Add m-esr52 at 52.6.0

1 files changed, 901 insertions, 0 deletions

diff --git a/js/src/gc/Memory.cpp b/js/src/gc/Memory.cpp
new file mode 100644
index 000000000..26da75469
--- /dev/null
+++ b/js/src/gc/Memory.cpp
@@ -0,0 +1,901 @@
+/* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 4 -*-
+ * vim: set ts=8 sts=4 et sw=4 tw=99:
+ * This Source Code Form is subject to the terms of the Mozilla Public
+ * License, v. 2.0. If a copy of the MPL was not distributed with this
+ * file, You can obtain one at http://mozilla.org/MPL/2.0/. */
+
+#include "gc/Memory.h"
+
+#include "mozilla/Atomics.h"
+#include "mozilla/TaggedAnonymousMemory.h"
+
+#include "js/HeapAPI.h"
+#include "vm/Runtime.h"
+
+#if defined(XP_WIN)
+
+#include "jswin.h"
+#include <psapi.h>
+
+#elif defined(SOLARIS)
+
+#include <sys/mman.h>
+#include <unistd.h>
+
+#elif defined(XP_UNIX)
+
+#include <algorithm>
+#include <errno.h>
+#include <sys/mman.h>
+#include <sys/resource.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <unistd.h>
+
+#endif
+
+namespace js {
+namespace gc {
+
+// The GC can only safely decommit memory when the page size of the
+// running process matches the compiled arena size.
+static size_t pageSize = 0;
+
+// The OS allocation granularity may not match the page size.
+static size_t allocGranularity = 0;
+
+#if defined(XP_UNIX)
+// The addresses handed out by mmap may grow up or down.
+static mozilla::Atomic<int, mozilla::Relaxed> growthDirection(0);
+#endif
+
+// Data from OOM crashes shows there may be up to 24 chunksized but unusable
+// chunks available in low memory situations. These chunks may all need to be
+// used up before we gain access to remaining *alignable* chunksized regions,
+// so we use a generous limit of 32 unusable chunks to ensure we reach them.
+static const int MaxLastDitchAttempts = 32;
+
+static void GetNewChunk(void** aAddress, void** aRetainedAddr, size_t size, size_t alignment);
+static void* MapAlignedPagesSlow(size_t size, size_t alignment);
+static void* MapAlignedPagesLastDitch(size_t size, size_t alignment);
+
+size_t
+SystemPageSize()
+{
+    return pageSize;
+}
+
+static bool
+DecommitEnabled()
+{
+    return pageSize == ArenaSize;
+}
+
+/*
+ * This returns the offset of address p from the nearest aligned address at
+ * or below p - or alternatively, the number of unaligned bytes at the end of
+ * the region starting at p (as we assert that allocation size is an integer
+ * multiple of the alignment).
+ */
+static inline size_t
+OffsetFromAligned(void* p, size_t alignment)
+{
+    return uintptr_t(p) % alignment;
+}
+
+void*
+TestMapAlignedPagesLastDitch(size_t size, size_t alignment)
+{
+    return MapAlignedPagesLastDitch(size, alignment);
+}
+
+
+#if defined(XP_WIN)
+
+void
+InitMemorySubsystem()
+{
+    if (pageSize == 0) {
+        SYSTEM_INFO sysinfo;
+        GetSystemInfo(&sysinfo);
+        pageSize = sysinfo.dwPageSize;
+        allocGranularity = sysinfo.dwAllocationGranularity;
+    }
+}
+
+#  if WINAPI_FAMILY_PARTITION(WINAPI_PARTITION_DESKTOP)
+
+static inline void*
+MapMemoryAt(void* desired, size_t length, int flags, int prot = PAGE_READWRITE)
+{
+    return VirtualAlloc(desired, length, flags, prot);
+}
+
+static inline void*
+MapMemory(size_t length, int flags, int prot = PAGE_READWRITE)
+{
+    return VirtualAlloc(nullptr, length, flags, prot);
+}
+
+void*
+MapAlignedPages(size_t size, size_t alignment)
+{
+    MOZ_ASSERT(size >= alignment);
+    MOZ_ASSERT(size >= allocGranularity);
+    MOZ_ASSERT(size % alignment == 0);
+    MOZ_ASSERT(size % pageSize == 0);
