Quelle nsMemoryReporterManager.cpp Sprache: C

/* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* vim: set ts=8 sts=2 et sw=2 tw=80: */
/* 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 "nsMemoryReporterManager.h"

#include "nsAtomTable.h"
#include "nsCOMPtr.h"
#include "nsCOMArray.h"
#include "nsPrintfCString.h"
#include "nsProxyRelease.h"
#include "nsServiceManagerUtils.h"
#include "nsITimer.h"
#include "nsThreadManager.h"
#include "nsThreadUtils.h"
#include "nsPIDOMWindow.h"
#include "nsIObserverService.h"
#include "nsIOService.h"
#include "nsIGlobalObject.h"
#include "nsIXPConnect.h"
#ifdef MOZ_GECKO_PROFILER
#  include "GeckoProfilerReporter.h"
#endif
#if defined(XP_UNIX) || defined(MOZ_DMD)
#  include "nsMemoryInfoDumper.h"
#endif
#include "nsNetCID.h"
#include "nsThread.h"
#include "VRProcessManager.h"
#include "mozilla/Attributes.h"
#include "mozilla/MemoryReportingProcess.h"
#include "mozilla/PodOperations.h"
#include "mozilla/Preferences.h"
#include "mozilla/RDDProcessManager.h"
#include "mozilla/ResultExtensions.h"
#include "mozilla/Services.h"
#include "mozilla/Telemetry.h"
#include "mozilla/UniquePtrExtensions.h"
#include "mozilla/dom/MemoryReportTypes.h"
#include "mozilla/dom/ContentParent.h"
#include "mozilla/gfx/GPUProcessManager.h"
#include "mozilla/ipc/UtilityProcessManager.h"
#include "mozilla/ipc/FileDescriptorUtils.h"
#ifdef MOZ_PHC
#  include "PHC.h"
#endif

#ifdef MOZ_WIDGET_ANDROID
#  include "mozilla/java/GeckoAppShellWrappers.h"
#  include "mozilla/jni/Utils.h"
#endif

#ifdef XP_WIN
#  include "mozilla/MemoryInfo.h"

#  include <process.h>
#  ifndef getpid
#    define getpid _getpid
#  endif
#else
#  include <unistd.h>
#endif

using namespace mozilla;
using namespace mozilla::ipc;
using namespace dom;

#if defined(XP_LINUX)

#  include "mozilla/MemoryMapping.h"

#  include <malloc.h>
#  include <string.h>
#  include <stdlib.h>

[[nodiscard]] static nsresult GetProcSelfStatmField(int aField, int64_t* aN) {
  // There are more than two fields, but we're only interested in the first
  // two.
  static const int MAX_FIELD = 2;
  size_t fields[MAX_FIELD];
  MOZ_ASSERT(aField < MAX_FIELD, "bad field number");
  FILE* f = fopen("/proc/self/statm", "r");
  if (f) {
    int nread = fscanf(f, "%zu %zu", &fields[0], &fields[1]);
    fclose(f);
    if (nread == MAX_FIELD) {
      *aN = fields[aField] * getpagesize();
      return NS_OK;
    }
  }
  return NS_ERROR_FAILURE;
}

[[nodiscard]] static nsresult GetProcSelfSmapsPrivate(int64_t* aN, pid_t aPid) {
  // You might be tempted to calculate USS by subtracting the "shared" value
  // from the "resident" value in /proc/<pid>/statm. But at least on Linux,
  // statm's "shared" value actually counts pages backed by files, which has
  // little to do with whether the pages are actually shared. /proc/self/smaps
  // on the other hand appears to give us the correct information.

  nsTArray<MemoryMapping> mappings(1024);
  MOZ_TRY(GetMemoryMappings(mappings, aPid));

  int64_t amount = 0;
  for (auto& mapping : mappings) {
    amount += mapping.Private_Clean();
    amount += mapping.Private_Dirty();
  }
  *aN = amount;
  return NS_OK;
}

#  define HAVE_VSIZE_AND_RESIDENT_REPORTERS 1
[[nodiscard]] static nsresult VsizeDistinguishedAmount(int64_t* aN) {
  return GetProcSelfStatmField(0, aN);
}

