]> git.saurik.com Git - apple/xnu.git/blobdiff - bsd/kern/kern_sysctl.c
xnu-2050.48.11.tar.gz
[apple/xnu.git] / bsd / kern / kern_sysctl.c
index 047eb70b1a38cb0dd8958324dd1a5629e86e302e..56782c39c70eae2c58a17e3269d7ffe92ed9d049 100644 (file)
@@ -1,23 +1,29 @@
 /*
- * Copyright (c) 2000 Apple Computer, Inc. All rights reserved.
+ * Copyright (c) 2000-2011 Apple Inc. All rights reserved.
  *
- * @APPLE_LICENSE_HEADER_START@
+ * @APPLE_OSREFERENCE_LICENSE_HEADER_START@
  * 
- * The contents of this file constitute Original Code as defined in and
- * are subject to the Apple Public Source License Version 1.1 (the
- * "License").  You may not use this file except in compliance with the
- * License.  Please obtain a copy of the License at
- * http://www.apple.com/publicsource and read it before using this file.
+ * This file contains Original Code and/or Modifications of Original Code
+ * as defined in and that are subject to the Apple Public Source License
+ * Version 2.0 (the 'License'). You may not use this file except in
+ * compliance with the License. The rights granted to you under the License
+ * may not be used to create, or enable the creation or redistribution of,
+ * unlawful or unlicensed copies of an Apple operating system, or to
+ * circumvent, violate, or enable the circumvention or violation of, any
+ * terms of an Apple operating system software license agreement.
  * 
- * This Original Code and all software distributed under the License are
- * distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY KIND, EITHER
+ * Please obtain a copy of the License at
+ * http://www.opensource.apple.com/apsl/ and read it before using this file.
+ * 
+ * The Original Code and all software distributed under the License are
+ * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
  * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
  * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE OR NON-INFRINGEMENT.  Please see the
- * License for the specific language governing rights and limitations
- * under the License.
+ * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
+ * Please see the License for the specific language governing rights and
+ * limitations under the License.
  * 
- * @APPLE_LICENSE_HEADER_END@
+ * @APPLE_OSREFERENCE_LICENSE_HEADER_END@
  */
 /* Copyright (c) 1995 NeXT Computer, Inc. All Rights Reserved */
 /*-
  *
  *     @(#)kern_sysctl.c       8.4 (Berkeley) 4/14/94
  */
+/*
+ * NOTICE: This file was modified by SPARTA, Inc. in 2005 to introduce
+ * support for mandatory and extensible security protections.  This notice
+ * is included in support of clause 2.2 (b) of the Apple Public License,
+ * Version 2.0.
+ */
 
 /*
- * sysctl system call.
+* DEPRECATED sysctl system call code
+ *
+ * Everything in this file is deprecated. Sysctls should be handled
+ * by the code in kern_newsysctl.c.
+ * The remaining "case" sections are supposed to be converted into
+ * SYSCTL_*-style definitions, and as soon as all of them are gone,
+ * this source file is supposed to die.
+ *
+ * DO NOT ADD ANY MORE "case" SECTIONS TO THIS FILE, instead define
+ * your sysctl with SYSCTL_INT, SYSCTL_PROC etc. in your source file.
  */
 
 #include <sys/param.h>
 #include <sys/systm.h>
 #include <sys/kernel.h>
 #include <sys/malloc.h>
-#include <sys/proc.h>
-#include <sys/file.h>
-#include <sys/vnode.h>
+#include <sys/proc_internal.h>
+#include <sys/kauth.h>
+#include <sys/file_internal.h>
+#include <sys/vnode_internal.h>
 #include <sys/unistd.h>
 #include <sys/buf.h>
 #include <sys/ioctl.h>
+#include <sys/namei.h>
 #include <sys/tty.h>
 #include <sys/disklabel.h>
 #include <sys/vm.h>
 #include <sys/sysctl.h>
+#include <sys/user.h>
+#include <sys/aio_kern.h>
+#include <sys/reboot.h>
+
+#include <security/audit/audit.h>
+#include <kern/kalloc.h>
+
 #include <mach/machine.h>
+#include <mach/mach_host.h>
 #include <mach/mach_types.h>
 #include <mach/vm_param.h>
+#include <kern/mach_param.h>
 #include <kern/task.h>
+#include <kern/thread.h>
+#include <kern/lock.h>
+#include <kern/processor.h>
+#include <kern/debug.h>
 #include <vm/vm_kern.h>
+#include <vm/vm_map.h>
 #include <mach/host_info.h>
 
-extern vm_map_t bsd_pageable_map;
-
-#include <sys/mount.h>
+#include <sys/mount_internal.h>
 #include <sys/kdebug.h>
+#include <sys/sysproto.h>
 
 #include <IOKit/IOPlatformExpert.h>
 #include <pexpert/pexpert.h>
 
-#if __ppc__
-#include <osfmk/ppc/machine_routines.h>
+#include <machine/machine_routines.h>
+#include <machine/exec.h>
+
+#include <vm/vm_protos.h>
+#include <sys/imgsrc.h>
+
+#if defined(__i386__) || defined(__x86_64__)
+#include <i386/cpuid.h>
 #endif
 
-sysctlfn kern_sysctl;
-sysctlfn hw_sysctl;
-#ifdef DEBUG
-sysctlfn debug_sysctl;
+#if CONFIG_FREEZE
+#include <sys/kern_memorystatus.h>
 #endif
-extern sysctlfn vm_sysctl;
-extern sysctlfn vfs_sysctl;
+
+/*
+ * deliberately setting max requests to really high number
+ * so that runaway settings do not cause MALLOC overflows
+ */
+#define AIO_MAX_REQUESTS (128 * CONFIG_AIO_MAX)
+
 extern sysctlfn net_sysctl;
 extern sysctlfn cpu_sysctl;
+extern int aio_max_requests;                           
+extern int aio_max_requests_per_process;       
+extern int aio_worker_threads;                         
+extern int lowpri_IO_window_msecs;
+extern int lowpri_IO_delay_msecs;
+extern int nx_enabled;
+extern int speculative_reads_disabled;
+extern int ignore_is_ssd;
+extern unsigned int speculative_prefetch_max;
+extern unsigned int speculative_prefetch_max_iosize;
+extern unsigned int preheat_pages_max;
+extern unsigned int preheat_pages_min;
+extern long numvnodes;
+
+extern unsigned int vm_max_delayed_work_limit;
+extern unsigned int vm_max_batch;
+
+extern unsigned int vm_page_free_min;
+extern unsigned int vm_page_free_target;
+extern unsigned int vm_page_free_reserved;
+extern unsigned int vm_page_speculative_percentage;
+extern unsigned int vm_page_speculative_q_age_ms;
 
+/*
+ * Conditionally allow dtrace to see these functions for debugging purposes.
+ */
+#ifdef STATIC
+#undef STATIC
+#endif
+#if 0
+#define STATIC
+#else
+#define STATIC static
+#endif
 
+extern boolean_t    mach_timer_coalescing_enabled;
+
+STATIC void
+fill_user32_eproc(proc_t, struct user32_eproc *__restrict);
+STATIC void
+fill_user32_externproc(proc_t, struct user32_extern_proc *__restrict);
+STATIC void
+fill_user64_eproc(proc_t, struct user64_eproc *__restrict);
+STATIC void
+fill_user64_proc(proc_t, struct user64_kinfo_proc *__restrict);
+STATIC void
+fill_user64_externproc(proc_t, struct user64_extern_proc *__restrict);
+STATIC void
+fill_user32_proc(proc_t, struct user32_kinfo_proc *__restrict);
+
+extern int 
+kdbg_control(int *name, u_int namelen, user_addr_t where, size_t * sizep);
+#if NFSCLIENT
+extern int 
+netboot_root(void);
+#endif
 int
-userland_sysctl(struct proc *p, int *name, u_int namelen, void *old, size_t 
-               *oldlenp, int inkernel, void *new, size_t newlen, size_t *retval);
+pcsamples_ops(int *name, u_int namelen, user_addr_t where, size_t *sizep, 
+              proc_t p);
+__private_extern__ kern_return_t
+reset_vmobjectcache(unsigned int val1, unsigned int val2);
+int
+sysctl_procargs(int *name, u_int namelen, user_addr_t where, 
+                               size_t *sizep, proc_t cur_proc);
+STATIC int
+sysctl_procargsx(int *name, u_int namelen, user_addr_t where, size_t *sizep, 
+                 proc_t cur_proc, int argc_yes);
+int
+sysctl_struct(user_addr_t oldp, size_t *oldlenp, user_addr_t newp, 
+              size_t newlen, void *sp, int len);
+
+STATIC int sysdoproc_filt_KERN_PROC_PID(proc_t p, void * arg);
+STATIC int sysdoproc_filt_KERN_PROC_PGRP(proc_t p, void * arg);
+STATIC int sysdoproc_filt_KERN_PROC_TTY(proc_t p, void * arg);
+STATIC int  sysdoproc_filt_KERN_PROC_UID(proc_t p, void * arg);
+STATIC int  sysdoproc_filt_KERN_PROC_RUID(proc_t p, void * arg);
+#if CONFIG_LCTX
+STATIC int  sysdoproc_filt_KERN_PROC_LCID(proc_t p, void * arg);
+#endif
+int sysdoproc_callback(proc_t p, void *arg);
+
+
+/* forward declarations for non-static STATIC */
+STATIC void fill_loadavg64(struct loadavg *la, struct user64_loadavg *la64);
+STATIC void fill_loadavg32(struct loadavg *la, struct user32_loadavg *la32);
+STATIC int sysctl_handle_kern_threadname(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_sched_stats(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_sched_stats_enable(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_kdebug_ops SYSCTL_HANDLER_ARGS;
+STATIC int sysctl_dotranslate SYSCTL_HANDLER_ARGS;
+STATIC int sysctl_doaffinity SYSCTL_HANDLER_ARGS;
+#if COUNT_SYSCALLS
+STATIC int sysctl_docountsyscalls SYSCTL_HANDLER_ARGS;
+#endif /* COUNT_SYSCALLS */
+#if !CONFIG_EMBEDDED
+STATIC int sysctl_doprocargs SYSCTL_HANDLER_ARGS;
+#endif /* !CONFIG_EMBEDDED */
+STATIC int sysctl_doprocargs2 SYSCTL_HANDLER_ARGS;
+STATIC int sysctl_prochandle SYSCTL_HANDLER_ARGS;
+#if DEBUG
+STATIC int sysctl_dodebug SYSCTL_HANDLER_ARGS;
+#endif
+STATIC int sysctl_aiomax(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_aioprocmax(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_aiothreads(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_maxproc(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_osversion(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_sysctl_bootargs(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_maxvnodes(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_securelvl(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_domainname(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_hostname(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_procname(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_boottime(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_symfile(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+#if NFSCLIENT
+STATIC int sysctl_netboot(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+#endif
+#ifdef CONFIG_IMGSRC_ACCESS
+STATIC int sysctl_imgsrcdev(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+#endif
+STATIC int sysctl_usrstack(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_usrstack64(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_coredump(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_suid_coredump(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_delayterm(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_rage_vnode(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_kern_check_openevt(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_nx(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_loadavg(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_vm_toggle_address_reuse(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_swapusage(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+#if defined(__i386__) || defined(__x86_64__)
+STATIC int sysctl_sysctl_exec_affinity(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+#endif
+STATIC int fetch_process_cputype( proc_t cur_proc, int *name, u_int namelen, cpu_type_t *cputype);
+STATIC int sysctl_sysctl_native(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_sysctl_cputype(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_safeboot(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_singleuser(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
+STATIC int sysctl_slide(struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req);
 
-void
-fill_proc(struct proc *p,struct kinfo_proc *kp, int doingzomb);
 
-void
-fill_externproc(struct proc *p, struct extern_proc *exp);
+extern void IORegistrySetOSBuildVersion(char * build_version); 
 
+STATIC void
+fill_loadavg64(struct loadavg *la, struct user64_loadavg *la64)
+{
+       la64->ldavg[0]  = la->ldavg[0];
+       la64->ldavg[1]  = la->ldavg[1];
+       la64->ldavg[2]  = la->ldavg[2];
+       la64->fscale    = (user64_long_t)la->fscale;
+}
 
+STATIC void
+fill_loadavg32(struct loadavg *la, struct user32_loadavg *la32)
+{
+       la32->ldavg[0]  = la->ldavg[0];
+       la32->ldavg[1]  = la->ldavg[1];
+       la32->ldavg[2]  = la->ldavg[2];
+       la32->fscale    = (user32_long_t)la->fscale;
+}
 
 /*
- * temporary location for vm_sysctl.  This should be machine independant
+ * sysctl_mem_hold
+ *
+ * Description:        Wire down the callers address map on behalf of sysctl's
+ *             that perform their own copy operations while holding
+ *             locks e.g. in the paging path, which could lead to a
+ *             deadlock, or while holding a spinlock.
+ *
+ * Parameters: addr                    User buffer address
+ *             len                     User buffer length
+ *
+ * Returns:    0                       Success
+ *     vslock:ENOMEM                   Insufficient physical pages to wire
+ *     vslock:EACCES                   Bad protection mode
+ *     vslock:EINVAL                   Invalid parameters
+ *
+ * Notes:      This code is invoked for the first OID element where the
+ *             CTLFLAG_LOCKED is not specified for a given OID node
+ *             element durng OID traversal, and is held for all
+ *             subsequent node traversals, and only released after the
+ *             leaf node handler invocation is complete.
+ *
+ * Legacy:     For legacy scyctl's provided by third party code which
+ *             expect funnel protection for calls into their code, this
+ *             routine will also take the funnel, which will also only
+ *             be released after the leaf node handler is complete.
+ *
+ *             This is to support legacy 32 bit BSD KEXTs and legacy 32
+ *             bit single threaded filesystem KEXTs and similar code
+ *             which relies on funnel protection, e.g. for things like
+ *             FSID based sysctl's.
+ *
+ *             NEW CODE SHOULD NOT RELY ON THIS BEHAVIOUR!  IT WILL BE
+ *             REMOVED IN A FUTURE RELASE OF Mac OS X!
+ *
+ * Bugs:       This routine does nothing with the new_addr and new_len
+ *             at present, but it should, since read from the user space
+ *             process adddress space which could potentially trigger
+ *             paging may also be occurring deep down.  This is due to
+ *             a current limitation of the vslock() routine, which will
+ *             always request a wired mapping be read/write, due to not
+ *             taking an access mode parameter.  Note that this could
+ *             also cause problems for output on architectures where
+ *             write access does not require read acccess if the current
+ *             mapping lacks read access.
+ *
+ * XXX:                To be moved to kern_newsysctl.c to avoid __private_extern__
  */
-vm_sysctl(name, namelen, oldp, oldlenp, newp, newlen, p)
-       int *name;
-       u_int namelen;
-       void *oldp;
-       size_t *oldlenp;
-       void *newp;
-       size_t newlen;
-       struct proc *p;
-{
-       int error, level, inthostid;
-       extern long avenrun[3], mach_factor[3];
-       struct loadavg loadinfo;
-
-       //if (namelen != 1 && !(name[0] == VM_LOADAVG))
-               //return (ENOTDIR);             /* overloaded */
-
-       switch (name[0]) {
-       case VM_LOADAVG:
-               loadinfo.ldavg[0] = avenrun[0];
-               loadinfo.ldavg[1] = avenrun[1];
-               loadinfo.ldavg[2] = avenrun[2];
-               loadinfo.fscale = LSCALE;
-               return (sysctl_struct(oldp, oldlenp, newp, newlen, &loadinfo, sizeof(struct loadavg)));
-       case VM_MACHFACTOR:
-               loadinfo.ldavg[0] = mach_factor[0];
-               loadinfo.ldavg[1] = mach_factor[1];
-               loadinfo.ldavg[2] = mach_factor[2];
-               loadinfo.fscale = LSCALE;
-               return (sysctl_struct(oldp, oldlenp, newp, newlen, &loadinfo, sizeof(struct loadavg)));
-       case VM_METER:
-               return (EOPNOTSUPP);
-       case VM_MAXID:
-               return (EOPNOTSUPP);
-       default:
-               return (EOPNOTSUPP);
-       }
-       /* NOTREACHED */
-       return (EOPNOTSUPP);
+int sysctl_mem_lock(user_addr_t old_addr, user_size_t old_len, user_addr_t new_addr, user_size_t new_len);
+int
+sysctl_mem_lock(__unused user_addr_t old_addr, __unused user_size_t old_len, __unused user_addr_t new_addr, __unused user_size_t new_len)
+{
+       return 0;
 }
 
 /*
  * Locking and stats
  */
-static struct sysctl_lock {
-       int     sl_lock;
-       int     sl_want;
-       int     sl_locked;
-} memlock;
-
-struct __sysctl_args {
-       int *name;
-       u_int namelen;
-       void *old;
-       size_t *oldlenp;
-       void *new;
-       size_t newlen;
-};
+
+/* sysctl() syscall */
 int
-__sysctl(p, uap, retval)
-       struct proc *p;
-       register struct __sysctl_args *uap;
-       register_t *retval;
-{
-       int error, dolock = 1;
-       size_t savelen, oldlen = 0;
-       sysctlfn *fn;
+__sysctl(proc_t p, struct __sysctl_args *uap, __unused int32_t *retval)
+{
+       boolean_t funnel_state = FALSE;         /* not held if unknown */
+       int error;
+       size_t savelen = 0, oldlen = 0, newlen;
        int name[CTL_MAXNAME];
-       int i;
+       int error1;
+       boolean_t vslock_taken = FALSE;
+       boolean_t funnel_taken = FALSE;
+#if CONFIG_MACF
+       kauth_cred_t my_cred;
+#endif
 
        /*
         * all top-level sysctl names are non-terminal
         */
        if (uap->namelen > CTL_MAXNAME || uap->namelen < 2)
                return (EINVAL);
-       if (error =
-           copyin(uap->name, &name, uap->namelen * sizeof(int)))
+       error = copyin(uap->name, &name[0], uap->namelen * sizeof(int));
+       if (error)
                return (error);
+               
+       AUDIT_ARG(ctlname, name, uap->namelen);
+
+       if (proc_is64bit(p)) {
+               /* uap->newlen is a size_t value which grows to 64 bits 
+                * when coming from a 64-bit process.  since it's doubtful we'll 
+                * have a sysctl newp buffer greater than 4GB we shrink it to size_t
+                */
+               newlen = CAST_DOWN(size_t, uap->newlen);
+       }
+       else {
+               newlen = uap->newlen;
+       }
 