+    MOZ_ASSERT_IF(alignment < allocGranularity, allocGranularity % alignment == 0);
+    MOZ_ASSERT_IF(alignment > allocGranularity, alignment % allocGranularity == 0);
+
+    void* p = MapMemory(size, MEM_COMMIT | MEM_RESERVE);
+
+    /* Special case: If we want allocation alignment, no further work is needed. */
+    if (alignment == allocGranularity)
+        return p;
+
+    if (OffsetFromAligned(p, alignment) == 0)
+        return p;
+
+    void* retainedAddr;
+    GetNewChunk(&p, &retainedAddr, size, alignment);
+    if (retainedAddr)
+        UnmapPages(retainedAddr, size);
+    if (p) {
+        if (OffsetFromAligned(p, alignment) == 0)
+            return p;
+        UnmapPages(p, size);
+    }
+
+    p = MapAlignedPagesSlow(size, alignment);
+    if (!p)
+        return MapAlignedPagesLastDitch(size, alignment);
+
+    MOZ_ASSERT(OffsetFromAligned(p, alignment) == 0);
+    return p;
+}
+
+static void*
+MapAlignedPagesSlow(size_t size, size_t alignment)
+{
+    /*
+     * Windows requires that there be a 1:1 mapping between VM allocation
+     * and deallocation operations.  Therefore, take care here to acquire the
+     * final result via one mapping operation.  This means unmapping any
+     * preliminary result that is not correctly aligned.
+     */
+    void* p;
+    do {
+        /*
+         * Over-allocate in order to map a memory region that is definitely
+         * large enough, then deallocate and allocate again the correct size,
+         * within the over-sized mapping.
+         *
+         * Since we're going to unmap the whole thing anyway, the first
+         * mapping doesn't have to commit pages.
+         */
+        size_t reserveSize = size + alignment - pageSize;
+        p = MapMemory(reserveSize, MEM_RESERVE);
+        if (!p)
+            return nullptr;
+        void* chunkStart = (void*)AlignBytes(uintptr_t(p), alignment);
+        UnmapPages(p, reserveSize);
+        p = MapMemoryAt(chunkStart, size, MEM_COMMIT | MEM_RESERVE);
+
+        /* Failure here indicates a race with another thread, so try again. */
+    } while (!p);
+
+    return p;
+}
+
+/*
+ * In a low memory or high fragmentation situation, alignable chunks of the
+ * desired size may still be available, even if there are no more contiguous
+ * free chunks that meet the |size + alignment - pageSize| requirement of
+ * MapAlignedPagesSlow. In this case, try harder to find an alignable chunk
+ * by temporarily holding onto the unaligned parts of each chunk until the
+ * allocator gives us a chunk that either is, or can be aligned.
+ */
+static void*
+MapAlignedPagesLastDitch(size_t size, size_t alignment)
+{
+    void* tempMaps[MaxLastDitchAttempts];
+    int attempt = 0;
+    void* p = MapMemory(size, MEM_COMMIT | MEM_RESERVE);
+    if (OffsetFromAligned(p, alignment) == 0)
+        return p;
+    for (; attempt < MaxLastDitchAttempts; ++attempt) {
+        GetNewChunk(&p, tempMaps + attempt, size, alignment);
+        if (OffsetFromAligned(p, alignment) == 0) {
+            if (tempMaps[attempt])
+                UnmapPages(tempMaps[attempt], size);
+            break;
+        }
+        if (!tempMaps[attempt])
+            break; /* Bail if GetNewChunk failed. */
+    }
+    if (OffsetFromAligned(p, alignment)) {
+        UnmapPages(p, size);
+        p = nullptr;
+    }
+    while (--attempt >= 0)
+        UnmapPages(tempMaps[attempt], size);
+    return p;
+}
+
+/*
+ * On Windows, map and unmap calls must be matched, so we deallocate the
+ * unaligned chunk, then reallocate the unaligned part to block off the
+ * old address and force the allocator to give us a new one.
