kernel - Fix some rare pmap races in i386 and x86_64.

[dragonfly.git] / sys / platform / vkernel / platform / pmap.c

/*
 * Copyright (c) 2006 The DragonFly Project.  All rights reserved.
 * Copyright (c) 1991 Regents of the University of California.
 * All rights reserved.
 * Copyright (c) 1994 John S. Dyson
 * All rights reserved.
 * Copyright (c) 1994 David Greenman
 * All rights reserved.
 * Copyright (c) 2004-2006 Matthew Dillon
 * All rights reserved.
 * 
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in
 *    the documentation and/or other materials provided with the
 *    distribution.
 * 3. Neither the name of The DragonFly Project nor the names of its
 *    contributors may be used to endorse or promote products derived
 *    from this software without specific, prior written permission.
 * 
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE
 * COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
 * INCIDENTAL, SPECIAL, EXEMPLARY OR CONSEQUENTIAL DAMAGES (INCLUDING,
 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
 * AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
 * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
 * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 * 
 * from:   @(#)pmap.c      7.7 (Berkeley)  5/12/91
 * $FreeBSD: src/sys/i386/i386/pmap.c,v 1.250.2.18 2002/03/06 22:48:53 silby Exp $
 * $DragonFly: src/sys/platform/vkernel/platform/pmap.c,v 1.31 2008/08/25 17:01:40 dillon Exp $
 */
/*
 * NOTE: PMAP_INVAL_ADD: In pc32 this function is called prior to adjusting
 * the PTE in the page table, because a cpu synchronization might be required.
 * The actual invalidation is delayed until the following call or flush.  In
 * the VKERNEL build this function is called prior to adjusting the PTE and
 * invalidates the table synchronously (not delayed), and is not SMP safe
 * as a consequence.
 */

#include <sys/types.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/stat.h>
#include <sys/mman.h>
#include <sys/vkernel.h>
#include <sys/proc.h>
#include <sys/thread.h>
#include <sys/user.h>
#include <sys/vmspace.h>

#include <vm/pmap.h>
#include <vm/vm_page.h>
#include <vm/vm_extern.h>
#include <vm/vm_kern.h>
#include <vm/vm_object.h>
#include <vm/vm_zone.h>
#include <vm/vm_pageout.h>

#include <machine/md_var.h>
#include <machine/pcb.h>
#include <machine/pmap_inval.h>
#include <machine/globaldata.h>

#include <sys/sysref2.h>

#include <assert.h>

struct pmap kernel_pmap;

static struct vm_zone pvzone;
static struct vm_object pvzone_obj;
static TAILQ_HEAD(,pmap) pmap_list = TAILQ_HEAD_INITIALIZER(pmap_list);
static int pv_entry_count;
static int pv_entry_max;
static int pv_entry_high_water;
static int pmap_pagedaemon_waken;
static boolean_t pmap_initialized = FALSE;
static int protection_codes[8];

static void i386_protection_init(void);
static void pmap_remove_all(vm_page_t m);
static int pmap_release_free_page(struct pmap *pmap, vm_page_t p);

#define MINPV   2048
#ifndef PMAP_SHPGPERPROC
#define PMAP_SHPGPERPROC 200
#endif

#define pmap_pde(m, v)  (&((m)->pm_pdir[(vm_offset_t)(v) >> PDRSHIFT]))

#define pte_prot(m, p) \
        (protection_codes[p & (VM_PROT_READ|VM_PROT_WRITE|VM_PROT_EXECUTE)])

void
pmap_init(void)
{
        int i;
        struct pv_entry *pvinit;

        for (i = 0; i < vm_page_array_size; i++) {
                vm_page_t m;

                m = &vm_page_array[i];
                TAILQ_INIT(&m->md.pv_list);
                m->md.pv_list_count = 0;
        }

        i = vm_page_array_size;
        if (i < MINPV)
                i = MINPV;
        pvinit = (struct pv_entry *)kmem_alloc(&kernel_map, i*sizeof(*pvinit));
        zbootinit(&pvzone, "PV ENTRY", sizeof(*pvinit), pvinit, i);
        pmap_initialized = TRUE;
}

void
pmap_init2(void)
{
        int shpgperproc = PMAP_SHPGPERPROC;

        TUNABLE_INT_FETCH("vm.pmap.shpgperproc", &shpgperproc);
        pv_entry_max = shpgperproc * maxproc + vm_page_array_size;
        TUNABLE_INT_FETCH("vm.pmap.pv_entries", &pv_entry_max);
        pv_entry_high_water = 9 * (pv_entry_max / 10);
        zinitna(&pvzone, &pvzone_obj, NULL, 0, pv_entry_max, ZONE_INTERRUPT, 1);
}

/*
 * Bootstrap the kernel_pmap so it can be used with pmap_enter().  
 *
 * NOTE! pm_pdir for the kernel pmap is offset so VA's translate
 * directly into PTD indexes (PTA is also offset for the same reason).
 * This is necessary because, for now, KVA is not mapped at address 0.
 *
 * Page table pages are not managed like they are in normal pmaps, so
 * no pteobj is needed.
 */
void
pmap_bootstrap(void)
{
        vm_pindex_t i = (vm_offset_t)KernelPTD >> PAGE_SHIFT;

        kernel_pmap.pm_pdir = KernelPTD - (KvaStart >> SEG_SHIFT);
        kernel_pmap.pm_pdirpte = KernelPTA[i];
        kernel_pmap.pm_count = 1;
        kernel_pmap.pm_active = (cpumask_t)-1 & ~CPUMASK_LOCK;
        TAILQ_INIT(&kernel_pmap.pm_pvlist);
        i386_protection_init();
}

/*
 * Initialize pmap0/vmspace0 .  Since process 0 never enters user mode we
 * just dummy it up so it works well enough for fork().
 *
 * In DragonFly, process pmaps may only be used to manipulate user address
 * space, never kernel address space.
 */
void
pmap_pinit0(struct pmap *pmap)
{
        pmap_pinit(pmap);
}

/************************************************************************
 *              Procedures to manage whole physical maps                *
 ************************************************************************
 *
 * Initialize a preallocated and zeroed pmap structure,
 * such as one in a vmspace structure.
 */
void
pmap_pinit(struct pmap *pmap)
{
        vm_page_t ptdpg;
        int npages;

        /*
         * No need to allocate page table space yet but we do need a valid
         * page directory table.
         */
        if (pmap->pm_pdir == NULL) {
                pmap->pm_pdir =
                    (pd_entry_t *)kmem_alloc_pageable(&kernel_map, PAGE_SIZE);
        }

        /*
         * allocate object for the pte array and page directory
         */
        npages = VPTE_PAGETABLE_SIZE +
                 (VM_MAX_USER_ADDRESS / PAGE_SIZE) * sizeof(vpte_t);
        npages = (npages + PAGE_MASK) / PAGE_SIZE;

        if (pmap->pm_pteobj == NULL)
                pmap->pm_pteobj = vm_object_allocate(OBJT_DEFAULT, npages);
        pmap->pm_pdindex = npages - 1;

        /*
         * allocate the page directory page
         */
        ptdpg = vm_page_grab(pmap->pm_pteobj, pmap->pm_pdindex,
                             VM_ALLOC_NORMAL | VM_ALLOC_RETRY);

        ptdpg->wire_count = 1;
        ++vmstats.v_wire_count;

        /* not usually mapped */
        vm_page_flag_clear(ptdpg, PG_MAPPED | PG_BUSY);
        ptdpg->valid = VM_PAGE_BITS_ALL;

        pmap_kenter((vm_offset_t)pmap->pm_pdir, VM_PAGE_TO_PHYS(ptdpg));
        pmap->pm_pdirpte = KernelPTA[(vm_offset_t)pmap->pm_pdir >> PAGE_SHIFT];
        if ((ptdpg->flags & PG_ZERO) == 0)
                bzero(pmap->pm_pdir, PAGE_SIZE);

        pmap->pm_count = 1;
        pmap->pm_active = 0;
        pmap->pm_ptphint = NULL;
        pmap->pm_cpucachemask = 0;
        TAILQ_INIT(&pmap->pm_pvlist);
        bzero(&pmap->pm_stats, sizeof pmap->pm_stats);
        pmap->pm_stats.resident_count = 1;
}

/*
 * Clean up a pmap structure so it can be physically freed
 */
void
pmap_puninit(pmap_t pmap)
{
        if (pmap->pm_pdir) {
                kmem_free(&kernel_map, (vm_offset_t)pmap->pm_pdir, PAGE_SIZE);
                pmap->pm_pdir = NULL;
        }
        if (pmap->pm_pteobj) {
                vm_object_deallocate(pmap->pm_pteobj);
                pmap->pm_pteobj = NULL;
        }
}


/*
 * Wire in kernel global address entries.  To avoid a race condition
 * between pmap initialization and pmap_growkernel, this procedure
 * adds the pmap to the master list (which growkernel scans to update),
 * then copies the template.
 *
 * In a virtual kernel there are no kernel global address entries.
 */
void
pmap_pinit2(struct pmap *pmap)
{
        crit_enter();
        TAILQ_INSERT_TAIL(&pmap_list, pmap, pm_pmnode);
        crit_exit();
}

/*
 * Release all resources held by the given physical map.
 *
 * Should only be called if the map contains no valid mappings.
 */
static int pmap_release_callback(struct vm_page *p, void *data);

void
pmap_release(struct pmap *pmap)
{
        struct mdglobaldata *gd = mdcpu;
        vm_object_t object = pmap->pm_pteobj;
        struct rb_vm_page_scan_info info;

        KKASSERT(pmap != &kernel_pmap);

#if defined(DIAGNOSTIC)
        if (object->ref_count != 1)
                panic("pmap_release: pteobj reference count != 1");
#endif
        /*
         * Once we destroy the page table, the mapping becomes invalid.
         * Don't waste time doing a madvise to invalidate the mapping, just
         * set cpucachemask to 0.
         */
        if (pmap->pm_pdir == gd->gd_PT1pdir) {
                gd->gd_PT1pdir = NULL;
                *gd->gd_PT1pde = 0;
                /* madvise(gd->gd_PT1map, SEG_SIZE, MADV_INVAL); */
        }
        if (pmap->pm_pdir == gd->gd_PT2pdir) {
                gd->gd_PT2pdir = NULL;
                *gd->gd_PT2pde = 0;
                /* madvise(gd->gd_PT2map, SEG_SIZE, MADV_INVAL); */
        }
        if (pmap->pm_pdir == gd->gd_PT3pdir) {
                gd->gd_PT3pdir = NULL;
                *gd->gd_PT3pde = 0;
                /* madvise(gd->gd_PT3map, SEG_SIZE, MADV_INVAL); */
        }
        
        info.pmap = pmap;
        info.object = object;
        crit_enter();
        TAILQ_REMOVE(&pmap_list, pmap, pm_pmnode);
        crit_exit();

        do {
                crit_enter();
                info.error = 0;
                info.mpte = NULL;
                info.limit = object->generation;

                vm_page_rb_tree_RB_SCAN(&object->rb_memq, NULL, 
                                        pmap_release_callback, &info);
                if (info.error == 0 && info.mpte) {
                        if (!pmap_release_free_page(pmap, info.mpte))
                                info.error = 1;
                }
                crit_exit();
        } while (info.error);

        /*
         * Leave the KVA reservation for pm_pdir cached for later reuse.
         */
        pmap->pm_pdirpte = 0;
        pmap->pm_cpucachemask = 0;
}

/*
 * Callback to release a page table page backing a directory
 * entry.
 */
static int
pmap_release_callback(struct vm_page *p, void *data)
{
        struct rb_vm_page_scan_info *info = data;

        if (p->pindex == info->pmap->pm_pdindex) {
                info->mpte = p;
                return(0);
        }
        if (!pmap_release_free_page(info->pmap, p)) {
                info->error = 1;
                return(-1);
        }
        if (info->object->generation != info->limit) {
                info->error = 1;
                return(-1);
        }
        return(0);
}

/*
 * Retire the given physical map from service.  Should only be called if
 * the map contains no valid mappings.
 */
void
pmap_destroy(pmap_t pmap)
{
        int count;

        if (pmap == NULL)
                return;

        count = --pmap->pm_count;
        if (count == 0) {
                pmap_release(pmap);
                panic("destroying a pmap is not yet implemented");
        }
}