+/*
+ * XXX TODO:   push down rights check for CTL_HW OIDs; most duplicate
+ * XXX         it anyway, which is a performance sink, and requires use
+ * XXX         of SUID root programs (see <rdar://3915692>).
+ *
+ * Note:       Opt out of non-leaf node enforcement by removing this
+ *             check for the top level OID value, and then adding
+ *             CTLFLAG_ANYBODY to the leaf nodes in question.  Enforce as
+ *             suser for writed in leaf nodes by omitting this flag.
+ *             Enforce with a higher granularity by making the leaf node
+ *             of type SYSCTL_PROC() in order to provide a procedural
+ *             enforcement call site.
+ *
+ * NOTE:       This function is called prior to any subfunctions being
+ *             called with a fallback to userland_sysctl(); as such, this
+ *             permissions check here will veto the fallback operation.
+ */
        /* CTL_UNSPEC is used to get oid to AUTO_OID */
-       if (uap->new != NULL && 
-               (((name[0] == CTL_KERN) && (name[1] != KERN_IPC)) || 
-                (name[0] == CTL_HW) || (name[0] == CTL_VM) ||
-                (name[0] == CTL_VFS)) &&
-           (error = suser(p->p_ucred, &p->p_acflag)))
+       if (uap->new != USER_ADDR_NULL
+           && ((name[0] == CTL_HW)
+           || (name[0] == CTL_VM))
+           && (error = suser(kauth_cred_get(), &p->p_acflag)))
                return (error);
 
-       switch (name[0]) {
-       case CTL_KERN:
-               fn = kern_sysctl;
-               if (name[1] != KERN_VNODE)      /* XXX */
-                       dolock = 0;
-               break;
-       case CTL_HW:
-               fn = hw_sysctl;
-               break;
-       case CTL_VM:
-               fn = vm_sysctl;
-               break;
-                
-       case CTL_VFS:
-               fn = vfs_sysctl;
-               break;
-#if FIXME  /* [ */
-       case CTL_MACHDEP:
-               fn = cpu_sysctl;
-               break;
-#endif  /* FIXME ] */
-#ifdef DEBUG
-       case CTL_DEBUG:
-               fn = debug_sysctl;
-               break;
+// XXX need to relocate into each terminal instead of leaving this here...
+// XXX macf preemptory check.
+#if CONFIG_MACF
+       my_cred = kauth_cred_proc_ref(p);
+       error = mac_system_check_sysctl(
+           my_cred, 
+           (int *) name,
+           uap->namelen,
+           uap->old,
+           uap->oldlenp,
+           0,          /* XXX 1 for CTL_KERN checks */
+           uap->new,
+           newlen
+       );
+       kauth_cred_unref(&my_cred);
+       if (error)
+               return (error);
 #endif
-       default:
-               fn = 0;
+
+       if (uap->oldlenp != USER_ADDR_NULL) {
+               uint64_t        oldlen64 = fuulong(uap->oldlenp);
+
+               oldlen = CAST_DOWN(size_t, oldlen64);
+               /*
+                * If more than 4G, clamp to 4G - useracc() below will catch
+                * with an EFAULT, if it's actually necessary.
+                */
+               if (oldlen64 > 0x00000000ffffffffULL)
+                       oldlen = 0xffffffffUL;
        }
 
-       if (uap->oldlenp &&
-           (error = copyin(uap->oldlenp, &oldlen, sizeof(oldlen))))
-               return (error);
+       if ((name[0] == CTL_VFS || name[0] == CTL_VM)) {
+               /*
+                * Always take the funnel for CTL_VFS and CTL_VM
+                *
+                * XXX We should also take it for any OID without the
+                * XXX CTLFLAG_LOCKED set on it; fix this later!
+                */
+               funnel_state = thread_funnel_set(kernel_flock, TRUE);
+               funnel_taken = TRUE;
+
+               /*
+                * XXX Take the vslock() only when we are copying out; this
+                * XXX erroneously assumes that the copy in will not cause
+                * XXX a fault if caled from the paging path due to the
+                * XXX having been recently touched in order to establish
+                * XXX the input data.  This is a bad assumption.
+                *
+                * Note:        This is overkill, but third parties might
+                *              already call sysctl internally in KEXTs that
+                *              implement mass storage drivers.  If you are
+                *              writing a new KEXT, don't do that.
+                */
+               if(uap->old != USER_ADDR_NULL) {
+                       if (!useracc(uap->old, (user_size_t)oldlen, B_WRITE)) {
+                               thread_funnel_set(kernel_flock, funnel_state);
+                               return (EFAULT);
+                       }
 
-       if (uap->old != NULL) {
-               if (!useracc(uap->old, oldlen, B_WRITE))
-                       return (EFAULT);
-
-               /* The pc sampling mechanism does not need to take this lock */
-               if (name[1] != KERN_PCSAMPLES) {
-                 while (memlock.sl_lock) {
-                       memlock.sl_want = 1;
-                       sleep((caddr_t)&memlock, PRIBIO+1);
-                       memlock.sl_locked++;
-                 }
-                 memlock.sl_lock = 1;
+                       if (oldlen) {
+                               if ((error = vslock(uap->old, (user_size_t)oldlen))) {
+                                       thread_funnel_set(kernel_flock, funnel_state);
+                                       return(error);
+                               }
+                               savelen = oldlen;
+                               vslock_taken = TRUE;
+                       }
                }
+       }
 
-               if (dolock)
-                       vslock(uap->old, oldlen);
-               savelen = oldlen;
+       /*
+        * XXX convert vfs_sysctl subelements to newsysctl; this is hard
+        * XXX because of VFS_NUMMNTOPS being top level.
+        */
+       error = ENOTSUP;
+       if (name[0] == CTL_VFS) {
+               error = vfs_sysctl(name + 1, uap->namelen - 1, uap->old,
+                       &oldlen, uap->new, newlen, p);
        }
 
-       if (fn)
-           error = (*fn)(name + 1, uap->namelen - 1, uap->old,
-                         &oldlen, uap->new, uap->newlen, p);
-       else
-           error = EOPNOTSUPP;
-
-       if ( (name[0] != CTL_VFS) && (error == EOPNOTSUPP))
-               error = userland_sysctl(p, name, uap->namelen,
-                                       uap->old, uap->oldlenp, 0,
-                                       uap->new, uap->newlen, &oldlen);
-
-       if (uap->old != NULL) {
-               if (dolock)
-                       vsunlock(uap->old, savelen, B_WRITE);
-               if (name[1] != KERN_PCSAMPLES) {
-                 memlock.sl_lock = 0;
-                 if (memlock.sl_want) {
-                       memlock.sl_want = 0;
-                       wakeup((caddr_t)&memlock);
-                 }
-               }
+       if (vslock_taken == TRUE) {
+               error1 = vsunlock(uap->old, (user_size_t)savelen, B_WRITE);
+               if (!error)
+                       error = error1;
+        }
+
+       if ( (name[0] != CTL_VFS) && (error == ENOTSUP) ) {
+               size_t  tmp = oldlen;
+               error = userland_sysctl(p, name, uap->namelen, uap->old, &tmp, 
+                                       uap->new, newlen, &oldlen);
        }
+
+       /*
+        * If we took the funnel, which we only do for CTL_VFS and CTL_VM on
+        * 32 bit architectures, then drop it.
+        *
+        * XXX the grabbing and dropping need to move into the leaf nodes,
+        * XXX for sysctl's that are not marked CTLFLAG_LOCKED, but this is
+        * XXX true for the vslock, as well.  We have a start at a routine
+        * to wrapper this (above), but it's not turned on.  The current code
+        * removed the funnel and the vslock() from all but these two top
+        * level OIDs.  Note that VFS only needs to take the funnel if the FS
+        * against which it's operating is not thread safe (but since an FS
+        * can be in the paging path, it still needs to take the vslock()).
+        */
+       if (funnel_taken)
+               thread_funnel_set(kernel_flock, funnel_state);
+
        if ((error) && (error != ENOMEM))
                return (error);
 
-       if (uap->oldlenp) {
-               i = copyout(&oldlen, uap->oldlenp, sizeof(oldlen));
-               if (i) 
-                   return i;
-       }
+       if (uap->oldlenp != USER_ADDR_NULL)
+               error = suulong(uap->oldlenp, oldlen);
 
        return (error);
 }
@@ -297,237 +552,315 @@ __sysctl(p, uap, retval)
 /*
  * Attributes stored in the kernel.
  */
-extern char hostname[MAXHOSTNAMELEN]; /* defined in bsd/kern/init_main.c */
-extern int hostnamelen;
-extern char domainname[MAXHOSTNAMELEN];
-extern int domainnamelen;
-extern long hostid;
+__private_extern__ char corefilename[MAXPATHLEN+1];
+__private_extern__ int do_coredump;
+__private_extern__ int sugid_coredump;
+
+#if COUNT_SYSCALLS
+__private_extern__ int do_count_syscalls;
+#endif
+
 #ifdef INSECURE
 int securelevel = -1;
 #else
 int securelevel;
 #endif
 
-int get_kernel_symfile( struct proc *p, char **symfile );
+STATIC int
+sysctl_doaffinity SYSCTL_HANDLER_ARGS
+{
+       __unused int cmd = oidp->oid_arg2;      /* subcommand*/
+       int *name = arg1;               /* oid element argument vector */
+       int namelen = arg2;             /* number of oid element arguments */
+       user_addr_t oldp = req->oldptr; /* user buffer copy out address */
+       size_t *oldlenp = &req->oldlen; /* user buffer copy out size */
+       user_addr_t newp = req->newptr; /* user buffer copy in address */
+//     size_t newlen = req->newlen;    /* user buffer copy in size */
+
+       int error = ENOTSUP;            /* Default to failure */
+
+       proc_t cur_proc = current_proc();
+
+       if (namelen < 1)
+               return (ENOTSUP);
+
+       if (name[0] == 0 && 1 == namelen) {
+               error = sysctl_rdint(oldp, oldlenp, newp,
+                                       (cur_proc->p_flag & P_AFFINITY) ? 1 : 0);
+       } else if (name[0] == 1 && 2 == namelen) {
+               if (name[1] == 0) {
+                       OSBitAndAtomic(~((uint32_t)P_AFFINITY), &cur_proc->p_flag);
+               } else {
+                       OSBitOrAtomic(P_AFFINITY, &cur_proc->p_flag);
+               }
+               error =  0;
+       }
+
+       /* adjust index so we return the right required/consumed amount */
+       if (!error)
+               req->oldidx += req->oldlen;
 
-/*
- * kernel related system variables.
- */
-kern_sysctl(name, namelen, oldp, oldlenp, newp, newlen, p)
-       int *name;
-       u_int namelen;
-       void *oldp;
-       size_t *oldlenp;
-       void *newp;
-       size_t newlen;
-       struct proc *p;
-{
-       int error, level, inthostid;
-       unsigned int oldval=0;
-       extern char ostype[], osrelease[], version[];
-
-       /* all sysctl names at this level are terminal */
-       if (namelen != 1 && !(name[0] == KERN_PROC || name[0] == KERN_PROF 
-               || name[0] == KERN_KDEBUG
-               || name[0] == KERN_PROCARGS
-                || name[0] == KERN_PCSAMPLES
-                || name[0] == KERN_IPC
-       ))
-               return (ENOTDIR);               /* overloaded */
+       return (error);
+}
+SYSCTL_PROC(_kern, KERN_AFFINITY, affinity, CTLTYPE_NODE|CTLFLAG_RD | CTLFLAG_LOCKED,
+       0,                      /* Pointer argument (arg1) */
+       0,                      /* Integer argument (arg2) */
+       sysctl_doaffinity,      /* Handler function */
+       NULL,                   /* Data pointer */
+       "");
+
+STATIC int
+sysctl_dotranslate SYSCTL_HANDLER_ARGS
+{
+       __unused int cmd = oidp->oid_arg2;      /* subcommand*/
+       int *name = arg1;               /* oid element argument vector */
+       int namelen = arg2;             /* number of oid element arguments */
+       user_addr_t oldp = req->oldptr; /* user buffer copy out address */
+       size_t *oldlenp = &req->oldlen; /* user buffer copy out size */
+       user_addr_t newp = req->newptr; /* user buffer copy in address */
+//     size_t newlen = req->newlen;    /* user buffer copy in size */
+       int error;
+
+       proc_t cur_proc = current_proc();
+       proc_t p;
+       int istranslated = 0;
+       kauth_cred_t my_cred;
+       uid_t uid;
 
-       switch (name[0]) {
-       case KERN_OSTYPE:
-               return (sysctl_rdstring(oldp, oldlenp, newp, ostype));
-       case KERN_OSRELEASE:
-               return (sysctl_rdstring(oldp, oldlenp, newp, osrelease));
-       case KERN_OSREV:
-               return (sysctl_rdint(oldp, oldlenp, newp, BSD));
-       case KERN_VERSION:
-               return (sysctl_rdstring(oldp, oldlenp, newp, version));
-       case KERN_MAXVNODES:
-               oldval = desiredvnodes;
-               error = sysctl_int(oldp, oldlenp, newp, 
-                               newlen, &desiredvnodes);
-               reset_vmobjectcache(oldval, desiredvnodes);
-               return(error);
-       case KERN_MAXPROC:
-               return (sysctl_int(oldp, oldlenp, newp, newlen, &maxproc));
-       case KERN_MAXFILES:
-               return (sysctl_int(oldp, oldlenp, newp, newlen, &maxfiles));
-       case KERN_ARGMAX:
-               return (sysctl_rdint(oldp, oldlenp, newp, ARG_MAX));
-       case KERN_SECURELVL:
-               level = securelevel;
-               if ((error = sysctl_int(oldp, oldlenp, newp, newlen, &level)) ||
-                   newp == NULL)
-                       return (error);
-               if (level < securelevel && p->p_pid != 1)
-                       return (EPERM);
-               securelevel = level;
-               return (0);
-       case KERN_HOSTNAME:
-               error = sysctl_string(oldp, oldlenp, newp, newlen,
-                   hostname, sizeof(hostname));
-               if (newp && !error)
-                       hostnamelen = newlen;
-               return (error);
-       case KERN_DOMAINNAME:
-               error = sysctl_string(oldp, oldlenp, newp, newlen,
-                   domainname, sizeof(domainname));
-               if (newp && !error)
-                       domainnamelen = newlen;
-               return (error);
-       case KERN_HOSTID:
-               inthostid = hostid;  /* XXX assumes sizeof long <= sizeof int */
-               error =  sysctl_int(oldp, oldlenp, newp, newlen, &inthostid);
-               hostid = inthostid;
-               return (error);
-       case KERN_CLOCKRATE:
-               return (sysctl_clockrate(oldp, oldlenp));
-       case KERN_BOOTTIME:
-               return (sysctl_rdstruct(oldp, oldlenp, newp, &boottime,
-                   sizeof(struct timeval)));
-       case KERN_VNODE:
-               return (sysctl_vnode(oldp, oldlenp));
-       case KERN_PROC:
-               return (sysctl_doproc(name + 1, namelen - 1, oldp, oldlenp));
-       case KERN_FILE:
-               return (sysctl_file(oldp, oldlenp));
-#ifdef GPROF
-       case KERN_PROF:
-               return (sysctl_doprof(name + 1, namelen - 1, oldp, oldlenp,
-                   newp, newlen));
-#endif
-       case KERN_POSIX1:
-               return (sysctl_rdint(oldp, oldlenp, newp, _POSIX_VERSION));
-       case KERN_NGROUPS:
-               return (sysctl_rdint(oldp, oldlenp, newp, NGROUPS_MAX));
-       case KERN_JOB_CONTROL:
-               return (sysctl_rdint(oldp, oldlenp, newp, 1));
-       case KERN_SAVED_IDS:
-#ifdef _POSIX_SAVED_IDS
-               return (sysctl_rdint(oldp, oldlenp, newp, 1));
-#else
-               return (sysctl_rdint(oldp, oldlenp, newp, 0));
-#endif
-#if FIXME  /* [ */
-       case KERN_MAXPARTITIONS:
-               return (sysctl_rdint(oldp, oldlenp, newp, MAXPARTITIONS));
-#endif  /* FIXME ] */
-       case KERN_KDEBUG:
-               return (kdebug_ops(name + 1, namelen - 1, oldp, oldlenp, p));
-       case KERN_PCSAMPLES:
-               return (pcsamples_ops(name + 1, namelen - 1, oldp, oldlenp, p));
-       case KERN_PROCARGS:
-               /* new one as it does not use kinfo_proc */
-               return (sysctl_procargs(name + 1, namelen - 1, oldp, oldlenp));
-        case KERN_SYMFILE:
-                {
-                    char *str;
-                    error = get_kernel_symfile( p, &str );
-                    if ( error ) return error;
-                    return (sysctl_rdstring(oldp, oldlenp, newp, str));
-                }    
-       default:
-               return (EOPNOTSUPP);
+       if (namelen != 1)
+               return (ENOTSUP);
+
+       p = proc_find(name[0]);
+       if (p == NULL)
+               return (EINVAL);
+
+       my_cred = kauth_cred_proc_ref(p);
+       uid = kauth_cred_getuid(my_cred);
+       kauth_cred_unref(&my_cred);
+       if ((uid != kauth_cred_getuid(kauth_cred_get())) 
+               && suser(kauth_cred_get(), &cur_proc->p_acflag)) {
+               proc_rele(p);
+               return (EPERM);
        }
-       /* NOTREACHED */
-}
 
+       istranslated = (p->p_flag & P_TRANSLATED);
+       proc_rele(p);
+       error =  sysctl_rdint(oldp, oldlenp, newp,
+                               (istranslated != 0) ? 1 : 0);
+
+       /* adjust index so we return the right required/consumed amount */
+       if (!error)
+               req->oldidx += req->oldlen;
+
+       return (error);
+}
 /*
- * hardware related system variables.
+ * XXX make CTLFLAG_RW so sysctl_rdint() will EPERM on attempts to write;
+ * XXX this may not be necessary.
  */
-hw_sysctl(name, namelen, oldp, oldlenp, newp, newlen, p)
-       int *name;
-       u_int namelen;
-       void *oldp;
-       size_t *oldlenp;
-       void *newp;
-       size_t newlen;
-       struct proc *p;
-{
-       char dummy[65];
-       int  epochTemp;
-       extern int vm_page_wire_count;
-#if __ppc__
-       ml_ppc_cpu_info_t cpu_info;
-
-       ml_ppc_get_info(&cpu_info);
-#endif
+SYSCTL_PROC(_kern, KERN_TRANSLATE, translate, CTLTYPE_NODE|CTLFLAG_RW | CTLFLAG_LOCKED,
+       0,                      /* Pointer argument (arg1) */
+       0,                      /* Integer argument (arg2) */
+       sysctl_dotranslate,     /* Handler function */
+       NULL,                   /* Data pointer */
+       "");
+
+STATIC int
+sysctl_handle_kern_threadname( __unused struct sysctl_oid *oidp, __unused void *arg1,
+             __unused int arg2, struct sysctl_req *req)
+{
+       int error;
+       struct uthread *ut = get_bsdthread_info(current_thread());
+       user_addr_t oldp=0, newp=0;
+       size_t *oldlenp=NULL;
+       size_t newlen=0;
+
+       oldp = req->oldptr;
+       oldlenp = &(req->oldlen);
+       newp = req->newptr;
+       newlen = req->newlen;
+
+       /* We want the current length, and maybe the string itself */
+       if(oldlenp) {
+               /* if we have no thread name yet tell'em we want MAXTHREADNAMESIZE - 1 */
+               size_t currlen = MAXTHREADNAMESIZE - 1;
+               
+               if(ut->pth_name)
+                       /* use length of current thread name */
+                       currlen = strlen(ut->pth_name);
+               if(oldp) {
+                       if(*oldlenp < currlen)
+                               return ENOMEM;
+                       /* NOTE - we do not copy the NULL terminator */
+                       if(ut->pth_name) {
+                               error = copyout(ut->pth_name,oldp,currlen);
+                               if(error)
+                                       return error;
+                       }
+               }       
+               /* return length of thread name minus NULL terminator (just like strlen)  */
+               req->oldidx = currlen;
+       }
 
