
struct kqueue *kq = kn->kn_kq;
if (kq->kq_p != p)
panic("%s: proc mismatch (kq->kq_p=%p != p=%p)",
__func__, kq->kq_p, p);
kqlock(kq);
proc_fdunlock(p);
if (kqlock2knotedrop(kq, kn)) {
kn->kn_fop->f_detach(kn);
knote_drop(kn, p);
}
proc_fdlock(p);
list = &fdp->fd_knlist[fd];
}
}
static int
knote_fdpattach(struct knote *kn, struct filedesc *fdp, struct proc *p)
{
struct klist *list = NULL;
if (! kn->kn_fop->f_isfd) {
if (fdp->fd_knhashmask == 0)
fdp->fd_knhash = hashinit(CONFIG_KN_HASHSIZE, M_KQUEUE,
&fdp->fd_knhashmask);
list = &fdp->fd_knhash[KN_HASH(kn->kn_id, fdp->fd_knhashmask)];
} else {
if ((u_int)fdp->fd_knlistsize <= kn->kn_id) {
u_int size = 0;
if (kn->kn_id >= (uint64_t)p->p_rlimit[RLIMIT_NOFILE].rlim_cur
|| kn->kn_id >= (uint64_t)maxfiles)
return (EINVAL);
size = fdp->fd_knlistsize;
while (size <= kn->kn_id)
size += KQEXTENT;
if (size >= (UINT_MAX/sizeof(struct klist *)))
return (EINVAL);
MALLOC(list, struct klist *,
size * sizeof(struct klist *), M_KQUEUE, M_WAITOK);
if (list == NULL)
return (ENOMEM);
bcopy((caddr_t)fdp->fd_knlist, (caddr_t)list,
fdp->fd_knlistsize * sizeof(struct klist *));
bzero((caddr_t)list +
fdp->fd_knlistsize * sizeof(struct klist *),
(size - fdp->fd_knlistsize) * sizeof(struct klist *));
FREE(fdp->fd_knlist, M_KQUEUE);
fdp->fd_knlist = list;
fdp->fd_knlistsize = size;
}
list = &fdp->fd_knlist[kn->kn_id];
}
SLIST_INSERT_HEAD(list, kn, kn_link);
return (0);
}
static void
knote_drop(struct knote *kn, __unused struct proc *ctxp)
{
struct kqueue *kq = kn->kn_kq;
struct proc *p = kq->kq_p;
struct filedesc *fdp = p->p_fd;
struct klist *list;
int needswakeup;
proc_fdlock(p);
if (kn->kn_fop->f_isfd)
list = &fdp->fd_knlist[kn->kn_id];
else
list = &fdp->fd_knhash[KN_HASH(kn->kn_id, fdp->fd_knhashmask)];
SLIST_REMOVE(list, kn, knote, kn_link);
kqlock(kq);
knote_dequeue(kn);
needswakeup = (kn->kn_status & KN_USEWAIT);
kqunlock(kq);
proc_fdunlock(p);
if (needswakeup)
wait_queue_wakeup_all((wait_queue_t)kq->kq_wqs, &kn->kn_status,
THREAD_AWAKENED);
if (kn->kn_fop->f_isfd)
fp_drop(p, kn->kn_id, kn->kn_fp, 0);
knote_free(kn);
}
static void
knote_activate(struct knote *kn, int propagate)
{
struct kqueue *kq = kn->kn_kq;
kn->kn_status |= KN_ACTIVE;
knote_enqueue(kn);
kqueue_wakeup(kq, 0);
if (propagate)
KNOTE(&kq->kq_sel.si_note, 0);
}
static void
knote_deactivate(struct knote *kn)
{
kn->kn_status &= ~KN_ACTIVE;
knote_dequeue(kn);
}
static void
knote_enqueue(struct knote *kn)
{
if ((kn->kn_status & (KN_QUEUED | KN_STAYQUEUED)) == KN_STAYQUEUED ||
(kn->kn_status & (KN_QUEUED | KN_STAYQUEUED | KN_DISABLED)) == 0) {