+ */
+static void
+GetNewChunk(void** aAddress, void** aRetainedAddr, size_t size, size_t alignment)
+{
+    void* address = *aAddress;
+    void* retainedAddr = nullptr;
+    do {
+        size_t retainedSize;
+        size_t offset = OffsetFromAligned(address, alignment);
+        if (!offset)
+            break;
+        UnmapPages(address, size);
+        retainedSize = alignment - offset;
+        retainedAddr = MapMemoryAt(address, retainedSize, MEM_RESERVE);
+        address = MapMemory(size, MEM_COMMIT | MEM_RESERVE);
+        /* If retainedAddr is null here, we raced with another thread. */
+    } while (!retainedAddr);
+    *aAddress = address;
+    *aRetainedAddr = retainedAddr;
+}
+
+void
+UnmapPages(void* p, size_t size)
+{
+    MOZ_ALWAYS_TRUE(VirtualFree(p, 0, MEM_RELEASE));
+}
+
+bool
+MarkPagesUnused(void* p, size_t size)
+{
+    if (!DecommitEnabled())
+        return true;
+
+    MOZ_ASSERT(OffsetFromAligned(p, pageSize) == 0);
+    LPVOID p2 = MapMemoryAt(p, size, MEM_RESET);
+    return p2 == p;
+}
+
+void
+MarkPagesInUse(void* p, size_t size)
+{
+    if (!DecommitEnabled())
+        return;
+
+    MOZ_ASSERT(OffsetFromAligned(p, pageSize) == 0);
+}
+
+size_t
+GetPageFaultCount()
+{
+    PROCESS_MEMORY_COUNTERS pmc;
+    if (!GetProcessMemoryInfo(GetCurrentProcess(), &pmc, sizeof(pmc)))
+        return 0;
+    return pmc.PageFaultCount;
+}
+
+void*
+AllocateMappedContent(int fd, size_t offset, size_t length, size_t alignment)
+{
+    MOZ_ASSERT(length && alignment);
+
+    // The allocation granularity and the requested offset
+    // must both be divisible by the requested alignment.
+    // Alignments larger than the allocation granularity are not supported.
+    if (allocGranularity % alignment != 0 || offset % alignment != 0)
+        return nullptr;
+
+    HANDLE hFile = reinterpret_cast<HANDLE>(intptr_t(fd));
+
+    // This call will fail if the file does not exist, which is what we want.
+    HANDLE hMap = CreateFileMapping(hFile, nullptr, PAGE_READONLY, 0, 0, nullptr);
+    if (!hMap)
+        return nullptr;
+
+    size_t alignedOffset = offset - (offset % allocGranularity);
+    size_t alignedLength = length + (offset % allocGranularity);
+
+    DWORD offsetH = uint32_t(uint64_t(alignedOffset) >> 32);
+    DWORD offsetL = uint32_t(alignedOffset);
+
+    // If the offset or length are out of bounds, this call will fail.
+    uint8_t* map = static_cast<uint8_t*>(MapViewOfFile(hMap, FILE_MAP_COPY, offsetH,
+                                                       offsetL, alignedLength));
+
+    // This just decreases the file mapping object's internal reference count;
+    // it won't actually be destroyed until we unmap the associated view.
+    CloseHandle(hMap);
+
+    if (!map)
+        return nullptr;
+
+#ifdef DEBUG
+    // Zero out data before and after the desired mapping to catch errors early.
+    if (offset != alignedOffset)
+        memset(map, 0, offset - alignedOffset);
+    if (alignedLength % pageSize)
+        memset(map + alignedLength, 0, pageSize - (alignedLength % pageSize));
+#endif
+
+    return map + (offset - alignedOffset);
+}
+
+void
+DeallocateMappedContent(void* p, size_t /*length*/)
+{
+    if (!p)
+        return;
+
+    // Calculate the address originally returned by MapViewOfFile.
+    // This is needed because AllocateMappedContent returns a pointer
+    // that might be offset from the view, as the beginning of a
+    // view must be aligned with the allocation granularity.
