]> git.saurik.com Git - apple/xnu.git/blobdiff - osfmk/i386/machine_routines.c
xnu-6153.101.6.tar.gz
[apple/xnu.git] / osfmk / i386 / machine_routines.c
index a90d68178a80f6dfc3ea6372a097d9778cba0aff..e27b01b22ccca5be8d6a9bf1bd2750c4ffe2a2d0 100644 (file)
@@ -2,7 +2,7 @@
  * Copyright (c) 2000-2012 Apple Inc. All rights reserved.
  *
  * @APPLE_OSREFERENCE_LICENSE_HEADER_START@
- * 
+ *
  * 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
  * 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.
- * 
+ *
  * 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,
@@ -22,7 +22,7 @@
  * 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_OSREFERENCE_LICENSE_HEADER_END@
  */
 
 #include <mach/processor.h>
 #include <kern/processor.h>
 #include <kern/machine.h>
-#include <kern/cpu_data.h>
+
 #include <kern/cpu_number.h>
 #include <kern/thread.h>
 #include <kern/thread_call.h>
+#include <kern/policy_internal.h>
+
 #include <prng/random.h>
 #include <i386/machine_cpu.h>
 #include <i386/lapic.h>
 #include <kern/kpc.h>
 #endif
 #include <architecture/i386/pio.h>
-
+#include <i386/cpu_data.h>
 #if DEBUG
-#define DBG(x...)      kprintf("DBG: " x)
+#define DBG(x...)       kprintf("DBG: " x)
 #else
 #define DBG(x...)
 #endif
 
-extern void    wakeup(void *);
+#if MONOTONIC
+#include <kern/monotonic.h>
+#endif /* MONOTONIC */
+
+extern void     wakeup(void *);
 
 static int max_cpus_initialized = 0;
 
-unsigned int   LockTimeOut;
-unsigned int   TLBTimeOut;
-unsigned int   LockTimeOutTSC;
-unsigned int   MutexSpin;
-uint64_t       LastDebuggerEntryAllowance;
-uint64_t       delay_spin_threshold;
+uint64_t        LockTimeOut;
+uint64_t        TLBTimeOut;
+uint64_t        LockTimeOutTSC;
+uint32_t        LockTimeOutUsec;
+uint64_t        MutexSpin;
+uint64_t        low_MutexSpin;
+int64_t         high_MutexSpin;
+uint64_t        LastDebuggerEntryAllowance;
+uint64_t        delay_spin_threshold;
 
 extern uint64_t panic_restart_timeout;
 
 boolean_t virtualized = FALSE;
 
-decl_simple_lock_data(static,  ml_timer_evaluation_slock);
+decl_simple_lock_data(static, ml_timer_evaluation_slock);
 uint32_t ml_timer_eager_evaluations;
 uint64_t ml_timer_eager_evaluation_max;
 static boolean_t ml_timer_evaluation_in_progress = FALSE;
@@ -90,24 +99,27 @@ static boolean_t ml_timer_evaluation_in_progress = FALSE;
 /* IO memory map services */
 
 /* Map memory map IO space */
-vm_offset_t ml_io_map(
-       vm_offset_t phys_addr, 
+vm_offset_t
+ml_io_map(
+       vm_offset_t phys_addr,
        vm_size_t size)
 {
-       return(io_map(phys_addr,size,VM_WIMG_IO));
+       return io_map(phys_addr, size, VM_WIMG_IO);
 }
 
 /* boot memory allocation */
-vm_offset_t ml_static_malloc(
-                            __unused vm_size_t size)
+vm_offset_t
+ml_static_malloc(
+       __unused vm_size_t size)
 {
-       return((vm_offset_t)NULL);
+       return (vm_offset_t)NULL;
 }
 
 
-void ml_get_bouncepool_info(vm_offset_t *phys_addr, vm_size_t *size)
+void
+ml_get_bouncepool_info(vm_offset_t *phys_addr, vm_size_t *size)
 {
-        *phys_addr = 0;
+       *phys_addr = 0;
        *size      = 0;
 }
 
@@ -121,12 +133,25 @@ ml_static_ptovirt(
 #else
        return (vm_offset_t)((paddr) | LINEAR_KERNEL_ADDRESS);
 #endif
-} 
+}
+
+vm_offset_t
+ml_static_slide(
+       vm_offset_t vaddr)
+{
+       return VM_KERNEL_SLIDE(vaddr);
+}
 
+vm_offset_t
+ml_static_unslide(
+       vm_offset_t vaddr)
+{
+       return VM_KERNEL_UNSLIDE(vaddr);
+}
 
 /*
- *     Routine:        ml_static_mfree
- *     Function:
+ * Reclaim memory, by virtual address, that was used in early boot that is no longer needed
+ * by the kernel.
  */
 void
 ml_static_mfree(
@@ -136,45 +161,66 @@ ml_static_mfree(
        addr64_t vaddr_cur;
        ppnum_t ppn;
        uint32_t freed_pages = 0;
+       vm_size_t map_size;
+
        assert(vaddr >= VM_MIN_KERNEL_ADDRESS);
 
-       assert((vaddr & (PAGE_SIZE-1)) == 0); /* must be page aligned */
+       assert((vaddr & (PAGE_SIZE - 1)) == 0); /* must be page aligned */
+
+       for (vaddr_cur = vaddr; vaddr_cur < round_page_64(vaddr + size);) {
+               map_size = pmap_query_pagesize(kernel_pmap, vaddr_cur);
+
+               /* just skip if nothing mapped here */
+               if (map_size == 0) {
+                       vaddr_cur += PAGE_SIZE;
+                       continue;
+               }
+
+               /*
+                * Can't free from the middle of a large page.
+                */
+               assert((vaddr_cur & (map_size - 1)) == 0);
 
