]> git.saurik.com Git - apple/xnu.git/blobdiff - osfmk/i386/trap.c
xnu-2050.48.11.tar.gz
[apple/xnu.git] / osfmk / i386 / trap.c
index bb395224973a2fda38d97edb6466f7340427573b..c6f921c6f140692c407cdb03512deab581baaaa1 100644 (file)
@@ -1,60 +1,65 @@
 /*
 /*
- * Copyright (c) 2000-2004 Apple Computer, Inc. All rights reserved.
+ * Copyright (c) 2000-2010 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,
  * 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@
  */
 /*
  */
 /*
- * @OSF_COPYRIGHT@
- */
+* @OSF_COPYRIGHT@
+*/
 /* 
 /* 
- * Mach Operating System
- * Copyright (c) 1991,1990,1989,1988 Carnegie Mellon University
- * All Rights Reserved.
- 
- * Permission to use, copy, modify and distribute this software and its
- * documentation is hereby granted, provided that both the copyright
- * notice and this permission notice appear in all copies of the
- * software, derivative works or modified versions, and any portions
- * thereof, and that both notices appear in supporting documentation.
- 
- * CARNEGIE MELLON ALLOWS FREE USE OF THIS SOFTWARE IN ITS "AS IS"
- * CONDITION.  CARNEGIE MELLON DISCLAIMS ANY LIABILITY OF ANY KIND FOR
- * ANY DAMAGES WHATSOEVER RESULTING FROM THE USE OF THIS SOFTWARE.
- 
- * Carnegie Mellon requests users of this software to return to
- 
- *  Software Distribution Coordinator  or  Software.Distribution@CS.CMU.EDU
- *  School of Computer Science
- *  Carnegie Mellon University
- *  Pittsburgh PA 15213-3890
- 
- * any improvements or extensions that they make and grant Carnegie Mellon
- * the rights to redistribute these changes.
- */
+* Mach Operating System
+* Copyright (c) 1991,1990,1989,1988 Carnegie Mellon University
+* All Rights Reserved.
+* 
+* Permission to use, copy, modify and distribute this software and its
+* documentation is hereby granted, provided that both the copyright
+* notice and this permission notice appear in all copies of the
+* software, derivative works or modified versions, and any portions
+* thereof, and that both notices appear in supporting documentation.
+* 
+* CARNEGIE MELLON ALLOWS FREE USE OF THIS SOFTWARE IN ITS "AS IS"
+* CONDITION.  CARNEGIE MELLON DISCLAIMS ANY LIABILITY OF ANY KIND FOR
+* ANY DAMAGES WHATSOEVER RESULTING FROM THE USE OF THIS SOFTWARE.
+* 
+* Carnegie Mellon requests users of this software to return to
+* 
+*  Software Distribution Coordinator  or  Software.Distribution@CS.CMU.EDU
+*  School of Computer Science
+*  Carnegie Mellon University
+*  Pittsburgh PA 15213-3890
+* 
+* any improvements or extensions that they make and grant Carnegie Mellon
+* the rights to redistribute these changes.
+*/
 /*
 /*
- */
+*/
+
 /*
 /*
- * Hardware trap/fault handler.
+* Hardware trap/fault handler.
  */
 
  */
 
-#include <mach_kdb.h>
-#include <mach_kgdb.h>
 #include <mach_kdp.h>
 #include <mach_ldebug.h>
 
 #include <mach_kdp.h>
 #include <mach_ldebug.h>
 
@@ -63,7 +68,8 @@
 #include <i386/trap.h>
 #include <i386/pmap.h>
 #include <i386/fpu.h>
 #include <i386/trap.h>
 #include <i386/pmap.h>
 #include <i386/fpu.h>
-#include <architecture/i386/pio.h> /* inb() */
+#include <i386/misc_protos.h> /* panic_io_port_read() */
+#include <i386/lapic.h>
 
 #include <mach/exception.h>
 #include <mach/kern_return.h>
 
 #include <mach/exception.h>
 #include <mach/kern_return.h>
 #include <kern/exception.h>
 #include <kern/spl.h>
 #include <kern/misc_protos.h>
 #include <kern/exception.h>
 #include <kern/spl.h>
 #include <kern/misc_protos.h>
+#include <kern/debug.h>
 
 #include <sys/kdebug.h>
 
 
 #include <sys/kdebug.h>
 
-#if    MACH_KGDB
-#include <kgdb/kgdb_defs.h>
-#endif /* MACH_KGDB */
-
-#if    MACH_KDB
-#include <debug.h>
-#include <ddb/db_watch.h>
-#include <ddb/db_run.h>
-#include <ddb/db_break.h>
-#include <ddb/db_trap.h>
-#endif /* MACH_KDB */
-
 #include <string.h>
 
 #include <string.h>
 
-#include <i386/io_emulate.h>
 #include <i386/postcode.h>
 #include <i386/mp_desc.h>
 #include <i386/proc_reg.h>
 #include <i386/postcode.h>
 #include <i386/mp_desc.h>
 #include <i386/proc_reg.h>
+#if CONFIG_MCA
+#include <i386/machine_check.h>
+#endif
 #include <mach/i386/syscall_sw.h>
 
 #include <mach/i386/syscall_sw.h>
 
+#include <libkern/OSDebug.h>
+#include <i386/cpu_threads.h>
+#include <machine/pal_routines.h>
+
+extern void throttle_lowpri_io(int);
+extern void kprint_state(x86_saved_state64_t *saved_state);
+
 /*
  * Forward declarations
  */
 static void user_page_fault_continue(kern_return_t kret);
 /*
  * Forward declarations
  */
 static void user_page_fault_continue(kern_return_t kret);
+#ifdef __i386__
 static void panic_trap(x86_saved_state32_t *saved_state);
 static void set_recovery_ip(x86_saved_state32_t *saved_state, vm_offset_t ip);
 static void panic_trap(x86_saved_state32_t *saved_state);
 static void set_recovery_ip(x86_saved_state32_t *saved_state, vm_offset_t ip);
+extern void panic_64(x86_saved_state_t *, int, const char *, boolean_t);
+#else
+static void panic_trap(x86_saved_state64_t *saved_state);
+static void set_recovery_ip(x86_saved_state64_t *saved_state, vm_offset_t ip);
+#endif
+
+volatile perfCallback perfTrapHook = NULL; /* Pointer to CHUD trap hook routine */
 
 
-perfCallback perfTrapHook = NULL; /* Pointer to CHUD trap hook routine */
-perfCallback perfASTHook  = NULL; /* Pointer to CHUD AST hook routine */
+#if CONFIG_DTRACE
+/* See <rdar://problem/4613924> */
+perfCallback tempDTraceTrapHook = NULL; /* Pointer to DTrace fbt trap hook routine */
+
+extern boolean_t dtrace_tally_fault(user_addr_t);
+#endif
+
+extern boolean_t pmap_smep_enabled;
 
 void
 thread_syscall_return(
         kern_return_t ret)
 {
         thread_t       thr_act = current_thread();
 
 void
 thread_syscall_return(
         kern_return_t ret)
 {
         thread_t       thr_act = current_thread();
+       boolean_t       is_mach;
+       int             code;
+
+       pal_register_cache_state(thr_act, DIRTY);
 
         if (thread_is_64bit(thr_act)) {
                x86_saved_state64_t     *regs;
                
                regs = USER_REGS64(thr_act);
 
 
         if (thread_is_64bit(thr_act)) {
                x86_saved_state64_t     *regs;
                
                regs = USER_REGS64(thr_act);
 
-               if (kdebug_enable && ((regs->rax & SYSCALL_CLASS_MASK) == (SYSCALL_CLASS_MACH << SYSCALL_CLASS_SHIFT))) {
+               code = (int) (regs->rax & SYSCALL_NUMBER_MASK);
+               is_mach = (regs->rax & SYSCALL_CLASS_MASK)
+                           == (SYSCALL_CLASS_MACH << SYSCALL_CLASS_SHIFT);
+               if (kdebug_enable && is_mach) {
                        /* Mach trap */
                        /* Mach trap */
-                       KERNEL_DEBUG_CONSTANT(
-                                             MACHDBG_CODE(DBG_MACH_EXCP_SC, ((int) (regs->rax & SYSCALL_NUMBER_MASK)))
-                                             | DBG_FUNC_END,
-                                             ret, 0, 0, 0, 0);
+                       KERNEL_DEBUG_CONSTANT_IST(KDEBUG_TRACE, 
+                             MACHDBG_CODE(DBG_MACH_EXCP_SC,code)|DBG_FUNC_END,
+                             ret, 0, 0, 0, 0);
                }
                regs->rax = ret;
                }
                regs->rax = ret;
-
+#if DEBUG
+               if (is_mach)
+                       DEBUG_KPRINT_SYSCALL_MACH(
+                               "thread_syscall_return: 64-bit mach ret=%u\n",
+                               ret);
+               else
+                       DEBUG_KPRINT_SYSCALL_UNIX(
+                               "thread_syscall_return: 64-bit unix ret=%u\n",
+                               ret);
+#endif
        } else {
                x86_saved_state32_t     *regs;
                
                regs = USER_REGS32(thr_act);
 
        } else {
                x86_saved_state32_t     *regs;
                
                regs = USER_REGS32(thr_act);
 
-               if (kdebug_enable && ((int) regs->eax < 0)) {
+               code = ((int) regs->eax);
+               is_mach = (code < 0);
+               if (kdebug_enable && is_mach) {
                        /* Mach trap */
                        /* Mach trap */
-                       KERNEL_DEBUG_CONSTANT(
-                                             MACHDBG_CODE(DBG_MACH_EXCP_SC, -((int) regs->eax))
-                                             | DBG_FUNC_END,
-                                             ret, 0, 0, 0, 0);
+                       KERNEL_DEBUG_CONSTANT_IST(KDEBUG_TRACE, 
+                             MACHDBG_CODE(DBG_MACH_EXCP_SC,-code)|DBG_FUNC_END,
+                             ret, 0, 0, 0, 0);
                }
                regs->eax = ret;
                }
                regs->eax = ret;
+#if DEBUG
+               if (is_mach)
+                       DEBUG_KPRINT_SYSCALL_MACH(
+                               "thread_syscall_return: 32-bit mach ret=%u\n",
+                               ret);
+               else
+                       DEBUG_KPRINT_SYSCALL_UNIX(
+                               "thread_syscall_return: 32-bit unix ret=%u\n",
+                               ret);
+#endif
        }
        }
-        thread_exception_return();
-        /*NOTREACHED*/
-}
-
-
-#if    MACH_KDB
-boolean_t      debug_all_traps_with_kdb = FALSE;
-extern struct db_watchpoint *db_watchpoint_list;
-extern boolean_t db_watchpoints_inserted;
-extern boolean_t db_breakpoints_inserted;
-
-void
-thread_kdb_return(void)
-{
-       thread_t                thr_act = current_thread();
-       x86_saved_state_t       *iss = USER_STATE(thr_act);
-
-        if (is_saved_state64(iss)) {
-               x86_saved_state64_t     *regs;
-               
-               regs = saved_state64(iss);
-
-               if (kdb_trap(regs->isf.trapno, (int)regs->isf.err, (void *)regs)) {
-                       thread_exception_return();
-                       /*NOTREACHED*/
-               }
-
-       } else {
-               x86_saved_state32_t     *regs;
-               
-               regs = saved_state32(iss);
+       throttle_lowpri_io(TRUE);
 