[[nodiscard]] static nsresult ResidentDistinguishedAmount(int64_t* aN) {
  return GetProcSelfStatmField(1, aN);
}

[[nodiscard]] static nsresult ResidentFastDistinguishedAmount(int64_t* aN) {
  return ResidentDistinguishedAmount(aN);
}

#  define HAVE_RESIDENT_UNIQUE_REPORTER 1
[[nodiscard]] static nsresult ResidentUniqueDistinguishedAmount(
    int64_t* aN, pid_t aPid = 0) {
  return GetProcSelfSmapsPrivate(aN, aPid);
}

#  ifdef HAVE_MALLINFO
#    define HAVE_SYSTEM_HEAP_REPORTER 1
[[nodiscard]] static nsresult SystemHeapSize(int64_t* aSizeOut) {
  struct mallinfo info = mallinfo();

  // The documentation in the glibc man page makes it sound like |uordblks|
  // would suffice, but that only gets the small allocations that are put in
  // the brk heap. We need |hblkhd| as well to get the larger allocations
  // that are mmapped.
  //
  // The fields in |struct mallinfo| are all |int|, <sigh>, so it is
  // unreliable if memory usage gets high. However, the system heap size on
  // Linux should usually be zero (so long as jemalloc is enabled) so that
  // shouldn't be a problem. Nonetheless, cast the |int|s to |size_t| before
  // adding them to provide a small amount of extra overflow protection.
  *aSizeOut = size_t(info.hblkhd) + size_t(info.uordblks);
  return NS_OK;
}
#  endif

#elif defined(__DragonFly__) || defined(__FreeBSD__) || defined(__NetBSD__) || \
    defined(__OpenBSD__) || defined(__FreeBSD_kernel__)

#  include <sys/param.h>
#  include <sys/sysctl.h>
#  if defined(__DragonFly__) || defined(__FreeBSD__) || \
      defined(__FreeBSD_kernel__)
#    include <sys/user.h>
#  endif

#  include <unistd.h>

#  if defined(__NetBSD__)
#    undef KERN_PROC
#    define KERN_PROC KERN_PROC2
#    define KINFO_PROC struct kinfo_proc2
#  else
#    define KINFO_PROC struct kinfo_proc
#  endif

#  if defined(__DragonFly__)
#    define KP_SIZE(kp) (kp.kp_vm_map_size)
#    define KP_RSS(kp) (kp.kp_vm_rssize * getpagesize())
#  elif defined(__FreeBSD__) || defined(__FreeBSD_kernel__)
#    define KP_SIZE(kp) (kp.ki_size)
#    define KP_RSS(kp) (kp.ki_rssize * getpagesize())
#  elif defined(__NetBSD__)
#    define KP_SIZE(kp) (kp.p_vm_msize * getpagesize())
#    define KP_RSS(kp) (kp.p_vm_rssize * getpagesize())
#  elif defined(__OpenBSD__)
#    define KP_SIZE(kp) \
      ((kp.p_vm_dsize + kp.p_vm_ssize + kp.p_vm_tsize) * getpagesize())
#    define KP_RSS(kp) (kp.p_vm_rssize * getpagesize())
#  endif

[[nodiscard]] static nsresult GetKinfoProcSelf(KINFO_PROC* aProc) {
#  if defined(__OpenBSD__) && defined(MOZ_SANDBOX)
  static LazyLogModule sPledgeLog("SandboxPledge");
  MOZ_LOG(sPledgeLog, LogLevel::Debug,
          ("%s called when pledged, returning NS_ERROR_FAILURE\n", __func__));
  return NS_ERROR_FAILURE;
#  endif
  int mib[] = {
      // clang-format off
    CTL_KERN,
    KERN_PROC,
    KERN_PROC_PID,
    getpid(),
#  if defined(__NetBSD__) || defined(__OpenBSD__)
    sizeof(KINFO_PROC),
    1,
#  endif
      // clang-format on
  };
  u_int miblen = sizeof(mib) / sizeof(mib[0]);
  size_t size = sizeof(KINFO_PROC);
  if (sysctl(mib, miblen, aProc, &size, nullptr, 0)) {
    return NS_ERROR_FAILURE;
  }
  return NS_OK;
}