-       /* all sysctl names at this level are terminal */
-       if (namelen != 1)
-               return (ENOTDIR);               /* overloaded */
+       /* We want to set the name to something */
+       if(newp) 
+       {
+               if(newlen > (MAXTHREADNAMESIZE - 1))
+                       return ENAMETOOLONG;
+               if(!ut->pth_name)
+               {
+                       ut->pth_name = (char*)kalloc( MAXTHREADNAMESIZE );
+                       if(!ut->pth_name)
+                               return ENOMEM;
+               }
+               bzero(ut->pth_name, MAXTHREADNAMESIZE);
+               error = copyin(newp, ut->pth_name, newlen);
+               if(error)
+                       return error;
+       }
+               
+       return 0;
+}
+
+SYSCTL_PROC(_kern, KERN_THREADNAME, threadname, CTLFLAG_ANYBODY | CTLTYPE_STRING | CTLFLAG_RW | CTLFLAG_LOCKED, 0, 0, sysctl_handle_kern_threadname,"A","");
 
-       switch (name[0]) {
-       case HW_MACHINE:
-               if(!PEGetMachineName(dummy,64))
-                       return(EINVAL);
-               return (sysctl_rdstring(oldp, oldlenp, newp, dummy));
-       case HW_MODEL:
-               if(!PEGetModelName(dummy,64))
-                       return(EINVAL);
-               return (sysctl_rdstring(oldp, oldlenp, newp, dummy));
-       case HW_NCPU: {
-               int numcpus=1;
-               host_basic_info_data_t hinfo;
-               kern_return_t kret;
-               int count= HOST_BASIC_INFO_COUNT;
 #define BSD_HOST 1
+STATIC int
+sysctl_sched_stats(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       host_basic_info_data_t hinfo;
+       kern_return_t kret;
+       uint32_t size;
+       int changed;
+       mach_msg_type_number_t count = HOST_BASIC_INFO_COUNT;
+       struct _processor_statistics_np *buf;
+       int error;
+
+       kret = host_info((host_t)BSD_HOST, HOST_BASIC_INFO, (host_info_t)&hinfo, &count);
+       if (kret != KERN_SUCCESS) {
+               return EINVAL;
+       }
+
+       size = sizeof(struct _processor_statistics_np) * (hinfo.logical_cpu_max + 2); /* One for RT Queue, One for Fair Share Queue */
+       
+       if (req->oldlen < size) {
+               return EINVAL;
+       }
+
+       MALLOC(buf, struct _processor_statistics_np*, size, M_TEMP, M_ZERO | M_WAITOK);
+       
+       kret = get_sched_statistics(buf, &size);
+       if (kret != KERN_SUCCESS) {
+               error = EINVAL;
+               goto out;
+       }
+
+       error = sysctl_io_opaque(req, buf, size, &changed);
+       if (error) {
+               goto out;
+       }
+
+       if (changed) {
+               panic("Sched info changed?!");
+       }
+out:
+       FREE(buf, M_TEMP);
+       return error;
+}
+
+SYSCTL_PROC(_kern, OID_AUTO, sched_stats, CTLFLAG_LOCKED, 0, 0, sysctl_sched_stats, "-", "");
+
+STATIC int
+sysctl_sched_stats_enable(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, __unused struct sysctl_req *req)
+{
+       boolean_t active;
+       int res;
+
+       if (req->newlen != sizeof(active)) {
+               return EINVAL;
+       }
+
+       res = copyin(req->newptr, &active, sizeof(active));
+       if (res != 0) {
+               return res;
+       }
+
+       return set_sched_stats_active(active);
+}
+
+SYSCTL_PROC(_kern, OID_AUTO, sched_stats_enable, CTLFLAG_LOCKED | CTLFLAG_WR, 0, 0, sysctl_sched_stats_enable, "-", "");
+
+extern int get_kernel_symfile(proc_t, char **);
 
-                       kret = host_info(BSD_HOST, HOST_BASIC_INFO, &hinfo, &count);
-                       if (kret == KERN_SUCCESS) {
-                               numcpus = hinfo.avail_cpus;
-                               return (sysctl_rdint(oldp, oldlenp, newp, numcpus));
-                       } else {
-                               return(EINVAL);
+#if COUNT_SYSCALLS
+#define KERN_COUNT_SYSCALLS (KERN_OSTYPE + 1000)
+
+extern int     nsysent;
+extern int syscalls_log[];
+extern const char *syscallnames[];
+
+STATIC int
+sysctl_docountsyscalls SYSCTL_HANDLER_ARGS
+{
+       __unused int cmd = oidp->oid_arg2;      /* subcommand*/
+       __unused int *name = arg1;      /* oid element argument vector */
+       __unused int namelen = arg2;    /* number of oid element arguments */
+       user_addr_t oldp = req->oldptr; /* user buffer copy out address */
+       size_t *oldlenp = &req->oldlen; /* user buffer copy out size */
+       user_addr_t newp = req->newptr; /* user buffer copy in address */
+       size_t newlen = req->newlen;    /* user buffer copy in size */
+       int error;
+
+       int tmp;
+
+       /* valid values passed in:
+        * = 0 means don't keep called counts for each bsd syscall
+        * > 0 means keep called counts for each bsd syscall
+        * = 2 means dump current counts to the system log
+        * = 3 means reset all counts
+        * for example, to dump current counts:  
+        *              sysctl -w kern.count_calls=2
+        */
+       error = sysctl_int(oldp, oldlenp, newp, newlen, &tmp);
+       if ( error != 0 ) {
+               return (error);
+       }
+               
+       if ( tmp == 1 ) {
+               do_count_syscalls = 1;
+       }
+       else if ( tmp == 0 || tmp == 2 || tmp == 3 ) {
+               int                     i;
+               for ( i = 0; i < nsysent; i++ ) {
+                       if ( syscalls_log[i] != 0 ) {
+                               if ( tmp == 2 ) {
+                                       printf("%d calls - name %s \n", syscalls_log[i], syscallnames[i]);
+                               }
+                               else {
+                                       syscalls_log[i] = 0;
+                               }
                        }
                }
-       case HW_BYTEORDER:
-               return (sysctl_rdint(oldp, oldlenp, newp, BYTE_ORDER));
-       case HW_PHYSMEM:
-               return (sysctl_rdint(oldp, oldlenp, newp, mem_size));
-       case HW_USERMEM:
-               return (sysctl_rdint(oldp, oldlenp, newp,
-                   (mem_size - vm_page_wire_count * page_size)));
-       case HW_PAGESIZE:
-               return (sysctl_rdint(oldp, oldlenp, newp, page_size));
-       case HW_EPOCH:
-               epochTemp = PEGetPlatformEpoch();
-               if (epochTemp == -1) return(EINVAL);
-               return (sysctl_rdint(oldp, oldlenp, newp, epochTemp));
-       case HW_BUS_FREQ:
-               return (sysctl_rdint(oldp, oldlenp, newp, gPEClockFrequencyInfo.bus_clock_rate_hz));
-       case HW_CPU_FREQ:
-               return (sysctl_rdint(oldp, oldlenp, newp, gPEClockFrequencyInfo.cpu_clock_rate_hz));
-#if __ppc__
-       case HW_VECTORUNIT:
-               return (sysctl_rdint(oldp, oldlenp, newp, cpu_info.vector_unit));
-       case HW_CACHELINE:
-               return (sysctl_rdint(oldp, oldlenp, newp, cpu_info.cache_line_size));
-       case HW_L1ICACHESIZE:
-               return (sysctl_rdint(oldp, oldlenp, newp, cpu_info.l1_icache_size));
-       case HW_L1DCACHESIZE:
-               return (sysctl_rdint(oldp, oldlenp, newp, cpu_info.l1_dcache_size));
-       case HW_L2SETTINGS:
-               if (cpu_info.l2_cache_size == 0xFFFFFFFF) return(EINVAL);
-               return (sysctl_rdint(oldp, oldlenp, newp, cpu_info.l2_settings));
-       case HW_L2CACHESIZE:
-               if (cpu_info.l2_cache_size == 0xFFFFFFFF) return(EINVAL);
-               return (sysctl_rdint(oldp, oldlenp, newp, cpu_info.l2_cache_size));
-       case HW_L3SETTINGS:
-               if (cpu_info.l3_cache_size == 0xFFFFFFFF) return(EINVAL);
-               return (sysctl_rdint(oldp, oldlenp, newp, cpu_info.l3_settings));
-       case HW_L3CACHESIZE:
-               if (cpu_info.l3_cache_size == 0xFFFFFFFF) return(EINVAL);
-               return (sysctl_rdint(oldp, oldlenp, newp, cpu_info.l3_cache_size));
-#endif
-       default:
-               return (EOPNOTSUPP);
+               if ( tmp != 0 ) {
+                       do_count_syscalls = 1;
+               }
        }
-}
 
-#ifdef DEBUG
+       /* adjust index so we return the right required/consumed amount */
+       if (!error)
+               req->oldidx += req->oldlen;
+
+       return (error);
+}
+SYSCTL_PROC(_kern, KERN_COUNT_SYSCALLS, count_syscalls, CTLTYPE_NODE|CTLFLAG_RD | CTLFLAG_LOCKED,
+       0,                      /* Pointer argument (arg1) */
+       0,                      /* Integer argument (arg2) */
+       sysctl_docountsyscalls, /* Handler function */
+       NULL,                   /* Data pointer */
+       "");
+#endif /* COUNT_SYSCALLS */
+
+#if DEBUG
 /*
  * Debugging related system variables.
  */
@@ -539,55 +872,92 @@ struct ctldebug debug2, debug3, debug4;
 struct ctldebug debug5, debug6, debug7, debug8, debug9;
 struct ctldebug debug10, debug11, debug12, debug13, debug14;
 struct ctldebug debug15, debug16, debug17, debug18, debug19;
-static struct ctldebug *debugvars[CTL_DEBUG_MAXID] = {
+STATIC struct ctldebug *debugvars[CTL_DEBUG_MAXID] = {
        &debug0, &debug1, &debug2, &debug3, &debug4,
        &debug5, &debug6, &debug7, &debug8, &debug9,
        &debug10, &debug11, &debug12, &debug13, &debug14,
        &debug15, &debug16, &debug17, &debug18, &debug19,
 };
-int
-debug_sysctl(name, namelen, oldp, oldlenp, newp, newlen, p)
-       int *name;
-       u_int namelen;
-       void *oldp;
-       size_t *oldlenp;
-       void *newp;
-       size_t newlen;
-       struct proc *p;
+STATIC int
+sysctl_dodebug SYSCTL_HANDLER_ARGS
 {
+       int cmd = oidp->oid_arg2;       /* subcommand*/
+       int *name = arg1;               /* oid element argument vector */
+       int namelen = arg2;             /* number of oid element arguments */
+       user_addr_t oldp = req->oldptr; /* user buffer copy out address */
+       size_t *oldlenp = &req->oldlen; /* user buffer copy out size */
+       user_addr_t newp = req->newptr; /* user buffer copy in address */
+       size_t newlen = req->newlen;    /* user buffer copy in size */
+       int error;
+
        struct ctldebug *cdp;
 
        /* all sysctl names at this level are name and field */
-       if (namelen != 2)
-               return (ENOTDIR);               /* overloaded */
-       cdp = debugvars[name[0]];
+       if (namelen != 1)
+               return (ENOTSUP);               /* overloaded */
+       if (cmd < 0 || cmd >= CTL_DEBUG_MAXID)
+               return (ENOTSUP);
+       cdp = debugvars[cmd];
        if (cdp->debugname == 0)
-               return (EOPNOTSUPP);
-       switch (name[1]) {
+               return (ENOTSUP);
+       switch (name[0]) {
        case CTL_DEBUG_NAME:
-               return (sysctl_rdstring(oldp, oldlenp, newp, cdp->debugname));
+               error = sysctl_rdstring(oldp, oldlenp, newp, cdp->debugname);
+               break;
        case CTL_DEBUG_VALUE:
-               return (sysctl_int(oldp, oldlenp, newp, newlen, cdp->debugvar));
+               error = sysctl_int(oldp, oldlenp, newp, newlen, cdp->debugvar);
+               break;
        default:
-               return (EOPNOTSUPP);
+               error = ENOTSUP;
+               break;
        }
-       /* NOTREACHED */
+
+       /* adjust index so we return the right required/consumed amount */
+       if (!error)
+               req->oldidx += req->oldlen;
+
+       return (error);
 }
+/*
+ * XXX We mark this RW instead of RD to let sysctl_rdstring() return the
+ * XXX historical error.
+ */
+SYSCTL_PROC(_debug, CTL_DEBUG_NAME, name, CTLTYPE_NODE|CTLFLAG_RW | CTLFLAG_LOCKED,
+       0,                      /* Pointer argument (arg1) */
+       CTL_DEBUG_NAME,         /* Integer argument (arg2) */
+       sysctl_dodebug,         /* Handler function */
+       NULL,                   /* Data pointer */
+       "Debugging");
+SYSCTL_PROC(_debug, CTL_DEBUG_VALUE, value, CTLTYPE_NODE|CTLFLAG_RW | CTLFLAG_LOCKED,
+       0,                      /* Pointer argument (arg1) */
+       CTL_DEBUG_VALUE,        /* Integer argument (arg2) */
+       sysctl_dodebug,         /* Handler function */
+       NULL,                   /* Data pointer */
+       "Debugging");
 #endif /* DEBUG */
 
+/*
+ * The following sysctl_* functions should not be used
+ * any more, as they can only cope with callers in
+ * user mode: Use new-style
+ *  sysctl_io_number()
+ *  sysctl_io_string()
+ *  sysctl_io_opaque()
+ * instead.
+ */
+
 /*
  * Validate parameters and get old / set new parameters
  * for an integer-valued sysctl function.
  */
-sysctl_int(oldp, oldlenp, newp, newlen, valp)
-       void *oldp;
-       size_t *oldlenp;
-       void *newp;
-       size_t newlen;
-       int *valp;
+int
+sysctl_int(user_addr_t oldp, size_t *oldlenp, 
+           user_addr_t newp, size_t newlen, int *valp)
 {
        int error = 0;
 
+       if (oldp != USER_ADDR_NULL && oldlenp == NULL)
+               return (EFAULT);
        if (oldp && *oldlenp < sizeof(int))
                return (ENOMEM);
        if (newp && newlen != sizeof(int))
@@ -595,22 +965,23 @@ sysctl_int(oldp, oldlenp, newp, newlen, valp)
        *oldlenp = sizeof(int);
        if (oldp)
                error = copyout(valp, oldp, sizeof(int));
-       if (error == 0 && newp)
+       if (error == 0 && newp) {
                error = copyin(newp, valp, sizeof(int));
+               AUDIT_ARG(value32, *valp);
+       }
        return (error);
 }
 
 /*
  * As above, but read-only.
  */
-sysctl_rdint(oldp, oldlenp, newp, val)
-       void *oldp;
-       size_t *oldlenp;
-       void *newp;
-       int val;
+int
+sysctl_rdint(user_addr_t oldp, size_t *oldlenp, user_addr_t newp, int val)
 {
        int error = 0;
 
+       if (oldp != USER_ADDR_NULL && oldlenp == NULL)
+               return (EFAULT);
        if (oldp && *oldlenp < sizeof(int))
                return (ENOMEM);
        if (newp)
@@ -621,32 +992,118 @@ sysctl_rdint(oldp, oldlenp, newp, val)
        return (error);
 }
 
+/*
+ * Validate parameters and get old / set new parameters
+ * for an quad(64bit)-valued sysctl function.
+ */
+int
+sysctl_quad(user_addr_t oldp, size_t *oldlenp, 
+            user_addr_t newp, size_t newlen, quad_t *valp)
+{
+       int error = 0;
+
+       if (oldp != USER_ADDR_NULL && oldlenp == NULL)
+               return (EFAULT);
+       if (oldp && *oldlenp < sizeof(quad_t))
+               return (ENOMEM);
+       if (newp && newlen != sizeof(quad_t))
+               return (EINVAL);
+       *oldlenp = sizeof(quad_t);
+       if (oldp)
+               error = copyout(valp, oldp, sizeof(quad_t));
+       if (error == 0 && newp)
+               error = copyin(newp, valp, sizeof(quad_t));
+       return (error);
+}
+
+/*
+ * As above, but read-only.
+ */
+int
+sysctl_rdquad(user_addr_t oldp, size_t *oldlenp, user_addr_t newp, quad_t val)
+{
+       int error = 0;
+
+       if (oldp != USER_ADDR_NULL && oldlenp == NULL)
+               return (EFAULT);
+       if (oldp && *oldlenp < sizeof(quad_t))
+               return (ENOMEM);
+       if (newp)
+               return (EPERM);
+       *oldlenp = sizeof(quad_t);
+       if (oldp)
+               error = copyout((caddr_t)&val, oldp, sizeof(quad_t));
+       return (error);
+}
+
+/*
+ * Validate parameters and get old / set new parameters
+ * for a string-valued sysctl function.  Unlike sysctl_string, if you
+ * give it a too small (but larger than 0 bytes) buffer, instead of
+ * returning ENOMEM, it truncates the returned string to the buffer
+ * size.  This preserves the semantics of some library routines
+ * implemented via sysctl, which truncate their returned data, rather
+ * than simply returning an error. The returned string is always NUL
+ * terminated.
+ */
+int
+sysctl_trstring(user_addr_t oldp, size_t *oldlenp, 
+              user_addr_t newp, size_t newlen, char *str, int maxlen)
+{
+       int len, copylen, error = 0;
+
+       if (oldp != USER_ADDR_NULL && oldlenp == NULL)
+               return (EFAULT);
+       copylen = len = strlen(str) + 1;
+       if (oldp && (len < 0 || *oldlenp < 1))
+               return (ENOMEM);
+       if (oldp && (*oldlenp < (size_t)len))
+               copylen = *oldlenp + 1;
+       if (newp && (maxlen < 0 || newlen >= (size_t)maxlen))
+               return (EINVAL);
+       *oldlenp = copylen - 1; /* deal with NULL strings correctly */
+       if (oldp) {
+               error = copyout(str, oldp, copylen);
+               if (!error) {
+                       unsigned char c = 0;
+                       /* NUL terminate */
+                       oldp += *oldlenp;
+                       error = copyout((void *)&c, oldp, sizeof(char));
+               }
+       }
+       if (error == 0 && newp) {
+               error = copyin(newp, str, newlen);
+               str[newlen] = 0;
+               AUDIT_ARG(text, (char *)str);
+       }
+       return (error);
+}
+
 /*
  * Validate parameters and get old / set new parameters
  * for a string-valued sysctl function.
  */
-sysctl_string(oldp, oldlenp, newp, newlen, str, maxlen)
-       void *oldp;
-       size_t *oldlenp;
-       void *newp;
-       size_t newlen;
-       char *str;
-       int maxlen;
+int
+sysctl_string(user_addr_t oldp, size_t *oldlenp, 
+              user_addr_t newp, size_t newlen, char *str, int maxlen)
 {
        int len, error = 0;
 