-       for (vaddr_cur = vaddr;
-            vaddr_cur < round_page_64(vaddr+size);
-            vaddr_cur += PAGE_SIZE) {
                ppn = pmap_find_phys(kernel_pmap, vaddr_cur);
-               if (ppn != (vm_offset_t)NULL) {
-                       kernel_pmap->stats.resident_count++;
-                       if (kernel_pmap->stats.resident_count >
-                           kernel_pmap->stats.resident_max) {
-                               kernel_pmap->stats.resident_max =
-                                       kernel_pmap->stats.resident_count;
+               assert(ppn != (ppnum_t)NULL);
+
+               pmap_remove(kernel_pmap, vaddr_cur, vaddr_cur + map_size);
+               while (map_size > 0) {
+                       if (++kernel_pmap->stats.resident_count > kernel_pmap->stats.resident_max) {
+                               kernel_pmap->stats.resident_max = kernel_pmap->stats.resident_count;
                        }
-                       pmap_remove(kernel_pmap, vaddr_cur, vaddr_cur+PAGE_SIZE);
+
                        assert(pmap_valid_page(ppn));
                        if (IS_MANAGED_PAGE(ppn)) {
-                               vm_page_create(ppn,(ppn+1));
+                               vm_page_create(ppn, (ppn + 1));
                                freed_pages++;
                        }
+                       map_size -= PAGE_SIZE;
+                       vaddr_cur += PAGE_SIZE;
+                       ppn++;
                }
        }
        vm_page_lockspin_queues();
        vm_page_wire_count -= freed_pages;
        vm_page_wire_count_initial -= freed_pages;
+       if (vm_page_wire_count_on_boot != 0) {
+               assert(vm_page_wire_count_on_boot >= freed_pages);
+               vm_page_wire_count_on_boot -= freed_pages;
+       }
        vm_page_unlock_queues();
 
-#if    DEBUG   
+#if     DEBUG
        kprintf("ml_static_mfree: Released 0x%x pages at VA %p, size:0x%llx, last ppn: 0x%x\n", freed_pages, (void *)vaddr, (uint64_t)size, ppn);
 #endif
 }
 
 
 /* virtual to physical on wired pages */
-vm_offset_t ml_vtophys(
+vm_offset_t
+ml_vtophys(
        vm_offset_t vaddr)
 {
-       return  (vm_offset_t)kvtophys(vaddr);
+       return (vm_offset_t)kvtophys(vaddr);
 }
 
 /*
@@ -187,24 +233,30 @@ vm_offset_t ml_vtophys(
  *                     the duration of the copy process.
  */
 
-vm_size_t ml_nofault_copy(
+vm_size_t
+ml_nofault_copy(
        vm_offset_t virtsrc, vm_offset_t virtdst, vm_size_t size)
 {
        addr64_t cur_phys_dst, cur_phys_src;
        uint32_t count, nbytes = 0;
 
        while (size > 0) {
-               if (!(cur_phys_src = kvtophys(virtsrc)))
+               if (!(cur_phys_src = kvtophys(virtsrc))) {
                        break;
-               if (!(cur_phys_dst = kvtophys(virtdst)))
+               }
+               if (!(cur_phys_dst = kvtophys(virtdst))) {
                        break;
-               if (!pmap_valid_page(i386_btop(cur_phys_dst)) || !pmap_valid_page(i386_btop(cur_phys_src)))
+               }
+               if (!pmap_valid_page(i386_btop(cur_phys_dst)) || !pmap_valid_page(i386_btop(cur_phys_src))) {
                        break;
+               }
                count = (uint32_t)(PAGE_SIZE - (cur_phys_src & PAGE_MASK));
-               if (count > (PAGE_SIZE - (cur_phys_dst & PAGE_MASK)))
+               if (count > (PAGE_SIZE - (cur_phys_dst & PAGE_MASK))) {
                        count = (uint32_t)(PAGE_SIZE - (cur_phys_dst & PAGE_MASK));
-               if (count > size)
+               }
+               if (count > size) {
                        count = (uint32_t)size;
+               }
 
                bcopy_phys(cur_phys_src, cur_phys_dst, count);
 
@@ -227,20 +279,24 @@ vm_size_t ml_nofault_copy(
  *                     FALSE otherwise.
  */
 
-boolean_t ml_validate_nofault(
+boolean_t
+ml_validate_nofault(
        vm_offset_t virtsrc, vm_size_t size)
 {
        addr64_t cur_phys_src;
        uint32_t count;
 
        while (size > 0) {
-               if (!(cur_phys_src = kvtophys(virtsrc)))
+               if (!(cur_phys_src = kvtophys(virtsrc))) {
                        return FALSE;
-               if (!pmap_valid_page(i386_btop(cur_phys_src)))
+               }
+               if (!pmap_valid_page(i386_btop(cur_phys_src))) {
                        return FALSE;
+               }
                count = (uint32_t)(PAGE_SIZE - (cur_phys_src & PAGE_MASK));
-               if (count > size)
+               if (count > size) {
                        count = (uint32_t)size;
+               }
 
                virtsrc += count;
                size -= count;
@@ -252,54 +308,76 @@ boolean_t ml_validate_nofault(
 /* Interrupt handling */
 
 /* Initialize Interrupts */
-void ml_init_interrupt(void)
+void
+ml_init_interrupt(void)
 {
        (void) ml_set_interrupts_enabled(TRUE);
 }
 
 
 /* Get Interrupts Enabled */
-boolean_t ml_get_interrupts_enabled(void)
+boolean_t
+ml_get_interrupts_enabled(void)
 {
-  unsigned long flags;
+       unsigned long flags;
 
-  __asm__ volatile("pushf; pop %0" :  "=r" (flags));
-  return (flags & EFL_IF) != 0;
+       __asm__ volatile ("pushf; pop   %0":  "=r" (flags));
+       return (flags & EFL_IF) != 0;
 }
 
 /* Set Interrupts Enabled */
-boolean_t ml_set_interrupts_enabled(boolean_t enable)
+boolean_t
+ml_set_interrupts_enabled(boolean_t enable)
 {
        unsigned long flags;
        boolean_t istate;
-       
-       __asm__ volatile("pushf; pop    %0" :  "=r" (flags));
+
+       __asm__ volatile ("pushf; pop   %0"  :  "=r" (flags));
 
        assert(get_interrupt_level() ? (enable == FALSE) : TRUE);
 
        istate = ((flags & EFL_IF) != 0);
 
        if (enable) {
-               __asm__ volatile("sti;nop");
+               __asm__ volatile ("sti;nop");
 