/*
 * Add a reference to the specified pmap.
 */
void
pmap_reference(pmap_t pmap)
{
        if (pmap != NULL) {
                pmap->pm_count++;
        }
}

/************************************************************************
 *                      VMSPACE MANAGEMENT                              *
 ************************************************************************
 *
 * The VMSPACE management we do in our virtual kernel must be reflected
 * in the real kernel.  This is accomplished by making vmspace system
 * calls to the real kernel.
 */
void
cpu_vmspace_alloc(struct vmspace *vm)
{
        int r;
        void *rp;

#define LAST_EXTENT     (VM_MAX_USER_ADDRESS - 0x80000000)

        if (vmspace_create(&vm->vm_pmap, 0, NULL) < 0)
                panic("vmspace_create() failed");

        rp = vmspace_mmap(&vm->vm_pmap, (void *)0x00000000, 0x40000000,
                          PROT_READ|PROT_WRITE,
                          MAP_FILE|MAP_SHARED|MAP_VPAGETABLE|MAP_FIXED,
                          MemImageFd, 0);
        if (rp == MAP_FAILED)
                panic("vmspace_mmap: failed1");
        vmspace_mcontrol(&vm->vm_pmap, (void *)0x00000000, 0x40000000,
                         MADV_NOSYNC, 0);
        rp = vmspace_mmap(&vm->vm_pmap, (void *)0x40000000, 0x40000000,
                          PROT_READ|PROT_WRITE,
                          MAP_FILE|MAP_SHARED|MAP_VPAGETABLE|MAP_FIXED,
                          MemImageFd, 0x40000000);
        if (rp == MAP_FAILED)
                panic("vmspace_mmap: failed2");
        vmspace_mcontrol(&vm->vm_pmap, (void *)0x40000000, 0x40000000,
                         MADV_NOSYNC, 0);
        rp = vmspace_mmap(&vm->vm_pmap, (void *)0x80000000, LAST_EXTENT,
                          PROT_READ|PROT_WRITE,
                          MAP_FILE|MAP_SHARED|MAP_VPAGETABLE|MAP_FIXED,
                          MemImageFd, 0x80000000);
        vmspace_mcontrol(&vm->vm_pmap, (void *)0x80000000, LAST_EXTENT,
                         MADV_NOSYNC, 0);
        if (rp == MAP_FAILED)
                panic("vmspace_mmap: failed3");

        r = vmspace_mcontrol(&vm->vm_pmap, (void *)0x00000000, 0x40000000, 
                             MADV_SETMAP, vmspace_pmap(vm)->pm_pdirpte);
        if (r < 0)
                panic("vmspace_mcontrol: failed1");
        r = vmspace_mcontrol(&vm->vm_pmap, (void *)0x40000000, 0x40000000,
                             MADV_SETMAP, vmspace_pmap(vm)->pm_pdirpte);
        if (r < 0)
                panic("vmspace_mcontrol: failed2");
        r = vmspace_mcontrol(&vm->vm_pmap, (void *)0x80000000, LAST_EXTENT,
                             MADV_SETMAP, vmspace_pmap(vm)->pm_pdirpte);
        if (r < 0)
                panic("vmspace_mcontrol: failed3");
}

void
cpu_vmspace_free(struct vmspace *vm)
{
        if (vmspace_destroy(&vm->vm_pmap) < 0)
                panic("vmspace_destroy() failed");
}

/************************************************************************
 *          Procedures which operate directly on the kernel PMAP        *
 ************************************************************************/

/*
 * This maps the requested page table and gives us access to it.
 *
 * This routine can be called from a potentially preempting interrupt
 * thread or from a normal thread.
 */
static vpte_t *
get_ptbase(struct pmap *pmap, vm_offset_t va)
{
        struct mdglobaldata *gd = mdcpu;

        if (pmap == &kernel_pmap) {
                KKASSERT(va >= KvaStart && va < KvaEnd);
                return(KernelPTA + (va >> PAGE_SHIFT));
        } else if (pmap->pm_pdir == gd->gd_PT1pdir) {
                if ((pmap->pm_cpucachemask & gd->mi.gd_cpumask) == 0) {
                        *gd->gd_PT1pde = pmap->pm_pdirpte;
                        madvise(gd->gd_PT1map, SEG_SIZE, MADV_INVAL);
                        atomic_set_int(&pmap->pm_cpucachemask, gd->mi.gd_cpumask);
                }
                return(gd->gd_PT1map + (va >> PAGE_SHIFT));
        } else if (pmap->pm_pdir == gd->gd_PT2pdir) {
                if ((pmap->pm_cpucachemask & gd->mi.gd_cpumask) == 0) {
                        *gd->gd_PT2pde = pmap->pm_pdirpte;
                        madvise(gd->gd_PT2map, SEG_SIZE, MADV_INVAL);
                        atomic_set_int(&pmap->pm_cpucachemask, gd->mi.gd_cpumask);
                }
                return(gd->gd_PT2map + (va >> PAGE_SHIFT));
        }

        /*
         * If we aren't running from a potentially preempting interrupt,
         * load a new page table directory into the page table cache
         */
        if (gd->mi.gd_intr_nesting_level == 0 &&
            (gd->mi.gd_curthread->td_flags & TDF_INTTHREAD) == 0) {
                /*
                 * Choose one or the other and map the page table
                 * in the KVA space reserved for it.
                 */
                if ((gd->gd_PTflip = 1 - gd->gd_PTflip) == 0) {
                        gd->gd_PT1pdir = pmap->pm_pdir;
                        *gd->gd_PT1pde = pmap->pm_pdirpte;
                        madvise(gd->gd_PT1map, SEG_SIZE, MADV_INVAL);
                        atomic_set_int(&pmap->pm_cpucachemask,
                                        gd->mi.gd_cpumask);
                        return(gd->gd_PT1map + (va >> PAGE_SHIFT));
                } else {
                        gd->gd_PT2pdir = pmap->pm_pdir;
                        *gd->gd_PT2pde = pmap->pm_pdirpte;
                        madvise(gd->gd_PT2map, SEG_SIZE, MADV_INVAL);
                        atomic_set_int(&pmap->pm_cpucachemask,
                                        gd->mi.gd_cpumask);
                        return(gd->gd_PT2map + (va >> PAGE_SHIFT));
                }
        }

        /*
         * If we are running from a preempting interrupt use a private
         * map.  The caller must be in a critical section.
         */
        KKASSERT(IN_CRITICAL_SECT(curthread));
        if (pmap->pm_pdir == gd->gd_PT3pdir) {
                if ((pmap->pm_cpucachemask & gd->mi.gd_cpumask) == 0) {
                        *gd->gd_PT3pde = pmap->pm_pdirpte;
                        madvise(gd->gd_PT3map, SEG_SIZE, MADV_INVAL);
                        atomic_set_int(&pmap->pm_cpucachemask,
                                        gd->mi.gd_cpumask);
                }
        } else {
                gd->gd_PT3pdir = pmap->pm_pdir;
                *gd->gd_PT3pde = pmap->pm_pdirpte;
                madvise(gd->gd_PT3map, SEG_SIZE, MADV_INVAL);
                atomic_set_int(&pmap->pm_cpucachemask,
                                gd->mi.gd_cpumask);
        }
        return(gd->gd_PT3map + (va >> PAGE_SHIFT));
}

static vpte_t *
get_ptbase1(struct pmap *pmap, vm_offset_t va)
{
        struct mdglobaldata *gd = mdcpu;

        if (pmap == &kernel_pmap) {
                KKASSERT(va >= KvaStart && va < KvaEnd);
                return(KernelPTA + (va >> PAGE_SHIFT));
        } else if (pmap->pm_pdir == gd->gd_PT1pdir) {
                if ((pmap->pm_cpucachemask & gd->mi.gd_cpumask) == 0) {
                        *gd->gd_PT1pde = pmap->pm_pdirpte;
                        madvise(gd->gd_PT1map, SEG_SIZE, MADV_INVAL);
                        atomic_set_int(&pmap->pm_cpucachemask, gd->mi.gd_cpumask);
                }
                return(gd->gd_PT1map + (va >> PAGE_SHIFT));
        }
        KKASSERT(gd->mi.gd_intr_nesting_level == 0 &&
                 (gd->mi.gd_curthread->td_flags & TDF_INTTHREAD) == 0);
        gd->gd_PT1pdir = pmap->pm_pdir;
        *gd->gd_PT1pde = pmap->pm_pdirpte;
        madvise(gd->gd_PT1map, SEG_SIZE, MADV_INVAL);
        return(gd->gd_PT1map + (va >> PAGE_SHIFT));
}

static vpte_t *
get_ptbase2(struct pmap *pmap, vm_offset_t va)
{
        struct mdglobaldata *gd = mdcpu;

        if (pmap == &kernel_pmap) {
                KKASSERT(va >= KvaStart && va < KvaEnd);
                return(KernelPTA + (va >> PAGE_SHIFT));
        } else if (pmap->pm_pdir == gd->gd_PT2pdir) {
                if ((pmap->pm_cpucachemask & gd->mi.gd_cpumask) == 0) {
                        *gd->gd_PT2pde = pmap->pm_pdirpte;
                        madvise(gd->gd_PT2map, SEG_SIZE, MADV_INVAL);
                        atomic_set_int(&pmap->pm_cpucachemask, gd->mi.gd_cpumask);
                }
                return(gd->gd_PT2map + (va >> PAGE_SHIFT));
        }
        KKASSERT(gd->mi.gd_intr_nesting_level == 0 &&
                 (gd->mi.gd_curthread->td_flags & TDF_INTTHREAD) == 0);
        gd->gd_PT2pdir = pmap->pm_pdir;
        *gd->gd_PT2pde = pmap->pm_pdirpte;
        madvise(gd->gd_PT2map, SEG_SIZE, MADV_INVAL);
        return(gd->gd_PT2map + (va >> PAGE_SHIFT));
}

/*
 * Return a pointer to the page table entry for the specified va in the
 * specified pmap.  NULL is returned if there is no valid page table page
 * for the VA.
 */
static __inline vpte_t *
pmap_pte(struct pmap *pmap, vm_offset_t va)
{
        vpte_t *ptep;

        ptep = &pmap->pm_pdir[va >> SEG_SHIFT];
        if (*ptep & VPTE_PS)
                return(ptep);
        if (*ptep)
                return (get_ptbase(pmap, va));
        return(NULL);
}


/*
 * Enter a mapping into kernel_pmap.  Mappings created in this fashion
 * are not managed.  Mappings must be immediately accessible on all cpus.
 *
 * Call pmap_inval_pte() to invalidate the virtual pte and clean out the
 * real pmap and handle related races before storing the new vpte.
 */
void
pmap_kenter(vm_offset_t va, vm_paddr_t pa)
{
        vpte_t *ptep;
        vpte_t npte;

        KKASSERT(va >= KvaStart && va < KvaEnd);
        npte = (vpte_t)pa | VPTE_R | VPTE_W | VPTE_V;
        ptep = KernelPTA + (va >> PAGE_SHIFT);
        if (*ptep & VPTE_V)
                pmap_inval_pte(ptep, &kernel_pmap, va);
        *ptep = npte;
}

/*
 * Synchronize a kvm mapping originally made for the private use on
 * some other cpu so it can be used on all cpus.
 *
 * XXX add MADV_RESYNC to improve performance.
 */
void
pmap_kenter_sync(vm_offset_t va)
{
        madvise((void *)va, PAGE_SIZE, MADV_INVAL);
}

/*
 * Synchronize a kvm mapping originally made for the private use on
 * some other cpu so it can be used on our cpu.  Turns out to be the
 * same madvise() call, because we have to sync the real pmaps anyway.
 *
 * XXX add MADV_RESYNC to improve performance.
 */
void
pmap_kenter_sync_quick(vm_offset_t va)
{
        madvise((void *)va, PAGE_SIZE, MADV_INVAL);
}

#if 0
/*
 * Make a previously read-only kernel mapping R+W (not implemented by
 * virtual kernels).
 */
void
pmap_kmodify_rw(vm_offset_t va)
{
        *pmap_kpte(va) |= VPTE_R | VPTE_W;
        madvise((void *)va, PAGE_SIZE, MADV_INVAL);
}

/*
 * Make a kernel mapping non-cacheable (not applicable to virtual kernels)
 */
void
pmap_kmodify_nc(vm_offset_t va)
{
        *pmap_kpte(va) |= VPTE_N;
        madvise((void *)va, PAGE_SIZE, MADV_INVAL);
}