+       if (oldp != USER_ADDR_NULL && oldlenp == NULL)
+               return (EFAULT);
        len = strlen(str) + 1;
-       if (oldp && *oldlenp < len)
+       if (oldp && (len < 0 || *oldlenp < (size_t)len))
                return (ENOMEM);
-       if (newp && newlen >= maxlen)
+       if (newp && (maxlen < 0 || newlen >= (size_t)maxlen))
                return (EINVAL);
+       *oldlenp = len -1; /* deal with NULL strings correctly */
        if (oldp) {
-               *oldlenp = len;
                error = copyout(str, oldp, len);
        }
        if (error == 0 && newp) {
                error = copyin(newp, str, newlen);
                str[newlen] = 0;
+               AUDIT_ARG(text, (char *)str);
        }
        return (error);
 }
@@ -654,16 +1111,16 @@ sysctl_string(oldp, oldlenp, newp, newlen, str, maxlen)
 /*
  * As above, but read-only.
  */
-sysctl_rdstring(oldp, oldlenp, newp, str)
-       void *oldp;
-       size_t *oldlenp;
-       void *newp;
-       char *str;
+int
+sysctl_rdstring(user_addr_t oldp, size_t *oldlenp, 
+                user_addr_t newp, char *str)
 {
        int len, error = 0;
 
+       if (oldp != USER_ADDR_NULL && oldlenp == NULL)
+               return (EFAULT);
        len = strlen(str) + 1;
-       if (oldp && *oldlenp < len)
+       if (oldp && *oldlenp < (size_t)len)
                return (ENOMEM);
        if (newp)
                return (EPERM);
@@ -677,19 +1134,17 @@ sysctl_rdstring(oldp, oldlenp, newp, str)
  * Validate parameters and get old / set new parameters
  * for a structure oriented sysctl function.
  */
-sysctl_struct(oldp, oldlenp, newp, newlen, sp, len)
-       void *oldp;
-       size_t *oldlenp;
-       void *newp;
-       size_t newlen;
-       void *sp;
-       int len;
+int
+sysctl_struct(user_addr_t oldp, size_t *oldlenp, 
+              user_addr_t newp, size_t newlen, void *sp, int len)
 {
        int error = 0;
 
-       if (oldp && *oldlenp < len)
+       if (oldp != USER_ADDR_NULL && oldlenp == NULL)
+               return (EFAULT);
+       if (oldp && (len < 0 || *oldlenp < (size_t)len))
                return (ENOMEM);
-       if (newp && newlen > len)
+       if (newp && (len < 0 || newlen > (size_t)len))
                return (EINVAL);
        if (oldp) {
                *oldlenp = len;
@@ -704,15 +1159,15 @@ sysctl_struct(oldp, oldlenp, newp, newlen, sp, len)
  * Validate parameters and get old parameters
  * for a structure oriented sysctl function.
  */
-sysctl_rdstruct(oldp, oldlenp, newp, sp, len)
-       void *oldp;
-       size_t *oldlenp;
-       void *newp, *sp;
-       int len;
+int
+sysctl_rdstruct(user_addr_t oldp, size_t *oldlenp, 
+                user_addr_t newp, void *sp, int len)
 {
        int error = 0;
 
-       if (oldp && *oldlenp < len)
+       if (oldp != USER_ADDR_NULL && oldlenp == NULL)
+               return (EFAULT);
+       if (oldp && (len < 0 || *oldlenp < (size_t)len))
                return (ENOMEM);
        if (newp)
                return (EPERM);
@@ -722,284 +1177,569 @@ sysctl_rdstruct(oldp, oldlenp, newp, sp, len)
        return (error);
 }
 
-/*
- * Get file structures.
- */
-sysctl_file(where, sizep)
-       char *where;
-       size_t *sizep;
+STATIC int
+sysdoproc_filt_KERN_PROC_PID(proc_t p, void * arg)
 {
-       int buflen, error;
-       struct file *fp;
-       char *start = where;
-
-       buflen = *sizep;
-       if (where == NULL) {
-               /*
-                * overestimate by 10 files
-                */
-               *sizep = sizeof(filehead) + (nfiles + 10) * sizeof(struct file);
-               return (0);
-       }
+       if (p->p_pid != (pid_t)*(int*)arg)
+               return(0);
+       else
+               return(1);
+}
 
-       /*
-        * first copyout filehead
-        */
-       if (buflen < sizeof(filehead)) {
-               *sizep = 0;
-               return (0);
-       }
-       if (error = copyout((caddr_t)&filehead, where, sizeof(filehead)))
-               return (error);
-       buflen -= sizeof(filehead);
-       where += sizeof(filehead);
+STATIC int
+sysdoproc_filt_KERN_PROC_PGRP(proc_t p, void * arg)
+{
+       if (p->p_pgrpid != (pid_t)*(int*)arg)
+               return(0);
+       else
+         return(1);
+}
 
-       /*
-        * followed by an array of file structures
-        */
-       for (fp = filehead.lh_first; fp != 0; fp = fp->f_list.le_next) {
-               if (buflen < sizeof(struct file)) {
-                       *sizep = where - start;
-                       return (ENOMEM);
-               }
-               if (error = copyout((caddr_t)fp, where, sizeof (struct file)))
-                       return (error);
-               buflen -= sizeof(struct file);
-               where += sizeof(struct file);
-       }
-       *sizep = where - start;
-       return (0);
+STATIC int
+sysdoproc_filt_KERN_PROC_TTY(proc_t p, void * arg)
+{
+       boolean_t funnel_state;
+       int retval;
+       struct tty *tp;
+
+       
+       funnel_state = thread_funnel_set(kernel_flock, TRUE);
+       /* This is very racy but list lock is held.. Hmmm. */
+       if ((p->p_flag & P_CONTROLT) == 0 ||
+               (p->p_pgrp == NULL) || (p->p_pgrp->pg_session == NULL) ||
+                       (tp = SESSION_TP(p->p_pgrp->pg_session)) == TTY_NULL ||
+                       tp->t_dev != (dev_t)*(int*)arg)
+                               retval = 0;
+       else
+               retval = 1;
+
+       thread_funnel_set(kernel_flock, funnel_state);
+
+       return(retval);
+}
+
+STATIC int
+sysdoproc_filt_KERN_PROC_UID(proc_t p, void * arg)
+{
+       kauth_cred_t my_cred;
+       uid_t uid;
+
+       if (p->p_ucred == NULL)
+               return(0);
+       my_cred = kauth_cred_proc_ref(p);
+       uid = kauth_cred_getuid(my_cred);
+       kauth_cred_unref(&my_cred);
+
+       if (uid != (uid_t)*(int*)arg)
+               return(0);
+       else
+               return(1);
+}
+
+
+STATIC int
+sysdoproc_filt_KERN_PROC_RUID(proc_t p, void * arg)
+{
+       kauth_cred_t my_cred;
+       uid_t ruid;
+
+       if (p->p_ucred == NULL)
+               return(0);
+       my_cred = kauth_cred_proc_ref(p);
+       ruid = kauth_cred_getruid(my_cred);
+       kauth_cred_unref(&my_cred);
+
+       if (ruid != (uid_t)*(int*)arg)
+               return(0);
+       else
+               return(1);
+}
+
+#if CONFIG_LCTX
+STATIC int
+sysdoproc_filt_KERN_PROC_LCID(proc_t p, void * arg)
+{
+       if ((p->p_lctx == NULL) ||
+               (p->p_lctx->lc_id != (pid_t)*(int*)arg))
+               return(0);
+       else
+               return(1);
 }
+#endif
 
 /*
  * try over estimating by 5 procs
  */
 #define KERN_PROCSLOP  (5 * sizeof (struct kinfo_proc))
+struct sysdoproc_args {
+       int     buflen;
+       void    *kprocp;
+       boolean_t is_64_bit;
+       user_addr_t     dp;
+       size_t needed;
+       int sizeof_kproc;
+       int *errorp;
+       int uidcheck;
+       int ruidcheck;
+       int ttycheck;
+       int uidval;
+};
 
-sysctl_doproc(name, namelen, where, sizep)
-       int *name;
-       u_int namelen;
-       char *where;
-       size_t *sizep;
-{
-       register struct proc *p;
-       register struct kinfo_proc *dp = (struct kinfo_proc *)where;
-       register int needed = 0;
-       int buflen = where != NULL ? *sizep : 0;
-       int doingzomb;
-       struct kinfo_proc kproc;
-       int error = 0;
+int
+sysdoproc_callback(proc_t p, void *arg)
+{
+       struct sysdoproc_args *args = arg;
+
+       if (args->buflen >= args->sizeof_kproc) {
+               if ((args->ruidcheck != 0) && (sysdoproc_filt_KERN_PROC_RUID(p, &args->uidval) == 0))
+                       return (PROC_RETURNED);
+               if ((args->uidcheck != 0) && (sysdoproc_filt_KERN_PROC_UID(p, &args->uidval) == 0))
+                       return (PROC_RETURNED);
+               if ((args->ttycheck != 0) && (sysdoproc_filt_KERN_PROC_TTY(p, &args->uidval) == 0))
+                       return (PROC_RETURNED);
+
+               bzero(args->kprocp, args->sizeof_kproc);
+               if (args->is_64_bit)
+                       fill_user64_proc(p, args->kprocp);
+               else
+                       fill_user32_proc(p, args->kprocp);
+               int error = copyout(args->kprocp, args->dp, args->sizeof_kproc);
+               if (error) {
+                       *args->errorp = error;
+                       return (PROC_RETURNED_DONE);
+               }
+               args->dp += args->sizeof_kproc;
+               args->buflen -= args->sizeof_kproc;
+       }
+       args->needed += args->sizeof_kproc;
+       return (PROC_RETURNED);
+}
 
-       if (namelen != 2 && !(namelen == 1 && name[0] == KERN_PROC_ALL))
+SYSCTL_NODE(_kern, KERN_PROC, proc, CTLFLAG_RD | CTLFLAG_LOCKED, 0, "");
+STATIC int
+sysctl_prochandle SYSCTL_HANDLER_ARGS
+{
+       int cmd = oidp->oid_arg2;       /* subcommand for multiple nodes */
+       int *name = arg1;               /* oid element argument vector */
+       int namelen = arg2;             /* number of oid element arguments */
+       user_addr_t where = req->oldptr;/* user buffer copy out address */
+
+       user_addr_t dp = where;
+       size_t needed = 0;
+       int buflen = where != USER_ADDR_NULL ? req->oldlen : 0;
+       int error = 0;
+       boolean_t is_64_bit = proc_is64bit(current_proc());
+       struct user32_kinfo_proc  user32_kproc;
+       struct user64_kinfo_proc  user_kproc;
+       int sizeof_kproc;
+       void *kprocp;
+       int (*filterfn)(proc_t, void *) = 0;
+       struct sysdoproc_args args;
+       int uidcheck = 0;
+       int ruidcheck = 0;
+       int ttycheck = 0;
+
+       if (namelen != 1 && !(namelen == 0 && cmd == KERN_PROC_ALL))
                return (EINVAL);
-       p = allproc.lh_first;
-       doingzomb = 0;
-again:
-       for (; p != 0; p = p->p_list.le_next) {
-               /*
-                * Skip embryonic processes.
-                */
-               if (p->p_stat == SIDL)
-                       continue;
-               /*
-                * TODO - make more efficient (see notes below).
-                * do by session.
-                */
-               switch (name[0]) {
+
+       if (is_64_bit) {
+               sizeof_kproc = sizeof(user_kproc);
+               kprocp = &user_kproc;
+       } else {
+               sizeof_kproc = sizeof(user32_kproc);
+               kprocp = &user32_kproc;
+       }
+
+       switch (cmd) {
 
                case KERN_PROC_PID:
-                       /* could do this with just a lookup */
-                       if (p->p_pid != (pid_t)name[1])
-                               continue;
+                       filterfn = sysdoproc_filt_KERN_PROC_PID;
                        break;
 
                case KERN_PROC_PGRP:
-                       /* could do this by traversing pgrp */
-                       if (p->p_pgrp->pg_id != (pid_t)name[1])
-                               continue;
+                       filterfn = sysdoproc_filt_KERN_PROC_PGRP;
                        break;
-
+       
                case KERN_PROC_TTY:
-                       if ( doingzomb || (p->p_flag & P_CONTROLT) == 0 ||
-                           p->p_session->s_ttyp == NULL ||
-                           p->p_session->s_ttyp->t_dev != (dev_t)name[1])
-                               continue;
+                       ttycheck = 1;
                        break;
 
                case KERN_PROC_UID:
-                       if (doingzomb || (p->p_ucred->cr_uid != (uid_t)name[1]))
-                               continue;
+                       uidcheck = 1;
                        break;
 
                case KERN_PROC_RUID:
-                       if ( doingzomb || (p->p_cred->p_ruid != (uid_t)name[1]))
-                               continue;
+                       ruidcheck = 1;
                        break;
-               }
-               if (buflen >= sizeof(struct kinfo_proc)) {
-                       bzero(&kproc, sizeof(struct kinfo_proc));
-                       fill_proc(p, &kproc, doingzomb);
-                       if (error = copyout((caddr_t)&kproc, &dp->kp_proc,
-                           sizeof(struct kinfo_proc)))
-                               return (error);
-                       dp++;
-                       buflen -= sizeof(struct kinfo_proc);
-               }
-               needed += sizeof(struct kinfo_proc);
-       }
-       if (doingzomb == 0) {
-               p = zombproc.lh_first;
-               doingzomb++;
-               goto again;
+
+#if CONFIG_LCTX
+               case KERN_PROC_LCID:
+                       filterfn = sysdoproc_filt_KERN_PROC_LCID;
+                       break;
+#endif
+               case KERN_PROC_ALL:
+                       break;
+
+               default:
+                       /* must be kern.proc.<unknown> */
+                       return (ENOTSUP);
        }
-       if (where != NULL) {
-               *sizep = (caddr_t)dp - where;
-               if (needed > *sizep)
+
+       error = 0;
+       args.buflen = buflen;
+       args.kprocp = kprocp;
+       args.is_64_bit = is_64_bit;
+       args.dp = dp;
+       args.needed = needed;
+       args.errorp = &error;
+       args.uidcheck = uidcheck;
+       args.ruidcheck = ruidcheck;
+       args.ttycheck = ttycheck;
+       args.sizeof_kproc = sizeof_kproc;
+       if (namelen)
+               args.uidval = name[0];
+
+       proc_iterate((PROC_ALLPROCLIST | PROC_ZOMBPROCLIST),
+           sysdoproc_callback, &args, filterfn, name);
+
+       if (error)
+               return (error);
+
+       dp = args.dp;
+       needed = args.needed;
+       
+       if (where != USER_ADDR_NULL) {
+               req->oldlen = dp - where;
+               if (needed > req->oldlen)
                        return (ENOMEM);
        } else {
                needed += KERN_PROCSLOP;
-               *sizep = needed;
+               req->oldlen = needed;
        }
+       /* adjust index so we return the right required/consumed amount */
+       req->oldidx += req->oldlen;
        return (0);
 }
 