struct kqtailq *tq = kn->kn_tq;
struct kqueue *kq = kn->kn_kq;
TAILQ_INSERT_TAIL(tq, kn, kn_tqe);
kn->kn_status |= KN_QUEUED;
kq->kq_count++;
}
}
static void
knote_dequeue(struct knote *kn)
{
struct kqueue *kq = kn->kn_kq;
if ((kn->kn_status & (KN_QUEUED | KN_STAYQUEUED)) == KN_QUEUED) {
struct kqtailq *tq = kn->kn_tq;
TAILQ_REMOVE(tq, kn, kn_tqe);
kn->kn_tq = &kq->kq_head;
kn->kn_status &= ~KN_QUEUED;
kq->kq_count--;
}
}
void
knote_init(void)
{
knote_zone = zinit(sizeof(struct knote), 8192*sizeof(struct knote),
8192, "knote zone");
kq_lck_grp_attr = lck_grp_attr_alloc_init();
kq_lck_grp = lck_grp_alloc_init("kqueue", kq_lck_grp_attr);
kq_lck_attr = lck_attr_alloc_init();
lck_mtx_init(&_filt_timerlock, kq_lck_grp, kq_lck_attr);
#if VM_PRESSURE_EVENTS
vm_pressure_init(kq_lck_grp, kq_lck_attr);
#endif
#if CONFIG_MEMORYSTATUS
memorystatus_kevent_init(kq_lck_grp, kq_lck_attr);
#endif
}
SYSINIT(knote, SI_SUB_PSEUDO, SI_ORDER_ANY, knote_init, NULL)
static struct knote *
knote_alloc(void)
{
return ((struct knote *)zalloc(knote_zone));
}
static void
knote_free(struct knote *kn)
{
zfree(knote_zone, kn);
}
#if SOCKETS
#include <sys/param.h>
#include <sys/socket.h>
#include <sys/protosw.h>
#include <sys/domain.h>
#include <sys/mbuf.h>
#include <sys/kern_event.h>
#include <sys/malloc.h>
#include <sys/sys_domain.h>
#include <sys/syslog.h>
#ifndef ROUNDUP64
#define ROUNDUP64(x) P2ROUNDUP((x), sizeof (u_int64_t))
#endif
#ifndef ADVANCE64
#define ADVANCE64(p, n) (void*)((char *)(p) + ROUNDUP64(n))
#endif
static lck_grp_attr_t *kev_lck_grp_attr;
static lck_attr_t *kev_lck_attr;
static lck_grp_t *kev_lck_grp;
static decl_lck_rw_data(,kev_lck_data);
static lck_rw_t *kev_rwlock = &kev_lck_data;
static int kev_attach(struct socket *so, int proto, struct proc *p);
static int kev_detach(struct socket *so);
static int kev_control(struct socket *so, u_long cmd, caddr_t data,
struct ifnet *ifp, struct proc *p);
static lck_mtx_t * event_getlock(struct socket *, int);
static int event_lock(struct socket *, int, void *);
static int event_unlock(struct socket *, int, void *);
static int event_sofreelastref(struct socket *);
static void kev_delete(struct kern_event_pcb *);
static struct pr_usrreqs event_usrreqs = {
.pru_attach = kev_attach,
.pru_control = kev_control,
.pru_detach = kev_detach,
.pru_soreceive = soreceive,
};
static struct protosw eventsw[] = {
{
.pr_type = SOCK_RAW,
.pr_protocol = SYSPROTO_EVENT,
.pr_flags = PR_ATOMIC,
.pr_usrreqs = &event_usrreqs,
.pr_lock = event_lock,
.pr_unlock = event_unlock,
.pr_getlock = event_getlock,