#  define HAVE_VSIZE_AND_RESIDENT_REPORTERS 1
[[nodiscard]] static nsresult VsizeDistinguishedAmount(int64_t* aN) {
  KINFO_PROC proc;
  nsresult rv = GetKinfoProcSelf(&proc);
  if (NS_SUCCEEDED(rv)) {
    *aN = KP_SIZE(proc);
  }
  return rv;
}

[[nodiscard]] static nsresult ResidentDistinguishedAmount(int64_t* aN) {
  KINFO_PROC proc;
  nsresult rv = GetKinfoProcSelf(&proc);
  if (NS_SUCCEEDED(rv)) {
    *aN = KP_RSS(proc);
  }
  return rv;
}

[[nodiscard]] static nsresult ResidentFastDistinguishedAmount(int64_t* aN) {
  return ResidentDistinguishedAmount(aN);
}

#  ifdef __FreeBSD__
#    include <libutil.h>
#    include <algorithm>

[[nodiscard]] static nsresult GetKinfoVmentrySelf(int64_t* aPrss,
                                                  uint64_t* aMaxreg) {
  int cnt;
  struct kinfo_vmentry* vmmap;
  struct kinfo_vmentry* kve;
  if (!(vmmap = kinfo_getvmmap(getpid(), &cnt))) {
    return NS_ERROR_FAILURE;
  }
  if (aPrss) {
    *aPrss = 0;
  }
  if (aMaxreg) {
    *aMaxreg = 0;
  }

  for (int i = 0; i < cnt; i++) {
    kve = &vmmap[i];
    if (aPrss) {
      *aPrss += kve->kve_private_resident;
    }
    if (aMaxreg) {
      *aMaxreg = std::max(*aMaxreg, kve->kve_end - kve->kve_start);
    }
  }

  free(vmmap);
  return NS_OK;
}

#    define HAVE_PRIVATE_REPORTER 1
[[nodiscard]] static nsresult PrivateDistinguishedAmount(int64_t* aN) {
  int64_t priv;
  nsresult rv = GetKinfoVmentrySelf(&priv, nullptr);
  NS_ENSURE_SUCCESS(rv, rv);
  *aN = priv * getpagesize();
  return NS_OK;
}

#    define HAVE_VSIZE_MAX_CONTIGUOUS_REPORTER 1
[[nodiscard]] static nsresult VsizeMaxContiguousDistinguishedAmount(
    int64_t* aN) {
  uint64_t biggestRegion;
  nsresult rv = GetKinfoVmentrySelf(nullptr, &biggestRegion);
  if (NS_SUCCEEDED(rv)) {
    *aN = biggestRegion;
  }
  return NS_OK;
}
#  endif  // FreeBSD

#elif defined(SOLARIS)

#  include <procfs.h>
#  include <fcntl.h>
#  include <unistd.h>

static void XMappingIter(int64_t& aVsize, int64_t& aResident,
                         int64_t& aShared) {
  aVsize = -1;
  aResident = -1;
  aShared = -1;
  int mapfd = open("/proc/self/xmap", O_RDONLY);
  struct stat st;
  prxmap_t* prmapp = nullptr;
  if (mapfd >= 0) {
    if (!fstat(mapfd, &st)) {
      int nmap = st.st_size / sizeof(prxmap_t);
      while (1) {
        // stat(2) on /proc/<pid>/xmap returns an incorrect value,
        // prior to the release of Solaris 11.
        // Here is a workaround for it.
        nmap *= 2;
        prmapp = (prxmap_t*)malloc((nmap + 1) * sizeof(prxmap_t));
        if (!prmapp) {
          // out of memory
          break;
        }
        int n = pread(mapfd, prmapp, (nmap + 1) * sizeof(prxmap_t), 0);
        if (n < 0) {
          break;
        }
        if (nmap >= n / sizeof(prxmap_t)) {
          aVsize = 0;
          aResident = 0;
          aShared = 0;
          for (int i = 0; i < n / sizeof(prxmap_t); i++) {
            aVsize += prmapp[i].pr_size;
            aResident += prmapp[i].pr_rss * prmapp[i].pr_pagesize;
            if (prmapp[i].pr_mflags & MA_SHARED) {
              aShared += prmapp[i].pr_rss * prmapp[i].pr_pagesize;
            }
          }
          break;
        }
        free(prmapp);
      }
      free(prmapp);
    }
    close(mapfd);
  }
}