#endif

/*
 * Map a contiguous range of physical memory to a KVM
 */
vm_offset_t
pmap_map(vm_offset_t *virtp, vm_paddr_t start, vm_paddr_t end, int prot)
{
        vm_offset_t     sva, virt;

        sva = virt = *virtp;
        while (start < end) {
                pmap_kenter(virt, start);
                virt += PAGE_SIZE;
                start += PAGE_SIZE;
        }
        *virtp = virt;
        return (sva);
}

vpte_t *
pmap_kpte(vm_offset_t va)
{
        vpte_t *ptep;

        KKASSERT(va >= KvaStart && va < KvaEnd);
        ptep = KernelPTA + (va >> PAGE_SHIFT);
        return(ptep);
}

/*
 * Enter an unmanaged KVA mapping for the private use of the current
 * cpu only.  pmap_kenter_sync() may be called to make the mapping usable
 * by other cpus.
 *
 * It is illegal for the mapping to be accessed by other cpus unleess
 * pmap_kenter_sync*() is called.
 */
void
pmap_kenter_quick(vm_offset_t va, vm_paddr_t pa)
{
        vpte_t *ptep;
        vpte_t npte;

        KKASSERT(va >= KvaStart && va < KvaEnd);

        npte = (vpte_t)pa | VPTE_R | VPTE_W | VPTE_V;
        ptep = KernelPTA + (va >> PAGE_SHIFT);
        if (*ptep & VPTE_V)
                pmap_inval_pte_quick(ptep, &kernel_pmap, va);
        *ptep = npte;
}

/*
 * Make a temporary mapping for a physical address.  This is only intended
 * to be used for panic dumps.
 */
void *
pmap_kenter_temporary(vm_paddr_t pa, int i)
{
        pmap_kenter(crashdumpmap + (i * PAGE_SIZE), pa);
        return ((void *)crashdumpmap);
}

/*
 * Remove an unmanaged mapping created with pmap_kenter*().
 */
void
pmap_kremove(vm_offset_t va)
{
        vpte_t *ptep;

        KKASSERT(va >= KvaStart && va < KvaEnd);

        ptep = KernelPTA + (va >> PAGE_SHIFT);
        if (*ptep & VPTE_V)
                pmap_inval_pte(ptep, &kernel_pmap, va);
        *ptep = 0;
}

/*
 * Remove an unmanaged mapping created with pmap_kenter*() but synchronize
 * only with this cpu.
 *
 * Unfortunately because we optimize new entries by testing VPTE_V later
 * on, we actually still have to synchronize with all the cpus.  XXX maybe
 * store a junk value and test against 0 in the other places instead?
 */
void
pmap_kremove_quick(vm_offset_t va)
{
        vpte_t *ptep;

        KKASSERT(va >= KvaStart && va < KvaEnd);

        ptep = KernelPTA + (va >> PAGE_SHIFT);
        if (*ptep & VPTE_V)
                pmap_inval_pte(ptep, &kernel_pmap, va); /* NOT _quick */
        *ptep = 0;
}

/*
 * Extract the physical address from the kernel_pmap that is associated
 * with the specified virtual address.
 */
vm_paddr_t
pmap_kextract(vm_offset_t va)
{
        vpte_t *ptep;
        vm_paddr_t pa;

        KKASSERT(va >= KvaStart && va < KvaEnd);

        ptep = KernelPTA + (va >> PAGE_SHIFT);
        pa = (vm_paddr_t)(*ptep & VPTE_FRAME) | (va & PAGE_MASK);
        return(pa);
}

/*
 * Map a set of unmanaged VM pages into KVM.
 */
void
pmap_qenter(vm_offset_t va, struct vm_page **m, int count)
{
        KKASSERT(va >= KvaStart && va + count * PAGE_SIZE < KvaEnd);
        while (count) {
                vpte_t *ptep;

                ptep = KernelPTA + (va >> PAGE_SHIFT);
                if (*ptep & VPTE_V)
                        pmap_inval_pte(ptep, &kernel_pmap, va);
                *ptep = (vpte_t)(*m)->phys_addr | VPTE_R | VPTE_W | VPTE_V;
                --count;
                ++m;
                va += PAGE_SIZE;
        }
}

/*
 * Map a set of VM pages to kernel virtual memory.  If a mapping changes
 * clear the supplied mask.  The caller handles any SMP interactions.
 * The mask is used to provide the caller with hints on what SMP interactions
 * might be needed.
 */
void
pmap_qenter2(vm_offset_t va, struct vm_page **m, int count, cpumask_t *mask)
{
        cpumask_t cmask = mycpu->gd_cpumask;

        KKASSERT(va >= KvaStart && va + count * PAGE_SIZE < KvaEnd);
        while (count) {
                vpte_t *ptep;
                vpte_t npte;

                ptep = KernelPTA + (va >> PAGE_SHIFT);
                npte = (vpte_t)(*m)->phys_addr | VPTE_R | VPTE_W | VPTE_V;
                if (*ptep != npte) {
                        *mask = 0;
                        pmap_inval_pte_quick(ptep, &kernel_pmap, va);
                        *ptep = npte;
                } else if ((*mask & cmask) == 0) {
                        pmap_kenter_sync_quick(va);
                }
                --count;
                ++m;
                va += PAGE_SIZE;
        }
        *mask |= cmask;
}

/*
 * Undo the effects of pmap_qenter*().
 */
void
pmap_qremove(vm_offset_t va, int count)
{
        KKASSERT(va >= KvaStart && va + count * PAGE_SIZE < KvaEnd);
        while (count) {
                vpte_t *ptep;

                ptep = KernelPTA + (va >> PAGE_SHIFT);
                if (*ptep & VPTE_V)
                        pmap_inval_pte(ptep, &kernel_pmap, va);
                *ptep = 0;
                --count;
                va += PAGE_SIZE;
        }
}

/************************************************************************
 *        Misc support glue called by machine independant code          *
 ************************************************************************
 *
 * These routines are called by machine independant code to operate on
 * certain machine-dependant aspects of processes, threads, and pmaps.
 */

/*
 * Initialize MD portions of the thread structure.
 */
void
pmap_init_thread(thread_t td)
{
        /* enforce pcb placement */
        td->td_pcb = (struct pcb *)(td->td_kstack + td->td_kstack_size) - 1;
        td->td_savefpu = &td->td_pcb->pcb_save;
        td->td_sp = (char *)td->td_pcb - 16;
}

/*
 * This routine directly affects the fork perf for a process.
 */
void
pmap_init_proc(struct proc *p)
{
}

/*
 * Destroy the UPAGES for a process that has exited and disassociate
 * the process from its thread.
 */
void
pmap_dispose_proc(struct proc *p)
{
        KASSERT(p->p_lock == 0, ("attempt to dispose referenced proc! %p", p));
}

/*
 * We pre-allocate all page table pages for kernel virtual memory so
 * this routine will only be called if KVM has been exhausted.
 */
void
pmap_growkernel(vm_offset_t addr)
{
        addr = (addr + PAGE_SIZE * NPTEPG) & ~(PAGE_SIZE * NPTEPG - 1);

        if (addr > virtual_end - SEG_SIZE)
                panic("KVM exhausted");
        kernel_vm_end = addr;
}

/*
 * The modification bit is not tracked for any pages in this range. XXX
 * such pages in this maps should always use pmap_k*() functions and not
 * be managed anyhow.
 *
 * XXX User and kernel address spaces are independant for virtual kernels,
 * this function only applies to the kernel pmap.
 */
static int
pmap_track_modified(pmap_t pmap, vm_offset_t va)
{
        if (pmap != &kernel_pmap)
                return 1;
        if ((va < clean_sva) || (va >= clean_eva))
                return 1;
        else
                return 0;
}

/************************************************************************
 *          Procedures supporting managed page table pages              *
 ************************************************************************
 *
 * These procedures are used to track managed page table pages.  These pages
 * use the page table page's vm_page_t to track PTEs in the page.  The
 * page table pages themselves are arranged in a VM object, pmap->pm_pteobj.
 *
 * This allows the system to throw away page table pages for user processes
 * at will and reinstantiate them on demand.
 */

/*
 * This routine works like vm_page_lookup() but also blocks as long as the
 * page is busy.  This routine does not busy the page it returns.
 *
 * Unless the caller is managing objects whos pages are in a known state,
 * the call should be made with a critical section held so the page's object
 * association remains valid on return.
 */
static vm_page_t
pmap_page_lookup(vm_object_t object, vm_pindex_t pindex)
{
        vm_page_t m;
                         
retry:
        m = vm_page_lookup(object, pindex);
        if (m && vm_page_sleep_busy(m, FALSE, "pplookp"))
                goto retry;
        return(m);
}

/*
 * This routine unholds page table pages, and if the hold count
 * drops to zero, then it decrements the wire count.
 *
 * We must recheck that this is the last hold reference after busy-sleeping
 * on the page.
 */
static int 
_pmap_unwire_pte_hold(pmap_t pmap, vm_page_t m) 
{
        while (vm_page_sleep_busy(m, FALSE, "pmuwpt"))
                ;
        KASSERT(m->queue == PQ_NONE,
                ("_pmap_unwire_pte_hold: %p->queue != PQ_NONE", m));

        if (m->hold_count == 1) {
                /*
                 * Unmap the page table page.  
                 */
                vm_page_busy(m);
                KKASSERT(pmap->pm_pdir[m->pindex] != 0);
                pmap_inval_pde(&pmap->pm_pdir[m->pindex], pmap, 
                                (vm_offset_t)m->pindex << SEG_SHIFT);
                KKASSERT(pmap->pm_stats.resident_count > 0);
                --pmap->pm_stats.resident_count;

                if (pmap->pm_ptphint == m)
                        pmap->pm_ptphint = NULL;

                /*
                 * This was our last hold, the page had better be unwired
                 * after we decrement wire_count.
                 *
                 * FUTURE NOTE: shared page directory page could result in
                 * multiple wire counts.
                 */
                vm_page_unhold(m);
                --m->wire_count;
                KKASSERT(m->wire_count == 0);
                --vmstats.v_wire_count;
                vm_page_flag_clear(m, PG_MAPPED | PG_WRITEABLE);
                vm_page_flash(m);
                vm_page_free_zero(m);
                return 1;
        }
        KKASSERT(m->hold_count > 1);
        vm_page_unhold(m);
        return 0;
}

static __inline int
pmap_unwire_pte_hold(pmap_t pmap, vm_page_t m)
{
        KKASSERT(m->hold_count > 0);
        if (m->hold_count > 1) {
                vm_page_unhold(m);
                return 0;
        } else {
                return _pmap_unwire_pte_hold(pmap, m);
        }
}

/*
 * After removing a page table entry, this routine is used to
 * conditionally free the page, and manage the hold/wire counts.
 */
static int
pmap_unuse_pt(pmap_t pmap, vm_offset_t va, vm_page_t mpte)
{
        unsigned ptepindex;

        if (mpte == NULL) {
                /*
                 * page table pages in the kernel_pmap are not managed.
                 */
                if (pmap == &kernel_pmap)
                        return(0);
                ptepindex = (va >> PDRSHIFT);
                if (pmap->pm_ptphint &&
                        (pmap->pm_ptphint->pindex == ptepindex)) {
                        mpte = pmap->pm_ptphint;
                } else {
                        mpte = pmap_page_lookup( pmap->pm_pteobj, ptepindex);
                        pmap->pm_ptphint = mpte;
                }
        }
        return pmap_unwire_pte_hold(pmap, mpte);
}

/*
 * Attempt to release and free the vm_page backing a page directory page
 * in a pmap.  Returns 1 on success, 0 on failure (if the procedure had
 * to sleep).
 */
static int
pmap_release_free_page(struct pmap *pmap, vm_page_t p)
{
        vpte_t *pde = pmap->pm_pdir;