 
-               if (kdb_trap(regs->trapno, regs->err, (void *)regs)) {
-                       thread_exception_return();
-                       /*NOTREACHED*/
-               }
-       }
+       thread_exception_return();
+        /*NOTREACHED*/
 }
 
 }
 
-#endif /* MACH_KDB */
 
 
-void
+static inline void
 user_page_fault_continue(
                         kern_return_t  kr)
 {
        thread_t        thread = current_thread();
 user_page_fault_continue(
                         kern_return_t  kr)
 {
        thread_t        thread = current_thread();
-       x86_saved_state_t *regs = USER_STATE(thread);
-       ast_t           *myast;
-       boolean_t       intr;
        user_addr_t     vaddr;
        user_addr_t     vaddr;
-#if    MACH_KDB
-       int             err;
-       int             trapno;
-#endif
 
 
-       assert((is_saved_state32(regs) && !thread_is_64bit(thread)) ||
-              (is_saved_state64(regs) &&  thread_is_64bit(thread)));
-
-        if (thread_is_64bit(thread)) {
-               x86_saved_state64_t     *uregs;
+       if (thread_is_64bit(thread)) {
+               x86_saved_state64_t     *uregs;
 
                uregs = USER_REGS64(thread);
 
 
                uregs = USER_REGS64(thread);
 
-#if    MACH_KDB
-               trapno = uregs->isf.trapno;
-               err = uregs->isf.err;
-#endif
                vaddr = (user_addr_t)uregs->cr2;
        } else {
                x86_saved_state32_t     *uregs;
 
                uregs = USER_REGS32(thread);
 
                vaddr = (user_addr_t)uregs->cr2;
        } else {
                x86_saved_state32_t     *uregs;
 
                uregs = USER_REGS32(thread);
 
-#if    MACH_KDB
-               trapno = uregs->trapno;
-               err = uregs->err;
-#endif
                vaddr = uregs->cr2;
        }
 
                vaddr = uregs->cr2;
        }
 
-       if ((kr == KERN_SUCCESS) || (kr == KERN_ABORTED)) {
-#if    MACH_KDB
-               if (!db_breakpoints_inserted) {
-                       db_set_breakpoints();
-               }
-               if (db_watchpoint_list &&
-                   db_watchpoints_inserted &&
-                   (err & T_PF_WRITE) &&
-                   db_find_watchpoint(thread->map,
-                                      (vm_offset_t)vaddr,
-                                      regs))
-                       kdb_trap(T_WATCHPOINT, 0, regs);
-#endif /* MACH_KDB */
-               intr = ml_set_interrupts_enabled(FALSE);
-               myast = ast_pending();
-               while (*myast & AST_ALL) {
-                       ast_taken(AST_ALL, intr);
-                       ml_set_interrupts_enabled(FALSE);
-                       myast = ast_pending();
-               }
-               ml_set_interrupts_enabled(intr);
 
 
-               thread_exception_return();
-               /*NOTREACHED*/
-       }
-
-#if    MACH_KDB
-       if (debug_all_traps_with_kdb &&
-           kdb_trap(trapno, err, regs)) {
-               thread_exception_return();
-               /*NOTREACHED*/
-       }
-#endif /* MACH_KDB */
+       /* PAL debug hook */
+       pal_dbg_page_fault( thread, vaddr, kr );
 
        i386_exception(EXC_BAD_ACCESS, kr, vaddr);
        /*NOTREACHED*/
 
        i386_exception(EXC_BAD_ACCESS, kr, vaddr);
        /*NOTREACHED*/
@@ -271,8 +233,8 @@ user_page_fault_continue(
  * Fault recovery in copyin/copyout routines.
  */
 struct recovery {
  * Fault recovery in copyin/copyout routines.
  */
 struct recovery {
-       uint32_t        fault_addr;
-       uint32_t        recover_addr;
+       uintptr_t       fault_addr;
+       uintptr_t       recover_addr;
 };
 
 extern struct recovery recover_table[];
 };
 
 extern struct recovery recover_table[];
@@ -281,31 +243,235 @@ extern struct recovery   recover_table_end[];
 const char *   trap_type[] = {TRAP_NAMES};
 unsigned       TRAP_TYPES = sizeof(trap_type)/sizeof(trap_type[0]);
 
 const char *   trap_type[] = {TRAP_NAMES};
 unsigned       TRAP_TYPES = sizeof(trap_type)/sizeof(trap_type[0]);
 
-extern unsigned panic_io_port;
+extern void    PE_incoming_interrupt(int interrupt);
+
+#if defined(__x86_64__) && DEBUG
+void
+kprint_state(x86_saved_state64_t       *saved_state)
+{
+       kprintf("current_cpu_datap() 0x%lx\n", (uintptr_t)current_cpu_datap());
+       kprintf("Current GS base MSR 0x%llx\n", rdmsr64(MSR_IA32_GS_BASE));
+       kprintf("Kernel  GS base MSR 0x%llx\n", rdmsr64(MSR_IA32_KERNEL_GS_BASE));
+       kprintf("state at 0x%lx:\n", (uintptr_t) saved_state);
+
+       kprintf("      rdi    0x%llx\n", saved_state->rdi);        
+       kprintf("      rsi    0x%llx\n", saved_state->rsi);    
+       kprintf("      rdx    0x%llx\n", saved_state->rdx);
+       kprintf("      r10    0x%llx\n", saved_state->r10);
+       kprintf("      r8     0x%llx\n", saved_state->r8);
+       kprintf("      r9     0x%llx\n", saved_state->r9);     
+       kprintf("      v_arg6 0x%llx\n", saved_state->v_arg6);
+       kprintf("      v_arg7 0x%llx\n", saved_state->v_arg7);
+       kprintf("      v_arg8 0x%llx\n", saved_state->v_arg8);
+
+       kprintf("      cr2    0x%llx\n", saved_state->cr2);
+       kprintf("real  cr2    0x%lx\n", get_cr2());
+       kprintf("      r15    0x%llx\n", saved_state->r15);
+       kprintf("      r14    0x%llx\n", saved_state->r14);
+       kprintf("      r13    0x%llx\n", saved_state->r13);
+       kprintf("      r12    0x%llx\n", saved_state->r12);
+       kprintf("      r11    0x%llx\n", saved_state->r11);
+       kprintf("      rbp    0x%llx\n", saved_state->rbp);
+       kprintf("      rbx    0x%llx\n", saved_state->rbx);
+       kprintf("      rcx    0x%llx\n", saved_state->rcx);
+       kprintf("      rax    0x%llx\n", saved_state->rax);
+
+       kprintf("      gs     0x%x\n", saved_state->gs);
+       kprintf("      fs     0x%x\n", saved_state->fs);
+
+       kprintf("  isf.trapno 0x%x\n", saved_state->isf.trapno);
+       kprintf("  isf._pad   0x%x\n", saved_state->isf._pad);
+       kprintf("  isf.trapfn 0x%llx\n", saved_state->isf.trapfn);
+       kprintf("  isf.err    0x%llx\n", saved_state->isf.err);
+       kprintf("  isf.rip    0x%llx\n", saved_state->isf.rip);
+       kprintf("  isf.cs     0x%llx\n", saved_state->isf.cs);
+       kprintf("  isf.rflags 0x%llx\n", saved_state->isf.rflags);
+       kprintf("  isf.rsp    0x%llx\n", saved_state->isf.rsp);
+       kprintf("  isf.ss     0x%llx\n", saved_state->isf.ss);
+}
+#endif
+
+
+/*
+ * Non-zero indicates latency assert is enabled and capped at valued
+ * absolute time units.
+ */
+   
+uint64_t interrupt_latency_cap = 0;
+boolean_t ilat_assert = FALSE;
+
+void
+interrupt_latency_tracker_setup(void) {
+       uint32_t ilat_cap_us;
+       if (PE_parse_boot_argn("interrupt_latency_cap_us", &ilat_cap_us, sizeof(ilat_cap_us))) {
+               interrupt_latency_cap = ilat_cap_us * NSEC_PER_USEC;
+               nanoseconds_to_absolutetime(interrupt_latency_cap, &interrupt_latency_cap);
+       } else {
+               interrupt_latency_cap = LockTimeOut;
+       }
+       PE_parse_boot_argn("-interrupt_latency_assert_enable", &ilat_assert, sizeof(ilat_assert));
+}
+
+void interrupt_reset_latency_stats(void) {
+       uint32_t i;
+       for (i = 0; i < real_ncpus; i++) {
+               cpu_data_ptr[i]->cpu_max_observed_int_latency =
+                   cpu_data_ptr[i]->cpu_max_observed_int_latency_vector = 0;
+       }
+}
+
+void interrupt_populate_latency_stats(char *buf, unsigned bufsize) {
+       uint32_t i, tcpu = ~0;
+       uint64_t cur_max = 0;
+
+       for (i = 0; i < real_ncpus; i++) {
+               if (cur_max < cpu_data_ptr[i]->cpu_max_observed_int_latency) {
+                       cur_max = cpu_data_ptr[i]->cpu_max_observed_int_latency;
+                       tcpu = i;
+               }
+       }
+
+       if (tcpu < real_ncpus)
+               snprintf(buf, bufsize, "0x%x 0x%x 0x%llx", tcpu, cpu_data_ptr[tcpu]->cpu_max_observed_int_latency_vector, cpu_data_ptr[tcpu]->cpu_max_observed_int_latency);
+}
+
+/*
+ * Handle interrupts:
+ *  - local APIC interrupts (IPIs, timers, etc) are handled by the kernel,
+ *  - device interrupts go to the platform expert.
+ */
+void
+interrupt(x86_saved_state_t *state)
+{
+       uint64_t        rip;
+       uint64_t        rsp;
+       int             interrupt_num;
+       boolean_t       user_mode = FALSE;
+       int             ipl;
+       int             cnum = cpu_number();
+       int             itype = 0;
+
+       if (is_saved_state64(state) == TRUE) {
+               x86_saved_state64_t     *state64;
+
+               state64 = saved_state64(state);
+               rip = state64->isf.rip;
+               rsp = state64->isf.rsp;
+               interrupt_num = state64->isf.trapno;
+#ifdef __x86_64__
+               if(state64->isf.cs & 0x03)
+#endif
+                       user_mode = TRUE;
+       } else {
+               x86_saved_state32_t     *state32;
+
+               state32 = saved_state32(state);
+               if (state32->cs & 0x03)
+                       user_mode = TRUE;
+               rip = state32->eip;
+               rsp = state32->uesp;
+               interrupt_num = state32->trapno;
+       }
+
+       if (cpu_data_ptr[cnum]->lcpu.package->num_idle == topoParms.nLThreadsPerPackage)
+               cpu_data_ptr[cnum]->cpu_hwIntpexits[interrupt_num]++;
+
+       if (interrupt_num == (LAPIC_DEFAULT_INTERRUPT_BASE + LAPIC_INTERPROCESSOR_INTERRUPT))
+               itype = 1;
+       else if (interrupt_num == (LAPIC_DEFAULT_INTERRUPT_BASE + LAPIC_TIMER_INTERRUPT))
+               itype = 2;
+       else
+               itype = 3;
+
+       KERNEL_DEBUG_CONSTANT_IST(KDEBUG_TRACE, 
+               MACHDBG_CODE(DBG_MACH_EXCP_INTR, 0) | DBG_FUNC_START,
+               interrupt_num,
+               (user_mode ? rip : VM_KERNEL_UNSLIDE(rip)),
+               user_mode, itype, 0);
+
+       SCHED_STATS_INTERRUPT(current_processor());
+
+       ipl = get_preemption_level();
+       
+       /*
+        * Handle local APIC interrupts
+        * else call platform expert for devices.
+        */
+       if (!lapic_interrupt(interrupt_num, state))
+               PE_incoming_interrupt(interrupt_num);
+
+       if (__improbable(get_preemption_level() != ipl)) {
+               panic("Preemption level altered by interrupt vector 0x%x: initial 0x%x, final: 0x%x\n", interrupt_num, ipl, get_preemption_level());
+       }
+
+
+       KERNEL_DEBUG_CONSTANT_IST(KDEBUG_TRACE, 
+               MACHDBG_CODE(DBG_MACH_EXCP_INTR, 0) | DBG_FUNC_END,
+               interrupt_num, 0, 0, 0, 0);
+
+       if (cpu_data_ptr[cnum]->cpu_nested_istack) {
+               cpu_data_ptr[cnum]->cpu_nested_istack_events++;
+       }
+       else  {
+               uint64_t int_latency = mach_absolute_time() - cpu_data_ptr[cnum]->cpu_int_event_time;
+               if (ilat_assert && (int_latency > interrupt_latency_cap) && !machine_timeout_suspended()) {
+                       panic("Interrupt vector 0x%x exceeded interrupt latency threshold, 0x%llx absolute time delta, prior signals: 0x%x, current signals: 0x%x", interrupt_num, int_latency, cpu_data_ptr[cnum]->cpu_prior_signals, cpu_data_ptr[cnum]->cpu_signals);
+               }
+               if (int_latency > cpu_data_ptr[cnum]->cpu_max_observed_int_latency) {
+                       cpu_data_ptr[cnum]->cpu_max_observed_int_latency = int_latency;
+                       cpu_data_ptr[cnum]->cpu_max_observed_int_latency_vector = interrupt_num;
+               }
+       }
+
+       /*
+        * Having serviced the interrupt first, look at the interrupted stack depth.
+        */
+       if (!user_mode) {
+               uint64_t depth = cpu_data_ptr[cnum]->cpu_kernel_stack
+                                + sizeof(struct x86_kernel_state)
+                                + sizeof(struct i386_exception_link *)
+                                - rsp;
+               if (depth > kernel_stack_depth_max) {
+                       kernel_stack_depth_max = (vm_offset_t)depth;
+                       KERNEL_DEBUG_CONSTANT(
+                               MACHDBG_CODE(DBG_MACH_SCHED, MACH_STACK_DEPTH),
+                               (long) depth, (long) VM_KERNEL_UNSLIDE(rip), 0, 0, 0);
+               }
+       }
+}
 