}
};
__private_extern__ int kevt_getstat SYSCTL_HANDLER_ARGS;
__private_extern__ int kevt_pcblist SYSCTL_HANDLER_ARGS;
SYSCTL_NODE(_net_systm, OID_AUTO, kevt,
CTLFLAG_RW|CTLFLAG_LOCKED, 0, "Kernel event family");
struct kevtstat kevtstat;
SYSCTL_PROC(_net_systm_kevt, OID_AUTO, stats,
CTLTYPE_STRUCT | CTLFLAG_RD | CTLFLAG_LOCKED, 0, 0,
kevt_getstat, "S,kevtstat", "");
SYSCTL_PROC(_net_systm_kevt, OID_AUTO, pcblist,
CTLTYPE_STRUCT | CTLFLAG_RD | CTLFLAG_LOCKED, 0, 0,
kevt_pcblist, "S,xkevtpcb", "");
static lck_mtx_t *
event_getlock(struct socket *so, int locktype)
{
#pragma unused(locktype)
struct kern_event_pcb *ev_pcb = (struct kern_event_pcb *)so->so_pcb;
if (so->so_pcb != NULL) {
if (so->so_usecount < 0)
panic("%s: so=%p usecount=%d lrh= %s\n", __func__,
so, so->so_usecount, solockhistory_nr(so));
} else {
panic("%s: so=%p NULL NO so_pcb %s\n", __func__,
so, solockhistory_nr(so));
}
return (&ev_pcb->evp_mtx);
}
static int
event_lock(struct socket *so, int refcount, void *lr)
{
void *lr_saved;
if (lr == NULL)
lr_saved = __builtin_return_address(0);
else
lr_saved = lr;
if (so->so_pcb != NULL) {
lck_mtx_lock(&((struct kern_event_pcb *)so->so_pcb)->evp_mtx);
} else {
panic("%s: so=%p NO PCB! lr=%p lrh= %s\n", __func__,
so, lr_saved, solockhistory_nr(so));
}
if (so->so_usecount < 0) {
panic("%s: so=%p so_pcb=%p lr=%p ref=%d lrh= %s\n", __func__,
so, so->so_pcb, lr_saved, so->so_usecount,
solockhistory_nr(so));
}
if (refcount)
so->so_usecount++;
so->lock_lr[so->next_lock_lr] = lr_saved;
so->next_lock_lr = (so->next_lock_lr+1) % SO_LCKDBG_MAX;
return (0);
}
static int
event_unlock(struct socket *so, int refcount, void *lr)
{
void *lr_saved;
lck_mtx_t *mutex_held;
if (lr == NULL)
lr_saved = __builtin_return_address(0);
else
lr_saved = lr;
if (refcount)
so->so_usecount--;
if (so->so_usecount < 0) {
panic("%s: so=%p usecount=%d lrh= %s\n", __func__,
so, so->so_usecount, solockhistory_nr(so));
}
if (so->so_pcb == NULL) {
panic("%s: so=%p NO PCB usecount=%d lr=%p lrh= %s\n", __func__,
so, so->so_usecount, (void *)lr_saved,
solockhistory_nr(so));
}
mutex_held = (&((struct kern_event_pcb *)so->so_pcb)->evp_mtx);
lck_mtx_assert(mutex_held, LCK_MTX_ASSERT_OWNED);
so->unlock_lr[so->next_unlock_lr] = lr_saved;
so->next_unlock_lr = (so->next_unlock_lr+1) % SO_LCKDBG_MAX;
if (so->so_usecount == 0) {
VERIFY(so->so_flags & SOF_PCBCLEARING);
event_sofreelastref(so);
} else {
lck_mtx_unlock(mutex_held);
}
return (0);
}
static int
event_sofreelastref(struct socket *so)
{
struct kern_event_pcb *ev_pcb = (struct kern_event_pcb *)so->so_pcb;