+    uintptr_t map = uintptr_t(p) - (uintptr_t(p) % allocGranularity);
+    MOZ_ALWAYS_TRUE(UnmapViewOfFile(reinterpret_cast<void*>(map)));
+}
+
+#  else // Various APIs are unavailable.
+
+void*
+MapAlignedPages(size_t size, size_t alignment)
+{
+    MOZ_ASSERT(size >= alignment);
+    MOZ_ASSERT(size >= allocGranularity);
+    MOZ_ASSERT(size % alignment == 0);
+    MOZ_ASSERT(size % pageSize == 0);
+    MOZ_ASSERT_IF(alignment < allocGranularity, allocGranularity % alignment == 0);
+    MOZ_ASSERT_IF(alignment > allocGranularity, alignment % allocGranularity == 0);
+
+    void* p = _aligned_malloc(size, alignment);
+
+    MOZ_ASSERT(OffsetFromAligned(p, alignment) == 0);
+    return p;
+}
+
+static void*
+MapAlignedPagesLastDitch(size_t size, size_t alignment)
+{
+    return nullptr;
+}
+
+void
+UnmapPages(void* p, size_t size)
+{
+    _aligned_free(p);
+}
+
+bool
+MarkPagesUnused(void* p, size_t size)
+{
+    MOZ_ASSERT(OffsetFromAligned(p, pageSize) == 0);
+    return true;
+}
+
+bool
+MarkPagesInUse(void* p, size_t size)
+{
+    MOZ_ASSERT(OffsetFromAligned(p, pageSize) == 0);
+}
+
+size_t
+GetPageFaultCount()
+{
+    // GetProcessMemoryInfo is unavailable.
+    return 0;
+}
+
+void*
+AllocateMappedContent(int fd, size_t offset, size_t length, size_t alignment)
+{
+    // Not implemented.
+    return nullptr;
+}
+
+// Deallocate mapped memory for object.
+void
+DeallocateMappedContent(void* p, size_t length)
+{
+    // Not implemented.
+}
+
+#  endif
+
+#elif defined(SOLARIS)
+
+#ifndef MAP_NOSYNC
+# define MAP_NOSYNC 0
+#endif
+
+void
+InitMemorySubsystem()
+{
+    if (pageSize == 0)
+        pageSize = allocGranularity = size_t(sysconf(_SC_PAGESIZE));
+}
+
+void*
+MapAlignedPages(size_t size, size_t alignment)
+{
+    MOZ_ASSERT(size >= alignment);
+    MOZ_ASSERT(size >= allocGranularity);
+    MOZ_ASSERT(size % alignment == 0);
+    MOZ_ASSERT(size % pageSize == 0);
+    MOZ_ASSERT_IF(alignment < allocGranularity, allocGranularity % alignment == 0);
+    MOZ_ASSERT_IF(alignment > allocGranularity, alignment % allocGranularity == 0);
+
+    int prot = PROT_READ | PROT_WRITE;
+    int flags = MAP_PRIVATE | MAP_ANON | MAP_ALIGN | MAP_NOSYNC;
+
+    void* p = mmap((caddr_t)alignment, size, prot, flags, -1, 0);
+    if (p == MAP_FAILED)
+        return nullptr;
+    return p;
+}
+
+static void*
+MapAlignedPagesLastDitch(size_t size, size_t alignment)
+{
+    return nullptr;
+}
+
+void
+UnmapPages(void* p, size_t size)
+{
+    MOZ_ALWAYS_TRUE(0 == munmap((caddr_t)p, size));
+}
+
+bool
+MarkPagesUnused(void* p, size_t size)
+{
+    MOZ_ASSERT(OffsetFromAligned(p, pageSize) == 0);
+    return true;
+}
+
+bool
+MarkPagesInUse(void* p, size_t size)
+{
+    if (!DecommitEnabled())
+        return;
+
+    MOZ_ASSERT(OffsetFromAligned(p, pageSize) == 0);
+}
+
+size_t
+GetPageFaultCount()
+{
+    return 0;
+}
+
+void*
+AllocateMappedContent(int fd, size_t offset, size_t length, size_t alignment)
+{
+    // Not implemented.