 static inline void
 reset_dr7(void)
 {
 
 static inline void
 reset_dr7(void)
 {
-       uint32_t dr7 = 0x400; /* magic dr7 reset value */
-       __asm__ volatile("movl %0,%%dr7" : : "r" (dr7));
+       long dr7 = 0x400; /* magic dr7 reset value; 32 bit on i386, 64 bit on x86_64 */
+       __asm__ volatile("mov %0,%%dr7" : : "r" (dr7));
 }
 #if MACH_KDP
 unsigned kdp_has_active_watchpoints = 0;
 }
 #if MACH_KDP
 unsigned kdp_has_active_watchpoints = 0;
+#define NO_WATCHPOINTS (!kdp_has_active_watchpoints)
+#else
+#define NO_WATCHPOINTS 1
 #endif
 /*
  * Trap from kernel mode.  Only page-fault errors are recoverable,
  * and then only in special circumstances.  All other errors are
  * fatal.  Return value indicates if trap was handled.
  */
 #endif
 /*
  * Trap from kernel mode.  Only page-fault errors are recoverable,
  * and then only in special circumstances.  All other errors are
  * fatal.  Return value indicates if trap was handled.
  */
+
 void
 kernel_trap(
 void
 kernel_trap(
-       x86_saved_state_t       *state)
+       x86_saved_state_t       *state,
+       uintptr_t *lo_spp)
 {
 {
+#ifdef __i386__
        x86_saved_state32_t     *saved_state;
        x86_saved_state32_t     *saved_state;
+#else
+       x86_saved_state64_t     *saved_state;
+#endif
        int                     code;
        user_addr_t             vaddr;
        int                     type;
        int                     code;
        user_addr_t             vaddr;
        int                     type;
-       vm_map_t                map;
+       vm_map_t                map = 0;        /* protected by T_PAGE_FAULT */
        kern_return_t           result = KERN_FAILURE;
        thread_t                thread;
        ast_t                   *myast;
        kern_return_t           result = KERN_FAILURE;
        thread_t                thread;
        ast_t                   *myast;
@@ -313,55 +479,87 @@ kernel_trap(
        vm_prot_t               prot;
         struct recovery                *rp;
        vm_offset_t             kern_ip;
        vm_prot_t               prot;
         struct recovery                *rp;
        vm_offset_t             kern_ip;
+#if NCOPY_WINDOWS > 0
        int                     fault_in_copy_window = -1;
        int                     fault_in_copy_window = -1;
+#endif
        int                     is_user = 0;
        int                     is_user = 0;
-#if MACH_KDB   
-       pt_entry_t              *pte;
-#endif /* MACH_KDB */
-
+       
        thread = current_thread();
 
        thread = current_thread();
 
-       if (is_saved_state64(state))
-               panic("kernel_trap(%p) with 64-bit state", state);
+#ifdef __i386__
+       if (__improbable(is_saved_state64(state))) {
+               panic_64(state, 0, "Kernel trap with 64-bit state", FALSE);
+       }
+
        saved_state = saved_state32(state);
 
        saved_state = saved_state32(state);
 
+       /* Record cpu where state was captured (trampolines don't set this) */
+       saved_state->cpu = cpu_number();
+
        vaddr = (user_addr_t)saved_state->cr2;
        type  = saved_state->trapno;
        code  = saved_state->err & 0xffff;
        intr  = (saved_state->efl & EFL_IF) != 0;       /* state of ints at trap */
        vaddr = (user_addr_t)saved_state->cr2;
        type  = saved_state->trapno;
        code  = saved_state->err & 0xffff;
        intr  = (saved_state->efl & EFL_IF) != 0;       /* state of ints at trap */
-
        kern_ip = (vm_offset_t)saved_state->eip;
        kern_ip = (vm_offset_t)saved_state->eip;
+#else
+       if (__improbable(is_saved_state32(state)))
+               panic("kernel_trap(%p) with 32-bit state", state);
+       saved_state = saved_state64(state);
+
+       /* Record cpu where state was captured */
+       saved_state->isf.cpu = cpu_number();
+
+       vaddr = (user_addr_t)saved_state->cr2;
+       type  = saved_state->isf.trapno;
+       code  = (int)(saved_state->isf.err & 0xffff);
+       intr  = (saved_state->isf.rflags & EFL_IF) != 0;        /* state of ints at trap */
+       kern_ip = (vm_offset_t)saved_state->isf.rip;
+#endif
 
        myast = ast_pending();
 
 
        myast = ast_pending();
 
-       if (perfASTHook) {
+       perfASTCallback astfn = perfASTHook;
+       if (__improbable(astfn != NULL)) {
                if (*myast & AST_CHUD_ALL)
                if (*myast & AST_CHUD_ALL)
-                       perfASTHook(type, NULL, 0, 0);
+                       astfn(AST_CHUD_ALL, myast);
        } else
                *myast &= ~AST_CHUD_ALL;
 
        /*
         * Is there a hook?
         */
        } else
                *myast &= ~AST_CHUD_ALL;
 
        /*
         * Is there a hook?
         */
-       if (perfTrapHook) {
-               if (perfTrapHook(type, NULL, 0, 0) == KERN_SUCCESS) {
+       perfCallback fn = perfTrapHook;
+       if (__improbable(fn != NULL)) {
+               if (fn(type, NULL, 0, 0) == KERN_SUCCESS) {
                        /*
                         * If it succeeds, we are done...
                         */
                        return;
                }
        }
                        /*
                         * If it succeeds, we are done...
                         */
                        return;
                }
        }
+
+#if CONFIG_DTRACE
+       if (__improbable(tempDTraceTrapHook != NULL)) {
+               if (tempDTraceTrapHook(type, state, lo_spp, 0) == KERN_SUCCESS) {
+                       /*
+                        * If it succeeds, we are done...
+                        */
+                       return;
+               }
+       }
+#endif /* CONFIG_DTRACE */
+
        /*
         * we come here with interrupts off as we don't want to recurse
         * on preemption below.  but we do want to re-enable interrupts
         * as soon we possibly can to hold latency down
         */
        /*
         * we come here with interrupts off as we don't want to recurse
         * on preemption below.  but we do want to re-enable interrupts
         * as soon we possibly can to hold latency down
         */
-       if (T_PREEMPT == type) {
-
-               KERNEL_DEBUG_CONSTANT((MACHDBG_CODE(DBG_MACH_EXCP_KTRAP_x86, type)) | DBG_FUNC_NONE,
-                                     0, 0, 0, kern_ip, 0);
-
+       if (__improbable(T_PREEMPT == type)) {
                ast_taken(AST_PREEMPTION, FALSE);
                ast_taken(AST_PREEMPTION, FALSE);
+
+               KERNEL_DEBUG_CONSTANT_IST(KDEBUG_TRACE, 
+                       (MACHDBG_CODE(DBG_MACH_EXCP_KTRAP_x86, type)) | DBG_FUNC_NONE,
+                       0, 0, 0, VM_KERNEL_UNSLIDE(kern_ip), 0);
                return;
        }
        
                return;
        }
        
@@ -371,24 +569,25 @@ kernel_trap(
                 */
                map = kernel_map;
 
                 */
                map = kernel_map;
 
-               if (thread != THREAD_NULL && thread->map != kernel_map) {
-                       vm_offset_t     copy_window_base;
+               if (__probable(thread != THREAD_NULL && thread->map != kernel_map)) {
+#if NCOPY_WINDOWS > 0
+                       vm_offset_t     copy_window_base;
                        vm_offset_t     kvaddr;
                        int             window_index;
 