#  define HAVE_VSIZE_AND_RESIDENT_REPORTERS 1
[[nodiscard]] static nsresult VsizeDistinguishedAmount(int64_t* aN) {
  int64_t vsize, resident, shared;
  XMappingIter(vsize, resident, shared);
  if (vsize == -1) {
    return NS_ERROR_FAILURE;
  }
  *aN = vsize;
  return NS_OK;
}

[[nodiscard]] static nsresult ResidentDistinguishedAmount(int64_t* aN) {
  int64_t vsize, resident, shared;
  XMappingIter(vsize, resident, shared);
  if (resident == -1) {
    return NS_ERROR_FAILURE;
  }
  *aN = resident;
  return NS_OK;
}

[[nodiscard]] static nsresult ResidentFastDistinguishedAmount(int64_t* aN) {
  return ResidentDistinguishedAmount(aN);
}

#  define HAVE_RESIDENT_UNIQUE_REPORTER 1
[[nodiscard]] static nsresult ResidentUniqueDistinguishedAmount(int64_t* aN) {
  int64_t vsize, resident, shared;
  XMappingIter(vsize, resident, shared);
  if (resident == -1) {
    return NS_ERROR_FAILURE;
  }
  *aN = resident - shared;
  return NS_OK;
}

#elif defined(XP_MACOSX)

#  include <mach/mach_init.h>
#  include <mach/mach_vm.h>
#  include <mach/shared_region.h>
#  include <mach/task.h>
#  include <sys/sysctl.h>

[[nodiscard]] static bool GetTaskBasicInfo(struct task_basic_info* aTi) {
  mach_msg_type_number_t count = TASK_BASIC_INFO_COUNT;
  kern_return_t kr =
      task_info(mach_task_self(), TASK_BASIC_INFO, (task_info_t)aTi, &count);
  return kr == KERN_SUCCESS;
}

// The VSIZE figure on Mac includes huge amounts of shared memory and is always
// absurdly high, eg. 2GB+ even at start-up.  But both 'top' and 'ps' report
// it, so we might as well too.
#  define HAVE_VSIZE_AND_RESIDENT_REPORTERS 1
[[nodiscard]] static nsresult VsizeDistinguishedAmount(int64_t* aN) {
  task_basic_info ti;
  if (!GetTaskBasicInfo(&ti)) {
    return NS_ERROR_FAILURE;
  }
  *aN = ti.virtual_size;
  return NS_OK;
}

// If we're using jemalloc on Mac, we need to instruct jemalloc to purge the
// pages it has madvise(MADV_FREE)'d before we read our RSS in order to get
// an accurate result.  The OS will take away MADV_FREE'd pages when there's
// memory pressure, so ideally, they shouldn't count against our RSS.
//
// Purging these pages can take a long time for some users (see bug 789975),
// so we provide the option to get the RSS without purging first.
[[nodiscard]] static nsresult ResidentDistinguishedAmountHelper(int64_t* aN,
                                                                bool aDoPurge) {
#  ifdef HAVE_JEMALLOC_STATS
  if (aDoPurge) {
    Telemetry::AutoTimer<Telemetry::MEMORY_FREE_PURGED_PAGES_MS> timer;
    jemalloc_purge_freed_pages();
  }
#  endif

  task_basic_info ti;
  if (!GetTaskBasicInfo(&ti)) {
    return NS_ERROR_FAILURE;
  }
  *aN = ti.resident_size;
  return NS_OK;
}

[[nodiscard]] static nsresult ResidentFastDistinguishedAmount(int64_t* aN) {
  return ResidentDistinguishedAmountHelper(aN, /* doPurge = */ false);
}

[[nodiscard]] static nsresult ResidentDistinguishedAmount(int64_t* aN) {
  return ResidentDistinguishedAmountHelper(aN, /* doPurge = */ true);
}