+    return nullptr;
+}
+
+// Deallocate mapped memory for object.
+void
+DeallocateMappedContent(void* p, size_t length)
+{
+    // Not implemented.
+}
+
+#elif defined(XP_UNIX)
+
+void
+InitMemorySubsystem()
+{
+    if (pageSize == 0)
+        pageSize = allocGranularity = size_t(sysconf(_SC_PAGESIZE));
+}
+
+static inline void*
+MapMemoryAt(void* desired, size_t length, int prot = PROT_READ | PROT_WRITE,
+            int flags = MAP_PRIVATE | MAP_ANON, int fd = -1, off_t offset = 0)
+{
+#if defined(__ia64__) || (defined(__sparc64__) && defined(__NetBSD__)) || defined(__aarch64__)
+    MOZ_ASSERT((0xffff800000000000ULL & (uintptr_t(desired) + length - 1)) == 0);
+#endif
+    void* region = mmap(desired, length, prot, flags, fd, offset);
+    if (region == MAP_FAILED)
+        return nullptr;
+    /*
+     * mmap treats the given address as a hint unless the MAP_FIXED flag is
+     * used (which isn't usually what you want, as this overrides existing
+     * mappings), so check that the address we got is the address we wanted.
+     */
+    if (region != desired) {
+        if (munmap(region, length))
+            MOZ_ASSERT(errno == ENOMEM);
+        return nullptr;
+    }
+    return region;
+}
+
+static inline void*
+MapMemory(size_t length, int prot = PROT_READ | PROT_WRITE,
+          int flags = MAP_PRIVATE | MAP_ANON, int fd = -1, off_t offset = 0)
+{
+#if defined(__ia64__) || (defined(__sparc64__) && defined(__NetBSD__))
+    /*
+     * The JS engine assumes that all allocated pointers have their high 17 bits clear,
+     * which ia64's mmap doesn't support directly. However, we can emulate it by passing
+     * mmap an "addr" parameter with those bits clear. The mmap will return that address,
+     * or the nearest available memory above that address, providing a near-guarantee
+     * that those bits are clear. If they are not, we return nullptr below to indicate
+     * out-of-memory.
+     *
+     * The addr is chosen as 0x0000070000000000, which still allows about 120TB of virtual
+     * address space.
+     *
+     * See Bug 589735 for more information.
+     */
+    void* region = mmap((void*)0x0000070000000000, length, prot, flags, fd, offset);
+    if (region == MAP_FAILED)
+        return nullptr;
+    /*
+     * If the allocated memory doesn't have its upper 17 bits clear, consider it
+     * as out of memory.
+     */
+    if ((uintptr_t(region) + (length - 1)) & 0xffff800000000000) {
+        if (munmap(region, length))
+            MOZ_ASSERT(errno == ENOMEM);
+        return nullptr;
+    }
+    return region;
+#elif defined(__aarch64__)
+   /*
+    * There might be similar virtual address issue on arm64 which depends on
+    * hardware and kernel configurations. But the work around is slightly
+    * different due to the different mmap behavior.
+    *
+    * TODO: Merge with the above code block if this implementation works for
+    * ia64 and sparc64.
+    */
+    const uintptr_t start = UINT64_C(0x0000070000000000);
+    const uintptr_t end   = UINT64_C(0x0000800000000000);
+    const uintptr_t step  = ChunkSize;
+   /*
+    * Optimization options if there are too many retries in practice:
+    * 1. Examine /proc/self/maps to find an available address. This file is
+    *    not always available, however. In addition, even if we examine
+    *    /proc/self/maps, we may still need to retry several times due to
+    *    racing with other threads.
+    * 2. Use a global/static variable with lock to track the addresses we have
+    *    allocated or tried.