                        kvaddr = (vm_offset_t)vaddr;
                        vm_offset_t     kvaddr;
                        int             window_index;
 
                        kvaddr = (vm_offset_t)vaddr;
-                       /*
+                       /*
                         * must determine if fault occurred in
                         * the copy window while pre-emption is
                         * disabled for this processor so that
                         * we only need to look at the window
                         * associated with this processor
                         */
                         * must determine if fault occurred in
                         * the copy window while pre-emption is
                         * disabled for this processor so that
                         * we only need to look at the window
                         * associated with this processor
                         */
-                       copy_window_base = current_cpu_datap()->cpu_copywindow_base;
+                       copy_window_base = current_cpu_datap()->cpu_copywindow_base;
 
                        if (kvaddr >= copy_window_base && kvaddr < (copy_window_base + (NBPDE * NCOPY_WINDOWS)) ) {
 
 
                        if (kvaddr >= copy_window_base && kvaddr < (copy_window_base + (NBPDE * NCOPY_WINDOWS)) ) {
 
-                               window_index = (kvaddr - copy_window_base) / NBPDE;
+                               window_index = (int)((kvaddr - copy_window_base) / NBPDE);
 
                                if (thread->machine.copy_window[window_index].user_base != (user_addr_t)-1) {
 
 
                                if (thread->machine.copy_window[window_index].user_base != (user_addr_t)-1) {
 
@@ -400,10 +599,44 @@ kernel_trap(
                                }
                                is_user = -1;
                        }
                                }
                                is_user = -1;
                        }
+#else
+                       if (__probable(vaddr < VM_MAX_USER_PAGE_ADDRESS)) {
+                               /* fault occurred in userspace */
+                               map = thread->map;
+                               is_user = -1;
+
+                               /* Intercept a potential Supervisor Mode Execute
+                                * Protection fault. These criteria identify
+                                * both NX faults and SMEP faults, but both
+                                * are fatal. We avoid checking PTEs (racy).
+                                * (The VM could just redrive a SMEP fault, hence
+                                * the intercept).
+                                */
+                               if (__improbable((code == (T_PF_PROT | T_PF_EXECUTE)) && (pmap_smep_enabled) && (saved_state->isf.rip == vaddr))) {
+                                       goto debugger_entry;
+                               }
+
+                               /*
+                                * If we're not sharing cr3 with the user
+                                * and we faulted in copyio,
+                                * then switch cr3 here and dismiss the fault.
+                                */
+                               if (no_shared_cr3 &&
+                                   (thread->machine.specFlags&CopyIOActive) &&
+                                   map->pmap->pm_cr3 != get_cr3_base()) {
+                                       pmap_assert(current_cpu_datap()->cpu_pmap_pcid_enabled == FALSE);
+                                       set_cr3_raw(map->pmap->pm_cr3);
+                                       return;
+                               }
+                       }
+#endif
                }
        }
                }
        }
-       KERNEL_DEBUG_CONSTANT((MACHDBG_CODE(DBG_MACH_EXCP_KTRAP_x86, type)) | DBG_FUNC_NONE,
-                             (int)(vaddr >> 32), (int)vaddr, is_user, kern_ip, 0);
+       user_addr_t     kd_vaddr = is_user ? vaddr : VM_KERNEL_UNSLIDE(vaddr);  
+       KERNEL_DEBUG_CONSTANT_IST(KDEBUG_TRACE, 
+               (MACHDBG_CODE(DBG_MACH_EXCP_KTRAP_x86, type)) | DBG_FUNC_NONE,
+               (unsigned)(kd_vaddr >> 32), (unsigned)kd_vaddr, is_user,
+               VM_KERNEL_UNSLIDE(kern_ip), 0);
 
 
        (void) ml_set_interrupts_enabled(intr);
 
 
        (void) ml_set_interrupts_enabled(intr);
@@ -426,8 +659,12 @@ kernel_trap(
                fpSSEexterrflt();
                return;
            case T_DEBUG:
                fpSSEexterrflt();
                return;
            case T_DEBUG:
-                   if ((saved_state->efl & EFL_TF) == 0
-                   && !kdp_has_active_watchpoints) {
+#ifdef __i386__
+                   if ((saved_state->efl & EFL_TF) == 0 && NO_WATCHPOINTS)
+#else
+                   if ((saved_state->isf.rflags & EFL_TF) == 0 && NO_WATCHPOINTS)
+#endif
+                   {
                            /* We've somehow encountered a debug
                             * register match that does not belong
                             * to the kernel debugger.
                            /* We've somehow encountered a debug
                             * register match that does not belong
                             * to the kernel debugger.
@@ -435,17 +672,27 @@ kernel_trap(
                             */
                            reset_dr7();
                            return;
                             */
                            reset_dr7();
                            return;
-                           }
+                   }
                    goto debugger_entry;
                    goto debugger_entry;
+#ifdef __x86_64__
+           case T_INT3:
+             goto debugger_entry;
+#endif
            case T_PAGE_FAULT:
            case T_PAGE_FAULT:
-               /*
-                * If the current map is a submap of the kernel map,
-                * and the address is within that map, fault on that
-                * map.  If the same check is done in vm_fault
-                * (vm_map_lookup), we may deadlock on the kernel map
-                * lock.
-                */
 
 
+#if CONFIG_DTRACE
+               if (thread != THREAD_NULL && thread->options & TH_OPT_DTRACE) { /* Executing under dtrace_probe? */
+                       if (dtrace_tally_fault(vaddr)) { /* Should a fault under dtrace be ignored? */
+                               /*
+                                * DTrace has "anticipated" the possibility of this fault, and has
+                                * established the suitable recovery state. Drop down now into the
+                                * recovery handling code in "case T_GENERAL_PROTECTION:". 
+                                */
+                               goto FALL_THROUGH;
+                       }
+               }
+#endif /* CONFIG_DTRACE */
+               
                prot = VM_PROT_READ;
 
                if (code & T_PF_WRITE)
                prot = VM_PROT_READ;
 
                if (code & T_PF_WRITE)
@@ -455,79 +702,29 @@ kernel_trap(
                        prot |= VM_PROT_EXECUTE;
 #endif
 
                        prot |= VM_PROT_EXECUTE;
 #endif
 
-#if    MACH_KDB
-               /*
-                * Check for watchpoint on kernel static data.
-                * vm_fault would fail in this case 
-                */
-               if (map == kernel_map && db_watchpoint_list && db_watchpoints_inserted &&
-                   (code & T_PF_WRITE) && vaddr < vm_map_max(map) &&
-                   ((*(pte = pmap_pte(kernel_pmap, (vm_map_offset_t)vaddr))) & INTEL_PTE_WRITE) == 0) {
-                       pmap_store_pte(
-                               pte,
-                               *pte | INTEL_PTE_VALID | INTEL_PTE_WRITE);
-                       /* XXX need invltlb here? */
-
-                       result = KERN_SUCCESS;
-                       goto look_for_watchpoints;
-               }
-#endif /* MACH_KDB */
-
                result = vm_fault(map,
                                  vm_map_trunc_page(vaddr),
                                  prot,
                                  FALSE, 
                                  THREAD_UNINT, NULL, 0);
 
                result = vm_fault(map,
                                  vm_map_trunc_page(vaddr),
                                  prot,
                                  FALSE, 
                                  THREAD_UNINT, NULL, 0);
 
-#if    MACH_KDB
-               if (result == KERN_SUCCESS) {
-                       /*
-                        * Look for watchpoints
-                        */
-look_for_watchpoints:
-                       if (map == kernel_map && db_watchpoint_list && db_watchpoints_inserted && (code & T_PF_WRITE) &&
-                           db_find_watchpoint(map, vaddr, saved_state))
-                               kdb_trap(T_WATCHPOINT, 0, saved_state);
-               }
-#endif /* MACH_KDB */
-
                if (result == KERN_SUCCESS) {
                if (result == KERN_SUCCESS) {
-
+#if NCOPY_WINDOWS > 0
                        if (fault_in_copy_window != -1) {
                        if (fault_in_copy_window != -1) {
-                               pt_entry_t      *updp;
-                               pt_entry_t      *kpdp;
-
-                               /*
-                                * in case there was no page table assigned
-                                * for the user base address and the pmap
-                                * got 'expanded' due to this fault, we'll
-                                * copy in the descriptor 
-                                *
-                                * we're either setting the page table descriptor
-                                * to the same value or it was 0... no need
-                                * for a TLB flush in either case
-                                */
-
-                               ml_set_interrupts_enabled(FALSE);
-                               updp = pmap_pde(map->pmap, thread->machine.copy_window[fault_in_copy_window].user_base);
-                               assert(updp);
-                               if (0 == updp) panic("trap: updp 0"); /* XXX DEBUG */
-                               kpdp = current_cpu_datap()->cpu_copywindow_pdp;
-                               kpdp += fault_in_copy_window;
-
-#if JOE_DEBUG
-                               if (*kpdp && (*kpdp & PG_FRAME) != (*updp & PG_FRAME))
-                                       panic("kernel_fault: user pdp doesn't match - updp = 0x%x, kpdp = 0x%x\n", updp, kpdp);
-#endif
-                               pmap_store_pte(kpdp, *updp);
-
+                               ml_set_interrupts_enabled(FALSE);
+                               copy_window_fault(thread, map,
+                                                 fault_in_copy_window);
                                (void) ml_set_interrupts_enabled(intr);
                        }
                                (void) ml_set_interrupts_enabled(intr);
                        }
+#endif /* NCOPY_WINDOWS > 0 */
                        return;
                }
                /*
                 * fall through
                 */
                        return;
                }
                /*
                 * fall through
                 */
+#if CONFIG_DTRACE
+FALL_THROUGH:
+#endif /* CONFIG_DTRACE */
 
            case T_GENERAL_PROTECTION:
                /*
 
            case T_GENERAL_PROTECTION:
                /*
@@ -544,8 +741,8 @@ look_for_watchpoints:
                /*
                 * Check thread recovery address also.
                 */
                /*
                 * Check thread recovery address also.
                 */
-               if (thread->recover) {
-                       set_recovery_ip(saved_state, thread->recover);
+               if (thread != THREAD_NULL && thread->recover) {
+                       set_recovery_ip(saved_state, thread->recover);
                        thread->recover = 0;
                        return;
                }
                        thread->recover = 0;
                        return;
                }
@@ -555,7 +752,6 @@ look_for_watchpoints:
                 *
                 * fall through...
                 */
                 *
                 * fall through...
                 */
-
            default:
                /*
                 * Exception 15 is reserved but some chips may generate it
            default:
                /*
                 * Exception 15 is reserved but some chips may generate it
@@ -571,29 +767,15 @@ debugger_entry:
                 * context at the moment of the trap, to facilitate
                 * access through the debugger.
                 */
                 * context at the moment of the trap, to facilitate
                 * access through the debugger.
                 */
-               sync_iss_to_iks(saved_state);
-#if MACH_KDB
-restart_debugger:
-#endif /* MACH_KDB */          
+               sync_iss_to_iks(state);
 #if  MACH_KDP
 #if  MACH_KDP
-                if (current_debugger != KDB_CUR_DB) {
-                       if (kdp_i386_trap(type, saved_state, result, vaddr))
+               if (current_debugger != KDB_CUR_DB) {
+                       if (kdp_i386_trap(type, saved_state, result, (vm_offset_t)vaddr))
                                return;
                }
                                return;
                }
-#endif /* MACH_KDP */
-#if MACH_KDB
-               else
-                       if (kdb_trap(type, code, saved_state)) {
-                               if (switch_debugger) {
-                                       current_debugger = KDP_CUR_DB;
-                                       switch_debugger = 0;
-                                       goto restart_debugger;
-                               }
-                               return;
-                       }
-#endif /* MACH_KDB */
+#endif
        }
        }
-
+       pal_cli();
        panic_trap(saved_state);
        /*
         * NO RETURN
        panic_trap(saved_state);
        /*
         * NO RETURN
@@ -601,24 +783,35 @@ restart_debugger:
 }
 