#  define HAVE_RESIDENT_UNIQUE_REPORTER 1

static bool InSharedRegion(mach_vm_address_t aAddr, cpu_type_t aType) {
  mach_vm_address_t base;
  mach_vm_address_t size;

  switch (aType) {
    case CPU_TYPE_ARM:
      base = SHARED_REGION_BASE_ARM;
      size = SHARED_REGION_SIZE_ARM;
      break;
    case CPU_TYPE_ARM64:
      base = SHARED_REGION_BASE_ARM64;
      size = SHARED_REGION_SIZE_ARM64;
      break;
    case CPU_TYPE_I386:
      base = SHARED_REGION_BASE_I386;
      size = SHARED_REGION_SIZE_I386;
      break;
    case CPU_TYPE_X86_64:
      base = SHARED_REGION_BASE_X86_64;
      size = SHARED_REGION_SIZE_X86_64;
      break;
    default:
      return false;
  }

  return base <= aAddr && aAddr < (base + size);
}

[[nodiscard]] static nsresult ResidentUniqueDistinguishedAmount(
    int64_t* aN, mach_port_t aPort = 0) {
  if (!aN) {
    return NS_ERROR_FAILURE;
  }

  cpu_type_t cpu_type;
  size_t len = sizeof(cpu_type);
  if (sysctlbyname("sysctl.proc_cputype", &cpu_type, &len, NULL, 0) != 0) {
    return NS_ERROR_FAILURE;
  }

  // Roughly based on libtop_update_vm_regions in
  // http://www.opensource.apple.com/source/top/top-100.1.2/libtop.c
  size_t privatePages = 0;
  mach_vm_size_t topSize = 0;
  for (mach_vm_address_t addr = MACH_VM_MIN_ADDRESS;; addr += topSize) {
    vm_region_top_info_data_t topInfo;
    mach_msg_type_number_t topInfoCount = VM_REGION_TOP_INFO_COUNT;
    mach_port_t topObjectName;

    kern_return_t kr = mach_vm_region(
        aPort ? aPort : mach_task_self(), &addr, &topSize, VM_REGION_TOP_INFO,
        reinterpret_cast<vm_region_info_t>(&topInfo), &topInfoCount,
        &topObjectName);
    if (kr == KERN_INVALID_ADDRESS) {
      // Done iterating VM regions.
      break;
    } else if (kr != KERN_SUCCESS) {
      return NS_ERROR_FAILURE;
    }

    if (InSharedRegion(addr, cpu_type) && topInfo.share_mode != SM_PRIVATE) {
      continue;
    }

    switch (topInfo.share_mode) {
      case SM_LARGE_PAGE:
        // NB: Large pages are not shareable and always resident.
      case SM_PRIVATE:
        privatePages += topInfo.private_pages_resident;
        privatePages += topInfo.shared_pages_resident;
        break;
      case SM_COW:
        privatePages += topInfo.private_pages_resident;
        if (topInfo.ref_count == 1) {
          // Treat copy-on-write pages as private if they only have one
          // reference.
          privatePages += topInfo.shared_pages_resident;
        }
        break;
      case SM_SHARED: {
        // Using mprotect() or similar to protect a page in the middle of a
        // mapping can create aliased mappings. They look like shared mappings
        // to the VM_REGION_TOP_INFO interface, so re-check with
        // VM_REGION_EXTENDED_INFO.

        mach_vm_size_t exSize = 0;
        vm_region_extended_info_data_t exInfo;
        mach_msg_type_number_t exInfoCount = VM_REGION_EXTENDED_INFO_COUNT;
        mach_port_t exObjectName;
        kr = mach_vm_region(aPort ? aPort : mach_task_self(), &addr, &exSize,
                            VM_REGION_EXTENDED_INFO,
                            reinterpret_cast<vm_region_info_t>(&exInfo),
                            &exInfoCount, &exObjectName);
        if (kr == KERN_INVALID_ADDRESS) {
          // Done iterating VM regions.
          break;
        } else if (kr != KERN_SUCCESS) {
          return NS_ERROR_FAILURE;
        }

        if (exInfo.share_mode == SM_PRIVATE_ALIASED) {
          privatePages += exInfo.pages_resident;
        }
        break;
      }
      default:
        break;
    }
  }

  vm_size_t pageSize;
  if (host_page_size(aPort ? aPort : mach_task_self(), &pageSize) !=
      KERN_SUCCESS) {
    pageSize = PAGE_SIZE;
  }