        /*
         * This code optimizes the case of freeing non-busy
         * page-table pages.  Those pages are zero now, and
         * might as well be placed directly into the zero queue.
         */
        if (vm_page_sleep_busy(p, FALSE, "pmaprl"))
                return 0;

        vm_page_busy(p);
        KKASSERT(pmap->pm_stats.resident_count > 0);
        --pmap->pm_stats.resident_count;

        if (p->hold_count)  {
                panic("pmap_release: freeing held page table page");
        }
        /*
         * Page directory pages need to have the kernel stuff cleared, so
         * they can go into the zero queue also.
         *
         * In virtual kernels there is no 'kernel stuff'.  For the moment
         * I just make sure the whole thing has been zero'd even though
         * it should already be completely zero'd.
         *
         * pmaps for vkernels do not self-map because they do not share
         * their address space with the vkernel.  Clearing of pde[] thus
         * only applies to page table pages and not to the page directory
         * page.
         */
        if (p->pindex == pmap->pm_pdindex) {
                bzero(pde, VPTE_PAGETABLE_SIZE);
                pmap_kremove((vm_offset_t)pmap->pm_pdir);
        } else {
                KKASSERT(pde[p->pindex] != 0);
                pmap_inval_pde(&pde[p->pindex], pmap, 
                                (vm_offset_t)p->pindex << SEG_SHIFT);
        }

        /*
         * Clear the matching hint
         */
        if (pmap->pm_ptphint && (pmap->pm_ptphint->pindex == p->pindex))
                pmap->pm_ptphint = NULL;

        /*
         * And throw the page away.  The page is completely zero'd out so
         * optimize the free call.
         */
        p->wire_count--;
        vmstats.v_wire_count--;
        vm_page_free_zero(p);
        return 1;
}

/*
 * This routine is called if the page table page is not mapped in the page
 * table directory.
 *
 * The routine is broken up into two parts for readability.
 *
 * It must return a held mpte and map the page directory page as required.
 * Because vm_page_grab() can block, we must re-check pm_pdir[ptepindex]
 */
static vm_page_t
_pmap_allocpte(pmap_t pmap, unsigned ptepindex)
{
        vm_paddr_t ptepa;
        vm_page_t m;

        /*
         * Find or fabricate a new pagetable page.  A busied page will be
         * returned.  This call may block.
         */
        m = vm_page_grab(pmap->pm_pteobj, ptepindex,
                         VM_ALLOC_NORMAL | VM_ALLOC_ZERO | VM_ALLOC_RETRY);

        KASSERT(m->queue == PQ_NONE,
                ("_pmap_allocpte: %p->queue != PQ_NONE", m));

        /*
         * Increment the hold count for the page we will be returning to
         * the caller.
         */
        m->hold_count++;

        /*
         * It is possible that someone else got in and mapped by the page
         * directory page while we were blocked, if so just unbusy and
         * return the held page.
         */
        if ((ptepa = pmap->pm_pdir[ptepindex]) != 0) {
                KKASSERT((ptepa & VPTE_FRAME) == VM_PAGE_TO_PHYS(m));
                vm_page_wakeup(m);
                return(m);
        }

        if (m->wire_count == 0)
                vmstats.v_wire_count++;
        m->wire_count++;

        /*
         * Map the pagetable page into the process address space, if
         * it isn't already there.
         */
        ++pmap->pm_stats.resident_count;

        ptepa = VM_PAGE_TO_PHYS(m);
        pmap->pm_pdir[ptepindex] = (vpte_t)ptepa | VPTE_R | VPTE_W | VPTE_V |
                                   VPTE_A | VPTE_M;

        /*
         * We are likely about to access this page table page, so set the
         * page table hint to reduce overhead.
         */
        pmap->pm_ptphint = m;

        /*
         * Try to use the new mapping, but if we cannot, then
         * do it with the routine that maps the page explicitly.
         */
        if ((m->flags & PG_ZERO) == 0)
                pmap_zero_page(ptepa);

        m->valid = VM_PAGE_BITS_ALL;
        vm_page_flag_clear(m, PG_ZERO);
        vm_page_flag_set(m, PG_MAPPED);
        vm_page_wakeup(m);

        return (m);
}

/*
 * Determine the page table page required to access the VA in the pmap
 * and allocate it if necessary.  Return a held vm_page_t for the page.
 *
 * Only used with user pmaps.
 */
static vm_page_t
pmap_allocpte(pmap_t pmap, vm_offset_t va)
{
        unsigned ptepindex;
        vm_offset_t ptepa;
        vm_page_t m;

        /*
         * Calculate pagetable page index
         */
        ptepindex = va >> PDRSHIFT;

        /*
         * Get the page directory entry
         */
        ptepa = (vm_offset_t) pmap->pm_pdir[ptepindex];

        /*
         * This supports switching from a 4MB page to a
         * normal 4K page.
         */
        if (ptepa & VPTE_PS) {
                KKASSERT(pmap->pm_pdir[ptepindex] != 0);
                pmap_inval_pde(&pmap->pm_pdir[ptepindex], pmap,
                               (vm_offset_t)ptepindex << SEG_SHIFT);
                ptepa = 0;
        }

        /*
         * If the page table page is mapped, we just increment the
         * hold count, and activate it.
         */
        if (ptepa) {
                /*
                 * In order to get the page table page, try the
                 * hint first.
                 */
                if (pmap->pm_ptphint &&
                        (pmap->pm_ptphint->pindex == ptepindex)) {
                        m = pmap->pm_ptphint;
                } else {
                        m = pmap_page_lookup( pmap->pm_pteobj, ptepindex);
                        pmap->pm_ptphint = m;
                }
                m->hold_count++;
                return m;
        }
        /*
         * Here if the pte page isn't mapped, or if it has been deallocated.
         */
        return _pmap_allocpte(pmap, ptepindex);
}

/************************************************************************
 *                      Managed pages in pmaps                          *
 ************************************************************************
 *
 * All pages entered into user pmaps and some pages entered into the kernel
 * pmap are managed, meaning that pmap_protect() and other related management
 * functions work on these pages.
 */

/*
 * free the pv_entry back to the free list.  This function may be
 * called from an interrupt.
 */
static __inline void
free_pv_entry(pv_entry_t pv)
{
        pv_entry_count--;
        zfree(&pvzone, pv);
}

/*
 * get a new pv_entry, allocating a block from the system
 * when needed.  This function may be called from an interrupt.
 */
static pv_entry_t
get_pv_entry(void)
{
        pv_entry_count++;
        if (pv_entry_high_water &&
            (pv_entry_count > pv_entry_high_water) &&
            (pmap_pagedaemon_waken == 0)) {
                pmap_pagedaemon_waken = 1;
                wakeup (&vm_pages_needed);
        }
        return zalloc(&pvzone);
}

/*
 * This routine is very drastic, but can save the system
 * in a pinch.
 */
void
pmap_collect(void)
{
        int i;
        vm_page_t m;
        static int warningdone=0;

        if (pmap_pagedaemon_waken == 0)
                return;
        pmap_pagedaemon_waken = 0;

        if (warningdone < 5) {
                kprintf("pmap_collect: collecting pv entries -- suggest increasing PMAP_SHPGPERPROC\n");
                warningdone++;
        }

        for(i = 0; i < vm_page_array_size; i++) {
                m = &vm_page_array[i];
                if (m->wire_count || m->hold_count || m->busy ||
                    (m->flags & PG_BUSY))
                        continue;
                pmap_remove_all(m);
        }
}
        
/*
 * If it is the first entry on the list, it is actually
 * in the header and we must copy the following entry up
 * to the header.  Otherwise we must search the list for
 * the entry.  In either case we free the now unused entry.
 */
static int
pmap_remove_entry(struct pmap *pmap, vm_page_t m, vm_offset_t va)
{
        pv_entry_t pv;
        int rtval;

        crit_enter();
        if (m->md.pv_list_count < pmap->pm_stats.resident_count) {
                TAILQ_FOREACH(pv, &m->md.pv_list, pv_list) {
                        if (pmap == pv->pv_pmap && va == pv->pv_va) 
                                break;
                }
        } else {
                TAILQ_FOREACH(pv, &pmap->pm_pvlist, pv_plist) {
                        if (va == pv->pv_va) 
                                break;
                }
        }

        /*
         * Note that pv_ptem is NULL if the page table page itself is not
         * managed, even if the page being removed IS managed.
         */
        rtval = 0;

        TAILQ_REMOVE(&m->md.pv_list, pv, pv_list);
        m->md.pv_list_count--;
        TAILQ_REMOVE(&pmap->pm_pvlist, pv, pv_plist);
        if (TAILQ_EMPTY(&m->md.pv_list))
                vm_page_flag_clear(m, PG_MAPPED | PG_WRITEABLE);
        ++pmap->pm_generation;
        rtval = pmap_unuse_pt(pmap, va, pv->pv_ptem);
        free_pv_entry(pv);

        crit_exit();
        return rtval;
}

/*
 * Create a pv entry for page at pa for (pmap, va).  If the page table page
 * holding the VA is managed, mpte will be non-NULL.
 */
static void
pmap_insert_entry(pmap_t pmap, vm_offset_t va, vm_page_t mpte, vm_page_t m)
{
        pv_entry_t pv;

        crit_enter();
        pv = get_pv_entry();
        pv->pv_va = va;
        pv->pv_pmap = pmap;
        pv->pv_ptem = mpte;

        TAILQ_INSERT_TAIL(&pmap->pm_pvlist, pv, pv_plist);
        TAILQ_INSERT_TAIL(&m->md.pv_list, pv, pv_list);
        ++pmap->pm_generation;
        m->md.pv_list_count++;

        crit_exit();
}

/*
 * pmap_remove_pte: do the things to unmap a page in a process
 */
static int
pmap_remove_pte(struct pmap *pmap, vpte_t *ptq, vm_offset_t va)
{
        vpte_t oldpte;
        vm_page_t m;

        oldpte = pmap_inval_loadandclear(ptq, pmap, va);
        if (oldpte & VPTE_WIRED)
                --pmap->pm_stats.wired_count;
        KKASSERT(pmap->pm_stats.wired_count >= 0);

#if 0
        /*
         * Machines that don't support invlpg, also don't support
         * VPTE_G.  XXX VPTE_G is disabled for SMP so don't worry about
         * the SMP case.
         */
        if (oldpte & VPTE_G)
                madvise((void *)va, PAGE_SIZE, MADV_INVAL);
#endif
        KKASSERT(pmap->pm_stats.resident_count > 0);
        --pmap->pm_stats.resident_count;
        if (oldpte & VPTE_MANAGED) {
                m = PHYS_TO_VM_PAGE(oldpte);
                if (oldpte & VPTE_M) {
#if defined(PMAP_DIAGNOSTIC)
                        if (pmap_nw_modified((pt_entry_t) oldpte)) {
                                kprintf(
        "pmap_remove: modified page not writable: va: 0x%x, pte: 0x%x\n",
                                    va, oldpte);
                        }
#endif
                        if (pmap_track_modified(pmap, va))
                                vm_page_dirty(m);
                }
                if (oldpte & VPTE_A)
                        vm_page_flag_set(m, PG_REFERENCED);
                return pmap_remove_entry(pmap, m, va);
        } else {
                return pmap_unuse_pt(pmap, va, NULL);
        }

        return 0;
}

/*
 * pmap_remove_page:
 *
 *      Remove a single page from a process address space.
 *
 *      This function may not be called from an interrupt if the pmap is
 *      not kernel_pmap.
 */
static void
pmap_remove_page(struct pmap *pmap, vm_offset_t va)
{
        vpte_t *ptq;

        /*
         * if there is no pte for this address, just skip it!!!  Otherwise
         * get a local va for mappings for this pmap and remove the entry.
         */
        if (*pmap_pde(pmap, va) != 0) {
                ptq = get_ptbase(pmap, va);
                if (*ptq) {
                        pmap_remove_pte(pmap, ptq, va);
                }
        }
}

/*
 * pmap_remove:
 *
 *      Remove the given range of addresses from the specified map.
 *
 *      It is assumed that the start and end are properly
 *      rounded to the page size.
 *
 *      This function may not be called from an interrupt if the pmap is
 *      not kernel_pmap.
 */
void
pmap_remove(struct pmap *pmap, vm_offset_t sva, vm_offset_t eva)
{
        vpte_t *ptbase;
        vm_offset_t pdnxt;
        vm_offset_t ptpaddr;
        vm_pindex_t sindex, eindex;

        if (pmap == NULL)
                return;

        KKASSERT(pmap->pm_stats.resident_count >= 0);
        if (pmap->pm_stats.resident_count == 0)
                return;