-               if ((get_preemption_level() == 0) && (*ast_pending() & AST_URGENT))
+               if ((get_preemption_level() == 0) && (*ast_pending() & AST_URGENT)) {
                        __asm__ volatile ("int %0" :: "N" (T_PREEMPT));
-       }
-       else {
-               if (istate)
-                       __asm__ volatile("cli");
+               }
+       } else {
+               if (istate) {
+                       __asm__ volatile ("cli");
+               }
        }
 
        return istate;
 }
 
+/* Early Set Interrupts Enabled */
+boolean_t
+ml_early_set_interrupts_enabled(boolean_t enable)
+{
+       if (enable == TRUE) {
+               kprintf("Caller attempted to enable interrupts too early in "
+                   "kernel startup. Halting.\n");
+               hlt();
+               /*NOTREACHED*/
+       }
+
+       /* On x86, do not allow interrupts to be enabled very early */
+       return FALSE;
+}
+
 /* Check if running at interrupt context */
-boolean_t ml_at_interrupt_context(void)
+boolean_t
+ml_at_interrupt_context(void)
 {
        return get_interrupt_level() != 0;
 }
 
-void ml_get_power_state(boolean_t *icp, boolean_t *pidlep) {
+void
+ml_get_power_state(boolean_t *icp, boolean_t *pidlep)
+{
        *icp = (get_interrupt_level() != 0);
        /* These will be technically inaccurate for interrupts that occur
         * successively within a single "idle exit" event, but shouldn't
@@ -309,7 +387,9 @@ void ml_get_power_state(boolean_t *icp, boolean_t *pidlep) {
 }
 
 /* Generate a fake interrupt */
-void ml_cause_interrupt(void)
+__dead2
+void
+ml_cause_interrupt(void)
 {
        panic("ml_cause_interrupt not defined yet on Intel");
 }
@@ -318,9 +398,10 @@ void ml_cause_interrupt(void)
  * TODO: transition users of this to kernel_thread_start_priority
  * ml_thread_policy is an unsupported KPI
  */
-void ml_thread_policy(
+void
+ml_thread_policy(
        thread_t thread,
-__unused       unsigned policy_id,
+       __unused        unsigned policy_id,
        unsigned policy_info)
 {
        if (policy_info & MACHINE_NETWORK_WORKLOOP) {
@@ -330,26 +411,27 @@ __unused  unsigned policy_id,
                info.importance = 1;
 
                kret = thread_policy_set_internal(thread, THREAD_PRECEDENCE_POLICY,
-                                                               (thread_policy_t)&info,
-                                                               THREAD_PRECEDENCE_POLICY_COUNT);
+                   (thread_policy_t)&info,
+                   THREAD_PRECEDENCE_POLICY_COUNT);
                assert(kret == KERN_SUCCESS);
        }
 }
 
 /* Initialize Interrupts */
-void ml_install_interrupt_handler(
+void
+ml_install_interrupt_handler(
        void *nub,
        int source,
        void *target,
        IOInterruptHandler handler,
-       void *refCon)  
+       void *refCon)
 {
        boolean_t current_state;
 
-       current_state = ml_get_interrupts_enabled();
+       current_state = ml_set_interrupts_enabled(FALSE);
 
        PE_install_interrupt_handler(nub, source, target,
-                                    (IOInterruptHandler) handler, refCon);
+           (IOInterruptHandler) handler, refCon);
 
        (void) ml_set_interrupts_enabled(current_state);
 
@@ -359,11 +441,12 @@ void ml_install_interrupt_handler(
 
 void
 machine_signal_idle(
-        processor_t processor)
+       processor_t processor)
 {
        cpu_interrupt(processor->cpu_id);
 }
 
+__dead2
 void
 machine_signal_idle_deferred(
        __unused processor_t processor)
@@ -371,6 +454,7 @@ machine_signal_idle_deferred(
        panic("Unimplemented");
 }
 
+__dead2
 void
 machine_signal_idle_cancel(
        __unused processor_t processor)
@@ -380,12 +464,12 @@ machine_signal_idle_cancel(
 
 static kern_return_t
 register_cpu(
-        uint32_t        lapic_id,
+       uint32_t        lapic_id,
        processor_t     *processor_out,
        boolean_t       boot_cpu )
 {
-       int             target_cpu;
-       cpu_data_t      *this_cpu_datap;
+       int             target_cpu;
+       cpu_data_t      *this_cpu_datap;
 
        this_cpu_datap = cpu_data_alloc(boot_cpu);
        if (this_cpu_datap == NULL) {
@@ -393,46 +477,47 @@ register_cpu(
        }
        target_cpu = this_cpu_datap->cpu_number;
        assert((boot_cpu && (target_cpu == 0)) ||
-             (!boot_cpu && (target_cpu != 0)));
+           (!boot_cpu && (target_cpu != 0)));
 
        lapic_cpu_map(lapic_id, target_cpu);
 
        /* The cpu_id is not known at registration phase. Just do
-        * lapic_id for now 
+        * lapic_id for now
         */
        this_cpu_datap->cpu_phys_number = lapic_id;
 
        this_cpu_datap->cpu_console_buf = console_cpu_alloc(boot_cpu);
-       if (this_cpu_datap->cpu_console_buf == NULL)
-               goto failed;
-
-       this_cpu_datap->cpu_chud = chudxnu_cpu_alloc(boot_cpu);
-       if (this_cpu_datap->cpu_chud == NULL)
+       if (this_cpu_datap->cpu_console_buf == NULL) {
                goto failed;
+       }
 
 #if KPC
-       if (kpc_register_cpu(this_cpu_datap) != TRUE)
+       if (kpc_register_cpu(this_cpu_datap) != TRUE) {
                goto failed;
+       }
 #endif
 
        if (!boot_cpu) {
                cpu_thread_alloc(this_cpu_datap->cpu_number);
-               if (this_cpu_datap->lcpu.core == NULL)
+               if (this_cpu_datap->lcpu.core == NULL) {
                        goto failed;
+               }
 