 
 }
 
 
+#ifdef __i386__
 static void
 set_recovery_ip(x86_saved_state32_t  *saved_state, vm_offset_t ip)
 {
         saved_state->eip = ip;
 }
 static void
 set_recovery_ip(x86_saved_state32_t  *saved_state, vm_offset_t ip)
 {
         saved_state->eip = ip;
 }
+#else
+static void
+set_recovery_ip(x86_saved_state64_t  *saved_state, vm_offset_t ip)
+{
+        saved_state->isf.rip = ip;
+}
+#endif
 
 
 
 
+#ifdef __i386__
 static void
 panic_trap(x86_saved_state32_t *regs)
 {
        const char *trapname = "Unknown";
 static void
 panic_trap(x86_saved_state32_t *regs)
 {
        const char *trapname = "Unknown";
-       uint32_t        cr0 = get_cr0();
-       uint32_t        cr2 = get_cr2();
-       uint32_t        cr3 = get_cr3();
-       uint32_t        cr4 = get_cr4();
+       pal_cr_t        cr0, cr2, cr3, cr4;
 
 
-       if (panic_io_port)
-         (void)inb(panic_io_port);
+       pal_get_control_registers( &cr0, &cr2, &cr3, &cr4 );
+
+       /*
+        * Issue an I/O port read if one has been requested - this is an
+        * event logic analyzers can use as a trigger point.
+        */
+       panic_io_port_read();
 
        kprintf("panic trap number 0x%x, eip 0x%x\n", regs->trapno, regs->eip);
        kprintf("cr0 0x%08x cr2 0x%08x cr3 0x%08x cr4 0x%08x\n",
 
        kprintf("panic trap number 0x%x, eip 0x%x\n", regs->trapno, regs->eip);
        kprintf("cr0 0x%08x cr2 0x%08x cr3 0x%08x cr4 0x%08x\n",
@@ -626,16 +819,18 @@ panic_trap(x86_saved_state32_t *regs)
 
        if (regs->trapno < TRAP_TYPES)
                trapname = trap_type[regs->trapno];
 
        if (regs->trapno < TRAP_TYPES)
                trapname = trap_type[regs->trapno];
-
-       panic("Unresolved kernel trap (CPU %d, Type %d=%s), registers:\n"
+#undef panic
+       panic("Kernel trap at 0x%08x, type %d=%s, registers:\n"
              "CR0: 0x%08x, CR2: 0x%08x, CR3: 0x%08x, CR4: 0x%08x\n"
              "EAX: 0x%08x, EBX: 0x%08x, ECX: 0x%08x, EDX: 0x%08x\n"
              "CR2: 0x%08x, EBP: 0x%08x, ESI: 0x%08x, EDI: 0x%08x\n"
              "CR0: 0x%08x, CR2: 0x%08x, CR3: 0x%08x, CR4: 0x%08x\n"
              "EAX: 0x%08x, EBX: 0x%08x, ECX: 0x%08x, EDX: 0x%08x\n"
              "CR2: 0x%08x, EBP: 0x%08x, ESI: 0x%08x, EDI: 0x%08x\n"
-             "EFL: 0x%08x, EIP: 0x%08x, CS:  0x%08x, DS:  0x%08x\n",
-             cpu_number(), regs->trapno, trapname, cr0, cr2, cr3, cr4,
+             "EFL: 0x%08x, EIP: 0x%08x, CS:  0x%08x, DS:  0x%08x\n"
+             "Error code: 0x%08x%s\n",
+             regs->eip, regs->trapno, trapname, cr0, cr2, cr3, cr4,
              regs->eax,regs->ebx,regs->ecx,regs->edx,
              regs->cr2,regs->ebp,regs->esi,regs->edi,
              regs->eax,regs->ebx,regs->ecx,regs->edx,
              regs->cr2,regs->ebp,regs->esi,regs->edi,
-             regs->efl,regs->eip,regs->cs, regs->ds);
+             regs->efl,regs->eip,regs->cs & 0xFFFF, regs->ds & 0xFFFF, regs->err,
+             virtualized ? " VMM" : "");
        /*
         * This next statement is not executed,
         * but it's needed to stop the compiler using tail call optimization
        /*
         * This next statement is not executed,
         * but it's needed to stop the compiler using tail call optimization
@@ -643,187 +838,73 @@ panic_trap(x86_saved_state32_t *regs)
         */
        cr0 = 0;
 }
         */
        cr0 = 0;
 }
+#else
 
 
-extern void     kprintf_break_lock(void);
 
 
-
-/*
- * Called from locore on a special reserved stack after a double-fault
- * is taken in kernel space.
- * Kernel stack overflow is one route here.
- */
-void
-panic_double_fault(int code)
-{
-       struct i386_tss *my_ktss = current_ktss();
-
-       /* Set postcode (DEBUG only) */
-       postcode(PANIC_DOUBLE_FAULT);
-
-/* Issue an I/O port read if one has been requested - this is an event logic
- * analyzers can use as a trigger point.
- */
-       if (panic_io_port)
-               (void)inb(panic_io_port);
-
-       /*
-        * Break kprintf lock in case of recursion,
-        * and record originally faulted instruction address.
-        */
-       kprintf_break_lock();
-
-#if MACH_KDP
-       /*
-        * Print backtrace leading to first fault:
-        */
-       panic_i386_backtrace((void *) my_ktss->ebp, 10);
-#endif
-
-       panic("Double fault (CPU:%d, thread:%p, code:0x%x),"
-             "registers:\n"
-             "CR0: 0x%08x, CR2: 0x%08x, CR3: 0x%08x, CR4: 0x%08x\n"
-             "EAX: 0x%08x, EBX: 0x%08x, ECX: 0x%08x, EDX: 0x%08x\n"
-             "ESP: 0x%08x, EBP: 0x%08x, ESI: 0x%08x, EDI: 0x%08x\n"
-             "EFL: 0x%08x, EIP: 0x%08x\n",
-             cpu_number(), current_thread(), code,
-             get_cr0(), get_cr2(), get_cr3(), get_cr4(),
-             my_ktss->eax, my_ktss->ebx, my_ktss->ecx, my_ktss->edx,
-             my_ktss->esp, my_ktss->ebp, my_ktss->esi, my_ktss->edi,
-             my_ktss->eflags, my_ktss->eip);
-}
-
-
-/*
- * Called from locore on a special reserved stack after a machine-check
- */
-void
-panic_machine_check(int code)
+static void
+panic_trap(x86_saved_state64_t *regs)
 {
 {
-       struct i386_tss *my_ktss = current_ktss();
-
-       /* Set postcode (DEBUG only) */
-       postcode(PANIC_MACHINE_CHECK);
+       const char      *trapname = "Unknown";
+       pal_cr_t        cr0, cr2, cr3, cr4;
+       boolean_t       potential_smep_fault = FALSE, potential_kernel_NX_fault = FALSE;
 
 
+       pal_get_control_registers( &cr0, &cr2, &cr3, &cr4 );
+       assert(ml_get_interrupts_enabled() == FALSE);
+       current_cpu_datap()->cpu_fatal_trap_state = regs;
        /*
        /*
-        * Break kprintf lock in case of recursion,
-        * and record originally faulted instruction address.
+        * Issue an I/O port read if one has been requested - this is an
+        * event logic analyzers can use as a trigger point.
         */
         */
-       kprintf_break_lock();
-       panic("Machine-check (CPU:%d, thread:%p, code:0x%x),"
-             "registers:\n"
-             "CR0: 0x%08x, CR2: 0x%08x, CR3: 0x%08x, CR4: 0x%08x\n"
-             "EAX: 0x%08x, EBX: 0x%08x, ECX: 0x%08x, EDX: 0x%08x\n"
-             "ESP: 0x%08x, EBP: 0x%08x, ESI: 0x%08x, EDI: 0x%08x\n"
-             "EFL: 0x%08x, EIP: 0x%08x\n",
-             cpu_number(), current_thread(), code,
-             get_cr0(), get_cr2(), get_cr3(), get_cr4(),
-             my_ktss->eax, my_ktss->ebx, my_ktss->ecx, my_ktss->edx,
-             my_ktss->esp, my_ktss->ebp, my_ktss->esi, my_ktss->edi,
-             my_ktss->eflags, my_ktss->eip);
-}
+       panic_io_port_read();
 
 
-void
-panic_double_fault64(x86_saved_state_t *esp)
-{
-       /* Set postcode (DEBUG only) */
-       postcode(PANIC_DOUBLE_FAULT);
+       kprintf("panic trap number 0x%x, rip 0x%016llx\n",
+               regs->isf.trapno, regs->isf.rip);
+       kprintf("cr0 0x%016llx cr2 0x%016llx cr3 0x%016llx cr4 0x%016llx\n",
+               cr0, cr2, cr3, cr4);
 
 
-       /*
-        * Break kprintf lock in case of recursion,
-        * and record originally faulted instruction address.
-        */
-       kprintf_break_lock();
+       if (regs->isf.trapno < TRAP_TYPES)
+               trapname = trap_type[regs->isf.trapno];
 