        /*
         * special handling of removing one page.  a very
         * common operation and easy to short circuit some
         * code.
         */
        if (((sva + PAGE_SIZE) == eva) && 
                ((pmap->pm_pdir[(sva >> PDRSHIFT)] & VPTE_PS) == 0)) {
                pmap_remove_page(pmap, sva);
                return;
        }

        /*
         * Get a local virtual address for the mappings that are being
         * worked with.
         *
         * XXX this is really messy because the kernel pmap is not relative
         * to address 0
         */
        sindex = (sva >> PAGE_SHIFT);
        eindex = (eva >> PAGE_SHIFT);

        for (; sindex < eindex; sindex = pdnxt) {
                vpte_t pdirindex;

                /*
                 * Calculate index for next page table.
                 */
                pdnxt = ((sindex + NPTEPG) & ~(NPTEPG - 1));
                if (pmap->pm_stats.resident_count == 0)
                        break;

                pdirindex = sindex / NPDEPG;
                if (((ptpaddr = pmap->pm_pdir[pdirindex]) & VPTE_PS) != 0) {
                        KKASSERT(pmap->pm_pdir[pdirindex] != 0);
                        pmap->pm_stats.resident_count -= NBPDR / PAGE_SIZE;
                        pmap_inval_pde(&pmap->pm_pdir[pdirindex], pmap,
                                (vm_offset_t)pdirindex << SEG_SHIFT);
                        continue;
                }

                /*
                 * Weed out invalid mappings. Note: we assume that the page
                 * directory table is always allocated, and in kernel virtual.
                 */
                if (ptpaddr == 0)
                        continue;

                /*
                 * Limit our scan to either the end of the va represented
                 * by the current page table page, or to the end of the
                 * range being removed.
                 */
                if (pdnxt > eindex)
                        pdnxt = eindex;

                /*
                 * NOTE: pmap_remove_pte() can block.
                 */
                for (; sindex != pdnxt; sindex++) {
                        vm_offset_t va;

                        ptbase = get_ptbase(pmap, sindex << PAGE_SHIFT);
                        if (*ptbase == 0)
                                continue;
                        va = i386_ptob(sindex);
                        if (pmap_remove_pte(pmap, ptbase, va))
                                break;
                }
        }
}

/*
 * pmap_remove_all:
 *
 * Removes this physical page from all physical maps in which it resides.
 * Reflects back modify bits to the pager.
 *
 * This routine may not be called from an interrupt.
 */
static void
pmap_remove_all(vm_page_t m)
{
        vpte_t *pte, tpte;
        pv_entry_t pv;

#if defined(PMAP_DIAGNOSTIC)
        /*
         * XXX this makes pmap_page_protect(NONE) illegal for non-managed
         * pages!
         */
        if (!pmap_initialized || (m->flags & PG_FICTITIOUS)) {
                panic("pmap_page_protect: illegal for unmanaged page, va: 0x%08llx", (long long)VM_PAGE_TO_PHYS(m));
        }
#endif

        crit_enter();
        while ((pv = TAILQ_FIRST(&m->md.pv_list)) != NULL) {
                KKASSERT(pv->pv_pmap->pm_stats.resident_count > 0);
                --pv->pv_pmap->pm_stats.resident_count;

                pte = pmap_pte(pv->pv_pmap, pv->pv_va);
                KKASSERT(pte != NULL);

                tpte = pmap_inval_loadandclear(pte, pv->pv_pmap, pv->pv_va);
                if (tpte & VPTE_WIRED)
                        --pv->pv_pmap->pm_stats.wired_count;
                KKASSERT(pv->pv_pmap->pm_stats.wired_count >= 0);

                if (tpte & VPTE_A)
                        vm_page_flag_set(m, PG_REFERENCED);

                /*
                 * Update the vm_page_t clean and reference bits.
                 */
                if (tpte & VPTE_M) {
#if defined(PMAP_DIAGNOSTIC)
                        if (pmap_nw_modified((pt_entry_t) tpte)) {
                                kprintf(
        "pmap_remove_all: modified page not writable: va: 0x%x, pte: 0x%x\n",
                                    pv->pv_va, tpte);
                        }
#endif
                        if (pmap_track_modified(pv->pv_pmap, pv->pv_va))
                                vm_page_dirty(m);
                }
                TAILQ_REMOVE(&m->md.pv_list, pv, pv_list);
                TAILQ_REMOVE(&pv->pv_pmap->pm_pvlist, pv, pv_plist);
                ++pv->pv_pmap->pm_generation;
                m->md.pv_list_count--;
                if (TAILQ_EMPTY(&m->md.pv_list))
                        vm_page_flag_clear(m, PG_MAPPED | PG_WRITEABLE);
                pmap_unuse_pt(pv->pv_pmap, pv->pv_va, pv->pv_ptem);
                free_pv_entry(pv);
        }
        KKASSERT((m->flags & (PG_MAPPED | PG_WRITEABLE)) == 0);
        crit_exit();
}

/*
 * pmap_protect:
 *
 *      Set the physical protection on the specified range of this map
 *      as requested.
 *
 *      This function may not be called from an interrupt if the map is
 *      not the kernel_pmap.
 */
void
pmap_protect(pmap_t pmap, vm_offset_t sva, vm_offset_t eva, vm_prot_t prot)
{
        vpte_t *ptbase;
        vpte_t *ptep;
        vm_offset_t pdnxt, ptpaddr;
        vm_pindex_t sindex, eindex;
        vm_pindex_t sbase;

        if (pmap == NULL)
                return;

        if ((prot & VM_PROT_READ) == VM_PROT_NONE) {
                pmap_remove(pmap, sva, eva);
                return;
        }

        if (prot & VM_PROT_WRITE)
                return;

        ptbase = get_ptbase(pmap, sva);

        sindex = (sva >> PAGE_SHIFT);
        eindex = (eva >> PAGE_SHIFT);
        sbase = sindex;

        for (; sindex < eindex; sindex = pdnxt) {

                unsigned pdirindex;

                pdnxt = ((sindex + NPTEPG) & ~(NPTEPG - 1));

                pdirindex = sindex / NPDEPG;

                /*
                 * Clear the modified and writable bits for a 4m page.
                 * Throw away the modified bit (?)
                 */
                if (((ptpaddr = pmap->pm_pdir[pdirindex]) & VPTE_PS) != 0) {
                        pmap_clean_pde(&pmap->pm_pdir[pdirindex], pmap,
                                        (vm_offset_t)pdirindex << SEG_SHIFT);
                        pmap->pm_stats.resident_count -= NBPDR / PAGE_SIZE;
                        continue;
                }

                /*
                 * Weed out invalid mappings. Note: we assume that the page
                 * directory table is always allocated, and in kernel virtual.
                 */
                if (ptpaddr == 0)
                        continue;

                if (pdnxt > eindex) {
                        pdnxt = eindex;
                }

                for (; sindex != pdnxt; sindex++) {
                        vpte_t pbits;
                        vm_page_t m;

                        /*
                         * Clean managed pages and also check the accessed
                         * bit.  Just remove write perms for unmanaged
                         * pages.  Be careful of races, turning off write
                         * access will force a fault rather then setting
                         * the modified bit at an unexpected time.
                         */
                        ptep = &ptbase[sindex - sbase];
                        if (*ptep & VPTE_MANAGED) {
                                pbits = pmap_clean_pte(ptep, pmap,
                                                       i386_ptob(sindex));
                                m = NULL;
                                if (pbits & VPTE_A) {
                                        m = PHYS_TO_VM_PAGE(pbits);
                                        vm_page_flag_set(m, PG_REFERENCED);
                                        atomic_clear_int(ptep, VPTE_A);
                                }
                                if (pbits & VPTE_M) {
                                        if (pmap_track_modified(pmap, i386_ptob(sindex))) {
                                                if (m == NULL)
                                                        m = PHYS_TO_VM_PAGE(pbits);
                                                vm_page_dirty(m);
                                        }
                                }
                        } else {
                                pbits = pmap_setro_pte(ptep, pmap,
                                                       i386_ptob(sindex));
                        }
                }
        }
}

/*
 * Enter a managed page into a pmap.  If the page is not wired related pmap
 * data can be destroyed at any time for later demand-operation.
 *
 * Insert the vm_page (m) at virtual address (v) in (pmap), with the
 * specified protection, and wire the mapping if requested.
 *
 * NOTE: This routine may not lazy-evaluate or lose information.  The
 * page must actually be inserted into the given map NOW.
 *
 * NOTE: When entering a page at a KVA address, the pmap must be the
 * kernel_pmap.
 */
void
pmap_enter(pmap_t pmap, vm_offset_t va, vm_page_t m, vm_prot_t prot,
           boolean_t wired)
{
        vm_paddr_t pa;
        vpte_t *pte;
        vm_paddr_t opa;
        vm_offset_t origpte, newpte;
        vm_page_t mpte;

        if (pmap == NULL)
                return;

        va &= VPTE_FRAME;

        /*
         * Get the page table page.   The kernel_pmap's page table pages
         * are preallocated and have no associated vm_page_t.
         */
        if (pmap == &kernel_pmap)
                mpte = NULL;
        else
                mpte = pmap_allocpte(pmap, va);

        pte = pmap_pte(pmap, va);

        /*
         * Page Directory table entry not valid, we need a new PT page
         * and pmap_allocpte() didn't give us one.  Oops!
         */
        if (pte == NULL) {
                panic("pmap_enter: invalid page directory pmap=%p, va=0x%p\n",
                      pmap, (void *)va);
        }

        /*
         * Deal with races on the original mapping (though don't worry
         * about VPTE_A races) by cleaning it.  This will force a fault
         * if an attempt is made to write to the page.
         */
        pa = VM_PAGE_TO_PHYS(m) & VPTE_FRAME;
        origpte = pmap_clean_pte(pte, pmap, va);
        opa = origpte & VPTE_FRAME;

        if (origpte & VPTE_PS)
                panic("pmap_enter: attempted pmap_enter on 4MB page");

        /*
         * Mapping has not changed, must be protection or wiring change.
         */
        if (origpte && (opa == pa)) {
                /*
                 * Wiring change, just update stats. We don't worry about
                 * wiring PT pages as they remain resident as long as there
                 * are valid mappings in them. Hence, if a user page is wired,
                 * the PT page will be also.
                 */
                if (wired && ((origpte & VPTE_WIRED) == 0))
                        ++pmap->pm_stats.wired_count;
                else if (!wired && (origpte & VPTE_WIRED))
                        --pmap->pm_stats.wired_count;
                KKASSERT(pmap->pm_stats.wired_count >= 0);

                /*
                 * Remove the extra pte reference.  Note that we cannot
                 * optimize the RO->RW case because we have adjusted the
                 * wiring count above and may need to adjust the wiring
                 * bits below.
                 */
                if (mpte)
                        mpte->hold_count--;

                /*
                 * We might be turning off write access to the page,
                 * so we go ahead and sense modify status.
                 */
                if (origpte & VPTE_MANAGED) {
                        if ((origpte & VPTE_M) &&
                            pmap_track_modified(pmap, va)) {
                                vm_page_t om;
                                om = PHYS_TO_VM_PAGE(opa);
                                vm_page_dirty(om);
                        }
                        pa |= VPTE_MANAGED;
                        KKASSERT(m->flags & PG_MAPPED);
                }
                goto validate;
        } 
        /*
         * Mapping has changed, invalidate old range and fall through to
         * handle validating new mapping.
         */
        while (opa) {
                int err;
                err = pmap_remove_pte(pmap, pte, va);
                if (err)
                        panic("pmap_enter: pte vanished, va: %p", (void *)va);
                pte = pmap_pte(pmap, va);
                origpte = pmap_clean_pte(pte, pmap, va);
                opa = origpte & VPTE_FRAME;
                if (opa) {
                        kprintf("pmap_enter: Warning, raced pmap %p va %p\n",
                                pmap, (void *)va);
                }
        }

        /*
         * Enter on the PV list if part of our managed memory. Note that we
         * raise IPL while manipulating pv_table since pmap_enter can be
         * called at interrupt time.
         */
        if (pmap_initialized && 
            (m->flags & (PG_FICTITIOUS|PG_UNMANAGED)) == 0) {
                pmap_insert_entry(pmap, va, mpte, m);
                pa |= VPTE_MANAGED;
                vm_page_flag_set(m, PG_MAPPED);
        }