 #if NCOPY_WINDOWS > 0
                this_cpu_datap->cpu_pmap = pmap_cpu_alloc(boot_cpu);
-               if (this_cpu_datap->cpu_pmap == NULL)
+               if (this_cpu_datap->cpu_pmap == NULL) {
                        goto failed;
+               }
 #endif
 
                this_cpu_datap->cpu_processor = cpu_processor_alloc(boot_cpu);
-               if (this_cpu_datap->cpu_processor == NULL)
+               if (this_cpu_datap->cpu_processor == NULL) {
                        goto failed;
+               }
                /*
                 * processor_init() deferred to topology start
                 * because "slot numbers" a.k.a. logical processor numbers
-                * are not yet finalized.
+                * are not yet finalized.
                 */
        }
 
@@ -445,14 +530,10 @@ failed:
 #if NCOPY_WINDOWS > 0
        pmap_cpu_free(this_cpu_datap->cpu_pmap);
 #endif
-       chudxnu_cpu_free(this_cpu_datap->cpu_chud);
        console_cpu_free(this_cpu_datap->cpu_console_buf);
 #if KPC
-       kpc_counterbuf_free(this_cpu_datap->cpu_kpc_buf[0]);
-       kpc_counterbuf_free(this_cpu_datap->cpu_kpc_buf[1]);
-       kpc_counterbuf_free(this_cpu_datap->cpu_kpc_shadow);
-       kpc_counterbuf_free(this_cpu_datap->cpu_kpc_reload);
-#endif
+       kpc_unregister_cpu(this_cpu_datap);
+#endif /* KPC */
 
        return KERN_FAILURE;
 }
@@ -460,157 +541,176 @@ failed:
 
 kern_return_t
 ml_processor_register(
-        cpu_id_t        cpu_id,
-        uint32_t        lapic_id,
-        processor_t     *processor_out,
-        boolean_t       boot_cpu,
+       cpu_id_t        cpu_id,
+       uint32_t        lapic_id,
+       processor_t     *processor_out,
+       boolean_t       boot_cpu,
        boolean_t       start )
 {
-    static boolean_t done_topo_sort = FALSE;
-    static uint32_t num_registered = 0;
+       static boolean_t done_topo_sort = FALSE;
+       static uint32_t num_registered = 0;
 
-    /* Register all CPUs first, and track max */
-    if( start == FALSE )
-    {
-       num_registered++;
+       /* Register all CPUs first, and track max */
+       if (start == FALSE) {
+               num_registered++;
 
-       DBG( "registering CPU lapic id %d\n", lapic_id );
+               DBG( "registering CPU lapic id %d\n", lapic_id );
 
-       return register_cpu( lapic_id, processor_out, boot_cpu );
-    }
+               return register_cpu( lapic_id, processor_out, boot_cpu );
+       }
 
-    /* Sort by topology before we start anything */
-    if( !done_topo_sort )
-    {
-       DBG( "about to start CPUs. %d registered\n", num_registered );
+       /* Sort by topology before we start anything */
+       if (!done_topo_sort) {
+               DBG( "about to start CPUs. %d registered\n", num_registered );
 
-       cpu_topology_sort( num_registered );
-       done_topo_sort = TRUE;
-    }
+               cpu_topology_sort( num_registered );
+               done_topo_sort = TRUE;
+       }
 
-    /* Assign the cpu ID */
-    uint32_t cpunum = -1;
-    cpu_data_t *this_cpu_datap = NULL;
+       /* Assign the cpu ID */
+       uint32_t cpunum = -1;
+       cpu_data_t  *this_cpu_datap = NULL;
 
-    /* find cpu num and pointer */
-    cpunum = ml_get_cpuid( lapic_id );
+       /* find cpu num and pointer */
+       cpunum = ml_get_cpuid( lapic_id );
 
-    if( cpunum == 0xFFFFFFFF ) /* never heard of it? */
-       panic( "trying to start invalid/unregistered CPU %d\n", lapic_id );
+       if (cpunum == 0xFFFFFFFF) { /* never heard of it? */
+               panic( "trying to start invalid/unregistered CPU %d\n", lapic_id );
+       }
 
-    this_cpu_datap = cpu_datap(cpunum);
+       this_cpu_datap = cpu_datap(cpunum);
 
-    /* fix the CPU id */
-    this_cpu_datap->cpu_id = cpu_id;
+       /* fix the CPU id */
+       this_cpu_datap->cpu_id = cpu_id;
 
-    /* allocate and initialize other per-cpu structures */
-    if (!boot_cpu) {
-       mp_cpus_call_cpu_init(cpunum);
-       prng_cpu_init(cpunum);
-    }
+       /* allocate and initialize other per-cpu structures */
+       if (!boot_cpu) {
+               mp_cpus_call_cpu_init(cpunum);
+               random_cpu_init(cpunum);
+       }
 
-    /* output arg */
-    *processor_out = this_cpu_datap->cpu_processor;
+       /* output arg */
+       *processor_out = this_cpu_datap->cpu_processor;
 
-    /* OK, try and start this CPU */
-    return cpu_topology_start_cpu( cpunum );
+       /* OK, try and start this CPU */
+       return cpu_topology_start_cpu( cpunum );
 }
 
 
 void
 ml_cpu_get_info(ml_cpu_info_t *cpu_infop)
 {
-       boolean_t       os_supports_sse;
+       boolean_t       os_supports_sse;
        i386_cpu_info_t *cpuid_infop;
 
-       if (cpu_infop == NULL)
+       if (cpu_infop == NULL) {
                return;
+       }
+
        /*
         * Are we supporting MMX/SSE/SSE2/SSE3?
         * As distinct from whether the cpu has these capabilities.
         */
        os_supports_sse = !!(get_cr4() & CR4_OSXMM);
 