  *aN = privatePages * pageSize;
  return NS_OK;
}

[[nodiscard]] static nsresult PhysicalFootprintAmount(int64_t* aN,
                                                      mach_port_t aPort = 0) {
  MOZ_ASSERT(aN);

  // The phys_footprint value (introduced in 10.11) of the TASK_VM_INFO data
  // matches the value in the 'Memory' column of the Activity Monitor.
  task_vm_info_data_t task_vm_info;
  mach_msg_type_number_t count = TASK_VM_INFO_COUNT;
  kern_return_t kr = task_info(aPort ? aPort : mach_task_self(), TASK_VM_INFO,
                               (task_info_t)&task_vm_info, &count);
  if (kr != KERN_SUCCESS) {
    return NS_ERROR_FAILURE;
  }

  *aN = task_vm_info.phys_footprint;
  return NS_OK;
}

#elif defined(XP_WIN)

#  include <windows.h>
#  include <psapi.h>
#  include <algorithm>

#  define HAVE_VSIZE_AND_RESIDENT_REPORTERS 1
[[nodiscard]] static nsresult VsizeDistinguishedAmount(int64_t* aN) {
  MEMORYSTATUSEX s;
  s.dwLength = sizeof(s);

  if (!GlobalMemoryStatusEx(&s)) {
    return NS_ERROR_FAILURE;
  }

  *aN = s.ullTotalVirtual - s.ullAvailVirtual;
  return NS_OK;
}

[[nodiscard]] static nsresult ResidentDistinguishedAmount(int64_t* aN) {
  PROCESS_MEMORY_COUNTERS pmc;
  pmc.cb = sizeof(PROCESS_MEMORY_COUNTERS);

  if (!GetProcessMemoryInfo(GetCurrentProcess(), &pmc, sizeof(pmc))) {
    return NS_ERROR_FAILURE;
  }

  *aN = pmc.WorkingSetSize;
  return NS_OK;
}

[[nodiscard]] static nsresult ResidentFastDistinguishedAmount(int64_t* aN) {
  return ResidentDistinguishedAmount(aN);
}

#  define HAVE_RESIDENT_UNIQUE_REPORTER 1

[[nodiscard]] static nsresult ResidentUniqueDistinguishedAmount(
    int64_t* aN, HANDLE aProcess = nullptr) {
  // Determine how many entries we need.
  PSAPI_WORKING_SET_INFORMATION tmp;
  DWORD tmpSize = sizeof(tmp);
  memset(&tmp, 0, tmpSize);

  HANDLE proc = aProcess ? aProcess : GetCurrentProcess();
  QueryWorkingSet(proc, &tmp, tmpSize);

  // Fudge the size in case new entries are added between calls.
  size_t entries = tmp.NumberOfEntries * 2;

  if (!entries) {
    return NS_ERROR_FAILURE;
  }

  DWORD infoArraySize = tmpSize + (entries * sizeof(PSAPI_WORKING_SET_BLOCK));
  UniqueFreePtr<PSAPI_WORKING_SET_INFORMATION> infoArray(
      static_cast<PSAPI_WORKING_SET_INFORMATION*>(malloc(infoArraySize)));

  if (!infoArray) {
    return NS_ERROR_FAILURE;
  }

  if (!QueryWorkingSet(proc, infoArray.get(), infoArraySize)) {
    return NS_ERROR_FAILURE;
  }

  entries = static_cast<size_t>(infoArray->NumberOfEntries);
  size_t privatePages = 0;
  for (size_t i = 0; i < entries; i++) {
    // Count shared pages that only one process is using as private.
    if (!infoArray->WorkingSetInfo[i].Shared ||
        infoArray->WorkingSetInfo[i].ShareCount <= 1) {
      privatePages++;
    }
  }

  SYSTEM_INFO si;
  GetSystemInfo(&si);