        /*
         * Increment counters
         */
        ++pmap->pm_stats.resident_count;
        if (wired)
                pmap->pm_stats.wired_count++;

validate:
        /*
         * Now validate mapping with desired protection/wiring.
         */
        newpte = (vm_offset_t) (pa | pte_prot(pmap, prot) | VPTE_V);

        if (wired)
                newpte |= VPTE_WIRED;
        if (pmap != &kernel_pmap)
                newpte |= VPTE_U;

        /*
         * If the mapping or permission bits are different from the
         * (now cleaned) original pte, an update is needed.  We've
         * already downgraded or invalidated the page so all we have
         * to do now is update the bits.
         *
         * XXX should we synchronize RO->RW changes to avoid another
         * fault?
         */
        if ((origpte & ~(VPTE_W|VPTE_M|VPTE_A)) != newpte) {
                *pte = newpte | VPTE_A;
                if (newpte & VPTE_W)
                        vm_page_flag_set(m, PG_WRITEABLE);
        }
        KKASSERT((newpte & VPTE_MANAGED) == 0 || m->flags & PG_MAPPED);
}

/*
 * This code works like pmap_enter() but assumes VM_PROT_READ and not-wired.
 *
 * Currently this routine may only be used on user pmaps, not kernel_pmap.
 */
void
pmap_enter_quick(pmap_t pmap, vm_offset_t va, vm_page_t m)
{
        vpte_t *pte;
        vm_paddr_t pa;
        vm_page_t mpte;
        unsigned ptepindex;
        vm_offset_t ptepa;

        KKASSERT(pmap != &kernel_pmap);

        KKASSERT(va >= VM_MIN_USER_ADDRESS && va < VM_MAX_USER_ADDRESS);

        /*
         * Calculate pagetable page (mpte), allocating it if necessary.
         *
         * A held page table page (mpte), or NULL, is passed onto the 
         * section following.
         */
        ptepindex = va >> PDRSHIFT;

        do {
                /*
                 * Get the page directory entry
                 */
                ptepa = (vm_offset_t) pmap->pm_pdir[ptepindex];

                /*
                 * If the page table page is mapped, we just increment
                 * the hold count, and activate it.
                 */
                if (ptepa) {
                        if (ptepa & VPTE_PS)
                                panic("pmap_enter_quick: unexpected mapping into 4MB page");
                        if (pmap->pm_ptphint &&
                                (pmap->pm_ptphint->pindex == ptepindex)) {
                                mpte = pmap->pm_ptphint;
                        } else {
                                mpte = pmap_page_lookup( pmap->pm_pteobj, ptepindex);
                                pmap->pm_ptphint = mpte;
                        }
                        if (mpte)
                                mpte->hold_count++;
                } else {
                        mpte = _pmap_allocpte(pmap, ptepindex);
                }
        } while (mpte == NULL);

        /*
         * Ok, now that the page table page has been validated, get the pte.
         * If the pte is already mapped undo mpte's hold_count and
         * just return.
         */
        pte = pmap_pte(pmap, va);
        if (*pte) {
                pmap_unwire_pte_hold(pmap, mpte);
                return;
        }

        /*
         * Enter on the PV list if part of our managed memory. Note that we
         * raise IPL while manipulating pv_table since pmap_enter can be
         * called at interrupt time.
         */
        if ((m->flags & (PG_FICTITIOUS|PG_UNMANAGED)) == 0) {
                pmap_insert_entry(pmap, va, mpte, m);
                vm_page_flag_set(m, PG_MAPPED);
        }

        /*
         * Increment counters
         */
        ++pmap->pm_stats.resident_count;

        pa = VM_PAGE_TO_PHYS(m);

        /*
         * Now validate mapping with RO protection
         */
        if (m->flags & (PG_FICTITIOUS|PG_UNMANAGED))
                *pte = (vpte_t)pa | VPTE_V | VPTE_U;
        else
                *pte = (vpte_t)pa | VPTE_V | VPTE_U | VPTE_MANAGED;
        /*pmap_inval_add(&info, pmap, va); shouldn't be needed 0->valid */
        /*pmap_inval_flush(&info); don't need for vkernel */
}

/*
 * Extract the physical address for the translation at the specified
 * virtual address in the pmap.
 */
vm_paddr_t
pmap_extract(pmap_t pmap, vm_offset_t va)
{
        vm_paddr_t rtval;
        vpte_t pte;

        if (pmap && (pte = pmap->pm_pdir[va >> SEG_SHIFT]) != 0) {
                if (pte & VPTE_PS) {
                        rtval = pte & ~((vpte_t)(1 << SEG_SHIFT) - 1);
                        rtval |= va & SEG_MASK;
                } else {
                        pte = *get_ptbase(pmap, va);
                        rtval = (pte & VPTE_FRAME) | (va & PAGE_MASK);
                }
                return(rtval);
        }
        return(0);
}

#define MAX_INIT_PT (96)

/*
 * This routine preloads the ptes for a given object into the specified pmap.
 * This eliminates the blast of soft faults on process startup and
 * immediately after an mmap.
 */
static int pmap_object_init_pt_callback(vm_page_t p, void *data);

void
pmap_object_init_pt(pmap_t pmap, vm_offset_t addr, vm_prot_t prot,
                    vm_object_t object, vm_pindex_t pindex, 
                    vm_size_t size, int limit)
{
        struct rb_vm_page_scan_info info;
        struct lwp *lp;
        int psize;

        /*
         * We can't preinit if read access isn't set or there is no pmap
         * or object.
         */
        if ((prot & VM_PROT_READ) == 0 || pmap == NULL || object == NULL)
                return;

        /*
         * We can't preinit if the pmap is not the current pmap
         */
        lp = curthread->td_lwp;
        if (lp == NULL || pmap != vmspace_pmap(lp->lwp_vmspace))
                return;

        psize = size >> PAGE_SHIFT;

        if ((object->type != OBJT_VNODE) ||
                ((limit & MAP_PREFAULT_PARTIAL) && (psize > MAX_INIT_PT) &&
                        (object->resident_page_count > MAX_INIT_PT))) {
                return;
        }

        if (psize + pindex > object->size) {
                if (object->size < pindex)
                        return;           
                psize = object->size - pindex;
        }

        if (psize == 0)
                return;

        /*
         * Use a red-black scan to traverse the requested range and load
         * any valid pages found into the pmap.
         *
         * We cannot safely scan the object's memq unless we are in a
         * critical section since interrupts can remove pages from objects.
         */
        info.start_pindex = pindex;
        info.end_pindex = pindex + psize - 1;
        info.limit = limit;
        info.mpte = NULL;
        info.addr = addr;
        info.pmap = pmap;

        crit_enter();
        vm_page_rb_tree_RB_SCAN(&object->rb_memq, rb_vm_page_scancmp,
                                pmap_object_init_pt_callback, &info);
        crit_exit();
}

static
int
pmap_object_init_pt_callback(vm_page_t p, void *data)
{
        struct rb_vm_page_scan_info *info = data;
        vm_pindex_t rel_index;
        /*
         * don't allow an madvise to blow away our really
         * free pages allocating pv entries.
         */
        if ((info->limit & MAP_PREFAULT_MADVISE) &&
                vmstats.v_free_count < vmstats.v_free_reserved) {
                    return(-1);
        }
        if (((p->valid & VM_PAGE_BITS_ALL) == VM_PAGE_BITS_ALL) &&
            (p->busy == 0) && (p->flags & (PG_BUSY | PG_FICTITIOUS)) == 0) {
                if ((p->queue - p->pc) == PQ_CACHE)
                        vm_page_deactivate(p);
                vm_page_busy(p);
                rel_index = p->pindex - info->start_pindex;
                pmap_enter_quick(info->pmap,
                                 info->addr + i386_ptob(rel_index), p);
                vm_page_wakeup(p);
        }
        return(0);
}

/*
 * Return TRUE if the pmap is in shape to trivially
 * pre-fault the specified address.
 *
 * Returns FALSE if it would be non-trivial or if a
 * pte is already loaded into the slot.
 */
int
pmap_prefault_ok(pmap_t pmap, vm_offset_t addr)
{
        vpte_t *pte;

        if ((*pmap_pde(pmap, addr)) == 0)
                return(0);

        pte = get_ptbase(pmap, addr);
        if (*pte)
                return(0);

        return(1);
}

/*
 *      Routine:        pmap_change_wiring
 *      Function:       Change the wiring attribute for a map/virtual-address
 *                      pair.
 *      In/out conditions:
 *                      The mapping must already exist in the pmap.
 */
void
pmap_change_wiring(pmap_t pmap, vm_offset_t va, boolean_t wired)
{
        vpte_t *pte;

        if (pmap == NULL)
                return;

        pte = get_ptbase(pmap, va);

        if (wired && (*pte & VPTE_WIRED) == 0)
                ++pmap->pm_stats.wired_count;
        else if (!wired && (*pte & VPTE_WIRED))
                --pmap->pm_stats.wired_count;
        KKASSERT(pmap->pm_stats.wired_count >= 0);

        /*
         * Wiring is not a hardware characteristic so there is no need to
         * invalidate TLB.  However, in an SMP environment we must use
         * a locked bus cycle to update the pte (if we are not using 
         * the pmap_inval_*() API that is)... it's ok to do this for simple
         * wiring changes.
         */
        if (wired)
                atomic_set_int(pte, VPTE_WIRED);
        else
                atomic_clear_int(pte, VPTE_WIRED);
}

/*
 *      Copy the range specified by src_addr/len
 *      from the source map to the range dst_addr/len
 *      in the destination map.
 *
 *      This routine is only advisory and need not do anything.
 */
void
pmap_copy(pmap_t dst_pmap, pmap_t src_pmap, vm_offset_t dst_addr, 
        vm_size_t len, vm_offset_t src_addr)
{
        vm_offset_t addr;
        vm_offset_t end_addr = src_addr + len;
        vm_offset_t pdnxt;
        vpte_t *src_frame;
        vpte_t *dst_frame;
        vm_page_t m;

        /*
         * XXX BUGGY.  Amoung other things srcmpte is assumed to remain
         * valid through blocking calls, and that's just not going to
         * be the case.
         *
         * FIXME!
         */
        return;

        if (dst_addr != src_addr)
                return;
        if (dst_pmap->pm_pdir == NULL)
                return;
        if (src_pmap->pm_pdir == NULL)
                return;

        crit_enter();

        src_frame = get_ptbase1(src_pmap, src_addr);
        dst_frame = get_ptbase2(dst_pmap, src_addr);

        /*
         * critical section protection is required to maintain the page/object
         * association, interrupts can free pages and remove them from 
         * their objects.
         */
        for (addr = src_addr; addr < end_addr; addr = pdnxt) {
                vpte_t *src_pte, *dst_pte;
                vm_page_t dstmpte, srcmpte;
                vm_offset_t srcptepaddr;
                unsigned ptepindex;

                if (addr >= VM_MAX_USER_ADDRESS)
                        panic("pmap_copy: invalid to pmap_copy page tables\n");

                /*
                 * Don't let optional prefaulting of pages make us go
                 * way below the low water mark of free pages or way
                 * above high water mark of used pv entries.
                 */
                if (vmstats.v_free_count < vmstats.v_free_reserved ||
                    pv_entry_count > pv_entry_high_water)
                        break;
                
                pdnxt = ((addr + PAGE_SIZE*NPTEPG) & ~(PAGE_SIZE*NPTEPG - 1));
                ptepindex = addr >> PDRSHIFT;

                srcptepaddr = (vm_offset_t) src_pmap->pm_pdir[ptepindex];
                if (srcptepaddr == 0)
                        continue;
                        
                if (srcptepaddr & VPTE_PS) {
                        if (dst_pmap->pm_pdir[ptepindex] == 0) {
                                dst_pmap->pm_pdir[ptepindex] = (pd_entry_t) srcptepaddr;
                                dst_pmap->pm_stats.resident_count += NBPDR / PAGE_SIZE;
                        }
                        continue;
                }

                srcmpte = vm_page_lookup(src_pmap->pm_pteobj, ptepindex);
                if ((srcmpte == NULL) || (srcmpte->hold_count == 0) ||
                    (srcmpte->flags & PG_BUSY)) {
                        continue;
                }

                if (pdnxt > end_addr)
                        pdnxt = end_addr;

                src_pte = src_frame + ((addr - src_addr) >> PAGE_SHIFT);
                dst_pte = dst_frame + ((addr - src_addr) >> PAGE_SHIFT);
                while (addr < pdnxt) {
                        vpte_t ptetemp;