+    */
+    uintptr_t hint;
+    void* region = MAP_FAILED;
+    for (hint = start; region == MAP_FAILED && hint + length <= end; hint += step) {
+        region = mmap((void*)hint, length, prot, flags, fd, offset);
+        if (region != MAP_FAILED) {
+            if ((uintptr_t(region) + (length - 1)) & 0xffff800000000000) {
+                if (munmap(region, length)) {
+                    MOZ_ASSERT(errno == ENOMEM);
+                }
+                region = MAP_FAILED;
+            }
+        }
+    }
+    return region == MAP_FAILED ? nullptr : region;
+#else
+    void* region = MozTaggedAnonymousMmap(nullptr, length, prot, flags, fd, offset, "js-gc-heap");
+    if (region == MAP_FAILED)
+        return nullptr;
+    return region;
+#endif
+}
+
+void*
+MapAlignedPages(size_t size, size_t alignment)
+{
+    MOZ_ASSERT(size >= alignment);
+    MOZ_ASSERT(size >= allocGranularity);
+    MOZ_ASSERT(size % alignment == 0);
+    MOZ_ASSERT(size % pageSize == 0);
+    MOZ_ASSERT_IF(alignment < allocGranularity, allocGranularity % alignment == 0);
+    MOZ_ASSERT_IF(alignment > allocGranularity, alignment % allocGranularity == 0);
+
+    void* p = MapMemory(size);
+
+    /* Special case: If we want page alignment, no further work is needed. */
+    if (alignment == allocGranularity)
+        return p;
+
+    if (OffsetFromAligned(p, alignment) == 0)
+        return p;
+
+    void* retainedAddr;
+    GetNewChunk(&p, &retainedAddr, size, alignment);
+    if (retainedAddr)
+        UnmapPages(retainedAddr, size);
+    if (p) {
+        if (OffsetFromAligned(p, alignment) == 0)
+            return p;
+        UnmapPages(p, size);
+    }
+
+    p = MapAlignedPagesSlow(size, alignment);
+    if (!p)
+        return MapAlignedPagesLastDitch(size, alignment);
+
+    MOZ_ASSERT(OffsetFromAligned(p, alignment) == 0);
+    return p;
+}
+
+static void*
+MapAlignedPagesSlow(size_t size, size_t alignment)
+{
+    /* Overallocate and unmap the region's edges. */
+    size_t reqSize = size + alignment - pageSize;
+    void* region = MapMemory(reqSize);
+    if (!region)
+        return nullptr;
+
+    void* regionEnd = (void*)(uintptr_t(region) + reqSize);
+    void* front;
+    void* end;
+    if (growthDirection <= 0) {
+        size_t offset = OffsetFromAligned(regionEnd, alignment);
+        end = (void*)(uintptr_t(regionEnd) - offset);
+        front = (void*)(uintptr_t(end) - size);
+    } else {
+        size_t offset = OffsetFromAligned(region, alignment);
+        front = (void*)(uintptr_t(region) + (offset ? alignment - offset : 0));
+        end = (void*)(uintptr_t(front) + size);
+    }
+
+    if (front != region)
+        UnmapPages(region, uintptr_t(front) - uintptr_t(region));
+    if (end != regionEnd)
+        UnmapPages(end, uintptr_t(regionEnd) - uintptr_t(end));
+
+    return front;
+}
+
+/*
+ * In a low memory or high fragmentation situation, alignable chunks of the
+ * desired size may still be available, even if there are no more contiguous
+ * free chunks that meet the |size + alignment - pageSize| requirement of
+ * MapAlignedPagesSlow. In this case, try harder to find an alignable chunk
+ * by temporarily holding onto the unaligned parts of each chunk until the
+ * allocator gives us a chunk that either is, or can be aligned.