-       if (ml_fpu_avx_enabled())
+       if (ml_fpu_avx_enabled()) {
                cpu_infop->vector_unit = 9;
-       else if ((cpuid_features() & CPUID_FEATURE_SSE4_2) && os_supports_sse)
+       } else if ((cpuid_features() & CPUID_FEATURE_SSE4_2) && os_supports_sse) {
                cpu_infop->vector_unit = 8;
-       else if ((cpuid_features() & CPUID_FEATURE_SSE4_1) && os_supports_sse)
+       } else if ((cpuid_features() & CPUID_FEATURE_SSE4_1) && os_supports_sse) {
                cpu_infop->vector_unit = 7;
-       else if ((cpuid_features() & CPUID_FEATURE_SSSE3) && os_supports_sse)
+       } else if ((cpuid_features() & CPUID_FEATURE_SSSE3) && os_supports_sse) {
                cpu_infop->vector_unit = 6;
-       else if ((cpuid_features() & CPUID_FEATURE_SSE3) && os_supports_sse)
+       } else if ((cpuid_features() & CPUID_FEATURE_SSE3) && os_supports_sse) {
                cpu_infop->vector_unit = 5;
-       else if ((cpuid_features() & CPUID_FEATURE_SSE2) && os_supports_sse)
+       } else if ((cpuid_features() & CPUID_FEATURE_SSE2) && os_supports_sse) {
                cpu_infop->vector_unit = 4;
-       else if ((cpuid_features() & CPUID_FEATURE_SSE) && os_supports_sse)
+       } else if ((cpuid_features() & CPUID_FEATURE_SSE) && os_supports_sse) {
                cpu_infop->vector_unit = 3;
-       else if (cpuid_features() & CPUID_FEATURE_MMX)
+       } else if (cpuid_features() & CPUID_FEATURE_MMX) {
                cpu_infop->vector_unit = 2;
-       else
+       } else {
                cpu_infop->vector_unit = 0;
+       }
 
        cpuid_infop  = cpuid_info();
 
-       cpu_infop->cache_line_size = cpuid_infop->cache_linesize; 
+       cpu_infop->cache_line_size = cpuid_infop->cache_linesize;
 
        cpu_infop->l1_icache_size = cpuid_infop->cache_size[L1I];
        cpu_infop->l1_dcache_size = cpuid_infop->cache_size[L1D];
-  
-        if (cpuid_infop->cache_size[L2U] > 0) {
-            cpu_infop->l2_settings = 1;
-            cpu_infop->l2_cache_size = cpuid_infop->cache_size[L2U];
-        } else {
-            cpu_infop->l2_settings = 0;
-            cpu_infop->l2_cache_size = 0xFFFFFFFF;
-        }
-
-        if (cpuid_infop->cache_size[L3U] > 0) {
-            cpu_infop->l3_settings = 1;
-            cpu_infop->l3_cache_size = cpuid_infop->cache_size[L3U];
-        } else {
-            cpu_infop->l3_settings = 0;
-            cpu_infop->l3_cache_size = 0xFFFFFFFF;
-        }
+
+       if (cpuid_infop->cache_size[L2U] > 0) {
+               cpu_infop->l2_settings = 1;
+               cpu_infop->l2_cache_size = cpuid_infop->cache_size[L2U];
+       } else {
+               cpu_infop->l2_settings = 0;
+               cpu_infop->l2_cache_size = 0xFFFFFFFF;
+       }
+
+       if (cpuid_infop->cache_size[L3U] > 0) {
+               cpu_infop->l3_settings = 1;
+               cpu_infop->l3_cache_size = cpuid_infop->cache_size[L3U];
+       } else {
+               cpu_infop->l3_settings = 0;
+               cpu_infop->l3_cache_size = 0xFFFFFFFF;
+       }
 }
 
 void
 ml_init_max_cpus(unsigned long max_cpus)
 {
-        boolean_t current_state;
+       boolean_t current_state;
 
-        current_state = ml_set_interrupts_enabled(FALSE);
-        if (max_cpus_initialized != MAX_CPUS_SET) {
-                if (max_cpus > 0 && max_cpus <= MAX_CPUS) {
+       current_state = ml_set_interrupts_enabled(FALSE);
+       if (max_cpus_initialized != MAX_CPUS_SET) {
+               if (max_cpus > 0 && max_cpus <= MAX_CPUS) {
                        /*
                         * Note: max_cpus is the number of enabled processors
                         * that ACPI found; max_ncpus is the maximum number
                         * that the kernel supports or that the "cpus="
                         * boot-arg has set. Here we take int minimum.
                         */
-                        machine_info.max_cpus = (integer_t)MIN(max_cpus, max_ncpus);
+                       machine_info.max_cpus = (integer_t)MIN(max_cpus, max_ncpus);
                }
-                if (max_cpus_initialized == MAX_CPUS_WAIT)
-                        wakeup((event_t)&max_cpus_initialized);
-                max_cpus_initialized = MAX_CPUS_SET;
-        }
-        (void) ml_set_interrupts_enabled(current_state);
+               if (max_cpus_initialized == MAX_CPUS_WAIT) {
+                       wakeup((event_t)&max_cpus_initialized);
+               }
+               max_cpus_initialized = MAX_CPUS_SET;
+       }
+       (void) ml_set_interrupts_enabled(current_state);
 }
 
 int
 ml_get_max_cpus(void)
 {
-        boolean_t current_state;
+       boolean_t current_state;
+
+       current_state = ml_set_interrupts_enabled(FALSE);
+       if (max_cpus_initialized != MAX_CPUS_SET) {
+               max_cpus_initialized = MAX_CPUS_WAIT;
+               assert_wait((event_t)&max_cpus_initialized, THREAD_UNINT);
+               (void)thread_block(THREAD_CONTINUE_NULL);
+       }
+       (void) ml_set_interrupts_enabled(current_state);
+       return machine_info.max_cpus;
+}
 