-void
-fill_proc(p,kp, doingzomb)
-       register struct proc *p;
-       register struct kinfo_proc *kp;
-       int doingzomb;
+/*
+ * We specify the subcommand code for multiple nodes as the 'req->arg2' value
+ * in the sysctl declaration itself, which comes into the handler function
+ * as 'oidp->oid_arg2'.
+ *
+ * For these particular sysctls, since they have well known OIDs, we could
+ * have just obtained it from the '((int *)arg1)[0]' parameter, but that would
+ * not demonstrate how to handle multiple sysctls that used OID_AUTO instead
+ * of a well known value with a common handler function.  This is desirable,
+ * because we want well known values to "go away" at some future date.
+ *
+ * It should be noted that the value of '((int *)arg1)[1]' is used for many
+ * an integer parameter to the subcommand for many of these sysctls; we'd
+ * rather have used '((int *)arg1)[0]' for that, or even better, an element
+ * in a structure passed in as the the 'newp' argument to sysctlbyname(3),
+ * and then use leaf-node permissions enforcement, but that would have
+ * necessitated modifying user space code to correspond to the interface
+ * change, and we are striving for binary backward compatibility here; even
+ * though these are SPI, and not intended for use by user space applications
+ * which are not themselves system tools or libraries, some applications
+ * have erroneously used them.
+ */
+SYSCTL_PROC(_kern_proc, KERN_PROC_ALL, all, CTLTYPE_NODE|CTLFLAG_RD | CTLFLAG_LOCKED,
+       0,                      /* Pointer argument (arg1) */
+       KERN_PROC_ALL,          /* Integer argument (arg2) */
+       sysctl_prochandle,      /* Handler function */
+       NULL,                   /* Data is size variant on ILP32/LP64 */
+       "");
+SYSCTL_PROC(_kern_proc, KERN_PROC_PID, pid, CTLTYPE_NODE|CTLFLAG_RD | CTLFLAG_LOCKED,
+       0,                      /* Pointer argument (arg1) */
+       KERN_PROC_PID,          /* Integer argument (arg2) */
+       sysctl_prochandle,      /* Handler function */
+       NULL,                   /* Data is size variant on ILP32/LP64 */
+       "");
+SYSCTL_PROC(_kern_proc, KERN_PROC_TTY, tty, CTLTYPE_NODE|CTLFLAG_RD | CTLFLAG_LOCKED,
+       0,                      /* Pointer argument (arg1) */
+       KERN_PROC_TTY,          /* Integer argument (arg2) */
+       sysctl_prochandle,      /* Handler function */
+       NULL,                   /* Data is size variant on ILP32/LP64 */
+       "");
+SYSCTL_PROC(_kern_proc, KERN_PROC_PGRP, pgrp, CTLTYPE_NODE|CTLFLAG_RD | CTLFLAG_LOCKED,
+       0,                      /* Pointer argument (arg1) */
+       KERN_PROC_PGRP,         /* Integer argument (arg2) */
+       sysctl_prochandle,      /* Handler function */
+       NULL,                   /* Data is size variant on ILP32/LP64 */
+       "");
+SYSCTL_PROC(_kern_proc, KERN_PROC_UID, uid, CTLTYPE_NODE|CTLFLAG_RD | CTLFLAG_LOCKED,
+       0,                      /* Pointer argument (arg1) */
+       KERN_PROC_UID,          /* Integer argument (arg2) */
+       sysctl_prochandle,      /* Handler function */
+       NULL,                   /* Data is size variant on ILP32/LP64 */
+       "");
+SYSCTL_PROC(_kern_proc, KERN_PROC_RUID, ruid, CTLTYPE_NODE|CTLFLAG_RD | CTLFLAG_LOCKED,
+       0,                      /* Pointer argument (arg1) */
+       KERN_PROC_RUID,         /* Integer argument (arg2) */
+       sysctl_prochandle,      /* Handler function */
+       NULL,                   /* Data is size variant on ILP32/LP64 */
+       "");
+SYSCTL_PROC(_kern_proc, KERN_PROC_LCID, lcid, CTLTYPE_NODE|CTLFLAG_RD | CTLFLAG_LOCKED,
+       0,                      /* Pointer argument (arg1) */
+       KERN_PROC_LCID,         /* Integer argument (arg2) */
+       sysctl_prochandle,      /* Handler function */
+       NULL,                   /* Data is size variant on ILP32/LP64 */
+       "");
+
+
+/*
+ * Fill in non-zero fields of an eproc structure for the specified process.
+ */
+STATIC void
+fill_user32_eproc(proc_t p, struct user32_eproc *__restrict ep)
 {
-       fill_externproc(p, &kp->kp_proc);
-       if (!doingzomb)
-               fill_eproc(p, &kp->kp_eproc);
+       struct tty *tp;
+       struct pgrp *pg;
+       struct session *sessp;
+       kauth_cred_t my_cred;
+
+       pg = proc_pgrp(p);
+       sessp = proc_session(p);
+
+       if (pg != PGRP_NULL) {
+               ep->e_pgid = p->p_pgrpid;
+               ep->e_jobc = pg->pg_jobc;
+               if (sessp != SESSION_NULL && sessp->s_ttyvp)
+                       ep->e_flag = EPROC_CTTY;
+       }
+#if CONFIG_LCTX
+       if (p->p_lctx)
+               ep->e_lcid = p->p_lctx->lc_id;
+#endif
+       ep->e_ppid = p->p_ppid;
+       if (p->p_ucred) {
+               my_cred = kauth_cred_proc_ref(p);
+
+               /* A fake historical pcred */
+               ep->e_pcred.p_ruid = kauth_cred_getruid(my_cred);
+               ep->e_pcred.p_svuid = kauth_cred_getsvuid(my_cred);
+               ep->e_pcred.p_rgid = kauth_cred_getrgid(my_cred);
+               ep->e_pcred.p_svgid = kauth_cred_getsvgid(my_cred);
+
+               /* A fake historical *kauth_cred_t */
+               ep->e_ucred.cr_ref = my_cred->cr_ref;
+               ep->e_ucred.cr_uid = kauth_cred_getuid(my_cred);
+               ep->e_ucred.cr_ngroups = posix_cred_get(my_cred)->cr_ngroups;
+               bcopy(posix_cred_get(my_cred)->cr_groups,
+                       ep->e_ucred.cr_groups, NGROUPS * sizeof (gid_t));
+
+               kauth_cred_unref(&my_cred);
+       }
+
+       if ((p->p_flag & P_CONTROLT) && (sessp != SESSION_NULL) &&
+            (tp = SESSION_TP(sessp))) {
+               ep->e_tdev = tp->t_dev;
+               ep->e_tpgid = sessp->s_ttypgrpid;
+       } else
+               ep->e_tdev = NODEV;
+
+       if (sessp != SESSION_NULL) {
+               if (SESS_LEADER(p, sessp))
+                       ep->e_flag |= EPROC_SLEADER;
+               session_rele(sessp);
+       }
+       if (pg != PGRP_NULL)
+               pg_rele(pg);
 }
+
 /*
- * Fill in an eproc structure for the specified process.
+ * Fill in non-zero fields of an LP64 eproc structure for the specified process.
  */
-void
-fill_eproc(p, ep)
-       register struct proc *p;
-       register struct eproc *ep;
+STATIC void
+fill_user64_eproc(proc_t p, struct user64_eproc *__restrict ep)
 {
-       register struct tty *tp;
+       struct tty *tp;
+       struct pgrp *pg;
+       struct session *sessp;
+       kauth_cred_t my_cred;
+       
+       pg = proc_pgrp(p);
+       sessp = proc_session(p);
+
+       if (pg != PGRP_NULL) {
+               ep->e_pgid = p->p_pgrpid;
+               ep->e_jobc = pg->pg_jobc;
+               if (sessp != SESSION_NULL && sessp->s_ttyvp)
+                       ep->e_flag = EPROC_CTTY;
+       }
+#if CONFIG_LCTX
+       if (p->p_lctx)
+               ep->e_lcid = p->p_lctx->lc_id;
+#endif
+       ep->e_ppid = p->p_ppid;
+       if (p->p_ucred) {
+               my_cred = kauth_cred_proc_ref(p);
+
+               /* A fake historical pcred */
+               ep->e_pcred.p_ruid = kauth_cred_getruid(my_cred);
+               ep->e_pcred.p_svuid = kauth_cred_getsvuid(my_cred);
+               ep->e_pcred.p_rgid = kauth_cred_getrgid(my_cred);
+               ep->e_pcred.p_svgid = kauth_cred_getsvgid(my_cred);
+
+               /* A fake historical *kauth_cred_t */
+               ep->e_ucred.cr_ref = my_cred->cr_ref;
+               ep->e_ucred.cr_uid = kauth_cred_getuid(my_cred);
+               ep->e_ucred.cr_ngroups = posix_cred_get(my_cred)->cr_ngroups;
+               bcopy(posix_cred_get(my_cred)->cr_groups,
+                       ep->e_ucred.cr_groups, NGROUPS * sizeof (gid_t));
+
+               kauth_cred_unref(&my_cred);
+       }
 
-       /*
-        * Skip zombie processes.
-        */
-       if (p->p_stat == SZOMB)
-               return;
-
-       ep->e_paddr = p;
-       ep->e_sess = p->p_pgrp->pg_session;
-       ep->e_pcred = *p->p_cred;
-       ep->e_ucred = *p->p_ucred;
-       if (p->p_stat == SIDL || p->p_stat == SZOMB) {
-               ep->e_vm.vm_rssize = 0;
-               ep->e_vm.vm_tsize = 0;
-               ep->e_vm.vm_dsize = 0;
-               ep->e_vm.vm_ssize = 0;
-               /* ep->e_vm.vm_pmap = XXX; */
-       } else {
-#if FIXME  /* [ */
-               register vm_map_t vm = ((task_t)p->task)->map;
-
-               ep->e_vm.vm_rssize = pmap_resident_count(vm->pmap); /*XXX*/
-//             ep->e_vm.vm_tsize = vm->vm_tsize;
-//             ep->e_vm.vm_dsize = vm->vm_dsize;
-//             ep->e_vm.vm_ssize = vm->vm_ssize;
-#else  /* FIXME ][ */
-               ep->e_vm.vm_rssize = 0; /*XXX*/
-#endif  /* FIXME ] */
-       }
-       if (p->p_pptr)
-               ep->e_ppid = p->p_pptr->p_pid;
-       else
-               ep->e_ppid = 0;
-       ep->e_pgid = p->p_pgrp->pg_id;
-       ep->e_jobc = p->p_pgrp->pg_jobc;
-       if ((p->p_flag & P_CONTROLT) &&
-            (tp = ep->e_sess->s_ttyp)) {
+       if ((p->p_flag & P_CONTROLT) && (sessp != SESSION_NULL) &&
+            (tp = SESSION_TP(sessp))) {
                ep->e_tdev = tp->t_dev;
-               ep->e_tpgid = tp->t_pgrp ? tp->t_pgrp->pg_id : NO_PID;
-               ep->e_tsess = tp->t_session;
+               ep->e_tpgid = sessp->s_ttypgrpid;
        } else
                ep->e_tdev = NODEV;
-       ep->e_flag = ep->e_sess->s_ttyvp ? EPROC_CTTY : 0;
-       if (SESS_LEADER(p))
-               ep->e_flag |= EPROC_SLEADER;
-       if (p->p_wmesg)
-               strncpy(ep->e_wmesg, p->p_wmesg, WMESGLEN);
-       ep->e_xsize = ep->e_xrssize = 0;
-       ep->e_xccount = ep->e_xswrss = 0;
+
+       if (sessp != SESSION_NULL) {
+               if (SESS_LEADER(p, sessp))
+                       ep->e_flag |= EPROC_SLEADER;
+               session_rele(sessp);
+       }
+       if (pg != PGRP_NULL)
+               pg_rele(pg);
 }
+
 /*
  * Fill in an eproc structure for the specified process.
+ * bzeroed by our caller, so only set non-zero fields.
+ */
+STATIC void
+fill_user32_externproc(proc_t p, struct user32_extern_proc *__restrict exp)
+{
+       exp->p_starttime.tv_sec = p->p_start.tv_sec;
+       exp->p_starttime.tv_usec = p->p_start.tv_usec;
+       exp->p_flag = p->p_flag;
+       if (p->p_lflag & P_LTRACED)
+               exp->p_flag |= P_TRACED;
+       if (p->p_lflag & P_LPPWAIT)
+               exp->p_flag |= P_PPWAIT;
+       if (p->p_lflag & P_LEXIT)
+               exp->p_flag |= P_WEXIT;
+       exp->p_stat = p->p_stat;
+       exp->p_pid = p->p_pid;
+       exp->p_oppid = p->p_oppid;
+       /* Mach related  */
+       exp->user_stack = p->user_stack;
+       exp->p_debugger = p->p_debugger;
+       exp->sigwait = p->sigwait;
+       /* scheduling */
+#ifdef _PROC_HAS_SCHEDINFO_
+       exp->p_estcpu = p->p_estcpu;
+       exp->p_pctcpu = p->p_pctcpu;
+       exp->p_slptime = p->p_slptime;
+#endif
+       exp->p_realtimer.it_interval.tv_sec =
+               (user32_time_t)p->p_realtimer.it_interval.tv_sec;
+       exp->p_realtimer.it_interval.tv_usec =
+               (__int32_t)p->p_realtimer.it_interval.tv_usec;
+
+       exp->p_realtimer.it_value.tv_sec =
+               (user32_time_t)p->p_realtimer.it_value.tv_sec;
+       exp->p_realtimer.it_value.tv_usec =
+               (__int32_t)p->p_realtimer.it_value.tv_usec;
+
+       exp->p_rtime.tv_sec = (user32_time_t)p->p_rtime.tv_sec;
+       exp->p_rtime.tv_usec = (__int32_t)p->p_rtime.tv_usec;
+
+       exp->p_sigignore = p->p_sigignore;
+       exp->p_sigcatch = p->p_sigcatch;
+       exp->p_priority = p->p_priority;
+       exp->p_nice = p->p_nice;
+       bcopy(&p->p_comm, &exp->p_comm, MAXCOMLEN);
+       exp->p_xstat = p->p_xstat;
+       exp->p_acflag = p->p_acflag;
+}
+
+/*
+ * Fill in an LP64 version of extern_proc structure for the specified process.
  */
-void
-fill_externproc(p, exp)
-       register struct proc *p;
-       register struct extern_proc *exp;
-{
-       exp->p_forw = exp->p_back = NULL;
-       exp->p_vmspace = NULL;
-       exp->p_sigacts = p->p_sigacts;
-       exp->p_flag  = p->p_flag;
-       exp->p_stat  = p->p_stat ;
-       exp->p_pid  = p->p_pid ;
-       exp->p_oppid  = p->p_oppid ;
-       exp->p_dupfd  = p->p_dupfd ;
+STATIC void
+fill_user64_externproc(proc_t p, struct user64_extern_proc *__restrict exp)
+{
+       exp->p_starttime.tv_sec = p->p_start.tv_sec;
+       exp->p_starttime.tv_usec = p->p_start.tv_usec;
+       exp->p_flag = p->p_flag;
+       if (p->p_lflag & P_LTRACED)
+               exp->p_flag |= P_TRACED;
+       if (p->p_lflag & P_LPPWAIT)
+               exp->p_flag |= P_PPWAIT;
+       if (p->p_lflag & P_LEXIT)
+               exp->p_flag |= P_WEXIT;
+       exp->p_stat = p->p_stat;
+       exp->p_pid = p->p_pid;
+       exp->p_oppid = p->p_oppid;
        /* Mach related  */
-       exp->user_stack  = p->user_stack ;
-       exp->exit_thread  = p->exit_thread ;
-       exp->p_debugger  = p->p_debugger ;
-       exp->sigwait  = p->sigwait ;
+       exp->user_stack = p->user_stack;
+       exp->p_debugger = p->p_debugger;
+       exp->sigwait = p->sigwait;
        /* scheduling */
-       exp->p_estcpu  = p->p_estcpu ;
-       exp->p_cpticks  = p->p_cpticks ;
-       exp->p_pctcpu  = p->p_pctcpu ;
-       exp->p_wchan  = p->p_wchan ;
-       exp->p_wmesg  = p->p_wmesg ;
-       exp->p_swtime  = p->p_swtime ;
-       exp->p_slptime  = p->p_slptime ;
-       bcopy(&p->p_realtimer, &exp->p_realtimer,sizeof(struct itimerval));
-       bcopy(&p->p_rtime, &exp->p_rtime,sizeof(struct timeval));
-       exp->p_uticks  = p->p_uticks ;
-       exp->p_sticks  = p->p_sticks ;
-       exp->p_iticks  = p->p_iticks ;
-       exp->p_traceflag  = p->p_traceflag ;
-       exp->p_tracep  = p->p_tracep ;
-       exp->p_siglist  = p->p_siglist ;
-       exp->p_textvp  = p->p_textvp ;
-       exp->p_holdcnt = 0 ;
-       exp->p_sigmask  = p->p_sigmask ;
-       exp->p_sigignore  = p->p_sigignore ;
-       exp->p_sigcatch  = p->p_sigcatch ;
-       exp->p_priority  = p->p_priority ;
-       exp->p_usrpri  = p->p_usrpri ;
-       exp->p_nice  = p->p_nice ;
-       bcopy(&p->p_comm, &exp->p_comm,MAXCOMLEN);
-       exp->p_comm[MAXCOMLEN] = '\0';
-       exp->p_pgrp  = p->p_pgrp ;
-       exp->p_addr  = NULL;
-       exp->p_xstat  = p->p_xstat ;
-       exp->p_acflag  = p->p_acflag ;
-       exp->p_ru  = p->p_ru ;
-}
-
-kdebug_ops(name, namelen, where, sizep, p)
-int *name;
-u_int namelen;
-char *where;
-size_t *sizep;
-struct proc *p;
-{
-int size=*sizep;
-int ret=0;
-extern int kdbg_control(int *name, u_int namelen, char * where,size_t * sizep);
-
-        if (ret = suser(p->p_ucred, &p->p_acflag))
-               return(ret);
+#ifdef _PROC_HAS_SCHEDINFO_
+       exp->p_estcpu = p->p_estcpu;
+       exp->p_pctcpu = p->p_pctcpu;
+       exp->p_slptime = p->p_slptime;
+#endif
+       exp->p_realtimer.it_interval.tv_sec = p->p_realtimer.it_interval.tv_sec;
+       exp->p_realtimer.it_interval.tv_usec = p->p_realtimer.it_interval.tv_usec;
+
+       exp->p_realtimer.it_value.tv_sec = p->p_realtimer.it_value.tv_sec;
+       exp->p_realtimer.it_value.tv_usec = p->p_realtimer.it_value.tv_usec;
+
+       exp->p_rtime.tv_sec = p->p_rtime.tv_sec;
+       exp->p_rtime.tv_usec = p->p_rtime.tv_usec;
+
+       exp->p_sigignore = p->p_sigignore;
+       exp->p_sigcatch = p->p_sigcatch;
+       exp->p_priority = p->p_priority;
+       exp->p_nice = p->p_nice;
+       bcopy(&p->p_comm, &exp->p_comm, MAXCOMLEN);
+       exp->p_xstat = p->p_xstat;
+       exp->p_acflag = p->p_acflag;
+}
+
+STATIC void
+fill_user32_proc(proc_t p, struct user32_kinfo_proc *__restrict kp)
+{
+       /* on a 64 bit kernel, 32 bit users get some truncated information */
+       fill_user32_externproc(p, &kp->kp_proc);
+       fill_user32_eproc(p, &kp->kp_eproc);
+}
+
+STATIC void
+fill_user64_proc(proc_t p, struct user64_kinfo_proc *__restrict kp)
+{
+       fill_user64_externproc(p, &kp->kp_proc);
+       fill_user64_eproc(p, &kp->kp_eproc);
+}
 
+STATIC int
+sysctl_kdebug_ops SYSCTL_HANDLER_ARGS
+{
+       __unused int cmd = oidp->oid_arg2;      /* subcommand*/
+       int *name = arg1;               /* oid element argument vector */
+       int namelen = arg2;             /* number of oid element arguments */
+       user_addr_t oldp = req->oldptr; /* user buffer copy out address */
+       size_t *oldlenp = &req->oldlen; /* user buffer copy out size */
+//     user_addr_t newp = req->newptr; /* user buffer copy in address */
+//     size_t newlen = req->newlen;    /* user buffer copy in size */
+
+       proc_t p = current_proc();
+       int ret=0;
+
+       if (namelen == 0)
+               return(ENOTSUP);
+       
+       ret = suser(kauth_cred_get(), &p->p_acflag);
+       if (ret)
+               return(ret);
+       
        switch(name[0]) {
        case KERN_KDEFLAGS:
        case KERN_KDDFLAGS:
@@ -1010,83 +1750,131 @@ extern int kdbg_control(int *name, u_int namelen, char * where,size_t * sizep);
        case KERN_KDSETREG:
        case KERN_KDGETREG:
        case KERN_KDREADTR:
+        case KERN_KDWRITETR:
+        case KERN_KDWRITEMAP:
        case KERN_KDPIDTR:
        case KERN_KDTHRMAP:
        case KERN_KDPIDEX:
        case KERN_KDSETRTCDEC:
        case KERN_KDSETBUF:
-               ret = kdbg_control(name, namelen, where, sizep);
+       case KERN_KDGETENTROPY:
+       case KERN_KDENABLE_BG_TRACE:
+       case KERN_KDDISABLE_BG_TRACE:
+       case KERN_KDSET_TYPEFILTER:
+
+               ret = kdbg_control(name, namelen, oldp, oldlenp);
                break;
        default:
-               ret= EOPNOTSUPP;
+               ret= ENOTSUP;
                break;
        }
-       return(ret);
+
+       /* adjust index so we return the right required/consumed amount */
+       if (!ret)
+               req->oldidx += req->oldlen;
+
+       return (ret);
 }
+SYSCTL_PROC(_kern, KERN_KDEBUG, kdebug, CTLTYPE_NODE|CTLFLAG_RD | CTLFLAG_LOCKED,
+       0,                      /* Pointer argument (arg1) */
+       0,                      /* Integer argument (arg2) */
+       sysctl_kdebug_ops,      /* Handler function */
+       NULL,                   /* Data pointer */
+       "");
 
-pcsamples_ops(name, namelen, where, sizep, p)
-int *name;
-u_int namelen;
-char *where;
-size_t *sizep;
-struct proc *p;
+
+#if !CONFIG_EMBEDDED
+/*
+ * Return the top *sizep bytes of the user stack, or the entire area of the
+ * user stack down through the saved exec_path, whichever is smaller.
+ */
+STATIC int
+sysctl_doprocargs SYSCTL_HANDLER_ARGS
 {
-int ret=0;
-extern int pcsamples_control(int *name, u_int namelen, char * where,size_t * sizep);
+       __unused int cmd = oidp->oid_arg2;      /* subcommand*/
+       int *name = arg1;               /* oid element argument vector */
+       int namelen = arg2;             /* number of oid element arguments */
+       user_addr_t oldp = req->oldptr; /* user buffer copy out address */
+       size_t *oldlenp = &req->oldlen; /* user buffer copy out size */
+//     user_addr_t newp = req->newptr; /* user buffer copy in address */
+//     size_t newlen = req->newlen;    /* user buffer copy in size */
+       int error;
 