lck_mtx_assert(&(ev_pcb->evp_mtx), LCK_MTX_ASSERT_OWNED);
so->so_pcb = NULL;
so->so_rcv.sb_flags &= ~SB_UPCALL;
so->so_snd.sb_flags &= ~SB_UPCALL;
so->so_event = sonullevent;
lck_mtx_unlock(&(ev_pcb->evp_mtx));
lck_mtx_assert(&(ev_pcb->evp_mtx), LCK_MTX_ASSERT_NOTOWNED);
lck_rw_lock_exclusive(kev_rwlock);
LIST_REMOVE(ev_pcb, evp_link);
kevtstat.kes_pcbcount--;
kevtstat.kes_gencnt++;
lck_rw_done(kev_rwlock);
kev_delete(ev_pcb);
sofreelastref(so, 1);
return (0);
}
static int event_proto_count = (sizeof (eventsw) / sizeof (struct protosw));
static
struct kern_event_head kern_event_head;
static u_int32_t static_event_id = 0;
#define EVPCB_ZONE_MAX 65536
#define EVPCB_ZONE_NAME "kerneventpcb"
static struct zone *ev_pcb_zone;
void
kern_event_init(struct domain *dp)
{
struct protosw *pr;
int i;
VERIFY(!(dp->dom_flags & DOM_INITIALIZED));
VERIFY(dp == systemdomain);
kev_lck_grp_attr = lck_grp_attr_alloc_init();
if (kev_lck_grp_attr == NULL) {
panic("%s: lck_grp_attr_alloc_init failed\n", __func__);
}
kev_lck_grp = lck_grp_alloc_init("Kernel Event Protocol",
kev_lck_grp_attr);
if (kev_lck_grp == NULL) {
panic("%s: lck_grp_alloc_init failed\n", __func__);
}
kev_lck_attr = lck_attr_alloc_init();
if (kev_lck_attr == NULL) {
panic("%s: lck_attr_alloc_init failed\n", __func__);
}
lck_rw_init(kev_rwlock, kev_lck_grp, kev_lck_attr);
if (kev_rwlock == NULL) {
panic("%s: lck_mtx_alloc_init failed\n", __func__);
}
for (i = 0, pr = &eventsw[0]; i < event_proto_count; i++, pr++)
net_add_proto(pr, dp, 1);
ev_pcb_zone = zinit(sizeof(struct kern_event_pcb),
EVPCB_ZONE_MAX * sizeof(struct kern_event_pcb), 0, EVPCB_ZONE_NAME);
if (ev_pcb_zone == NULL) {
panic("%s: failed allocating ev_pcb_zone", __func__);
}
zone_change(ev_pcb_zone, Z_EXPAND, TRUE);
zone_change(ev_pcb_zone, Z_CALLERACCT, TRUE);
}
static int
kev_attach(struct socket *so, __unused int proto, __unused struct proc *p)
{
int error = 0;
struct kern_event_pcb *ev_pcb;
error = soreserve(so, KEV_SNDSPACE, KEV_RECVSPACE);
if (error != 0)
return (error);
if ((ev_pcb = (struct kern_event_pcb *)zalloc(ev_pcb_zone)) == NULL) {
return (ENOBUFS);
}
bzero(ev_pcb, sizeof(struct kern_event_pcb));
lck_mtx_init(&ev_pcb->evp_mtx, kev_lck_grp, kev_lck_attr);
ev_pcb->evp_socket = so;
ev_pcb->evp_vendor_code_filter = 0xffffffff;
so->so_pcb = (caddr_t) ev_pcb;
lck_rw_lock_exclusive(kev_rwlock);
LIST_INSERT_HEAD(&kern_event_head, ev_pcb, evp_link);
kevtstat.kes_pcbcount++;
kevtstat.kes_gencnt++;
lck_rw_done(kev_rwlock);
return (error);
}
static void
kev_delete(struct kern_event_pcb *ev_pcb)
{
VERIFY(ev_pcb != NULL);