+ */
+static void*
+MapAlignedPagesLastDitch(size_t size, size_t alignment)
+{
+    void* tempMaps[MaxLastDitchAttempts];
+    int attempt = 0;
+    void* p = MapMemory(size);
+    if (OffsetFromAligned(p, alignment) == 0)
+        return p;
+    for (; attempt < MaxLastDitchAttempts; ++attempt) {
+        GetNewChunk(&p, tempMaps + attempt, size, alignment);
+        if (OffsetFromAligned(p, alignment) == 0) {
+            if (tempMaps[attempt])
+                UnmapPages(tempMaps[attempt], size);
+            break;
+        }
+        if (!tempMaps[attempt])
+            break; /* Bail if GetNewChunk failed. */
+    }
+    if (OffsetFromAligned(p, alignment)) {
+        UnmapPages(p, size);
+        p = nullptr;
+    }
+    while (--attempt >= 0)
+        UnmapPages(tempMaps[attempt], size);
+    return p;
+}
+
+/*
+ * mmap calls don't have to be matched with calls to munmap, so we can unmap
+ * just the pages we don't need. However, as we don't know a priori if addresses
+ * are handed out in increasing or decreasing order, we have to try both
+ * directions (depending on the environment, one will always fail).
+ */
+static void
+GetNewChunk(void** aAddress, void** aRetainedAddr, size_t size, size_t alignment)
+{
+    void* address = *aAddress;
+    void* retainedAddr = nullptr;
+    bool addrsGrowDown = growthDirection <= 0;
+    int i = 0;
+    for (; i < 2; ++i) {
+        /* Try the direction indicated by growthDirection. */
+        if (addrsGrowDown) {
+            size_t offset = OffsetFromAligned(address, alignment);
+            void* head = (void*)((uintptr_t)address - offset);
+            void* tail = (void*)((uintptr_t)head + size);
+            if (MapMemoryAt(head, offset)) {
+                UnmapPages(tail, offset);
+                if (growthDirection >= -8)
+                    --growthDirection;
+                address = head;
+                break;
+            }
+        } else {
+            size_t offset = alignment - OffsetFromAligned(address, alignment);
+            void* head = (void*)((uintptr_t)address + offset);
+            void* tail = (void*)((uintptr_t)address + size);
+            if (MapMemoryAt(tail, offset)) {
+                UnmapPages(address, offset);
+                if (growthDirection <= 8)
+                    ++growthDirection;
+                address = head;
+                break;
+            }
+        }
+        /* If we're confident in the growth direction, don't try the other. */
+        if (growthDirection < -8 || growthDirection > 8)
+            break;
+        /* If that failed, try the opposite direction. */
+        addrsGrowDown = !addrsGrowDown;
+    }
+    /* If our current chunk cannot be aligned, see if the next one is aligned. */
+    if (OffsetFromAligned(address, alignment)) {
+        retainedAddr = address;
+        address = MapMemory(size);
+    }
+    *aAddress = address;
+    *aRetainedAddr = retainedAddr;
+}
+
+void
+UnmapPages(void* p, size_t size)
+{
+    if (munmap(p, size))
+        MOZ_ASSERT(errno == ENOMEM);
+}
+
+bool
+MarkPagesUnused(void* p, size_t size)
+{
+    if (!DecommitEnabled())
+        return false;
+
+    MOZ_ASSERT(OffsetFromAligned(p, pageSize) == 0);
+    int result = madvise(p, size, MADV_DONTNEED);
+    return result != -1;
+}
+
+void
+MarkPagesInUse(void* p, size_t size)
+{
+    if (!DecommitEnabled())
+        return;
+
+    MOZ_ASSERT(OffsetFromAligned(p, pageSize) == 0);
+}
+
+size_t
+GetPageFaultCount()
+{
+    struct rusage usage;
+    int err = getrusage(RUSAGE_SELF, &usage);
+    if (err)
+        return 0;
+    return usage.ru_majflt;
+}
+
+void*
+AllocateMappedContent(int fd, size_t offset, size_t length, size_t alignment)
+{
+    MOZ_ASSERT(length && alignment);
+
+    // The allocation granularity and the requested offset
+    // must both be divisible by the requested alignment.
+    // Alignments larger than the allocation granularity are not supported.
+    if (allocGranularity % alignment != 0 || offset % alignment != 0)
+        return nullptr;
+
+    // Sanity check the offset and size, as mmap does not do this for us.