-        if (ret = suser(p->p_ucred, &p->p_acflag))
-               return(ret);
+       error =  sysctl_procargsx( name, namelen, oldp, oldlenp, current_proc(), 0);
 
-       switch(name[0]) {
-       case KERN_PCDISABLE:
-       case KERN_PCGETBUF:
-       case KERN_PCSETUP:
-       case KERN_PCREMOVE:
-       case KERN_PCREADBUF:
-       case KERN_PCSETREG:
-       case KERN_PCSETBUF:
-       case KERN_PCCOMM:
-               ret = pcsamples_control(name, namelen, where, sizep);
-               break;
-       default:
-               ret= EOPNOTSUPP;
-               break;
-       }
-       return(ret);
+       /* adjust index so we return the right required/consumed amount */
+       if (!error)
+               req->oldidx += req->oldlen;
+
+       return (error);
 }
+SYSCTL_PROC(_kern, KERN_PROCARGS, procargs, CTLTYPE_NODE|CTLFLAG_RD | CTLFLAG_LOCKED,
+       0,                      /* Pointer argument (arg1) */
+       0,                      /* Integer argument (arg2) */
+       sysctl_doprocargs,      /* Handler function */
+       NULL,                   /* Data pointer */
+       "");
+#endif /* !CONFIG_EMBEDDED */
+
+STATIC int
+sysctl_doprocargs2 SYSCTL_HANDLER_ARGS
+{
+       __unused int cmd = oidp->oid_arg2;      /* subcommand*/
+       int *name = arg1;               /* oid element argument vector */
+       int namelen = arg2;             /* number of oid element arguments */
+       user_addr_t oldp = req->oldptr; /* user buffer copy out address */
+       size_t *oldlenp = &req->oldlen; /* user buffer copy out size */
+//     user_addr_t newp = req->newptr; /* user buffer copy in address */
+//     size_t newlen = req->newlen;    /* user buffer copy in size */
+       int error;
 
-/*
- *     Returns the top N bytes of the user stack, with
- *     everything below the first argument character
- *     zeroed for security reasons.
- *     Odd data structure is for compatibility.
- */
-sysctl_procargs(name, namelen, where, sizep)
-       int *name;
-       u_int namelen;
-       char *where;
-       size_t *sizep;
-{
-       register struct proc *p;
-       register int needed = 0;
-       int buflen = where != NULL ? *sizep : 0;
+       error = sysctl_procargsx( name, namelen, oldp, oldlenp, current_proc(), 1);
+
+       /* adjust index so we return the right required/consumed amount */
+       if (!error)
+               req->oldidx += req->oldlen;
+
+       return (error);
+}
+SYSCTL_PROC(_kern, KERN_PROCARGS2, procargs2, CTLTYPE_NODE|CTLFLAG_RD | CTLFLAG_LOCKED,
+       0,                      /* Pointer argument (arg1) */
+       0,                      /* Integer argument (arg2) */
+       sysctl_doprocargs2,     /* Handler function */
+       NULL,                   /* Data pointer */
+       "");
+
+STATIC int
+sysctl_procargsx(int *name, u_int namelen, user_addr_t where, 
+                 size_t *sizep, proc_t cur_proc, int argc_yes)
+{
+       proc_t p;
+       int buflen = where != USER_ADDR_NULL ? *sizep : 0;
        int error = 0;
-       struct vm_map *proc_map;
+       struct _vm_map *proc_map;
        struct task * task;
        vm_map_copy_t   tmp;
-       vm_offset_t     arg_addr;
-       vm_size_t       arg_size;
+       user_addr_t     arg_addr;
+       size_t          arg_size;
        caddr_t data;
-       unsigned size;
+       size_t argslen=0;
+       int size;
        vm_offset_t     copy_start, copy_end;
-       vm_offset_t     dealloc_start;  /* area to remove from kernel map */
-       vm_offset_t     dealloc_end;
-       int             *ip;
        kern_return_t ret;
        int pid;
+       kauth_cred_t my_cred;
+       uid_t uid;
 
-
-       if ((buflen <= 0) || (buflen > (PAGE_SIZE << 1))) {
+       if ( namelen < 1 )
+               return(EINVAL);
+       
+       if (argc_yes)
+               buflen -= sizeof(int);          /* reserve first word to return argc */
+
+       /* we only care about buflen when where (oldp from sysctl) is not NULL. */
+       /* when where (oldp from sysctl) is NULL and sizep (oldlenp from sysctl */
+       /* is not NULL then the caller wants us to return the length needed to */
+       /* hold the data we would return */ 
+       if (where != USER_ADDR_NULL && (buflen <= 0 || buflen > ARG_MAX)) {
                return(EINVAL);
        }
        arg_size = buflen;
@@ -1095,8 +1883,7 @@ sysctl_procargs(name, namelen, where, sizep)
         *      Lookup process by pid
         */
        pid = name[0];
-
-       p = pfind(pid);
+       p = proc_find(pid);
        if (p == NULL) {
                return(EINVAL);
        }
@@ -1111,10 +1898,49 @@ sysctl_procargs(name, namelen, where, sizep)
         *      size.
         */
 
-       if (!p->user_stack)
+       if (!p->user_stack) {
+               proc_rele(p);
                return(EINVAL);
+       }
+
+       if (where == USER_ADDR_NULL) {
+               /* caller only wants to know length of proc args data */
+               if (sizep == NULL) {
+                       proc_rele(p);
+                       return(EFAULT);
+               }
+                       
+                size = p->p_argslen;
+               proc_rele(p);
+                if (argc_yes) {
+                       size += sizeof(int);
+                }
+                else {
+                       /*
+                        * old PROCARGS will return the executable's path and plus some
+                        * extra space for work alignment and data tags
+                        */
+                       size += PATH_MAX + (6 * sizeof(int));
+                }
+               size += (size & (sizeof(int) - 1)) ? (sizeof(int) - (size & (sizeof(int) - 1))) : 0;
+               *sizep = size;
+               return (0);
+       }
+       
+       my_cred = kauth_cred_proc_ref(p);
+       uid = kauth_cred_getuid(my_cred);
+       kauth_cred_unref(&my_cred);
+
+       if ((uid != kauth_cred_getuid(kauth_cred_get())) 
+               && suser(kauth_cred_get(), &cur_proc->p_acflag)) {
+               proc_rele(p);
+               return (EINVAL);
+       }
 
-       arg_addr = (vm_offset_t)(p->user_stack - arg_size);
+       if ((u_int)arg_size > p->p_argslen)
+               arg_size = round_page(p->p_argslen);
+
+       arg_addr = p->user_stack - arg_size;
 
 
        /*
@@ -1123,23 +1949,39 @@ sysctl_procargs(name, namelen, where, sizep)
         *      that by getting a reference on the task itself.
         */
        task = p->task;
-       if (task == NULL)
+       if (task == NULL) {
+               proc_rele(p);
                return(EINVAL);
+       }
        
+       argslen = p->p_argslen;
+       /*
+        * Once we have a task reference we can convert that into a
+        * map reference, which we will use in the calls below.  The
+        * task/process may change its map after we take this reference
+        * (see execve), but the worst that will happen then is a return
+        * of stale info (which is always a possibility).
+        */
        task_reference(task);
+       proc_rele(p);
+       proc_map = get_task_map_reference(task);
+       task_deallocate(task);
+       
+       if (proc_map == NULL)
+               return(EINVAL);
+
 
        ret = kmem_alloc(kernel_map, &copy_start, round_page(arg_size));
        if (ret != KERN_SUCCESS) {
-               task_deallocate(task);
+               vm_map_deallocate(proc_map);
                return(ENOMEM);
        }
 
-       proc_map = get_task_map(task);
        copy_end = round_page(copy_start + arg_size);
 
-       if( vm_map_copyin(proc_map, trunc_page(arg_addr), round_page(arg_size)
-                       FALSE, &tmp) != KERN_SUCCESS) {
-                       task_deallocate(task);
+       if( vm_map_copyin(proc_map, (vm_map_address_t)arg_addr
+                         (vm_map_size_t)arg_size, FALSE, &tmp) != KERN_SUCCESS) {
+                       vm_map_deallocate(proc_map);
                        kmem_free(kernel_map, copy_start,
                                        round_page(arg_size));
                        return (EIO);
@@ -1149,72 +1991,1248 @@ sysctl_procargs(name, namelen, where, sizep)
         *      Now that we've done the copyin from the process'
         *      map, we can release the reference to it.
         */
-       task_deallocate(task);
+       vm_map_deallocate(proc_map);
 
-       if( vm_map_copy_overwrite(kernel_map, copy_start, 
-               tmp, FALSE) != KERN_SUCCESS) {
+       if( vm_map_copy_overwrite(kernel_map, 
+                                 (vm_map_address_t)copy_start, 
+                                 tmp, FALSE) != KERN_SUCCESS) {
                        kmem_free(kernel_map, copy_start,
                                        round_page(arg_size));
                        return (EIO);
        }
 
-       data = (caddr_t) (copy_end - arg_size);
-       ip = (int *) copy_end;          
-       size = arg_size;
-
-       /*
-        *      Now look down the stack for the bottom of the
-        *      argument list.  Since this call is otherwise
-        *      unprotected, we can't let the nosy user see
-        *      anything else on the stack.
-        *
-        *      The arguments are pushed on the stack by
-        *      execve() as:
-        *
-        *              .long   0
-        *              arg 0   (null-terminated)
-        *              arg 1
-        *              ...
-        *              arg N
-        *              .long   0
-        *
-        */
-
-       ip -= 2; /*skip trailing 0 word and assume at least one
-                 argument.  The last word of argN may be just
-                 the trailing 0, in which case we'd stop
-                 there */
-       while (*--ip)
-               if (ip == (int *)data)
-                       break;
-        /* 
-         *  To account for saved path name and not having a null after that
-         *  Run the sweep again. If we have already sweeped entire range skip this
-         */
-         if (ip != (int *)data) {
-                while (*--ip)
-                    if (ip == (int *)data)
-                            break;
-        }
-        
-       bzero(data, (unsigned) ((int)ip - (int)data));
-
-       dealloc_start = copy_start;
-       dealloc_end = copy_end;
+       if (arg_size > argslen) {
+               data = (caddr_t) (copy_end - argslen);
+               size = argslen;
+       } else {
+               data = (caddr_t) (copy_end - arg_size);
+               size = arg_size;
+       }
 
+       if (argc_yes) {
+               /* Put processes argc as the first word in the copyout buffer */
+               suword(where, p->p_argc);
+               error = copyout(data, (where + sizeof(int)), size);
+               size += sizeof(int);
+       } else {
+               error = copyout(data, where, size);
 
-       size = MIN(size, buflen);
-       error = copyout(data, where, size);
+               /*
+                * Make the old PROCARGS work to return the executable's path
+                * But, only if there is enough space in the provided buffer
+                *
+                * on entry: data [possibily] points to the beginning of the path 
+                * 
+                * Note: we keep all pointers&sizes aligned to word boundries
+                */
+               if ( (! error) && (buflen > 0 && (u_int)buflen > argslen) )
+               {
+                       int binPath_sz, alignedBinPath_sz = 0;
+                       int extraSpaceNeeded, addThis;
+                       user_addr_t placeHere;
+                       char * str = (char *) data;
+                       int max_len = size;
+
+                       /* Some apps are really bad about messing up their stacks
+                          So, we have to be extra careful about getting the length
+                          of the executing binary.  If we encounter an error, we bail.
+                       */
+
+                       /* Limit ourselves to PATH_MAX paths */
+                       if ( max_len > PATH_MAX ) max_len = PATH_MAX;
+
+                       binPath_sz = 0;
+
+                       while ( (binPath_sz < max_len-1) && (*str++ != 0) )
+                               binPath_sz++;
+
+                       /* If we have a NUL terminator, copy it, too */
+                       if (binPath_sz < max_len-1) binPath_sz += 1;
+
+                       /* Pre-Flight the space requiremnts */
+
+                       /* Account for the padding that fills out binPath to the next word */
+                       alignedBinPath_sz += (binPath_sz & (sizeof(int)-1)) ? (sizeof(int)-(binPath_sz & (sizeof(int)-1))) : 0;
+
+                       placeHere = where + size;
+
+                       /* Account for the bytes needed to keep placeHere word aligned */ 
+                       addThis = (placeHere & (sizeof(int)-1)) ? (sizeof(int)-(placeHere & (sizeof(int)-1))) : 0;
+
+                       /* Add up all the space that is needed */
+                       extraSpaceNeeded = alignedBinPath_sz + addThis + binPath_sz + (4 * sizeof(int));
+
+                       /* is there is room to tack on argv[0]? */
+                       if ( (buflen & ~(sizeof(int)-1)) >= ( argslen + extraSpaceNeeded ))
+                       {
+                               placeHere += addThis;
+                               suword(placeHere, 0);
+                               placeHere += sizeof(int);
+                               suword(placeHere, 0xBFFF0000);
+                               placeHere += sizeof(int);
+                               suword(placeHere, 0);
+                               placeHere += sizeof(int);
+                               error = copyout(data, placeHere, binPath_sz);
+                               if ( ! error )
+                               {
+                                       placeHere += binPath_sz;
+                                       suword(placeHere, 0);
+                                       size += extraSpaceNeeded;
+                               }
+                       }
+               }
+       }
 
-       if (dealloc_start != (vm_offset_t) 0) {
-               kmem_free(kernel_map, dealloc_start,
-                       dealloc_end - dealloc_start);
+       if (copy_start != (vm_offset_t) 0) {
+               kmem_free(kernel_map, copy_start, copy_end - copy_start);
        }
        if (error) {
                return(error);
        }
 