-        current_state = ml_set_interrupts_enabled(FALSE);
-        if (max_cpus_initialized != MAX_CPUS_SET) {
-                max_cpus_initialized = MAX_CPUS_WAIT;
-                assert_wait((event_t)&max_cpus_initialized, THREAD_UNINT);
-                (void)thread_block(THREAD_CONTINUE_NULL);
-        }
-        (void) ml_set_interrupts_enabled(current_state);
-        return(machine_info.max_cpus);
+boolean_t
+ml_wants_panic_trap_to_debugger(void)
+{
+       return FALSE;
+}
+
+void
+ml_panic_trap_to_debugger(__unused const char *panic_format_str,
+    __unused va_list *panic_args,
+    __unused unsigned int reason,
+    __unused void *ctx,
+    __unused uint64_t panic_options_mask,
+    __unused unsigned long panic_caller)
+{
+       return;
 }
 
 /*
@@ -620,23 +720,28 @@ ml_get_max_cpus(void)
 void
 ml_init_lock_timeout(void)
 {
-       uint64_t        abstime;
-       uint32_t        mtxspin;
+       uint64_t        abstime;
+       uint32_t        mtxspin;
 #if DEVELOPMENT || DEBUG
-       uint64_t        default_timeout_ns = NSEC_PER_SEC>>2;
+       uint64_t        default_timeout_ns = NSEC_PER_SEC >> 2;
 #else
-       uint64_t        default_timeout_ns = NSEC_PER_SEC>>1;
+       uint64_t        default_timeout_ns = NSEC_PER_SEC >> 1;
 #endif
-       uint32_t        slto;
-       uint32_t        prt;
+       uint32_t        slto;
+       uint32_t        prt;
 
-       if (PE_parse_boot_argn("slto_us", &slto, sizeof (slto)))
+       if (PE_parse_boot_argn("slto_us", &slto, sizeof(slto))) {
                default_timeout_ns = slto * NSEC_PER_USEC;
+       }
 
-       /* LockTimeOut is absolutetime, LockTimeOutTSC is in TSC ticks */
+       /*
+        * LockTimeOut is absolutetime, LockTimeOutTSC is in TSC ticks,
+        * and LockTimeOutUsec is in microseconds and it's 32-bits.
+        */
+       LockTimeOutUsec = (uint32_t) (default_timeout_ns / NSEC_PER_USEC);
        nanoseconds_to_absolutetime(default_timeout_ns, &abstime);
-       LockTimeOut = (uint32_t) abstime;
-       LockTimeOutTSC = (uint32_t) tmrCvt(abstime, tscFCvtn2t);
+       LockTimeOut = abstime;
+       LockTimeOutTSC = tmrCvt(abstime, tscFCvtn2t);
 
        /*
         * TLBTimeOut dictates the TLB flush timeout period. It defaults to
@@ -644,7 +749,7 @@ ml_init_lock_timeout(void)
         * zero value inhibits the timeout-panic and cuts a trace evnt instead
         * - see pmap_flush_tlbs().
         */
-       if (PE_parse_boot_argn("tlbto_us", &slto, sizeof (slto))) {
+       if (PE_parse_boot_argn("tlbto_us", &slto, sizeof(slto))) {
                default_timeout_ns = slto * NSEC_PER_USEC;
                nanoseconds_to_absolutetime(default_timeout_ns, &abstime);
                TLBTimeOut = (uint32_t) abstime;
@@ -652,25 +757,84 @@ ml_init_lock_timeout(void)
                TLBTimeOut = LockTimeOut;
        }
 
-       if (PE_parse_boot_argn("phyreadmaxus", &slto, sizeof (slto))) {
+#if DEVELOPMENT || DEBUG
+       reportphyreaddelayabs = LockTimeOut >> 1;
+#endif
+       if (PE_parse_boot_argn("phyreadmaxus", &slto, sizeof(slto))) {
                default_timeout_ns = slto * NSEC_PER_USEC;
                nanoseconds_to_absolutetime(default_timeout_ns, &abstime);
                reportphyreaddelayabs = abstime;
        }
 
-       if (PE_parse_boot_argn("mtxspin", &mtxspin, sizeof (mtxspin))) {
-               if (mtxspin > USEC_PER_SEC>>4)
-                       mtxspin =  USEC_PER_SEC>>4;
-               nanoseconds_to_absolutetime(mtxspin*NSEC_PER_USEC, &abstime);
+       if (PE_parse_boot_argn("phywritemaxus", &slto, sizeof(slto))) {
+               nanoseconds_to_absolutetime((uint64_t)slto * NSEC_PER_USEC, &abstime);
+               reportphywritedelayabs = abstime;
+       }
+
+       if (PE_parse_boot_argn("tracephyreadus", &slto, sizeof(slto))) {
+               nanoseconds_to_absolutetime((uint64_t)slto * NSEC_PER_USEC, &abstime);
+               tracephyreaddelayabs = abstime;
+       }
+
+       if (PE_parse_boot_argn("tracephywriteus", &slto, sizeof(slto))) {
+               nanoseconds_to_absolutetime((uint64_t)slto * NSEC_PER_USEC, &abstime);
+               tracephywritedelayabs = abstime;
+       }
+
+       if (PE_parse_boot_argn("mtxspin", &mtxspin, sizeof(mtxspin))) {
+               if (mtxspin > USEC_PER_SEC >> 4) {
+                       mtxspin =  USEC_PER_SEC >> 4;
+               }
+               nanoseconds_to_absolutetime(mtxspin * NSEC_PER_USEC, &abstime);
        } else {
-               nanoseconds_to_absolutetime(10*NSEC_PER_USEC, &abstime);
+               nanoseconds_to_absolutetime(10 * NSEC_PER_USEC, &abstime);
        }
        MutexSpin = (unsigned int)abstime;
+       low_MutexSpin = MutexSpin;
+       /*
+        * high_MutexSpin should be initialized as low_MutexSpin * real_ncpus, but
+        * real_ncpus is not set at this time
+        */
+       high_MutexSpin = -1;
 