+    struct stat st;
+    if (fstat(fd, &st) || offset >= uint64_t(st.st_size) || length > uint64_t(st.st_size) - offset)
+        return nullptr;
+
+    size_t alignedOffset = offset - (offset % allocGranularity);
+    size_t alignedLength = length + (offset % allocGranularity);
+
+    uint8_t* map = static_cast<uint8_t*>(MapMemory(alignedLength, PROT_READ | PROT_WRITE,
+                                                   MAP_PRIVATE, fd, alignedOffset));
+    if (!map)
+        return nullptr;
+
+#ifdef DEBUG
+    // Zero out data before and after the desired mapping to catch errors early.
+    if (offset != alignedOffset)
+        memset(map, 0, offset - alignedOffset);
+    if (alignedLength % pageSize)
+        memset(map + alignedLength, 0, pageSize - (alignedLength % pageSize));
+#endif
+
+    return map + (offset - alignedOffset);
+}
+
+void
+DeallocateMappedContent(void* p, size_t length)
+{
+    if (!p)
+        return;
+
+    // Calculate the address originally returned by mmap.
+    // This is needed because AllocateMappedContent returns a pointer
+    // that might be offset from the mapping, as the beginning of a
+    // mapping must be aligned with the allocation granularity.
+    uintptr_t map = uintptr_t(p) - (uintptr_t(p) % allocGranularity);
+    size_t alignedLength = length + (uintptr_t(p) % allocGranularity);
+    UnmapPages(reinterpret_cast<void*>(map), alignedLength);
+}
+
+#else
+#error "Memory mapping functions are not defined for your OS."
+#endif
+
+void
+ProtectPages(void* p, size_t size)
+{
+    MOZ_ASSERT(size % pageSize == 0);
+    MOZ_RELEASE_ASSERT(size > 0);
+    MOZ_RELEASE_ASSERT(p);
+#if defined(XP_WIN)
+    DWORD oldProtect;
+    if (!VirtualProtect(p, size, PAGE_NOACCESS, &oldProtect)) {
+        MOZ_CRASH_UNSAFE_PRINTF("VirtualProtect(PAGE_NOACCESS) failed! Error code: %u",
+                                GetLastError());
+    }
+    MOZ_ASSERT(oldProtect == PAGE_READWRITE);
+#else  // assume Unix
+    if (mprotect(p, size, PROT_NONE))
+        MOZ_CRASH("mprotect(PROT_NONE) failed");
+#endif
+}
+
+void
+MakePagesReadOnly(void* p, size_t size)
+{
+    MOZ_ASSERT(size % pageSize == 0);
+    MOZ_RELEASE_ASSERT(size > 0);
+    MOZ_RELEASE_ASSERT(p);
+#if defined(XP_WIN)
+    DWORD oldProtect;
+    if (!VirtualProtect(p, size, PAGE_READONLY, &oldProtect)) {
+        MOZ_CRASH_UNSAFE_PRINTF("VirtualProtect(PAGE_READONLY) failed! Error code: %u",
+                                GetLastError());
+    }
+    MOZ_ASSERT(oldProtect == PAGE_READWRITE);
+#else  // assume Unix
+    if (mprotect(p, size, PROT_READ))
+        MOZ_CRASH("mprotect(PROT_READ) failed");
+#endif
+}
+
+void
+UnprotectPages(void* p, size_t size)
+{
+    MOZ_ASSERT(size % pageSize == 0);
+    MOZ_RELEASE_ASSERT(size > 0);
+    MOZ_RELEASE_ASSERT(p);
+#if defined(XP_WIN)
+    DWORD oldProtect;
+    if (!VirtualProtect(p, size, PAGE_READWRITE, &oldProtect)) {
+        MOZ_CRASH_UNSAFE_PRINTF("VirtualProtect(PAGE_READWRITE) failed! Error code: %u",
+                                GetLastError());
+    }
+    MOZ_ASSERT(oldProtect == PAGE_NOACCESS || oldProtect == PAGE_READONLY);
+#else  // assume Unix
+    if (mprotect(p, size, PROT_READ | PROT_WRITE))
+        MOZ_CRASH("mprotect(PROT_READ | PROT_WRITE) failed");
+#endif
+}
+
+} // namespace gc
+} // namespace js