-       if (where != NULL)
+       if (where != USER_ADDR_NULL)
                *sizep = size;
        return (0);
 }
+
+
+/*
+ * Max number of concurrent aio requests
+ */
+STATIC int
+sysctl_aiomax
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       int new_value, changed;
+       int error = sysctl_io_number(req, aio_max_requests, sizeof(int), &new_value, &changed);
+       if (changed) {
+                /* make sure the system-wide limit is greater than the per process limit */
+               if (new_value >= aio_max_requests_per_process && new_value <= AIO_MAX_REQUESTS)
+                       aio_max_requests = new_value;
+               else
+                       error = EINVAL;
+       }
+       return(error);
+}
+
+
+/*
+ * Max number of concurrent aio requests per process
+ */
+STATIC int
+sysctl_aioprocmax
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       int new_value, changed;
+       int error = sysctl_io_number(req, aio_max_requests_per_process, sizeof(int), &new_value, &changed);
+       if (changed) {
+               /* make sure per process limit is less than the system-wide limit */
+               if (new_value <= aio_max_requests && new_value >= AIO_LISTIO_MAX)
+                       aio_max_requests_per_process = new_value;
+               else
+                       error = EINVAL;
+       }
+       return(error);
+}
+
+
+/*
+ * Max number of async IO worker threads
+ */
+STATIC int
+sysctl_aiothreads
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       int new_value, changed;
+       int error = sysctl_io_number(req, aio_worker_threads, sizeof(int), &new_value, &changed);
+       if (changed) {
+               /* we only allow an increase in the number of worker threads */
+               if (new_value > aio_worker_threads ) {
+                       _aio_create_worker_threads((new_value - aio_worker_threads));
+                       aio_worker_threads = new_value;
+               }
+               else
+                       error = EINVAL;
+       }
+       return(error);
+}
+
+
+/*
+ * System-wide limit on the max number of processes
+ */
+STATIC int
+sysctl_maxproc
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       int new_value, changed;
+       int error = sysctl_io_number(req, maxproc, sizeof(int), &new_value, &changed);
+       if (changed) {
+               AUDIT_ARG(value32, new_value);
+               /* make sure the system-wide limit is less than the configured hard
+                  limit set at kernel compilation */
+               if (new_value <= hard_maxproc && new_value > 0)
+                       maxproc = new_value;
+               else
+                       error = EINVAL;
+       }
+       return(error);
+}
+
+SYSCTL_STRING(_kern, KERN_OSTYPE, ostype, 
+               CTLFLAG_RD | CTLFLAG_KERN | CTLFLAG_LOCKED, 
+               ostype, 0, "");
+SYSCTL_STRING(_kern, KERN_OSRELEASE, osrelease, 
+               CTLFLAG_RD | CTLFLAG_KERN | CTLFLAG_LOCKED, 
+               osrelease, 0, "");
+SYSCTL_INT(_kern, KERN_OSREV, osrevision, 
+               CTLFLAG_RD | CTLFLAG_KERN | CTLFLAG_LOCKED, 
+               (int *)NULL, BSD, "");
+SYSCTL_STRING(_kern, KERN_VERSION, version, 
+               CTLFLAG_RD | CTLFLAG_KERN | CTLFLAG_LOCKED, 
+               version, 0, "");
+SYSCTL_STRING(_kern, OID_AUTO, uuid, 
+               CTLFLAG_RD | CTLFLAG_KERN | CTLFLAG_LOCKED, 
+               &kernel_uuid[0], 0, "");
+
+#if DEBUG
+int debug_kprint_syscall = 0;
+char debug_kprint_syscall_process[MAXCOMLEN+1];
+
+/* Thread safe: bits and string value are not used to reclaim state */
+SYSCTL_INT (_debug, OID_AUTO, kprint_syscall,
+           CTLFLAG_RW | CTLFLAG_LOCKED, &debug_kprint_syscall, 0, "kprintf syscall tracing");
+SYSCTL_STRING(_debug, OID_AUTO, kprint_syscall_process, 
+                         CTLFLAG_RW | CTLFLAG_LOCKED, debug_kprint_syscall_process, sizeof(debug_kprint_syscall_process),
+                         "name of process for kprintf syscall tracing");
+
+int debug_kprint_current_process(const char **namep)
+{
+       struct proc *p = current_proc();
+
+       if (p == NULL) {
+               return 0;
+       }
+
+       if (debug_kprint_syscall_process[0]) {
+               /* user asked to scope tracing to a particular process name */
+               if(0 == strncmp(debug_kprint_syscall_process,
+                                               p->p_comm, sizeof(debug_kprint_syscall_process))) {
+                       /* no value in telling the user that we traced what they asked */
+                       if(namep) *namep = NULL;
+
+                       return 1;
+               } else {
+                       return 0;
+               }
+       }
+
+       /* trace all processes. Tell user what we traced */
+       if (namep) {
+               *namep = p->p_comm;
+       }
+
+       return 1;
+}
+#endif
+
+/* PR-5293665: need to use a callback function for kern.osversion to set
+ * osversion in IORegistry */
+
+STATIC int
+sysctl_osversion(__unused struct sysctl_oid *oidp, void *arg1, int arg2, struct sysctl_req *req)
+{
+    int rval = 0;
+
+    rval = sysctl_handle_string(oidp, arg1, arg2, req);
+
+    if (req->newptr) {
+        IORegistrySetOSBuildVersion((char *)arg1); 
+    }
+
+    return rval;
+}
+
+SYSCTL_PROC(_kern, KERN_OSVERSION, osversion,
+        CTLFLAG_RW | CTLFLAG_KERN | CTLTYPE_STRING | CTLFLAG_LOCKED,
+        osversion, 256 /* OSVERSIZE*/, 
+        sysctl_osversion, "A", "");
+
+STATIC int
+sysctl_sysctl_bootargs
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       int error;
+       char buf[256];
+
+       strlcpy(buf, PE_boot_args(), 256);
+       error = sysctl_io_string(req, buf, 256, 0, NULL);
+       return(error);
+}
+
+SYSCTL_PROC(_kern, OID_AUTO, bootargs,
+       CTLFLAG_LOCKED | CTLFLAG_RD | CTLFLAG_KERN | CTLTYPE_STRING,
+       NULL, 0,
+       sysctl_sysctl_bootargs, "A", "bootargs");
+
+SYSCTL_INT(_kern, KERN_MAXFILES, maxfiles, 
+               CTLFLAG_RW | CTLFLAG_KERN | CTLFLAG_LOCKED,
+               &maxfiles, 0, "");
+SYSCTL_INT(_kern, KERN_ARGMAX, argmax, 
+               CTLFLAG_RD | CTLFLAG_KERN | CTLFLAG_LOCKED, 
+               (int *)NULL, ARG_MAX, "");
+SYSCTL_INT(_kern, KERN_POSIX1, posix1version, 
+               CTLFLAG_RD | CTLFLAG_KERN | CTLFLAG_LOCKED, 
+               (int *)NULL, _POSIX_VERSION, "");
+SYSCTL_INT(_kern, KERN_NGROUPS, ngroups, 
+               CTLFLAG_RD | CTLFLAG_KERN | CTLFLAG_LOCKED, 
+               (int *)NULL, NGROUPS_MAX, "");
+SYSCTL_INT(_kern, KERN_JOB_CONTROL, job_control, 
+               CTLFLAG_RD | CTLFLAG_KERN | CTLFLAG_LOCKED, 
+               (int *)NULL, 1, "");
+#if 1  /* _POSIX_SAVED_IDS from <unistd.h> */
+SYSCTL_INT(_kern, KERN_SAVED_IDS, saved_ids, 
+               CTLFLAG_RD | CTLFLAG_KERN | CTLFLAG_LOCKED, 
+               (int *)NULL, 1, "");
+#else
+SYSCTL_INT(_kern, KERN_SAVED_IDS, saved_ids, 
+               CTLFLAG_RD | CTLFLAG_KERN | CTLFLAG_LOCKED, 
+               NULL, 0, "");
+#endif
+SYSCTL_INT(_kern, OID_AUTO, num_files, 
+               CTLFLAG_RD | CTLFLAG_LOCKED, 
+               &nfiles, 0, "");
+SYSCTL_COMPAT_INT(_kern, OID_AUTO, num_vnodes, 
+               CTLFLAG_RD | CTLFLAG_LOCKED, 
+               &numvnodes, 0, "");
+SYSCTL_INT(_kern, OID_AUTO, num_tasks, 
+               CTLFLAG_RD | CTLFLAG_LOCKED, 
+               &task_max, 0, "");
+SYSCTL_INT(_kern, OID_AUTO, num_threads, 
+               CTLFLAG_RD | CTLFLAG_LOCKED, 
+               &thread_max, 0, "");
+SYSCTL_INT(_kern, OID_AUTO, num_taskthreads, 
+               CTLFLAG_RD | CTLFLAG_LOCKED, 
+               &task_threadmax, 0, "");
+
+STATIC int
+sysctl_maxvnodes (__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       int oldval = desiredvnodes;
+       int error = sysctl_io_number(req, desiredvnodes, sizeof(int), &desiredvnodes, NULL);
+
+       if (oldval != desiredvnodes) {
+               reset_vmobjectcache(oldval, desiredvnodes);
+               resize_namecache(desiredvnodes);
+       }
+
+       return(error);
+}
+
+SYSCTL_INT(_kern, OID_AUTO, namecache_disabled, 
+               CTLFLAG_RW | CTLFLAG_LOCKED, 
+               &nc_disabled, 0, ""); 
+
+SYSCTL_PROC(_kern, KERN_MAXVNODES, maxvnodes,
+               CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_LOCKED,
+               0, 0, sysctl_maxvnodes, "I", "");
+
+SYSCTL_PROC(_kern, KERN_MAXPROC, maxproc,
+               CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_LOCKED,
+               0, 0, sysctl_maxproc, "I", "");
+
+SYSCTL_PROC(_kern, KERN_AIOMAX, aiomax,
+               CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_LOCKED,
+               0, 0, sysctl_aiomax, "I", "");
+
+SYSCTL_PROC(_kern, KERN_AIOPROCMAX, aioprocmax,
+               CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_LOCKED,
+               0, 0, sysctl_aioprocmax, "I", "");
+
+SYSCTL_PROC(_kern, KERN_AIOTHREADS, aiothreads,
+               CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_LOCKED,
+               0, 0, sysctl_aiothreads, "I", "");
+
+STATIC int
+sysctl_securelvl
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       int new_value, changed;
+       int error = sysctl_io_number(req, securelevel, sizeof(int), &new_value, &changed);
+       if (changed) {
+               if (!(new_value < securelevel && req->p->p_pid != 1)) {
+                       proc_list_lock();
+                       securelevel = new_value;
+                       proc_list_unlock();
+               } else {
+                       error = EPERM;
+               }
+       }
+       return(error);
+}
+
+SYSCTL_PROC(_kern, KERN_SECURELVL, securelevel,
+               CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_LOCKED,
+               0, 0, sysctl_securelvl, "I", "");
+
+
+STATIC int
+sysctl_domainname
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       int error, changed;
+       error = sysctl_io_string(req, domainname, sizeof(domainname), 0, &changed);
+       if (changed) {
+               domainnamelen = strlen(domainname);
+       }
+       return(error);
+}
+
+SYSCTL_PROC(_kern, KERN_DOMAINNAME, nisdomainname,
+               CTLTYPE_STRING | CTLFLAG_RW | CTLFLAG_LOCKED,
+               0, 0, sysctl_domainname, "A", "");
+
+SYSCTL_COMPAT_INT(_kern, KERN_HOSTID, hostid, 
+               CTLFLAG_RW | CTLFLAG_KERN | CTLFLAG_LOCKED,
+               &hostid, 0, "");
+
+STATIC int
+sysctl_hostname
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       int error, changed;
+       error = sysctl_io_string(req, hostname, sizeof(hostname), 1, &changed);
+       if (changed) {
+               hostnamelen = req->newlen;
+       }
+       return(error);
+}
+
+
+SYSCTL_PROC(_kern, KERN_HOSTNAME, hostname,
+               CTLTYPE_STRING | CTLFLAG_RW | CTLFLAG_LOCKED,
+               0, 0, sysctl_hostname, "A", "");
+
+STATIC int
+sysctl_procname
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       /* Original code allowed writing, I'm copying this, although this all makes
+          no sense to me. Besides, this sysctl is never used. */
+       return sysctl_io_string(req, &req->p->p_name[0], (2*MAXCOMLEN+1), 1, NULL);
+}
+
+SYSCTL_PROC(_kern, KERN_PROCNAME, procname,
+               CTLTYPE_STRING | CTLFLAG_RW | CTLFLAG_ANYBODY | CTLFLAG_LOCKED,
+               0, 0, sysctl_procname, "A", "");
+
+SYSCTL_INT(_kern, KERN_SPECULATIVE_READS, speculative_reads_disabled, 
+               CTLFLAG_RW | CTLFLAG_KERN | CTLFLAG_LOCKED,
+               &speculative_reads_disabled, 0, "");
+
+SYSCTL_INT(_kern, OID_AUTO, ignore_is_ssd, 
+               CTLFLAG_RW | CTLFLAG_KERN | CTLFLAG_LOCKED,
+               &ignore_is_ssd, 0, "");
+
+SYSCTL_UINT(_kern, OID_AUTO, preheat_pages_max, 
+               CTLFLAG_RW | CTLFLAG_KERN | CTLFLAG_LOCKED,
+               &preheat_pages_max, 0, "");
+
+SYSCTL_UINT(_kern, OID_AUTO, preheat_pages_min, 
+               CTLFLAG_RW | CTLFLAG_KERN | CTLFLAG_LOCKED,
+               &preheat_pages_min, 0, "");
+
+SYSCTL_UINT(_kern, OID_AUTO, speculative_prefetch_max, 
+               CTLFLAG_RW | CTLFLAG_KERN | CTLFLAG_LOCKED,
+               &speculative_prefetch_max, 0, "");
+
+SYSCTL_UINT(_kern, OID_AUTO, speculative_prefetch_max_iosize, 
+               CTLFLAG_RW | CTLFLAG_KERN | CTLFLAG_LOCKED,
+               &speculative_prefetch_max_iosize, 0, "");
+
+SYSCTL_UINT(_kern, OID_AUTO, vm_page_free_target,
+               CTLFLAG_RW | CTLFLAG_KERN | CTLFLAG_LOCKED,
+               &vm_page_free_target, 0, "");
+
+SYSCTL_UINT(_kern, OID_AUTO, vm_page_free_min,
+               CTLFLAG_RW | CTLFLAG_KERN | CTLFLAG_LOCKED,
+               &vm_page_free_min, 0, "");
+
+SYSCTL_UINT(_kern, OID_AUTO, vm_page_free_reserved,
+               CTLFLAG_RW | CTLFLAG_KERN | CTLFLAG_LOCKED,
+               &vm_page_free_reserved, 0, "");
+
+SYSCTL_UINT(_kern, OID_AUTO, vm_page_speculative_percentage,
+               CTLFLAG_RW | CTLFLAG_KERN | CTLFLAG_LOCKED,
+               &vm_page_speculative_percentage, 0, "");
+
+SYSCTL_UINT(_kern, OID_AUTO, vm_page_speculative_q_age_ms,
+               CTLFLAG_RW | CTLFLAG_KERN | CTLFLAG_LOCKED,
+               &vm_page_speculative_q_age_ms, 0, "");
+
+SYSCTL_UINT(_kern, OID_AUTO, vm_max_delayed_work_limit,
+               CTLFLAG_RW | CTLFLAG_KERN | CTLFLAG_LOCKED,
+               &vm_max_delayed_work_limit, 0, "");
+
+SYSCTL_UINT(_kern, OID_AUTO, vm_max_batch,
+               CTLFLAG_RW | CTLFLAG_KERN | CTLFLAG_LOCKED,
+               &vm_max_batch, 0, "");
+
+
+STATIC int
+sysctl_boottime
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       time_t tv_sec = boottime_sec();
+       struct proc *p = req->p;
+
+       if (proc_is64bit(p)) {
+               struct user64_timeval t;
+               t.tv_sec = tv_sec;
+               t.tv_usec = 0;
+               return sysctl_io_opaque(req, &t, sizeof(t), NULL);
+       } else {
+               struct user32_timeval t;
+               t.tv_sec = tv_sec;
+               t.tv_usec = 0;
+               return sysctl_io_opaque(req, &t, sizeof(t), NULL);
+       }
+}
+
+SYSCTL_PROC(_kern, KERN_BOOTTIME, boottime,
+               CTLTYPE_STRUCT | CTLFLAG_RD | CTLFLAG_LOCKED,
+               0, 0, sysctl_boottime, "S,timeval", "");
+
+STATIC int
+sysctl_symfile
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       char *str;
+       int error = get_kernel_symfile(req->p, &str);
+       if (error)
+               return (error);
+       return sysctl_io_string(req, str, 0, 0, NULL);
+}
+
+
+SYSCTL_PROC(_kern, KERN_SYMFILE, symfile,
+               CTLTYPE_STRING | CTLFLAG_RD | CTLFLAG_LOCKED,
+               0, 0, sysctl_symfile, "A", "");
+
+#if NFSCLIENT
+STATIC int
+sysctl_netboot
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       return sysctl_io_number(req, netboot_root(), sizeof(int), NULL, NULL);
+}
+
+SYSCTL_PROC(_kern, KERN_NETBOOT, netboot,
+               CTLTYPE_INT | CTLFLAG_RD | CTLFLAG_LOCKED,
+               0, 0, sysctl_netboot, "I", "");
+#endif
+
+#ifdef CONFIG_IMGSRC_ACCESS
+/*
+ * Legacy--act as if only one layer of nesting is possible.
+ */
+STATIC int
+sysctl_imgsrcdev 
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       vfs_context_t ctx = vfs_context_current();
+       vnode_t devvp;
+       int result;
+
+       if (!vfs_context_issuser(ctx)) {
+               return EPERM;
+       }    
+
+       if (imgsrc_rootvnodes[0] == NULL) {
+               return ENOENT;
+       }    
+
+       result = vnode_getwithref(imgsrc_rootvnodes[0]);
+       if (result != 0) {
+               return result;
+       }
+       
+       devvp = vnode_mount(imgsrc_rootvnodes[0])->mnt_devvp;
+       result = vnode_getwithref(devvp);
+       if (result != 0) {
+               goto out;
+       }
+
+       result = sysctl_io_number(req, vnode_specrdev(devvp), sizeof(dev_t), NULL, NULL);
+
+       vnode_put(devvp);
+out:
+       vnode_put(imgsrc_rootvnodes[0]);
+       return result;
+}
+
+SYSCTL_PROC(_kern, OID_AUTO, imgsrcdev,
+               CTLTYPE_INT | CTLFLAG_RD | CTLFLAG_LOCKED,
+               0, 0, sysctl_imgsrcdev, "I", ""); 
+
+STATIC int
+sysctl_imgsrcinfo
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       int error;
+       struct imgsrc_info info[MAX_IMAGEBOOT_NESTING]; /* 2 for now, no problem */
+       uint32_t i;
+       vnode_t rvp, devvp;
+
+       if (imgsrc_rootvnodes[0] == NULLVP) {
+               return ENXIO;
+       }
+
+       for (i = 0; i < MAX_IMAGEBOOT_NESTING; i++) {
+               /*
+                * Go get the root vnode.
+                */
+               rvp = imgsrc_rootvnodes[i];
+               if (rvp == NULLVP) {
+                       break;
+               }
+
+               error = vnode_get(rvp);
+               if (error != 0) {
+                       return error;
+               }
+
+               /* 
+                * For now, no getting at a non-local volume.
+                */
+               devvp = vnode_mount(rvp)->mnt_devvp;
+               if (devvp == NULL) {
+                       vnode_put(rvp);
+                       return EINVAL;  
+               }
+
+               error = vnode_getwithref(devvp);
+               if (error != 0) {
+                       vnode_put(rvp);
+                       return error;
+               }
+
+               /*
+                * Fill in info.
+                */
+               info[i].ii_dev = vnode_specrdev(devvp);
+               info[i].ii_flags = 0;
+               info[i].ii_height = i;
+               bzero(info[i].ii_reserved, sizeof(info[i].ii_reserved));
+
+               vnode_put(devvp);
+               vnode_put(rvp);
+       }
+
+       return sysctl_io_opaque(req, info, i * sizeof(info[0]), NULL);
+}
+
+SYSCTL_PROC(_kern, OID_AUTO, imgsrcinfo,
+               CTLTYPE_OPAQUE | CTLFLAG_RD | CTLFLAG_LOCKED,
+               0, 0, sysctl_imgsrcinfo, "I", ""); 
+
+#endif /* CONFIG_IMGSRC_ACCESS */
+
+SYSCTL_INT(_kern, OID_AUTO, timer_coalescing_enabled, 
+               CTLFLAG_RW | CTLFLAG_LOCKED,
+               &mach_timer_coalescing_enabled, 0, "");
+
+STATIC int
+sysctl_usrstack
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       return sysctl_io_number(req, (int)req->p->user_stack, sizeof(int), NULL, NULL);
+}
+
+SYSCTL_PROC(_kern, KERN_USRSTACK32, usrstack,
+               CTLTYPE_INT | CTLFLAG_RD | CTLFLAG_LOCKED,
+               0, 0, sysctl_usrstack, "I", "");
+
+STATIC int
+sysctl_usrstack64
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       return sysctl_io_number(req, req->p->user_stack, sizeof(req->p->user_stack), NULL, NULL);
+}
+
+SYSCTL_PROC(_kern, KERN_USRSTACK64, usrstack64,
+               CTLTYPE_QUAD | CTLFLAG_RD | CTLFLAG_LOCKED,
+               0, 0, sysctl_usrstack64, "Q", "");
+
+SYSCTL_STRING(_kern, KERN_COREFILE, corefile, 
+               CTLFLAG_RW | CTLFLAG_KERN | CTLFLAG_LOCKED,
+               corefilename, sizeof(corefilename), "");
+
+STATIC int
+sysctl_coredump
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+#ifdef SECURE_KERNEL
+       return (ENOTSUP);