  *aN = privatePages * si.dwPageSize;
  return NS_OK;
}

#  define HAVE_VSIZE_MAX_CONTIGUOUS_REPORTER 1
[[nodiscard]] static nsresult VsizeMaxContiguousDistinguishedAmount(
    int64_t* aN) {
  SIZE_T biggestRegion = 0;
  MEMORY_BASIC_INFORMATION vmemInfo = {0};
  for (size_t currentAddress = 0;;) {
    if (!VirtualQuery((LPCVOID)currentAddress, &vmemInfo, sizeof(vmemInfo))) {
      // Something went wrong, just return whatever we've got already.
      break;
    }

    if (vmemInfo.State == MEM_FREE) {
      biggestRegion = std::max(biggestRegion, vmemInfo.RegionSize);
    }

    SIZE_T lastAddress = currentAddress;
    currentAddress += vmemInfo.RegionSize;

    // If we overflow, we've examined all of the address space.
    if (currentAddress < lastAddress) {
      break;
    }
  }

  *aN = biggestRegion;
  return NS_OK;
}

#  define HAVE_PRIVATE_REPORTER 1
[[nodiscard]] static nsresult PrivateDistinguishedAmount(int64_t* aN) {
  PROCESS_MEMORY_COUNTERS_EX pmcex;
  pmcex.cb = sizeof(PROCESS_MEMORY_COUNTERS_EX);

  if (!GetProcessMemoryInfo(GetCurrentProcess(),
                            (PPROCESS_MEMORY_COUNTERS)&pmcex, sizeof(pmcex))) {
    return NS_ERROR_FAILURE;
  }

  *aN = pmcex.PrivateUsage;
  return NS_OK;
}

#  define HAVE_SYSTEM_HEAP_REPORTER 1
// Windows can have multiple separate heaps, but we should not touch non-default
// heaps because they may be destroyed at anytime while we hold a handle.  So we
// count only the default heap.
[[nodiscard]] static nsresult SystemHeapSize(int64_t* aSizeOut) {
  HANDLE heap = GetProcessHeap();

  NS_ENSURE_TRUE(HeapLock(heap), NS_ERROR_FAILURE);

  int64_t heapSize = 0;
  PROCESS_HEAP_ENTRY entry;
  entry.lpData = nullptr;
  while (HeapWalk(heap, &entry)) {
    // We don't count entry.cbOverhead, because we just want to measure the
    // space available to the program.
    if (entry.wFlags & PROCESS_HEAP_ENTRY_BUSY) {
      heapSize += entry.cbData;
    }
  }

  // Check this result only after unlocking the heap, so that we don't leave
  // the heap locked if there was an error.
  DWORD lastError = GetLastError();

  // I have no idea how things would proceed if unlocking this heap failed...
  NS_ENSURE_TRUE(HeapUnlock(heap), NS_ERROR_FAILURE);

  NS_ENSURE_TRUE(lastError == ERROR_NO_MORE_ITEMS, NS_ERROR_FAILURE);

  *aSizeOut = heapSize;
  return NS_OK;
}

struct SegmentKind {
  DWORD mState;
  DWORD mType;
  DWORD mProtect;
  int mIsStack;
};

struct SegmentEntry : public PLDHashEntryHdr {
  static PLDHashNumber HashKey(const void* aKey) {
    auto kind = static_cast<const SegmentKind*>(aKey);
    return mozilla::HashGeneric(kind->mState, kind->mType, kind->mProtect,
                                kind->mIsStack);
  }

  static bool MatchEntry(const PLDHashEntryHdr* aEntry, const void* aKey) {
    auto kind = static_cast<const SegmentKind*>(aKey);
    auto entry = static_cast<const SegmentEntry*>(aEntry);
    return kind->mState == entry->mKind.mState &&
           kind->mType == entry->mKind.mType &&
           kind->mProtect == entry->mKind.mProtect &&
           kind->mIsStack == entry->mKind.mIsStack;
  }

  static void InitEntry(PLDHashEntryHdr* aEntry, const void* aKey) {
    auto kind = static_cast<const SegmentKind*>(aKey);
    auto entry = static_cast<SegmentEntry*>(aEntry);
    entry->mKind = *kind;
    entry->mCount = 0;
    entry->mSize = 0;
  }

  static const PLDHashTableOps Ops;

  SegmentKind mKind;  // The segment kind.
  uint32_t mCount;    // The number of segments of this kind.
  size_t mSize;       // The combined size of segments of this kind.
};