                        ptetemp = *src_pte;
                        /*
                         * we only virtual copy managed pages
                         */
                        if ((ptetemp & VPTE_MANAGED) != 0) {
                                /*
                                 * We have to check after allocpte for the
                                 * pte still being around...  allocpte can
                                 * block.
                                 *
                                 * pmap_allocpte can block, unfortunately
                                 * we have to reload the tables.
                                 */
                                dstmpte = pmap_allocpte(dst_pmap, addr);
                                src_frame = get_ptbase1(src_pmap, src_addr);
                                dst_frame = get_ptbase2(dst_pmap, src_addr);

                                if ((*dst_pte == 0) && (ptetemp = *src_pte) &&
                                    (ptetemp & VPTE_MANAGED) != 0) {
                                        /*
                                         * Clear the modified and accessed
                                         * (referenced) bits during the copy.
                                         *
                                         * We do not have to clear the write
                                         * bit to force a fault-on-modify
                                         * because the real kernel's target
                                         * pmap is empty and will fault anyway.
                                         */
                                        m = PHYS_TO_VM_PAGE(ptetemp);
                                        *dst_pte = ptetemp & ~(VPTE_M | VPTE_A);
                                        ++dst_pmap->pm_stats.resident_count;
                                        pmap_insert_entry(dst_pmap, addr,
                                                dstmpte, m);
                                        KKASSERT(m->flags & PG_MAPPED);
                                } else {
                                        pmap_unwire_pte_hold(dst_pmap, dstmpte);
                                }
                                if (dstmpte->hold_count >= srcmpte->hold_count)
                                        break;
                        }
                        addr += PAGE_SIZE;
                        src_pte++;
                        dst_pte++;
                }
        }
        crit_exit();
}       

/*
 * pmap_zero_page:
 *
 *      Zero the specified PA by mapping the page into KVM and clearing its
 *      contents.
 *
 *      This function may be called from an interrupt and no locking is
 *      required.
 */
void
pmap_zero_page(vm_paddr_t phys)
{
        struct mdglobaldata *gd = mdcpu;

        crit_enter();
        if (*gd->gd_CMAP3)
                panic("pmap_zero_page: CMAP3 busy");
        *gd->gd_CMAP3 = VPTE_V | VPTE_R | VPTE_W | (phys & VPTE_FRAME) | VPTE_A | VPTE_M;
        madvise(gd->gd_CADDR3, PAGE_SIZE, MADV_INVAL);

        bzero(gd->gd_CADDR3, PAGE_SIZE);
        *gd->gd_CMAP3 = 0;
        crit_exit();
}

/*
 * pmap_page_assertzero:
 *
 *      Assert that a page is empty, panic if it isn't.
 */
void
pmap_page_assertzero(vm_paddr_t phys)
{
        struct mdglobaldata *gd = mdcpu;
        int i;

        crit_enter();
        if (*gd->gd_CMAP3)
                panic("pmap_zero_page: CMAP3 busy");
        *gd->gd_CMAP3 = VPTE_V | VPTE_R | VPTE_W |
                        (phys & VPTE_FRAME) | VPTE_A | VPTE_M;
        madvise(gd->gd_CADDR3, PAGE_SIZE, MADV_INVAL);
        for (i = 0; i < PAGE_SIZE; i += 4) {
            if (*(int *)((char *)gd->gd_CADDR3 + i) != 0) {
                panic("pmap_page_assertzero() @ %p not zero!\n",
                    (void *)gd->gd_CADDR3);
            }
        }
        *gd->gd_CMAP3 = 0;
        crit_exit();
}

/*
 * pmap_zero_page:
 *
 *      Zero part of a physical page by mapping it into memory and clearing
 *      its contents with bzero.
 *
 *      off and size may not cover an area beyond a single hardware page.
 */
void
pmap_zero_page_area(vm_paddr_t phys, int off, int size)
{
        struct mdglobaldata *gd = mdcpu;

        crit_enter();
        if (*gd->gd_CMAP3)
                panic("pmap_zero_page: CMAP3 busy");
        *gd->gd_CMAP3 = VPTE_V | VPTE_R | VPTE_W |
                        (phys & VPTE_FRAME) | VPTE_A | VPTE_M;
        madvise(gd->gd_CADDR3, PAGE_SIZE, MADV_INVAL);

        bzero((char *)gd->gd_CADDR3 + off, size);
        *gd->gd_CMAP3 = 0;
        crit_exit();
}

/*
 * pmap_copy_page:
 *
 *      Copy the physical page from the source PA to the target PA.
 *      This function may be called from an interrupt.  No locking
 *      is required.
 */
void
pmap_copy_page(vm_paddr_t src, vm_paddr_t dst)
{
        struct mdglobaldata *gd = mdcpu;

        crit_enter();
        if (*(int *) gd->gd_CMAP1)
                panic("pmap_copy_page: CMAP1 busy");
        if (*(int *) gd->gd_CMAP2)
                panic("pmap_copy_page: CMAP2 busy");

        *(int *) gd->gd_CMAP1 = VPTE_V | VPTE_R | (src & PG_FRAME) | VPTE_A;
        *(int *) gd->gd_CMAP2 = VPTE_V | VPTE_R | VPTE_W | (dst & VPTE_FRAME) | VPTE_A | VPTE_M;

        madvise(gd->gd_CADDR1, PAGE_SIZE, MADV_INVAL);
        madvise(gd->gd_CADDR2, PAGE_SIZE, MADV_INVAL);

        bcopy(gd->gd_CADDR1, gd->gd_CADDR2, PAGE_SIZE);

        *(int *) gd->gd_CMAP1 = 0;
        *(int *) gd->gd_CMAP2 = 0;
        crit_exit();
}

/*
 * pmap_copy_page_frag:
 *
 *      Copy the physical page from the source PA to the target PA.
 *      This function may be called from an interrupt.  No locking
 *      is required.
 */
void
pmap_copy_page_frag(vm_paddr_t src, vm_paddr_t dst, size_t bytes)
{
        struct mdglobaldata *gd = mdcpu;

        crit_enter();
        if (*(int *) gd->gd_CMAP1)
                panic("pmap_copy_page: CMAP1 busy");
        if (*(int *) gd->gd_CMAP2)
                panic("pmap_copy_page: CMAP2 busy");

        *(int *) gd->gd_CMAP1 = VPTE_V | (src & VPTE_FRAME) | VPTE_A;
        *(int *) gd->gd_CMAP2 = VPTE_V | VPTE_R | VPTE_W | (dst & VPTE_FRAME) | VPTE_A | VPTE_M;

        madvise(gd->gd_CADDR1, PAGE_SIZE, MADV_INVAL);
        madvise(gd->gd_CADDR2, PAGE_SIZE, MADV_INVAL);

        bcopy((char *)gd->gd_CADDR1 + (src & PAGE_MASK),
              (char *)gd->gd_CADDR2 + (dst & PAGE_MASK),
              bytes);

        *(int *) gd->gd_CMAP1 = 0;
        *(int *) gd->gd_CMAP2 = 0;
        crit_exit();
}

/*
 * Returns true if the pmap's pv is one of the first
 * 16 pvs linked to from this page.  This count may
 * be changed upwards or downwards in the future; it
 * is only necessary that true be returned for a small
 * subset of pmaps for proper page aging.
 */
boolean_t
pmap_page_exists_quick(pmap_t pmap, vm_page_t m)
{
        pv_entry_t pv;
        int loops = 0;

        if (!pmap_initialized || (m->flags & PG_FICTITIOUS))
                return FALSE;

        crit_enter();

        TAILQ_FOREACH(pv, &m->md.pv_list, pv_list) {
                if (pv->pv_pmap == pmap) {
                        crit_exit();
                        return TRUE;
                }
                loops++;
                if (loops >= 16)
                        break;
        }
        crit_exit();
        return (FALSE);
}

/*
 * Remove all pages from specified address space
 * this aids process exit speeds.  Also, this code
 * is special cased for current process only, but
 * can have the more generic (and slightly slower)
 * mode enabled.  This is much faster than pmap_remove
 * in the case of running down an entire address space.
 */
void
pmap_remove_pages(pmap_t pmap, vm_offset_t sva, vm_offset_t eva)
{
        vpte_t *pte, tpte;
        pv_entry_t pv, npv;
        vm_page_t m;
        int iscurrentpmap;
        int32_t save_generation;
        struct lwp *lp;

        lp = curthread->td_lwp;
        if (lp && pmap == vmspace_pmap(lp->lwp_vmspace))
                iscurrentpmap = 1;
        else
                iscurrentpmap = 0;

        crit_enter();
        for (pv = TAILQ_FIRST(&pmap->pm_pvlist); pv; pv = npv) {
                if (pv->pv_va >= eva || pv->pv_va < sva) {
                        npv = TAILQ_NEXT(pv, pv_plist);
                        continue;
                }

                KKASSERT(pmap == pv->pv_pmap);

                pte = pmap_pte(pmap, pv->pv_va);

                /*
                 * We cannot remove wired pages from a process' mapping
                 * at this time
                 */
                if (*pte & VPTE_WIRED) {
                        npv = TAILQ_NEXT(pv, pv_plist);
                        continue;
                }
                tpte = pmap_inval_loadandclear(pte, pmap, pv->pv_va);

                m = PHYS_TO_VM_PAGE(tpte);

                KASSERT(m < &vm_page_array[vm_page_array_size],
                        ("pmap_remove_pages: bad tpte %x", tpte));

                KKASSERT(pmap->pm_stats.resident_count > 0);
                --pmap->pm_stats.resident_count;

                /*
                 * Update the vm_page_t clean and reference bits.
                 */
                if (tpte & VPTE_M) {
                        vm_page_dirty(m);
                }

                npv = TAILQ_NEXT(pv, pv_plist);
                TAILQ_REMOVE(&pmap->pm_pvlist, pv, pv_plist);
                save_generation = ++pmap->pm_generation;

                m->md.pv_list_count--;
                TAILQ_REMOVE(&m->md.pv_list, pv, pv_list);
                if (TAILQ_FIRST(&m->md.pv_list) == NULL)
                        vm_page_flag_clear(m, PG_MAPPED | PG_WRITEABLE);

                pmap_unuse_pt(pmap, pv->pv_va, pv->pv_ptem);
                free_pv_entry(pv);

                /*
                 * Restart the scan if we blocked during the unuse or free
                 * calls and other removals were made.
                 */
                if (save_generation != pmap->pm_generation) {
                        kprintf("Warning: pmap_remove_pages race-A avoided\n");
                        pv = TAILQ_FIRST(&pmap->pm_pvlist);
                }
        }
        crit_exit();
}

/*
 * pmap_testbit tests bits in active mappings of a VM page.
 */
static boolean_t
pmap_testbit(vm_page_t m, int bit)
{
        pv_entry_t pv;
        vpte_t *pte;

        if (!pmap_initialized || (m->flags & PG_FICTITIOUS))
                return FALSE;

        if (TAILQ_FIRST(&m->md.pv_list) == NULL)
                return FALSE;

        crit_enter();

        TAILQ_FOREACH(pv, &m->md.pv_list, pv_list) {
                /*
                 * if the bit being tested is the modified bit, then
                 * mark clean_map and ptes as never
                 * modified.
                 */
                if (bit & (VPTE_A|VPTE_M)) {
                        if (!pmap_track_modified(pv->pv_pmap, pv->pv_va))
                                continue;
                }

#if defined(PMAP_DIAGNOSTIC)
                if (!pv->pv_pmap) {
                        kprintf("Null pmap (tb) at va: 0x%x\n", pv->pv_va);
                        continue;
                }
#endif
                pte = pmap_pte(pv->pv_pmap, pv->pv_va);
                if (*pte & bit) {
                        crit_exit();
                        return TRUE;
                }
        }
        crit_exit();
        return (FALSE);
}

/*
 * This routine is used to clear bits in ptes.  Certain bits require special
 * handling, in particular (on virtual kernels) the VPTE_M (modify) bit.
 *
 * This routine is only called with certain VPTE_* bit combinations.
 */
static __inline void
pmap_clearbit(vm_page_t m, int bit)
{
        pv_entry_t pv;
        vpte_t *pte;
        vpte_t pbits;

        if (!pmap_initialized || (m->flags & PG_FICTITIOUS))
                return;

        crit_enter();