 
-       /*
-        * Dump the interrupt stack frame at last kernel entry.
-        */
-       if (is_saved_state64(esp)) {
-               x86_saved_state64_t     *ss64p = saved_state64(esp);
-               panic("Double fault (CPU:%d, thread:%p, trapno:0x%x, err:0x%qx),"
-                     "registers:\n"
-                     "CR0: 0x%08x, CR2: 0x%08x, CR3: 0x%08x, CR4: 0x%08x\n"
-                     "RAX: 0x%016qx, RBX: 0x%016qx, RCX: 0x%016qx, RDX: 0x%016qx\n"
-                     "RSP: 0x%016qx, RBP: 0x%016qx, RSI: 0x%016qx, RDI: 0x%016qx\n"
-                     "R8:  0x%016qx, R9:  0x%016qx, R10: 0x%016qx, R11: 0x%016qx\n"
-                     "R12: 0x%016qx, R13: 0x%016qx, R14: 0x%016qx, R15: 0x%016qx\n"
-                     "RFL: 0x%016qx, RIP: 0x%016qx\n",
-                     cpu_number(), current_thread(), ss64p->isf.trapno, ss64p->isf.err,
-                     get_cr0(), get_cr2(), get_cr3(), get_cr4(),
-                     ss64p->rax, ss64p->rbx, ss64p->rcx, ss64p->rdx,
-                     ss64p->isf.rsp, ss64p->rbp, ss64p->rsi, ss64p->rdi,
-                     ss64p->r8, ss64p->r9, ss64p->r10, ss64p->r11,
-                     ss64p->r12, ss64p->r13, ss64p->r14, ss64p->r15,
-                     ss64p->isf.rflags, ss64p->isf.rip);
-       } else {
-               x86_saved_state32_t     *ss32p = saved_state32(esp);
-               panic("Double fault (CPU:%d, thread:%p, trapno:0x%x, err:0x%x),"
-                     "registers:\n"
-                     "CR0: 0x%08x, CR2: 0x%08x, CR3: 0x%08x, CR4: 0x%08x\n"
-                     "EAX: 0x%08x, EBX: 0x%08x, ECX: 0x%08x, EDX: 0x%08x\n"
-                     "ESP: 0x%08x, EBP: 0x%08x, ESI: 0x%08x, EDI: 0x%08x\n"
-                     "EFL: 0x%08x, EIP: 0x%08x\n",
-                     cpu_number(), current_thread(), ss32p->trapno, ss32p->err,
-                     get_cr0(), get_cr2(), get_cr3(), get_cr4(),
-                     ss32p->eax, ss32p->ebx, ss32p->ecx, ss32p->edx,
-                     ss32p->uesp, ss32p->ebp, ss32p->esi, ss32p->edi,
-                     ss32p->efl, ss32p->eip);
+       if ((regs->isf.trapno == T_PAGE_FAULT) && (regs->isf.err == (T_PF_PROT | T_PF_EXECUTE)) && (regs->isf.rip == regs->cr2)) {
+               if (pmap_smep_enabled && (regs->isf.rip < VM_MAX_USER_PAGE_ADDRESS)) {
+                       potential_smep_fault = TRUE;
+               } else if (regs->isf.rip >= VM_MIN_KERNEL_AND_KEXT_ADDRESS) {
+                       potential_kernel_NX_fault = TRUE;
+               }
        }
        }
-}
-
-/*
- * Simplistic machine check handler.
- * We could peruse all those MSRs but we only dump register state as we do for
- * the double fault exception.
- * Note: the machine check registers are non-volatile across warm boot - so
- * they'll be around when we return.
- */
-void
-panic_machine_check64(x86_saved_state_t *esp)
-{
-       /* Set postcode (DEBUG only) */
-       postcode(PANIC_MACHINE_CHECK);
-
-       /*
-        * Break kprintf lock in case of recursion,
-        * and record originally faulted instruction address.
-        */
-       kprintf_break_lock();
 
 
+#undef panic
+       panic("Kernel trap at 0x%016llx, type %d=%s, registers:\n"
+             "CR0: 0x%016llx, CR2: 0x%016llx, CR3: 0x%016llx, CR4: 0x%016llx\n"
+             "RAX: 0x%016llx, RBX: 0x%016llx, RCX: 0x%016llx, RDX: 0x%016llx\n"
+             "RSP: 0x%016llx, RBP: 0x%016llx, RSI: 0x%016llx, RDI: 0x%016llx\n"
+             "R8:  0x%016llx, R9:  0x%016llx, R10: 0x%016llx, R11: 0x%016llx\n"
+             "R12: 0x%016llx, R13: 0x%016llx, R14: 0x%016llx, R15: 0x%016llx\n"
+             "RFL: 0x%016llx, RIP: 0x%016llx, CS:  0x%016llx, SS:  0x%016llx\n"
+             "Fault CR2: 0x%016llx, Error code: 0x%016llx, Fault CPU: 0x%x%s%s%s\n",
+             regs->isf.rip, regs->isf.trapno, trapname,
+             cr0, cr2, cr3, cr4,
+             regs->rax, regs->rbx, regs->rcx, regs->rdx,
+             regs->isf.rsp, regs->rbp, regs->rsi, regs->rdi,
+             regs->r8,  regs->r9,  regs->r10, regs->r11,
+             regs->r12, regs->r13, regs->r14, regs->r15,
+             regs->isf.rflags, regs->isf.rip, regs->isf.cs & 0xFFFF,
+             regs->isf.ss & 0xFFFF,regs->cr2, regs->isf.err, regs->isf.cpu,
+             virtualized ? " VMM" : "",
+             potential_kernel_NX_fault ? " Kernel NX fault" : "",
+             potential_smep_fault ? " SMEP/User NX fault" : "");
        /*
        /*
-        * Dump the interrupt stack frame at last kernel entry.
+        * This next statement is not executed,
+        * but it's needed to stop the compiler using tail call optimization
+        * for the panic call - which confuses the subsequent backtrace.
         */
         */
-       if (is_saved_state64(esp)) {
-               x86_saved_state64_t     *ss64p = saved_state64(esp);
-               panic("Machine Check (CPU:%d, thread:%p, trapno:0x%x, err:0x%qx),"
-                     "registers:\n"
-                     "CR0: 0x%08x, CR2: 0x%08x, CR3: 0x%08x, CR4: 0x%08x\n"
-                     "RAX: 0x%016qx, RBX: 0x%016qx, RCX: 0x%016qx, RDX: 0x%016qx\n"
-                     "RSP: 0x%016qx, RBP: 0x%016qx, RSI: 0x%016qx, RDI: 0x%016qx\n"
-                     "R8:  0x%016qx, R9:  0x%016qx, R10: 0x%016qx, R11: 0x%016qx\n"
-                     "R12: 0x%016qx, R13: 0x%016qx, R14: 0x%016qx, R15: 0x%016qx\n"
-                     "RFL: 0x%016qx, RIP: 0x%016qx\n",
-                     cpu_number(), current_thread(), ss64p->isf.trapno, ss64p->isf.err,
-                     get_cr0(), get_cr2(), get_cr3(), get_cr4(),
-                     ss64p->rax, ss64p->rbx, ss64p->rcx, ss64p->rdx,
-                     ss64p->isf.rsp, ss64p->rbp, ss64p->rsi, ss64p->rdi,
-                     ss64p->r8, ss64p->r9, ss64p->r10, ss64p->r11,
-                     ss64p->r12, ss64p->r13, ss64p->r14, ss64p->r15,
-                     ss64p->isf.rflags, ss64p->isf.rip);
-       } else {
-               x86_saved_state32_t     *ss32p = saved_state32(esp);
-               panic("Machine Check (CPU:%d, thread:%p, trapno:0x%x, err:0x%x),"
-                     "registers:\n"
-                     "CR0: 0x%08x, CR2: 0x%08x, CR3: 0x%08x, CR4: 0x%08x\n"
-                     "EAX: 0x%08x, EBX: 0x%08x, ECX: 0x%08x, EDX: 0x%08x\n"
-                     "ESP: 0x%08x, EBP: 0x%08x, ESI: 0x%08x, EDI: 0x%08x\n"
-                     "EFL: 0x%08x, EIP: 0x%08x\n",
-                     cpu_number(), current_thread(), ss32p->trapno, ss32p->err,
-                     get_cr0(), get_cr2(), get_cr3(), get_cr4(),
-                     ss32p->eax, ss32p->ebx, ss32p->ecx, ss32p->edx,
-                     ss32p->uesp, ss32p->ebp, ss32p->esi, ss32p->edi,
-                     ss32p->efl, ss32p->eip);
-       }
+       cr0 = 0;
 }
 }
+#endif
+
+#if CONFIG_DTRACE
+extern kern_return_t dtrace_user_probe(x86_saved_state_t *);
+#endif
 
 /*
  *     Trap from user mode.
 
 /*
  *     Trap from user mode.
@@ -832,18 +913,18 @@ void
 user_trap(
        x86_saved_state_t *saved_state)
 {
 user_trap(
        x86_saved_state_t *saved_state)
 {
-       int             exc;
-       int             code;
-       int             err;
-       unsigned int    subcode;
-       int             type;
-       user_addr_t     vaddr;
-       vm_prot_t       prot;
-       thread_t        thread = current_thread();
-       ast_t           *myast;
-       boolean_t       intr;
-       kern_return_t   kret;
-       user_addr_t     rip;
+       int                     exc;
+       int                     err;
+       mach_exception_code_t   code;
+       mach_exception_subcode_t subcode;
+       int                     type;
+       user_addr_t             vaddr;
+       vm_prot_t               prot;
+       thread_t                thread = current_thread();
+       ast_t                   *myast;
+       kern_return_t           kret;
+       user_addr_t             rip;
+       unsigned long           dr6 = 0; /* 32 bit for i386, 64 bit for x86_64 */
 
        assert((is_saved_state32(saved_state) && !thread_is_64bit(thread)) ||
               (is_saved_state64(saved_state) &&  thread_is_64bit(thread)));
 
        assert((is_saved_state32(saved_state) && !thread_is_64bit(thread)) ||
               (is_saved_state64(saved_state) &&  thread_is_64bit(thread)));
@@ -853,23 +934,42 @@ user_trap(
 
                regs = saved_state64(saved_state);
 
 
                regs = saved_state64(saved_state);
 
+               /* Record cpu where state was captured */
+               regs->isf.cpu = cpu_number();
+
                type = regs->isf.trapno;
                type = regs->isf.trapno;
-               err  = regs->isf.err & 0xffff;
+               err  = (int)regs->isf.err & 0xffff;
                vaddr = (user_addr_t)regs->cr2;
                rip   = (user_addr_t)regs->isf.rip;
        } else {
                vaddr = (user_addr_t)regs->cr2;
                rip   = (user_addr_t)regs->isf.rip;
        } else {
-               x86_saved_state32_t     *regs;
+               x86_saved_state32_t     *regs;
 
                regs = saved_state32(saved_state);
 
 
                regs = saved_state32(saved_state);
 
+               /* Record cpu where state was captured */
+               regs->cpu = cpu_number();
+
                type  = regs->trapno;
                err   = regs->err & 0xffff;
                vaddr = (user_addr_t)regs->cr2;
                rip   = (user_addr_t)regs->eip;
        }
 
                type  = regs->trapno;
                err   = regs->err & 0xffff;
                vaddr = (user_addr_t)regs->cr2;
                rip   = (user_addr_t)regs->eip;
        }
 