        nanoseconds_to_absolutetime(4ULL * NSEC_PER_SEC, &LastDebuggerEntryAllowance);
-       if (PE_parse_boot_argn("panic_restart_timeout", &prt, sizeof (prt)))
+       if (PE_parse_boot_argn("panic_restart_timeout", &prt, sizeof(prt))) {
                nanoseconds_to_absolutetime(prt * NSEC_PER_SEC, &panic_restart_timeout);
+       }
+
        virtualized = ((cpuid_features() & CPUID_FEATURE_VMM) != 0);
+       if (virtualized) {
+               int     vti;
+
+               if (!PE_parse_boot_argn("vti", &vti, sizeof(vti))) {
+                       vti = 6;
+               }
+               printf("Timeouts adjusted for virtualization (<<%d)\n", vti);
+               kprintf("Timeouts adjusted for virtualization (<<%d):\n", vti);
+#define VIRTUAL_TIMEOUT_INFLATE64(_timeout)                     \
+MACRO_BEGIN                                                     \
+       kprintf("%24s: 0x%016llx ", #_timeout, _timeout);       \
+       _timeout <<= vti;                                       \
+       kprintf("-> 0x%016llx\n",  _timeout);                   \
+MACRO_END
+#define VIRTUAL_TIMEOUT_INFLATE32(_timeout)                     \
+MACRO_BEGIN                                                     \
+       kprintf("%24s:         0x%08x ", #_timeout, _timeout);  \
+       if ((_timeout <<vti) >> vti == _timeout)                \
+               _timeout <<= vti;                               \
+       else                                                    \
+               _timeout = ~0; /* cap rather than overflow */   \
+       kprintf("-> 0x%08x\n",  _timeout);                      \
+MACRO_END
+               VIRTUAL_TIMEOUT_INFLATE32(LockTimeOutUsec);
+               VIRTUAL_TIMEOUT_INFLATE64(LockTimeOut);
+               VIRTUAL_TIMEOUT_INFLATE64(LockTimeOutTSC);
+               VIRTUAL_TIMEOUT_INFLATE64(TLBTimeOut);
+               VIRTUAL_TIMEOUT_INFLATE64(MutexSpin);
+               VIRTUAL_TIMEOUT_INFLATE64(low_MutexSpin);
+               VIRTUAL_TIMEOUT_INFLATE64(reportphyreaddelayabs);
+       }
+
        interrupt_latency_tracker_setup();
        simple_lock_init(&ml_timer_evaluation_slock, 0);
 }
@@ -692,6 +856,18 @@ ml_delay_should_spin(uint64_t interval)
        return (interval < delay_spin_threshold) ? TRUE : FALSE;
 }
 
+uint32_t yield_delay_us = 0;
+
+void
+ml_delay_on_yield(void)
+{
+#if DEVELOPMENT || DEBUG
+       if (yield_delay_us) {
+               delay(yield_delay_us);
+       }
+#endif
+}
+
 /*
  * This is called from the machine-independent layer
  * to perform machine-dependent info updates. Defer to cpu_thread_init().
@@ -718,110 +894,148 @@ ml_cpu_down(void)
  * The following are required for parts of the kernel
  * that cannot resolve these functions as inlines:
  */
-extern thread_t current_act(void);
+extern thread_t current_act(void) __attribute__((const));
 thread_t
 current_act(void)
 {
-  return(current_thread_fast());
+       return current_thread_fast();
 }
 
 #undef current_thread
-extern thread_t current_thread(void);
+extern thread_t current_thread(void) __attribute__((const));
 thread_t
 current_thread(void)
 {
-  return(current_thread_fast());
+       return current_thread_fast();
 }
 
 
-boolean_t ml_is64bit(void) {
-
-        return (cpu_mode_is64bit());
+boolean_t
+ml_is64bit(void)
+{
+       return cpu_mode_is64bit();
 }
 
 
-boolean_t ml_thread_is64bit(thread_t thread) {
-  
-        return (thread_is_64bit(thread));
+boolean_t
+ml_thread_is64bit(thread_t thread)
+{
+       return thread_is_64bit_addr(thread);
 }
 
 
-boolean_t ml_state_is64bit(void *saved_state) {
-
+boolean_t
+ml_state_is64bit(void *saved_state)
+{
        return is_saved_state64(saved_state);
 }
 
-void ml_cpu_set_ldt(int selector)
+void
+ml_cpu_set_ldt(int selector)
 {
        /*
         * Avoid loading the LDT
         * if we're setting the KERNEL LDT and it's already set.
         */
        if (selector == KERNEL_LDT &&
-           current_cpu_datap()->cpu_ldt == KERNEL_LDT)
+           current_cpu_datap()->cpu_ldt == KERNEL_LDT) {
                return;
+       }
 
        lldt(selector);
        current_cpu_datap()->cpu_ldt = selector;
 }
 
-void ml_fp_setvalid(boolean_t value)
+void
+ml_fp_setvalid(boolean_t value)
 {
-        fp_setvalid(value);
+       fp_setvalid(value);
 }
 
-uint64_t ml_cpu_int_event_time(void)
+uint64_t
+ml_cpu_int_event_time(void)
 {
        return current_cpu_datap()->cpu_int_event_time;
 }
 
-vm_offset_t ml_stack_remaining(void)
+vm_offset_t
+ml_stack_remaining(void)
 {
        uintptr_t local = (uintptr_t) &local;
 
        if (ml_at_interrupt_context() != 0) {
-           return (local - (current_cpu_datap()->cpu_int_stack_top - INTSTACK_SIZE));
+               return local - (current_cpu_datap()->cpu_int_stack_top - INTSTACK_SIZE);
+       } else {
+               return local - current_thread()->kernel_stack;
+       }
+}
+
+#if KASAN
+vm_offset_t ml_stack_base(void);
+vm_size_t ml_stack_size(void);
+
+vm_offset_t
+ml_stack_base(void)
+{
+       if (ml_at_interrupt_context()) {
+               return current_cpu_datap()->cpu_int_stack_top - INTSTACK_SIZE;
        } else {
-           return (local - current_thread()->kernel_stack);
+               return current_thread()->kernel_stack;
        }
 }
 