lck_mtx_destroy(&ev_pcb->evp_mtx, kev_lck_grp);
zfree(ev_pcb_zone, ev_pcb);
}
static int
kev_detach(struct socket *so)
{
struct kern_event_pcb *ev_pcb = (struct kern_event_pcb *) so->so_pcb;
if (ev_pcb != NULL) {
soisdisconnected(so);
so->so_flags |= SOF_PCBCLEARING;
}
return (0);
}
errno_t kev_vendor_code_find(
const char *string,
u_int32_t *out_vendor_code)
{
if (strlen(string) >= KEV_VENDOR_CODE_MAX_STR_LEN) {
return (EINVAL);
}
return (net_str_id_find_internal(string, out_vendor_code,
NSI_VENDOR_CODE, 1));
}
errno_t
kev_msg_post(struct kev_msg *event_msg)
{
mbuf_tag_id_t min_vendor, max_vendor;
net_str_id_first_last(&min_vendor, &max_vendor, NSI_VENDOR_CODE);
if (event_msg == NULL)
return (EINVAL);
if (event_msg->vendor_code < min_vendor ||
event_msg->vendor_code > max_vendor) {
OSIncrementAtomic64((SInt64 *)&kevtstat.kes_badvendor);
return (EINVAL);
}
return (kev_post_msg(event_msg));
}
int
kev_post_msg(struct kev_msg *event_msg)
{
struct mbuf *m, *m2;
struct kern_event_pcb *ev_pcb;
struct kern_event_msg *ev;
char *tmp;
u_int32_t total_size;
int i;
total_size = KEV_MSG_HEADER_SIZE;
for (i = 0; i < 5; i++) {
if (event_msg->dv[i].data_length == 0)
break;
total_size += event_msg->dv[i].data_length;
}
if (total_size > MLEN) {
OSIncrementAtomic64((SInt64 *)&kevtstat.kes_toobig);
return (EMSGSIZE);
}
m = m_get(M_DONTWAIT, MT_DATA);
if (m == 0) {
OSIncrementAtomic64((SInt64 *)&kevtstat.kes_nomem);
return (ENOMEM);
}
ev = mtod(m, struct kern_event_msg *);
total_size = KEV_MSG_HEADER_SIZE;
tmp = (char *) &ev->event_data[0];
for (i = 0; i < 5; i++) {
if (event_msg->dv[i].data_length == 0)
break;
total_size += event_msg->dv[i].data_length;
bcopy(event_msg->dv[i].data_ptr, tmp,
event_msg->dv[i].data_length);
tmp += event_msg->dv[i].data_length;
}
ev->id = ++static_event_id;
ev->total_size = total_size;
ev->vendor_code = event_msg->vendor_code;
ev->kev_class = event_msg->kev_class;
ev->kev_subclass = event_msg->kev_subclass;
ev->event_code = event_msg->event_code;
m->m_len = total_size;
lck_rw_lock_shared(kev_rwlock);
for (ev_pcb = LIST_FIRST(&kern_event_head);
ev_pcb;
ev_pcb = LIST_NEXT(ev_pcb, evp_link)) {
lck_mtx_lock(&ev_pcb->evp_mtx);
if (ev_pcb->evp_socket->so_pcb == NULL) {
lck_mtx_unlock(&ev_pcb->evp_mtx);
continue;
}
if (ev_pcb->evp_vendor_code_filter != KEV_ANY_VENDOR) {
if (ev_pcb->evp_vendor_code_filter != ev->vendor_code) {
lck_mtx_unlock(&ev_pcb->evp_mtx);
continue;
}
if (ev_pcb->evp_class_filter != KEV_ANY_CLASS) {
if (ev_pcb->evp_class_filter != ev->kev_class) {
lck_mtx_unlock(&ev_pcb->evp_mtx);
continue;