-       KERNEL_DEBUG_CONSTANT((MACHDBG_CODE(DBG_MACH_EXCP_UTRAP_x86, type)) | DBG_FUNC_NONE,
-                             (int)(vaddr>>32), (int)vaddr, (int)(rip>>32), (int)rip, 0);
+       if ((type == T_DEBUG) && thread->machine.ids) {
+               unsigned long clear = 0;
+               /* Stash and clear this processor's DR6 value, in the event
+                * this was a debug register match
+                */
+               __asm__ volatile ("mov %%db6, %0" : "=r" (dr6)); 
+               __asm__ volatile ("mov %0, %%db6" : : "r" (clear));
+       }
+
+       pal_sti();
+
+       KERNEL_DEBUG_CONSTANT_IST(KDEBUG_TRACE, 
+               (MACHDBG_CODE(DBG_MACH_EXCP_UTRAP_x86, type)) | DBG_FUNC_NONE,
+               (unsigned)(vaddr>>32), (unsigned)vaddr,
+               (unsigned)(rip>>32), (unsigned)rip, 0);
 
        code = 0;
        subcode = 0;
 
        code = 0;
        subcode = 0;
@@ -879,21 +979,31 @@ user_trap(
        kprintf("user_trap(0x%08x) type=%d vaddr=0x%016llx\n",
                saved_state, type, vaddr);
 #endif
        kprintf("user_trap(0x%08x) type=%d vaddr=0x%016llx\n",
                saved_state, type, vaddr);
 #endif
-       myast = ast_pending();
-       if (perfASTHook) {
+
+       perfASTCallback astfn = perfASTHook;
+       if (__improbable(astfn != NULL)) {
+               myast = ast_pending();
                if (*myast & AST_CHUD_ALL) {
                if (*myast & AST_CHUD_ALL) {
-                       perfASTHook(type, saved_state, 0, 0);
+                       astfn(AST_CHUD_ALL, myast);
                }
                }
-       } else {
-               *myast &= ~AST_CHUD_ALL;
        }
 
        /* Is there a hook? */
        }
 
        /* Is there a hook? */
-       if (perfTrapHook) {
-               if (perfTrapHook(type, saved_state, 0, 0) == KERN_SUCCESS)
+       perfCallback fn = perfTrapHook;
+       if (__improbable(fn != NULL)) {
+               if (fn(type, saved_state, 0, 0) == KERN_SUCCESS)
                        return; /* If it succeeds, we are done... */
        }
 
                        return; /* If it succeeds, we are done... */
        }
 
+       /*
+        * DTrace does not consume all user traps, only INT_3's for now.
+        * Avoid needlessly calling tempDTraceTrapHook here, and let the
+        * INT_3 case handle them.
+        */
+       DEBUG_KPRINT_SYSCALL_MASK(1,
+               "user_trap: type=0x%x(%s) err=0x%x cr2=%p rip=%p\n",
+               type, trap_type[type], err, (void *)(long) vaddr, (void *)(long) rip);
+       
        switch (type) {
 
            case T_DIVIDE_ERROR:
        switch (type) {
 
            case T_DIVIDE_ERROR:
@@ -904,12 +1014,11 @@ user_trap(
            case T_DEBUG:
                {
                        pcb_t   pcb;
            case T_DEBUG:
                {
                        pcb_t   pcb;
-                       unsigned int clear = 0;
                        /*
                        /*
-                        * get dr6 and set it in the thread's pcb before
-                        * returning to userland
+                        * Update the PCB with this processor's DR6 value
+                        * in the event this was a debug register match.
                         */
                         */
-                       pcb = thread->machine.pcb;
+                       pcb = THREAD_TO_PCB(thread);
                        if (pcb->ids) {
                                /*
                                 * We can get and set the status register
                        if (pcb->ids) {
                                /*
                                 * We can get and set the status register
@@ -918,22 +1027,22 @@ user_trap(
                                 * used on x86_64
                                 */
                                if (thread_is_64bit(thread)) {
                                 * used on x86_64
                                 */
                                if (thread_is_64bit(thread)) {
-                                       uint32_t dr6;
                                        x86_debug_state64_t *ids = pcb->ids;
                                        x86_debug_state64_t *ids = pcb->ids;
-                                       dr6 = (uint32_t)ids->dr6;
-                                       __asm__ volatile ("movl %%db6, %0" : "=r" (dr6));
                                        ids->dr6 = dr6;
                                } else { /* 32 bit thread */
                                        x86_debug_state32_t *ids = pcb->ids;
                                        ids->dr6 = dr6;
                                } else { /* 32 bit thread */
                                        x86_debug_state32_t *ids = pcb->ids;
-                                       __asm__ volatile ("movl %%db6, %0" : "=r" (ids->dr6));
+                                       ids->dr6 = (uint32_t) dr6;
                                }
                                }
-                               __asm__ volatile ("movl %0, %%db6" : : "r" (clear));
                        }
                        exc = EXC_BREAKPOINT;
                        code = EXC_I386_SGL;
                        break;
                }
            case T_INT3:
                        }
                        exc = EXC_BREAKPOINT;
                        code = EXC_I386_SGL;
                        break;
                }
            case T_INT3:
+#if CONFIG_DTRACE
+               if (dtrace_user_probe(saved_state) == KERN_SUCCESS)
+                       return; /* If it succeeds, we are done... */
+#endif
                exc = EXC_BREAKPOINT;
                code = EXC_I386_BPT;
                break;
                exc = EXC_BREAKPOINT;
                code = EXC_I386_BPT;
                break;
@@ -954,15 +1063,14 @@ user_trap(
                break;
 
            case T_NO_FPU:
                break;
 
            case T_NO_FPU:
-           case 32:            /* XXX */
                fpnoextflt();
                return;
 
            case T_FPU_FAULT:
                fpnoextflt();
                return;
 
            case T_FPU_FAULT:
-               fpextovrflt();
+               fpextovrflt(); /* Propagates exception directly, doesn't return */
                return;
 
                return;
 
-           case 10:            /* invalid TSS == iret with NT flag set */
+           case T_INVALID_TSS: /* invalid TSS == iret with NT flag set */
                exc = EXC_BAD_INSTRUCTION;
                code = EXC_I386_INVTSSFLT;
                subcode = err;
                exc = EXC_BAD_INSTRUCTION;
                code = EXC_I386_INVTSSFLT;
                subcode = err;
@@ -981,68 +1089,84 @@ user_trap(
                break;
 
            case T_GENERAL_PROTECTION:
                break;
 
            case T_GENERAL_PROTECTION:
-               exc = EXC_BAD_INSTRUCTION;
+               /*
+                * There's a wide range of circumstances which generate this
+                * class of exception. From user-space, many involve bad
+                * addresses (such as a non-canonical 64-bit address).
+                * So we map this to EXC_BAD_ACCESS (and thereby SIGSEGV).
+                * The trouble is cr2 doesn't contain the faulting address;
+                * we'd need to decode the faulting instruction to really
+                * determine this. We'll leave that to debuggers.
+                * However, attempted execution of privileged instructions
+                * (e.g. cli) also generate GP faults and so we map these to
+                * to EXC_BAD_ACCESS (and thence SIGSEGV) also - rather than
+                * EXC_BAD_INSTRUCTION which is more accurate. We just can't
+                * win!
+                */ 
+               exc = EXC_BAD_ACCESS;
                code = EXC_I386_GPFLT;
                subcode = err;
                break;
 
            case T_PAGE_FAULT:
                code = EXC_I386_GPFLT;
                subcode = err;
                break;
 
            case T_PAGE_FAULT:
+           {
                prot = VM_PROT_READ;
 
                if (err & T_PF_WRITE)
                        prot |= VM_PROT_WRITE;
 #if     PAE
                prot = VM_PROT_READ;
 
                if (err & T_PF_WRITE)
                        prot |= VM_PROT_WRITE;
 #if     PAE
-               if (err & T_PF_EXECUTE)
+               if (__improbable(err & T_PF_EXECUTE))
                        prot |= VM_PROT_EXECUTE;
 #endif
                kret = vm_fault(thread->map, vm_map_trunc_page(vaddr),
                                 prot, FALSE,
                                 THREAD_ABORTSAFE, NULL, 0);
 
                        prot |= VM_PROT_EXECUTE;
 #endif
                kret = vm_fault(thread->map, vm_map_trunc_page(vaddr),
                                 prot, FALSE,
                                 THREAD_ABORTSAFE, NULL, 0);
 
-               user_page_fault_continue(kret);
-       
+               if (__probable((kret == KERN_SUCCESS) || (kret == KERN_ABORTED))) {
+                       thread_exception_return();
                /* NOTREACHED */
                /* NOTREACHED */
+               }
+
+               user_page_fault_continue(kret);
+           }   /* NOTREACHED */
                break;
 
            case T_SSE_FLOAT_ERROR:
                break;
 
            case T_SSE_FLOAT_ERROR:
-               fpSSEexterrflt();
+               fpSSEexterrflt(); /* Propagates exception directly, doesn't return */
                return;
 
 
            case T_FLOATING_POINT_ERROR:
                return;
 
 
            case T_FLOATING_POINT_ERROR:
-               fpexterrflt();
+               fpexterrflt(); /* Propagates exception directly, doesn't return */
                return;
 
                return;
 
+           case T_DTRACE_RET:
+#if CONFIG_DTRACE
+               if (dtrace_user_probe(saved_state) == KERN_SUCCESS)
+                       return; /* If it succeeds, we are done... */
+#endif
+               /*
+                * If we get an INT 0x7f when we do not expect to,
+                * treat it as an illegal instruction
+                */
+               exc = EXC_BAD_INSTRUCTION;
+               code = EXC_I386_INVOP;
+               break;
+
            default:
            default:
-#if     MACH_KGDB
-               Debugger("Unanticipated user trap");
-               return;
-#endif  /* MACH_KGDB */
-#if    MACH_KDB
-               if (kdb_trap(type, err, saved_state))
-                   return;
-#endif /* MACH_KDB */
-               panic("user trap");
+               panic("Unexpected user trap, type %d", type);
                return;
        }
                return;
        }
-       intr = ml_set_interrupts_enabled(FALSE);
-       myast = ast_pending();
-       while (*myast & AST_ALL) {
-               ast_taken(AST_ALL, intr);
-               ml_set_interrupts_enabled(FALSE);
-               myast = ast_pending();
-       }
-       ml_set_interrupts_enabled(intr);
-
+       /* Note: Codepaths that directly return from user_trap() have pending
+        * ASTs processed in locore
+        */
        i386_exception(exc, code, subcode);
        i386_exception(exc, code, subcode);
-       /*NOTREACHED*/
+       /* NOTREACHED */
 }
 
 
 /*
  * Handle AST traps for i386.
 }
 
 
 /*
  * Handle AST traps for i386.
- * Check for delayed floating-point exception from
- * AT-bus machines.
  */
 
 extern void     log_thread_action (thread_t, char *);
  */
 
 extern void     log_thread_action (thread_t, char *);
@@ -1076,11 +1200,13 @@ i386_astintr(int preemption)
 void
 i386_exception(
        int     exc,
 void
 i386_exception(
        int     exc,
-       int     code,
-       int     subcode)
+       mach_exception_code_t code,
+       mach_exception_subcode_t subcode)
 {
 {
-       exception_data_type_t   codes[EXCEPTION_CODE_MAX];
+       mach_exception_data_type_t   codes[EXCEPTION_CODE_MAX];
 