+vm_size_t
+ml_stack_size(void)
+{
+       if (ml_at_interrupt_context()) {
+               return INTSTACK_SIZE;
+       } else {
+               return kernel_stack_size;
+       }
+}
+#endif
+
 void
 kernel_preempt_check(void)
 {
-       boolean_t       intr;
+       boolean_t       intr;
        unsigned long flags;
 
        assert(get_preemption_level() == 0);
 
-       __asm__ volatile("pushf; pop    %0" :  "=r" (flags));
-
-       intr = ((flags & EFL_IF) != 0);
-
-       if ((*ast_pending() & AST_URGENT) && intr == TRUE) {
+       if (__improbable(*ast_pending() & AST_URGENT)) {
                /*
-                * can handle interrupts and preemptions 
+                * can handle interrupts and preemptions
                 * at this point
                 */
+               __asm__ volatile ("pushf; pop   %0"  :  "=r" (flags));
+
+               intr = ((flags & EFL_IF) != 0);
 
                /*
                 * now cause the PRE-EMPTION trap
                 */
-               __asm__ volatile ("int %0" :: "N" (T_PREEMPT));
+               if (intr == TRUE) {
+                       __asm__ volatile ("int %0" :: "N" (T_PREEMPT));
+               }
        }
 }
 
-boolean_t machine_timeout_suspended(void) {
-       return (virtualized || pmap_tlb_flush_timeout || spinlock_timed_out || panic_active() || mp_recent_debugger_activity() || ml_recent_wake());
+boolean_t
+machine_timeout_suspended(void)
+{
+       return pmap_tlb_flush_timeout || spinlock_timed_out || panic_active() || mp_recent_debugger_activity() || ml_recent_wake();
 }
 
 /* Eagerly evaluate all pending timer and thread callouts
  */
-void ml_timer_evaluate(void) {
-       KERNEL_DEBUG_CONSTANT(DECR_TIMER_RESCAN|DBG_FUNC_START, 0, 0, 0, 0, 0);
+void
+ml_timer_evaluate(void)
+{
+       KERNEL_DEBUG_CONSTANT(DECR_TIMER_RESCAN | DBG_FUNC_START, 0, 0, 0, 0, 0);
 
        uint64_t te_end, te_start = mach_absolute_time();
-       simple_lock(&ml_timer_evaluation_slock);
+       simple_lock(&ml_timer_evaluation_slock, LCK_GRP_NULL);
        ml_timer_evaluation_in_progress = TRUE;
        thread_call_delayed_timer_rescan_all();
        mp_cpus_call(CPUMASK_ALL, ASYNC, timer_queue_expire_rescan, NULL);
@@ -831,45 +1045,95 @@ void ml_timer_evaluate(void) {
        ml_timer_eager_evaluation_max = MAX(ml_timer_eager_evaluation_max, (te_end - te_start));
        simple_unlock(&ml_timer_evaluation_slock);
 
-       KERNEL_DEBUG_CONSTANT(DECR_TIMER_RESCAN|DBG_FUNC_END, 0, 0, 0, 0, 0);
+       KERNEL_DEBUG_CONSTANT(DECR_TIMER_RESCAN | DBG_FUNC_END, 0, 0, 0, 0, 0);
 }
 
 boolean_t
-ml_timer_forced_evaluation(void) {
+ml_timer_forced_evaluation(void)
+{
        return ml_timer_evaluation_in_progress;
 }
 
 /* 32-bit right-rotate n bits */
-static inline uint32_t ror32(uint32_t val, const unsigned int n)
-{      
-       __asm__ volatile("rorl %%cl,%0" : "=r" (val) : "0" (val), "c" (n));
+static inline uint32_t
+ror32(uint32_t val, const unsigned int n)
+{
+       __asm__ volatile ("rorl %%cl,%0" : "=r" (val) : "0" (val), "c" (n));
        return val;
 }
 
 void
 ml_entropy_collect(void)
 {
-       uint32_t        tsc_lo, tsc_hi;
-       uint32_t        *ep;
+       uint32_t        tsc_lo, tsc_hi;
+       uint32_t        *ep;
 
        assert(cpu_number() == master_cpu);
 
        /* update buffer pointer cyclically */
-       if (EntropyData.index_ptr - EntropyData.buffer == ENTROPY_BUFFER_SIZE)
-               ep = EntropyData.index_ptr = EntropyData.buffer;
-       else
-               ep = EntropyData.index_ptr++;
+       ep = EntropyData.buffer + (EntropyData.sample_count & ENTROPY_BUFFER_INDEX_MASK);
+       EntropyData.sample_count += 1;
 
        rdtsc_nofence(tsc_lo, tsc_hi);
        *ep = ror32(*ep, 9) ^ tsc_lo;
 }
 
+uint64_t
+ml_energy_stat(__unused thread_t t)
+{
+       return 0;
+}
+
 void
-ml_gpu_stat_update(uint64_t gpu_ns_delta) {
+ml_gpu_stat_update(uint64_t gpu_ns_delta)
+{
        current_thread()->machine.thread_gpu_ns += gpu_ns_delta;
 }
 
 uint64_t
-ml_gpu_stat(thread_t t) {
+ml_gpu_stat(thread_t t)
+{
        return t->machine.thread_gpu_ns;
 }
+
+int plctrace_enabled = 0;
+
+void
+_disable_preemption(void)
+{
+       disable_preemption_internal();
+}
+
+void
+_enable_preemption(void)
+{
+       enable_preemption_internal();
+}
+
+void
+plctrace_disable(void)
+{
+       plctrace_enabled = 0;
+}
+
+static boolean_t ml_quiescing;
+
+void
+ml_set_is_quiescing(boolean_t quiescing)
+{
+       assert(FALSE == ml_get_interrupts_enabled());
+       ml_quiescing = quiescing;
+}
+
+boolean_t
+ml_is_quiescing(void)
+{
+       assert(FALSE == ml_get_interrupts_enabled());
+       return ml_quiescing;
+}
+
+uint64_t
+ml_get_booter_memory_size(void)
+{
+       return 0;
+}