        /*
         * Loop over all current mappings setting/clearing as appropos If
         * setting RO do we need to clear the VAC?
         */
        TAILQ_FOREACH(pv, &m->md.pv_list, pv_list) {
                /*
                 * don't write protect pager mappings
                 */
                if (bit == VPTE_W) {
                        if (!pmap_track_modified(pv->pv_pmap, pv->pv_va))
                                continue;
                }

#if defined(PMAP_DIAGNOSTIC)
                if (!pv->pv_pmap) {
                        kprintf("Null pmap (cb) at va: 0x%x\n", pv->pv_va);
                        continue;
                }
#endif

                /*
                 * Careful here.  We can use a locked bus instruction to
                 * clear VPTE_A or VPTE_M safely but we need to synchronize
                 * with the target cpus when we mess with VPTE_W.
                 *
                 * On virtual kernels we must force a new fault-on-write
                 * in the real kernel if we clear the Modify bit ourselves,
                 * otherwise the real kernel will not get a new fault and
                 * will never set our Modify bit again. 
                 */
                pte = pmap_pte(pv->pv_pmap, pv->pv_va);
                if (*pte & bit) {
                        if (bit == VPTE_W) {
                                /*
                                 * We must also clear VPTE_M when clearing
                                 * VPTE_W
                                 */
                                pbits = pmap_clean_pte(pte, pv->pv_pmap,
                                                       pv->pv_va);
                                if (pbits & VPTE_M)
                                        vm_page_dirty(m);
                        } else if (bit == VPTE_M) {
                                /*
                                 * We do not have to make the page read-only
                                 * when clearing the Modify bit.  The real
                                 * kernel will make the real PTE read-only
                                 * or otherwise detect the write and set
                                 * our VPTE_M again simply by us invalidating
                                 * the real kernel VA for the pmap (as we did
                                 * above).  This allows the real kernel to
                                 * handle the write fault without forwarding
                                 * the fault to us.
                                 */
                                atomic_clear_int(pte, VPTE_M);
                        } else if ((bit & (VPTE_W|VPTE_M)) == (VPTE_W|VPTE_M)) {
                                /*
                                 * We've been asked to clear W & M, I guess
                                 * the caller doesn't want us to update
                                 * the dirty status of the VM page.
                                 */
                                pmap_clean_pte(pte, pv->pv_pmap, pv->pv_va);
                        } else {
                                /*
                                 * We've been asked to clear bits that do
                                 * not interact with hardware.
                                 */
                                atomic_clear_int(pte, bit);
                        }
                }
        }
        crit_exit();
}

/*
 *      pmap_page_protect:
 *
 *      Lower the permission for all mappings to a given page.
 */
void
pmap_page_protect(vm_page_t m, vm_prot_t prot)
{
        if ((prot & VM_PROT_WRITE) == 0) {
                if (prot & (VM_PROT_READ | VM_PROT_EXECUTE)) {
                        pmap_clearbit(m, VPTE_W);
                        vm_page_flag_clear(m, PG_WRITEABLE);
                } else {
                        pmap_remove_all(m);
                }
        }
}

vm_paddr_t
pmap_phys_address(vm_pindex_t ppn)
{
        return (i386_ptob(ppn));
}

/*
 *      pmap_ts_referenced:
 *
 *      Return a count of reference bits for a page, clearing those bits.
 *      It is not necessary for every reference bit to be cleared, but it
 *      is necessary that 0 only be returned when there are truly no
 *      reference bits set.
 *
 *      XXX: The exact number of bits to check and clear is a matter that
 *      should be tested and standardized at some point in the future for
 *      optimal aging of shared pages.
 */
int
pmap_ts_referenced(vm_page_t m)
{
        pv_entry_t pv, pvf, pvn;
        vpte_t *pte;
        int rtval = 0;

        if (!pmap_initialized || (m->flags & PG_FICTITIOUS))
                return (rtval);

        crit_enter();

        if ((pv = TAILQ_FIRST(&m->md.pv_list)) != NULL) {

                pvf = pv;

                do {
                        pvn = TAILQ_NEXT(pv, pv_list);

                        TAILQ_REMOVE(&m->md.pv_list, pv, pv_list);

                        TAILQ_INSERT_TAIL(&m->md.pv_list, pv, pv_list);

                        if (!pmap_track_modified(pv->pv_pmap, pv->pv_va))
                                continue;

                        pte = pmap_pte(pv->pv_pmap, pv->pv_va);

                        if (pte && (*pte & VPTE_A)) {
#ifdef SMP
                                atomic_clear_int(pte, VPTE_A);
#else
                                atomic_clear_int_nonlocked(pte, VPTE_A);
#endif
                                rtval++;
                                if (rtval > 4) {
                                        break;
                                }
                        }
                } while ((pv = pvn) != NULL && pv != pvf);
        }
        crit_exit();

        return (rtval);
}

/*
 *      pmap_is_modified:
 *
 *      Return whether or not the specified physical page was modified
 *      in any physical maps.
 */
boolean_t
pmap_is_modified(vm_page_t m)
{
        return pmap_testbit(m, VPTE_M);
}

/*
 *      Clear the modify bits on the specified physical page.
 */
void
pmap_clear_modify(vm_page_t m)
{
        pmap_clearbit(m, VPTE_M);
}

/*
 *      pmap_clear_reference:
 *
 *      Clear the reference bit on the specified physical page.
 */
void
pmap_clear_reference(vm_page_t m)
{
        pmap_clearbit(m, VPTE_A);
}

/*
 * Miscellaneous support routines follow
 */

static void
i386_protection_init(void)
{
        int *kp, prot;

        kp = protection_codes;
        for (prot = 0; prot < 8; prot++) {
                if (prot & VM_PROT_READ)
                        *kp |= VPTE_R;
                if (prot & VM_PROT_WRITE)
                        *kp |= VPTE_W;
                if (prot & VM_PROT_EXECUTE)
                        *kp |= VPTE_X;
                ++kp;
        }
}

#if 0

/*
 * Map a set of physical memory pages into the kernel virtual
 * address space. Return a pointer to where it is mapped. This
 * routine is intended to be used for mapping device memory,
 * NOT real memory.
 *
 * NOTE: we can't use pgeflag unless we invalidate the pages one at
 * a time.
 */
void *
pmap_mapdev(vm_paddr_t pa, vm_size_t size)
{
        vm_offset_t va, tmpva, offset;
        vpte_t *pte;

        offset = pa & PAGE_MASK;
        size = roundup(offset + size, PAGE_SIZE);

        va = kmem_alloc_nofault(&kernel_map, size);
        if (!va)
                panic("pmap_mapdev: Couldn't alloc kernel virtual memory");

        pa = pa & VPTE_FRAME;
        for (tmpva = va; size > 0;) {
                pte = KernelPTA + (tmpva >> PAGE_SHIFT);
                *pte = pa | VPTE_R | VPTE_W | VPTE_V; /* | pgeflag; */
                size -= PAGE_SIZE;
                tmpva += PAGE_SIZE;
                pa += PAGE_SIZE;
        }
        cpu_invltlb();
        smp_invltlb();

        return ((void *)(va + offset));
}

void
pmap_unmapdev(vm_offset_t va, vm_size_t size)
{
        vm_offset_t base, offset;

        base = va & VPTE_FRAME;
        offset = va & PAGE_MASK;
        size = roundup(offset + size, PAGE_SIZE);
        pmap_qremove(va, size >> PAGE_SHIFT);
        kmem_free(&kernel_map, base, size);
}

#endif

/*
 * perform the pmap work for mincore
 */
int
pmap_mincore(pmap_t pmap, vm_offset_t addr)
{
        vpte_t *ptep, pte;
        vm_page_t m;
        int val = 0;
        
        ptep = pmap_pte(pmap, addr);
        if (ptep == 0) {
                return 0;
        }

        if ((pte = *ptep) != 0) {
                vm_offset_t pa;

                val = MINCORE_INCORE;
                if ((pte & VPTE_MANAGED) == 0)
                        return val;

                pa = pte & VPTE_FRAME;

                m = PHYS_TO_VM_PAGE(pa);

                /*
                 * Modified by us
                 */
                if (pte & VPTE_M)
                        val |= MINCORE_MODIFIED|MINCORE_MODIFIED_OTHER;
                /*
                 * Modified by someone
                 */
                else if (m->dirty || pmap_is_modified(m))
                        val |= MINCORE_MODIFIED_OTHER;
                /*
                 * Referenced by us
                 */
                if (pte & VPTE_A)
                        val |= MINCORE_REFERENCED|MINCORE_REFERENCED_OTHER;

                /*
                 * Referenced by someone
                 */
                else if ((m->flags & PG_REFERENCED) || pmap_ts_referenced(m)) {
                        val |= MINCORE_REFERENCED_OTHER;
                        vm_page_flag_set(m, PG_REFERENCED);
                }
        } 
        return val;
}

void
pmap_replacevm(struct proc *p, struct vmspace *newvm, int adjrefs)
{
        struct vmspace *oldvm;
        struct lwp *lp;

        oldvm = p->p_vmspace;
        crit_enter();
        if (oldvm != newvm) {
                p->p_vmspace = newvm;
                KKASSERT(p->p_nthreads == 1);
                lp = RB_ROOT(&p->p_lwp_tree);
                pmap_setlwpvm(lp, newvm);
                if (adjrefs) {
                        sysref_get(&newvm->vm_sysref);
                        sysref_put(&oldvm->vm_sysref);
                }
        }
        crit_exit();
}

void
pmap_setlwpvm(struct lwp *lp, struct vmspace *newvm)
{
        struct vmspace *oldvm;
        struct pmap *pmap;

        crit_enter();
        oldvm = lp->lwp_vmspace;

        if (oldvm != newvm) {
                lp->lwp_vmspace = newvm;
                if (curthread->td_lwp == lp) {
                        pmap = vmspace_pmap(newvm);
#if defined(SMP)
                        atomic_set_int(&pmap->pm_active, mycpu->gd_cpumask);
#else
                        pmap->pm_active |= 1;
#endif
#if defined(SWTCH_OPTIM_STATS)
                        tlb_flush_count++;
#endif
                        pmap = vmspace_pmap(oldvm);
#if defined(SMP)
                        atomic_clear_int(&pmap->pm_active, mycpu->gd_cpumask);
#else
                        pmap->pm_active &= ~1;
#endif
                }
        }
        crit_exit();
}


vm_offset_t
pmap_addr_hint(vm_object_t obj, vm_offset_t addr, vm_size_t size)
{

        if ((obj == NULL) || (size < NBPDR) || (obj->type != OBJT_DEVICE)) {
                return addr;
        }

        addr = (addr + (NBPDR - 1)) & ~(NBPDR - 1);
        return addr;
}


#if defined(DEBUG)

static void     pads (pmap_t pm);
void            pmap_pvdump (vm_paddr_t pa);

/* print address space of pmap*/
static void
pads(pmap_t pm)
{
        vm_offset_t va;
        int i, j;
        vpte_t *ptep;

        if (pm == &kernel_pmap)
                return;
        for (i = 0; i < 1024; i++)
                if (pm->pm_pdir[i])
                        for (j = 0; j < 1024; j++) {
                                va = (i << PDRSHIFT) + (j << PAGE_SHIFT);
                                if (pm == &kernel_pmap && va < KERNBASE)
                                        continue;
                                if (pm != &kernel_pmap && va > UPT_MAX_ADDRESS)
                                        continue;
                                ptep = pmap_pte(pm, va);
                                if (ptep && (*ptep & VPTE_V)) {
                                        kprintf("%p:%x ",
                                                (void *)va, (unsigned)*ptep);
                                }
                        };

}

void
pmap_pvdump(vm_paddr_t pa)
{
        pv_entry_t pv;
        vm_page_t m;

        kprintf("pa %08llx", (long long)pa);
        m = PHYS_TO_VM_PAGE(pa);
        TAILQ_FOREACH(pv, &m->md.pv_list, pv_list) {
#ifdef used_to_be
                kprintf(" -> pmap %p, va %x, flags %x",
                    (void *)pv->pv_pmap, pv->pv_va, pv->pv_flags);
#endif
                kprintf(" -> pmap %p, va %x", (void *)pv->pv_pmap, pv->pv_va);
                pads(pv->pv_pmap);
        }
        kprintf(" ");
}
#endif