 
+       DEBUG_KPRINT_SYSCALL_MACH("i386_exception: exc=%d code=0x%llx subcode=0x%llx\n",
+                                                         exc, code, subcode);
        codes[0] = code;                /* new exception interface */
        codes[1] = subcode;
        exception_triage(exc, codes, 2);
        codes[0] = code;                /* new exception interface */
        codes[1] = subcode;
        exception_triage(exc, codes, 2);
@@ -1088,76 +1214,6 @@ i386_exception(
 }
 
 
 }
 
 
-void
-kernel_preempt_check(void)
-{
-       ast_t           *myast;
-       boolean_t       intr;
-
-       /*
-        * disable interrupts to both prevent pre-emption
-        * and to keep the ast state from changing via
-        * an interrupt handler making something runnable
-        */
-       intr = ml_set_interrupts_enabled(FALSE);
-
-       myast = ast_pending();
-
-       if ((*myast & AST_URGENT) && intr == TRUE && get_interrupt_level() == 0) {
-               /*
-                * can handle interrupts and preemptions 
-                * at this point
-                */
-               ml_set_interrupts_enabled(intr);
-
-               /*
-                * now cause the PRE-EMPTION trap
-                */
-               __asm__ volatile ("     int     $0xff");
-       } else {
-               /*
-                * if interrupts were already disabled or
-                * we're in an interrupt context, we can't
-                * preempt...  of course if AST_URGENT
-                * isn't set we also don't want to
-                */
-               ml_set_interrupts_enabled(intr);
-       }
-}
-
-#if    MACH_KDB
-
-extern void    db_i386_state(x86_saved_state32_t *regs);
-
-#include <ddb/db_output.h>
-
-void 
-db_i386_state(
-       x86_saved_state32_t *regs)
-{
-       db_printf("eip  %8x\n", regs->eip);
-       db_printf("trap %8x\n", regs->trapno);
-       db_printf("err  %8x\n", regs->err);
-       db_printf("efl  %8x\n", regs->efl);
-       db_printf("ebp  %8x\n", regs->ebp);
-       db_printf("esp  %8x\n", regs->cr2);
-       db_printf("uesp %8x\n", regs->uesp);
-       db_printf("cs   %8x\n", regs->cs & 0xff);
-       db_printf("ds   %8x\n", regs->ds & 0xff);
-       db_printf("es   %8x\n", regs->es & 0xff);
-       db_printf("fs   %8x\n", regs->fs & 0xff);
-       db_printf("gs   %8x\n", regs->gs & 0xff);
-       db_printf("ss   %8x\n", regs->ss & 0xff);
-       db_printf("eax  %8x\n", regs->eax);
-       db_printf("ebx  %8x\n", regs->ebx);
-       db_printf("ecx  %8x\n", regs->ecx);
-       db_printf("edx  %8x\n", regs->edx);
-       db_printf("esi  %8x\n", regs->esi);
-       db_printf("edi  %8x\n", regs->edi);
-}
-
-#endif /* MACH_KDB */
-
 /* Synchronize a thread's i386_kernel_state (if any) with the given
  * i386_saved_state_t obtained from the trap/IPI handler; called in
  * kernel_trap() prior to entering the debugger, and when receiving
 /* Synchronize a thread's i386_kernel_state (if any) with the given
  * i386_saved_state_t obtained from the trap/IPI handler; called in
  * kernel_trap() prior to entering the debugger, and when receiving
@@ -1165,22 +1221,27 @@ db_i386_state(
  */
   
 void
  */
   
 void
-sync_iss_to_iks(x86_saved_state32_t *saved_state)
+sync_iss_to_iks(x86_saved_state_t *saved_state)
 {
 {
-       struct x86_kernel_state32 *iks;
+       struct x86_kernel_state *iks;
        vm_offset_t kstack;
        boolean_t record_active_regs = FALSE;
 
        vm_offset_t kstack;
        boolean_t record_active_regs = FALSE;
 
-       if ((kstack = current_thread()->kernel_stack) != 0) {
-               x86_saved_state32_t     *regs;
+       /* The PAL may have a special way to sync registers */
+       if( saved_state->flavor == THREAD_STATE_NONE )
+               pal_get_kern_regs( saved_state );
 
 
-               regs = saved_state;
+       if ((kstack = current_thread()->kernel_stack) != 0) {
+#ifdef __i386__
+               x86_saved_state32_t     *regs = saved_state32(saved_state);
+#else
+               x86_saved_state64_t     *regs = saved_state64(saved_state);
+#endif
 
                iks = STACK_IKS(kstack);
 
 
                iks = STACK_IKS(kstack);
 
-                /*
-                 * Did we take the trap/interrupt in kernel mode?
-                 */
+               /* Did we take the trap/interrupt in kernel mode? */
+#ifdef __i386__
                if (regs == USER_REGS32(current_thread()))
                        record_active_regs = TRUE;
                else {
                if (regs == USER_REGS32(current_thread()))
                        record_active_regs = TRUE;
                else {
@@ -1191,21 +1252,47 @@ sync_iss_to_iks(x86_saved_state32_t *saved_state)
                        iks->k_esi = regs->esi;
                        iks->k_eip = regs->eip;
                }
                        iks->k_esi = regs->esi;
                        iks->k_eip = regs->eip;
                }
+#else
+               if (regs == USER_REGS64(current_thread()))
+                       record_active_regs = TRUE;
+               else {
+                       iks->k_rbx = regs->rbx;
+                       iks->k_rsp = regs->isf.rsp;
+                       iks->k_rbp = regs->rbp;
+                       iks->k_r12 = regs->r12;
+                       iks->k_r13 = regs->r13;
+                       iks->k_r14 = regs->r14;
+                       iks->k_r15 = regs->r15;
+                       iks->k_rip = regs->isf.rip;
+               }
+#endif
        }
 
        if (record_active_regs == TRUE) {
        }
 
        if (record_active_regs == TRUE) {
-               /*
-                * Show the trap handler path
-                */
+#ifdef __i386__
+               /* Show the trap handler path */
                __asm__ volatile("movl %%ebx, %0" : "=m" (iks->k_ebx));
                __asm__ volatile("movl %%esp, %0" : "=m" (iks->k_esp));
                __asm__ volatile("movl %%ebp, %0" : "=m" (iks->k_ebp));
                __asm__ volatile("movl %%edi, %0" : "=m" (iks->k_edi));
                __asm__ volatile("movl %%esi, %0" : "=m" (iks->k_esi));
                __asm__ volatile("movl %%ebx, %0" : "=m" (iks->k_ebx));
                __asm__ volatile("movl %%esp, %0" : "=m" (iks->k_esp));
                __asm__ volatile("movl %%ebp, %0" : "=m" (iks->k_ebp));
                __asm__ volatile("movl %%edi, %0" : "=m" (iks->k_edi));
                __asm__ volatile("movl %%esi, %0" : "=m" (iks->k_esi));
-               /*
-                * "Current" instruction pointer
-                */
+               /* "Current" instruction pointer */
                __asm__ volatile("movl $1f, %0\n1:" : "=m" (iks->k_eip));
                __asm__ volatile("movl $1f, %0\n1:" : "=m" (iks->k_eip));
+#else
+               /* Show the trap handler path */
+               __asm__ volatile("movq %%rbx, %0" : "=m" (iks->k_rbx));
+               __asm__ volatile("movq %%rsp, %0" : "=m" (iks->k_rsp));
+               __asm__ volatile("movq %%rbp, %0" : "=m" (iks->k_rbp));
+               __asm__ volatile("movq %%r12, %0" : "=m" (iks->k_r12));
+               __asm__ volatile("movq %%r13, %0" : "=m" (iks->k_r13));
+               __asm__ volatile("movq %%r14, %0" : "=m" (iks->k_r14));
+               __asm__ volatile("movq %%r15, %0" : "=m" (iks->k_r15));
+               /* "Current" instruction pointer */
+               __asm__ volatile("leaq 1f(%%rip), %%rax; mov %%rax, %0\n1:"
+                                : "=m" (iks->k_rip)
+                                :
+                                : "rax");
+#endif
        }
 }
 
        }
 }
 
@@ -1216,26 +1303,32 @@ sync_iss_to_iks(x86_saved_state32_t *saved_state)
  * or user space.
  */
 void
  * or user space.
  */
 void
-sync_iss_to_iks_unconditionally(__unused x86_saved_state32_t *saved_state) {
-       struct x86_kernel_state32 *iks;
+sync_iss_to_iks_unconditionally(__unused x86_saved_state_t *saved_state) {
+       struct x86_kernel_state *iks;
        vm_offset_t kstack;
        vm_offset_t kstack;
-       boolean_t record_active_regs = FALSE;
 
        if ((kstack = current_thread()->kernel_stack) != 0) {
 
        if ((kstack = current_thread()->kernel_stack) != 0) {
-
                iks = STACK_IKS(kstack);
                iks = STACK_IKS(kstack);
-                               /*
-                * Show the trap handler path
-                */
+#ifdef __i386__
+               /* Display the trap handler path */
                __asm__ volatile("movl %%ebx, %0" : "=m" (iks->k_ebx));
                __asm__ volatile("movl %%esp, %0" : "=m" (iks->k_esp));
                __asm__ volatile("movl %%ebp, %0" : "=m" (iks->k_ebp));
                __asm__ volatile("movl %%edi, %0" : "=m" (iks->k_edi));
                __asm__ volatile("movl %%esi, %0" : "=m" (iks->k_esi));
                __asm__ volatile("movl %%ebx, %0" : "=m" (iks->k_ebx));
                __asm__ volatile("movl %%esp, %0" : "=m" (iks->k_esp));
                __asm__ volatile("movl %%ebp, %0" : "=m" (iks->k_ebp));
                __asm__ volatile("movl %%edi, %0" : "=m" (iks->k_edi));
                __asm__ volatile("movl %%esi, %0" : "=m" (iks->k_esi));
-               /*
-                * "Current" instruction pointer
-                */
+               /* "Current" instruction pointer */
                __asm__ volatile("movl $1f, %0\n1:" : "=m" (iks->k_eip));
                __asm__ volatile("movl $1f, %0\n1:" : "=m" (iks->k_eip));
-
+#else
+               /* Display the trap handler path */
+               __asm__ volatile("movq %%rbx, %0" : "=m" (iks->k_rbx));
+               __asm__ volatile("movq %%rsp, %0" : "=m" (iks->k_rsp));
+               __asm__ volatile("movq %%rbp, %0" : "=m" (iks->k_rbp));
+               __asm__ volatile("movq %%r12, %0" : "=m" (iks->k_r12));
+               __asm__ volatile("movq %%r13, %0" : "=m" (iks->k_r13));
+               __asm__ volatile("movq %%r14, %0" : "=m" (iks->k_r14));
+               __asm__ volatile("movq %%r15, %0" : "=m" (iks->k_r15));
+               /* "Current" instruction pointer */
+               __asm__ volatile("leaq 1f(%%rip), %%rax; mov %%rax, %0\n1:" : "=m" (iks->k_rip)::"rax");
+#endif
        }
 }
        }
 }