+#endif
+       int new_value, changed;
+       int error = sysctl_io_number(req, do_coredump, sizeof(int), &new_value, &changed);
+       if (changed) {
+               if ((new_value == 0) || (new_value == 1))
+                       do_coredump = new_value;
+               else
+                       error = EINVAL;
+       }
+       return(error);
+}
+
+SYSCTL_PROC(_kern, KERN_COREDUMP, coredump,
+               CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_LOCKED,
+               0, 0, sysctl_coredump, "I", "");
+
+STATIC int
+sysctl_suid_coredump
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+#ifdef SECURE_KERNEL
+       return (ENOTSUP);
+#endif
+       int new_value, changed;
+       int error = sysctl_io_number(req, sugid_coredump, sizeof(int), &new_value, &changed);
+       if (changed) {
+               if ((new_value == 0) || (new_value == 1))
+                       sugid_coredump = new_value;
+               else
+                       error = EINVAL;
+       }
+       return(error);
+}
+
+SYSCTL_PROC(_kern, KERN_SUGID_COREDUMP, sugid_coredump,
+               CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_LOCKED,
+               0, 0, sysctl_suid_coredump, "I", "");
+
+STATIC int
+sysctl_delayterm
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       struct proc *p = req->p;
+       int new_value, changed;
+       int error = sysctl_io_number(req, (req->p->p_lflag & P_LDELAYTERM)? 1: 0, sizeof(int), &new_value, &changed);
+       if (changed) {
+               proc_lock(p);
+               if (new_value)
+                       req->p->p_lflag |=  P_LDELAYTERM;
+               else
+                       req->p->p_lflag &=  ~P_LDELAYTERM;
+               proc_unlock(p);
+       }
+       return(error);
+}
+
+SYSCTL_PROC(_kern, KERN_PROCDELAYTERM, delayterm,
+               CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_LOCKED,
+               0, 0, sysctl_delayterm, "I", "");
+
+
+STATIC int
+sysctl_rage_vnode
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       struct proc *p = req->p;
+        struct  uthread *ut;
+       int new_value, old_value, changed;
+       int error;
+
+       ut = get_bsdthread_info(current_thread());
+
+       if (ut->uu_flag & UT_RAGE_VNODES)
+               old_value = KERN_RAGE_THREAD;
+       else if (p->p_lflag & P_LRAGE_VNODES)
+               old_value = KERN_RAGE_PROC;
+       else
+               old_value = 0;
+
+       error = sysctl_io_number(req, old_value, sizeof(int), &new_value, &changed);
+
+       if (error == 0) {
+               switch (new_value) {
+               case KERN_RAGE_PROC:
+                       proc_lock(p);
+                       p->p_lflag |= P_LRAGE_VNODES;
+                       proc_unlock(p);
+                       break;
+               case KERN_UNRAGE_PROC:
+                       proc_lock(p);
+                       p->p_lflag &= ~P_LRAGE_VNODES;
+                       proc_unlock(p);
+                       break;
+
+               case KERN_RAGE_THREAD:
+                       ut->uu_flag |= UT_RAGE_VNODES;
+                       break;
+               case KERN_UNRAGE_THREAD:
+                       ut = get_bsdthread_info(current_thread());
+                       ut->uu_flag &= ~UT_RAGE_VNODES;
+                       break;
+               }
+       }
+       return(error);
+}
+
+SYSCTL_PROC(_kern, KERN_RAGEVNODE, rage_vnode,
+               CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_ANYBODY | CTLFLAG_LOCKED,
+               0, 0, sysctl_rage_vnode, "I", "");
+
+/* XXX move this interface into libproc and remove this sysctl */
+STATIC int
+sysctl_setthread_cpupercent
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       int new_value, old_value;
+       int error = 0;
+       kern_return_t kret = KERN_SUCCESS;
+       uint8_t percent = 0;
+       int ms_refill = 0;
+
+       old_value = 0;
+
+       if ((error = sysctl_io_number(req, old_value, sizeof(old_value), &new_value, NULL)) != 0)
+               return (error);
+
+       percent = new_value & 0xff;                     /* low 8 bytes for perent */
+       ms_refill = (new_value >> 8) & 0xffffff;        /* upper 24bytes represent ms refill value */
+       if (percent > 100)
+               return (EINVAL);
+
+       /*
+        * If the caller is specifying a percentage of 0, this will unset the CPU limit, if present.
+        */
+       if ((kret = thread_set_cpulimit(THREAD_CPULIMIT_BLOCK, percent, ms_refill * NSEC_PER_MSEC)) != 0)
+               return (EIO);
+       
+       return (0);
+}
+
+SYSCTL_PROC(_kern, OID_AUTO, setthread_cpupercent,
+               CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_ANYBODY,
+               0, 0, sysctl_setthread_cpupercent, "I", "set thread cpu percentage limit");
+
+
+STATIC int
+sysctl_kern_check_openevt
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       struct proc *p = req->p;
+       int new_value, old_value, changed;
+       int error;
+
+       if (p->p_flag & P_CHECKOPENEVT) {
+               old_value = KERN_OPENEVT_PROC;
+       } else {
+               old_value = 0;
+       }
+
+       error = sysctl_io_number(req, old_value, sizeof(int), &new_value, &changed);
+
+       if (error == 0) {
+               switch (new_value) {
+               case KERN_OPENEVT_PROC:
+                       OSBitOrAtomic(P_CHECKOPENEVT, &p->p_flag);
+                       break;
+
+               case KERN_UNOPENEVT_PROC:
+                       OSBitAndAtomic(~((uint32_t)P_CHECKOPENEVT), &p->p_flag);
+                       break;
+
+               default:
+                       error = EINVAL;
+               }
+       }
+       return(error);
+}
+
+SYSCTL_PROC(_kern, KERN_CHECKOPENEVT, check_openevt, CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_ANYBODY | CTLFLAG_LOCKED,
+            0, 0, sysctl_kern_check_openevt, "I", "set the per-process check-open-evt flag");
+
+
+
+STATIC int
+sysctl_nx
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+#ifdef SECURE_KERNEL
+       return ENOTSUP;
+#endif
+       int new_value, changed;
+       int error;
+
+       error = sysctl_io_number(req, nx_enabled, sizeof(nx_enabled), &new_value, &changed);
+       if (error)
+               return error;
+
+       if (changed) {
+#if defined(__i386__) || defined(__x86_64__)
+               /*
+                * Only allow setting if NX is supported on the chip
+                */
+               if (!(cpuid_extfeatures() & CPUID_EXTFEATURE_XD))
+                       return ENOTSUP;
+#endif
+               nx_enabled = new_value;
+       }
+       return(error);
+}
+
+
+
+SYSCTL_PROC(_kern, KERN_NX_PROTECTION, nx, 
+               CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_KERN | CTLFLAG_LOCKED,
+               0, 0, sysctl_nx, "I", "");
+
+STATIC int
+sysctl_loadavg
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+               if (proc_is64bit(req->p)) {
+                       struct user64_loadavg loadinfo64;
+                       fill_loadavg64(&averunnable, &loadinfo64);
+                       return sysctl_io_opaque(req, &loadinfo64, sizeof(loadinfo64), NULL);
+               } else {
+                       struct user32_loadavg loadinfo32;
+                       fill_loadavg32(&averunnable, &loadinfo32);
+                       return sysctl_io_opaque(req, &loadinfo32, sizeof(loadinfo32), NULL);
+               }
+}
+
+SYSCTL_PROC(_vm, VM_LOADAVG, loadavg,
+               CTLTYPE_STRUCT | CTLFLAG_RD | CTLFLAG_LOCKED,
+               0, 0, sysctl_loadavg, "S,loadavg", "");
+
+/*
+ * Note:       Thread safe; vm_map_lock protects in  vm_toggle_entry_reuse()
+ */
+STATIC int
+sysctl_vm_toggle_address_reuse(__unused struct sysctl_oid *oidp, __unused void *arg1,
+             __unused int arg2, struct sysctl_req *req)
+{
+       int old_value=0, new_value=0, error=0;
+       
+       if(vm_toggle_entry_reuse( VM_TOGGLE_GETVALUE, &old_value ))
+               return(error);
+       error = sysctl_io_number(req, old_value, sizeof(int), &new_value, NULL);
+       if (!error) {
+               return (vm_toggle_entry_reuse(new_value, NULL));
+       }
+       return(error);
+}
+
+SYSCTL_PROC(_debug, OID_AUTO, toggle_address_reuse, CTLFLAG_ANYBODY | CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_LOCKED, 0, 0, sysctl_vm_toggle_address_reuse,"I","");
+
+STATIC int
+sysctl_swapusage
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+               int                     error;
+               uint64_t                swap_total;
+               uint64_t                swap_avail;
+               vm_size_t               swap_pagesize;
+               boolean_t               swap_encrypted;
+               struct xsw_usage        xsu;
+
+               error = macx_swapinfo(&swap_total,
+                                     &swap_avail,
+                                     &swap_pagesize,
+                                     &swap_encrypted);
+               if (error)
+                       return error;
+
+               xsu.xsu_total = swap_total;
+               xsu.xsu_avail = swap_avail;
+               xsu.xsu_used = swap_total - swap_avail;
+               xsu.xsu_pagesize = swap_pagesize;
+               xsu.xsu_encrypted = swap_encrypted;
+               return sysctl_io_opaque(req, &xsu, sizeof(xsu), NULL);
+}
+
+
+
+SYSCTL_PROC(_vm, VM_SWAPUSAGE, swapusage,
+               CTLTYPE_STRUCT | CTLFLAG_RD | CTLFLAG_LOCKED,
+               0, 0, sysctl_swapusage, "S,xsw_usage", "");
+
+#if CONFIG_FREEZE
+extern void vm_page_reactivate_all_throttled(void);
+
+static int
+sysctl_freeze_enabled SYSCTL_HANDLER_ARGS
+{
+#pragma unused(arg1, arg2)
+       int error, val = memorystatus_freeze_enabled ? 1 : 0;
+       boolean_t disabled;
+
+       error = sysctl_handle_int(oidp, &val, 0, req);
+       if (error || !req->newptr)
+               return (error);
+       
+       /* 
+        * If freeze is being disabled, we need to move dirty pages out from the throttle to the active queue. 
+        */
+       disabled = (!val && memorystatus_freeze_enabled);
+       
+       memorystatus_freeze_enabled = val ? TRUE : FALSE;
+       
+       if (disabled) {
+               vm_page_reactivate_all_throttled();
+       }
+       
+       return (0);
+}
+
+SYSCTL_PROC(_vm, OID_AUTO, freeze_enabled, CTLTYPE_INT|CTLFLAG_RW, &memorystatus_freeze_enabled, 0, sysctl_freeze_enabled, "I", "");
+#endif /* CONFIG_FREEZE */
+
+/* this kernel does NOT implement shared_region_make_private_np() */
+SYSCTL_INT(_kern, KERN_SHREG_PRIVATIZABLE, shreg_private, 
+               CTLFLAG_RD | CTLFLAG_LOCKED, 
+               (int *)NULL, 0, "");
+
+#if defined(__i386__) || defined(__x86_64__)
+STATIC int
+sysctl_sysctl_exec_affinity(__unused struct sysctl_oid *oidp,
+                          __unused void *arg1, __unused int arg2,
+                          struct sysctl_req *req)
+{
+       proc_t cur_proc = req->p;
+       int error;
+       
+       if (req->oldptr != USER_ADDR_NULL) {
+               cpu_type_t oldcputype = (cur_proc->p_flag & P_AFFINITY) ? CPU_TYPE_POWERPC : CPU_TYPE_I386;
+               if ((error = SYSCTL_OUT(req, &oldcputype, sizeof(oldcputype))))
+                       return error;
+       }
+
+       if (req->newptr != USER_ADDR_NULL) {
+               cpu_type_t newcputype;
+               if ((error = SYSCTL_IN(req, &newcputype, sizeof(newcputype))))
+                       return error;
+               if (newcputype == CPU_TYPE_I386)
+                       OSBitAndAtomic(~((uint32_t)P_AFFINITY), &cur_proc->p_flag);
+               else if (newcputype == CPU_TYPE_POWERPC)
+                       OSBitOrAtomic(P_AFFINITY, &cur_proc->p_flag);
+               else
+                       return (EINVAL);
+       }
+       
+       return 0;
+}
+SYSCTL_PROC(_sysctl, OID_AUTO, proc_exec_affinity, CTLTYPE_INT|CTLFLAG_RW|CTLFLAG_ANYBODY | CTLFLAG_LOCKED, 0, 0, sysctl_sysctl_exec_affinity ,"I","proc_exec_affinity");
+#endif
+
+STATIC int
+fetch_process_cputype(
+       proc_t cur_proc,
+       int *name,
+       u_int namelen,
+       cpu_type_t *cputype)
+{
+       proc_t p = PROC_NULL;
+       int refheld = 0;
+       cpu_type_t ret = 0;
+       int error = 0;
+       
+       if (namelen == 0)
+               p = cur_proc;
+       else if (namelen == 1) {
+               p = proc_find(name[0]);
+               if (p == NULL)
+                       return (EINVAL);
+               refheld = 1;
+       } else {
+               error = EINVAL;
+               goto out;
+       }
+
+#if defined(__i386__) || defined(__x86_64__)
+       if (p->p_flag & P_TRANSLATED) {
+               ret = CPU_TYPE_POWERPC;
+       }
+       else
+#endif
+       {
+               ret = cpu_type();
+               if (IS_64BIT_PROCESS(p))
+                       ret |= CPU_ARCH_ABI64;
+       }
+       *cputype = ret;
+       
+       if (refheld != 0)
+               proc_rele(p);
+out:
+       return (error);
+}
+
+STATIC int
+sysctl_sysctl_native(__unused struct sysctl_oid *oidp, void *arg1, int arg2,
+                   struct sysctl_req *req)
+{
+       int error;
+       cpu_type_t proc_cputype = 0;
+       if ((error = fetch_process_cputype(req->p, (int *)arg1, arg2, &proc_cputype)) != 0)
+               return error;
+       int res = 1;
+       if ((proc_cputype & ~CPU_ARCH_MASK) != (cpu_type() & ~CPU_ARCH_MASK))
+               res = 0;
+       return SYSCTL_OUT(req, &res, sizeof(res));
+}      
+SYSCTL_PROC(_sysctl, OID_AUTO, proc_native, CTLTYPE_NODE|CTLFLAG_RD | CTLFLAG_LOCKED, 0, 0, sysctl_sysctl_native ,"I","proc_native");
+
+STATIC int
+sysctl_sysctl_cputype(__unused struct sysctl_oid *oidp, void *arg1, int arg2,
+                    struct sysctl_req *req)
+{
+       int error;
+       cpu_type_t proc_cputype = 0;
+       if ((error = fetch_process_cputype(req->p, (int *)arg1, arg2, &proc_cputype)) != 0)
+               return error;
+       return SYSCTL_OUT(req, &proc_cputype, sizeof(proc_cputype));
+}
+SYSCTL_PROC(_sysctl, OID_AUTO, proc_cputype, CTLTYPE_NODE|CTLFLAG_RD | CTLFLAG_LOCKED, 0, 0, sysctl_sysctl_cputype ,"I","proc_cputype");
+
+STATIC int
+sysctl_safeboot
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       return sysctl_io_number(req, boothowto & RB_SAFEBOOT ? 1 : 0, sizeof(int), NULL, NULL);
+}
+
+SYSCTL_PROC(_kern, KERN_SAFEBOOT, safeboot,
+               CTLTYPE_INT | CTLFLAG_RD | CTLFLAG_LOCKED,
+               0, 0, sysctl_safeboot, "I", "");
+
+STATIC int
+sysctl_singleuser
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       return sysctl_io_number(req, boothowto & RB_SINGLE ? 1 : 0, sizeof(int), NULL, NULL);
+}
+
+SYSCTL_PROC(_kern, OID_AUTO, singleuser,
+               CTLTYPE_INT | CTLFLAG_RD | CTLFLAG_LOCKED,
+               0, 0, sysctl_singleuser, "I", "");
+
+/*
+ * Controls for debugging affinity sets - see osfmk/kern/affinity.c
+ */
+extern boolean_t       affinity_sets_enabled;
+extern int             affinity_sets_mapping;
+
+SYSCTL_INT (_kern, OID_AUTO, affinity_sets_enabled,
+           CTLFLAG_RW | CTLFLAG_LOCKED, (int *) &affinity_sets_enabled, 0, "hinting enabled");
+SYSCTL_INT (_kern, OID_AUTO, affinity_sets_mapping,
+           CTLFLAG_RW | CTLFLAG_LOCKED, &affinity_sets_mapping, 0, "mapping policy");
+
+/*
+ * Boolean indicating if KASLR is active.
+ */
+STATIC int
+sysctl_slide
+(__unused struct sysctl_oid *oidp, __unused void *arg1, __unused int arg2, struct sysctl_req *req)
+{
+       uint32_t        slide;
+
+       slide = vm_kernel_slide ? 1 : 0;
+
+       return sysctl_io_number( req, slide, sizeof(int), NULL, NULL);
+}
+
+SYSCTL_PROC(_kern, OID_AUTO, slide,
+               CTLTYPE_INT | CTLFLAG_RD | CTLFLAG_LOCKED,
+               0, 0, sysctl_slide, "I", "");
+
+/*
+ * Limit on total memory users can wire.
+ *
+ * vm_global_user_wire_limit - system wide limit on wired memory from all processes combined. 
+ *
+ * vm_user_wire_limit - per address space limit on wired memory.  This puts a cap on the process's rlimit value.
+ *
+ * These values are initialized to reasonable defaults at boot time based on the available physical memory in
+ * kmem_init().
+ *
+ * All values are in bytes.
+ */
+
+vm_map_size_t  vm_global_no_user_wire_amount;
+vm_map_size_t  vm_global_user_wire_limit;
+vm_map_size_t  vm_user_wire_limit;
+
+/*
+ * There needs to be a more automatic/elegant way to do this
+ */
+SYSCTL_QUAD(_vm, OID_AUTO, global_no_user_wire_amount, CTLFLAG_RW | CTLFLAG_LOCKED, &vm_global_no_user_wire_amount, "");
+SYSCTL_QUAD(_vm, OID_AUTO, global_user_wire_limit, CTLFLAG_RW | CTLFLAG_LOCKED, &vm_global_user_wire_limit, "");
+SYSCTL_QUAD(_vm, OID_AUTO, user_wire_limit, CTLFLAG_RW | CTLFLAG_LOCKED, &vm_user_wire_limit, "");
+
+extern int vm_map_copy_overwrite_aligned_src_not_internal;
+extern int vm_map_copy_overwrite_aligned_src_not_symmetric;
+extern int vm_map_copy_overwrite_aligned_src_large;
+SYSCTL_INT(_vm, OID_AUTO, vm_copy_src_not_internal, CTLFLAG_RD | CTLFLAG_LOCKED, &vm_map_copy_overwrite_aligned_src_not_internal, 0, "");
+SYSCTL_INT(_vm, OID_AUTO, vm_copy_src_not_symmetric, CTLFLAG_RD | CTLFLAG_LOCKED, &vm_map_copy_overwrite_aligned_src_not_symmetric, 0, "");
+SYSCTL_INT(_vm, OID_AUTO, vm_copy_src_large, CTLFLAG_RD | CTLFLAG_LOCKED, &vm_map_copy_overwrite_aligned_src_large, 0, "");
+
+
+/*
+ * enable back trace events for thread blocks
+ */
+
+extern uint32_t        kdebug_thread_block;
+
+SYSCTL_INT (_kern, OID_AUTO, kdebug_thread_block,
+           CTLFLAG_RW | CTLFLAG_LOCKED, &kdebug_thread_block, 0, "kdebug thread_block");
+
+/*
+ * Kernel stack size and depth
+ */
+SYSCTL_INT (_kern, OID_AUTO, stack_size,
+           CTLFLAG_RD | CTLFLAG_LOCKED, (int *) &kernel_stack_size, 0, "Kernel stack size");
+SYSCTL_INT (_kern, OID_AUTO, stack_depth_max,
+           CTLFLAG_RD | CTLFLAG_LOCKED, (int *) &kernel_stack_depth_max, 0, "Max kernel stack depth at interrupt or context switch");
+
+/*
+ * enable back trace for port allocations
+ */
+extern int ipc_portbt;
+
+SYSCTL_INT(_kern, OID_AUTO, ipc_portbt, 
+               CTLFLAG_RW | CTLFLAG_KERN | CTLFLAG_LOCKED,
+               &ipc_portbt, 0, "");
+
+/*
+ * Scheduler sysctls
+ */
+
+/*
+ * See osfmk/kern/sched_prim.c for the corresponding definition
+ * in osfmk/. If either version changes, update the other.
+ */
+#define SCHED_STRING_MAX_LENGTH (48)
+
+extern char sched_string[SCHED_STRING_MAX_LENGTH];
+SYSCTL_STRING(_kern, OID_AUTO, sched,
+                         CTLFLAG_RD | CTLFLAG_KERN | CTLFLAG_LOCKED,
+                         sched_string, sizeof(sched_string),
+                         "Timeshare scheduler implementation");
+
+/*
+ * Only support runtime modification on embedded platforms
+ * with development config enabled
+ */
+#if CONFIG_EMBEDDED
+#if !SECURE_KERNEL
+extern int precise_user_kernel_time;
+SYSCTL_INT(_kern, OID_AUTO, precise_user_kernel_time, 
+               CTLFLAG_RW | CTLFLAG_LOCKED,
+               &precise_user_kernel_time, 0, "Precise accounting of kernel vs. user time");
+#endif
+#endif