}
if ((ev_pcb->evp_subclass_filter !=
KEV_ANY_SUBCLASS) &&
(ev_pcb->evp_subclass_filter !=
ev->kev_subclass)) {
lck_mtx_unlock(&ev_pcb->evp_mtx);
continue;
}
}
}
m2 = m_copym(m, 0, m->m_len, M_NOWAIT);
if (m2 == 0) {
OSIncrementAtomic64((SInt64 *)&kevtstat.kes_nomem);
m_free(m);
lck_mtx_unlock(&ev_pcb->evp_mtx);
lck_rw_done(kev_rwlock);
return (ENOMEM);
}
if (sbappendrecord(&ev_pcb->evp_socket->so_rcv, m2)) {
so_inc_recv_data_stat(ev_pcb->evp_socket,
1, m->m_len, SO_TC_BE);
sorwakeup(ev_pcb->evp_socket);
OSIncrementAtomic64((SInt64 *)&kevtstat.kes_posted);
} else {
OSIncrementAtomic64((SInt64 *)&kevtstat.kes_fullsock);
}
lck_mtx_unlock(&ev_pcb->evp_mtx);
}
m_free(m);
lck_rw_done(kev_rwlock);
return (0);
}
static int
kev_control(struct socket *so,
u_long cmd,
caddr_t data,
__unused struct ifnet *ifp,
__unused struct proc *p)
{
struct kev_request *kev_req = (struct kev_request *) data;
struct kern_event_pcb *ev_pcb;
struct kev_vendor_code *kev_vendor;
u_int32_t *id_value = (u_int32_t *) data;
switch (cmd) {
case SIOCGKEVID:
*id_value = static_event_id;
break;
case SIOCSKEVFILT:
ev_pcb = (struct kern_event_pcb *) so->so_pcb;
ev_pcb->evp_vendor_code_filter = kev_req->vendor_code;
ev_pcb->evp_class_filter = kev_req->kev_class;
ev_pcb->evp_subclass_filter = kev_req->kev_subclass;
break;
case SIOCGKEVFILT:
ev_pcb = (struct kern_event_pcb *) so->so_pcb;
kev_req->vendor_code = ev_pcb->evp_vendor_code_filter;
kev_req->kev_class = ev_pcb->evp_class_filter;
kev_req->kev_subclass = ev_pcb->evp_subclass_filter;
break;
case SIOCGKEVVENDOR:
kev_vendor = (struct kev_vendor_code *)data;
kev_vendor->vendor_string[KEV_VENDOR_CODE_MAX_STR_LEN-1] = 0;
return (net_str_id_find_internal(kev_vendor->vendor_string,
&kev_vendor->vendor_code, NSI_VENDOR_CODE, 0));
default:
return (ENOTSUP);
}
return (0);
}
int
kevt_getstat SYSCTL_HANDLER_ARGS
{
#pragma unused(oidp, arg1, arg2)
int error = 0;
lck_rw_lock_shared(kev_rwlock);
if (req->newptr != USER_ADDR_NULL) {
error = EPERM;
goto done;
}
if (req->oldptr == USER_ADDR_NULL) {
req->oldidx = sizeof(struct kevtstat);
goto done;
}
error = SYSCTL_OUT(req, &kevtstat,
MIN(sizeof(struct kevtstat), req->oldlen));
done:
lck_rw_done(kev_rwlock);
return (error);
}
__private_extern__ int
kevt_pcblist SYSCTL_HANDLER_ARGS
{
#pragma unused(oidp, arg1, arg2)
int error = 0;
int n, i;
struct xsystmgen xsg;
void *buf = NULL;
size_t item_size = ROUNDUP64(sizeof (struct xkevtpcb)) +
ROUNDUP64(sizeof (struct xsocket_n)) +
2 * ROUNDUP64(sizeof (struct xsockbuf_n)) +
ROUNDUP64(sizeof (struct xsockstat_n));
struct kern_event_pcb *ev_pcb;
buf = _MALLOC(item_size, M_TEMP, M_WAITOK | M_ZERO);
if (buf == NULL)
return (ENOMEM);
lck_rw_lock_shared(kev_rwlock);
n = kevtstat.kes_pcbcount;
if (req->oldptr == USER_ADDR_NULL) {
req->oldidx = (n + n/8) * item_size;
goto done;
}
if (req->newptr != USER_ADDR_NULL) {
error = EPERM;
goto done;
}
bzero(&xsg, sizeof (xsg));
xsg.xg_len = sizeof (xsg);
xsg.xg_count = n;
xsg.xg_gen = kevtstat.kes_gencnt;
xsg.xg_sogen = so_gencnt;
error = SYSCTL_OUT(req, &xsg, sizeof (xsg));
if (error) {
goto done;
}
if (n == 0) {
goto done;
}
i = 0;
for (i = 0, ev_pcb = LIST_FIRST(&kern_event_head);
i < n && ev_pcb != NULL;
i++, ev_pcb = LIST_NEXT(ev_pcb, evp_link)) {
struct xkevtpcb *xk = (struct xkevtpcb *)buf;
struct xsocket_n *xso = (struct xsocket_n *)
ADVANCE64(xk, sizeof (*xk));
struct xsockbuf_n *xsbrcv = (struct xsockbuf_n *)
ADVANCE64(xso, sizeof (*xso));
struct xsockbuf_n *xsbsnd = (struct xsockbuf_n *)
ADVANCE64(xsbrcv, sizeof (*xsbrcv));
struct xsockstat_n *xsostats = (struct xsockstat_n *)
ADVANCE64(xsbsnd, sizeof (*xsbsnd));
bzero(buf, item_size);
lck_mtx_lock(&ev_pcb->evp_mtx);
xk->kep_len = sizeof(struct xkevtpcb);
xk->kep_kind = XSO_EVT;
xk->kep_evtpcb = (uint64_t)VM_KERNEL_ADDRPERM(ev_pcb);
xk->kep_vendor_code_filter = ev_pcb->evp_vendor_code_filter;
xk->kep_class_filter = ev_pcb->evp_class_filter;
xk->kep_subclass_filter = ev_pcb->evp_subclass_filter;
sotoxsocket_n(ev_pcb->evp_socket, xso);
sbtoxsockbuf_n(ev_pcb->evp_socket ?
&ev_pcb->evp_socket->so_rcv : NULL, xsbrcv);
sbtoxsockbuf_n(ev_pcb->evp_socket ?
&ev_pcb->evp_socket->so_snd : NULL, xsbsnd);
sbtoxsockstat_n(ev_pcb->evp_socket, xsostats);
lck_mtx_unlock(&ev_pcb->evp_mtx);
error = SYSCTL_OUT(req, buf, item_size);
}
if (error == 0) {
bzero(&xsg, sizeof (xsg));
xsg.xg_len = sizeof (xsg);
xsg.xg_count = n;
xsg.xg_gen = kevtstat.kes_gencnt;
xsg.xg_sogen = so_gencnt;
error = SYSCTL_OUT(req, &xsg, sizeof (xsg));
if (error) {
goto done;
}
}
done:
lck_rw_done(kev_rwlock);
return (error);
}
#endif
int
fill_kqueueinfo(struct kqueue *kq, struct kqueue_info * kinfo)
{
struct vinfo_stat * st;
st = &kinfo->kq_stat;
st->vst_size = kq->kq_count;
if (kq->kq_state & KQ_KEV64)
st->vst_blksize = sizeof(struct kevent64_s);
else
st->vst_blksize = sizeof(struct kevent);
st->vst_mode = S_IFIFO;
if (kq->kq_state & KQ_SEL)
kinfo->kq_state |= PROC_KQUEUE_SELECT;
if (kq->kq_state & KQ_SLEEP)
kinfo->kq_state |= PROC_KQUEUE_SLEEP;
return (0);
}
void
knote_markstayqueued(struct knote *kn)
{
kqlock(kn->kn_kq);
kn->kn_status |= KN_STAYQUEUED;
knote_enqueue(kn);
kqunlock(kn->kn_kq);
}