/* static */ const PLDHashTableOps SegmentEntry::Ops = {
    SegmentEntry::HashKey, SegmentEntry::MatchEntry,
    PLDHashTable::MoveEntryStub, PLDHashTable::ClearEntryStub,
    SegmentEntry::InitEntry};

class WindowsAddressSpaceReporter final : public nsIMemoryReporter {
  ~WindowsAddressSpaceReporter() {}

public:
  NS_DECL_ISUPPORTS

  NS_IMETHOD CollectReports(nsIHandleReportCallback* aHandleReport,
                            nsISupports* aData, bool aAnonymize) override {
    // First iterate over all the segments and record how many of each kind
    // there were and their aggregate sizes. We use a hash table for this
    // because there are a couple of dozen different kinds possible.

    PLDHashTable table(&SegmentEntry::Ops, sizeof(SegmentEntry));
    MEMORY_BASIC_INFORMATION info = {0};
    bool isPrevSegStackGuard = false;
    for (size_t currentAddress = 0;;) {
      if (!VirtualQuery((LPCVOID)currentAddress, &info, sizeof(info))) {
        // Something went wrong, just return whatever we've got already.
        break;
      }

      size_t size = info.RegionSize;

      // Note that |type| and |protect| are ignored in some cases.
      DWORD state = info.State;
      DWORD type =
          (state == MEM_RESERVE || state == MEM_COMMIT) ? info.Type : 0;
      DWORD protect = (state == MEM_COMMIT) ? info.Protect : 0;
      bool isStack = isPrevSegStackGuard && state == MEM_COMMIT &&
                     type == MEM_PRIVATE && protect == PAGE_READWRITE;

      SegmentKind kind = {state, type, protect, isStack ? 1 : 0};
      auto entry =
          static_cast<SegmentEntry*>(table.Add(&kind, mozilla::fallible));
      if (entry) {
        entry->mCount += 1;
        entry->mSize += size;
      }

      isPrevSegStackGuard = info.State == MEM_COMMIT &&
                            info.Type == MEM_PRIVATE &&
                            info.Protect == (PAGE_READWRITE | PAGE_GUARD);

      size_t lastAddress = currentAddress;
      currentAddress += size;

      // If we overflow, we've examined all of the address space.
      if (currentAddress < lastAddress) {
        break;
      }
    }

    // Then iterate over the hash table and report the details for each segment
    // kind.

    for (auto iter = table.Iter(); !iter.Done(); iter.Next()) {
      // For each range of pages, we consider one or more of its State, Type
      // and Protect values. These are documented at
      // https://msdn.microsoft.com/en-us/library/windows/desktop/aa366775%28v=vs.85%29.aspx
      // (for State and Type) and
      // https://msdn.microsoft.com/en-us/library/windows/desktop/aa366786%28v=vs.85%29.aspx
      // (for Protect).
      //
      // Not all State values have accompanying Type and Protection values.
      bool doType = false;
      bool doProtect = false;

      auto entry = static_cast<const SegmentEntry*>(iter.Get());

      nsCString path("address-space");

      switch (entry->mKind.mState) {
        case MEM_FREE:
          path.AppendLiteral("/free");
          break;

        case MEM_RESERVE:
          path.AppendLiteral("/reserved");
          doType = true;
          break;

        case MEM_COMMIT:
          path.AppendLiteral("/commit");
          doType = true;
          doProtect = true;
          break;

        default:
          // Should be impossible, but handle it just in case.
          path.AppendLiteral("/???");
          break;
      }

      if (doType) {
        switch (entry->mKind.mType) {
          case MEM_IMAGE:
            path.AppendLiteral("/image");
            break;

          case MEM_MAPPED:
            path.AppendLiteral("/mapped");
            break;

          case MEM_PRIVATE:
            path.AppendLiteral("/private");
            break;

          default:
            // Should be impossible, but handle it just in case.
            path.AppendLiteral("/???");
            break;
        }
      }

      if (doProtect) {
        DWORD protect = entry->mKind.mProtect;
        // Basic attributes. Exactly one of these should be set.
        if (protect & PAGE_EXECUTE) {
          path.AppendLiteral("/execute");
        }
        if (protect & PAGE_EXECUTE_READ) {
          path.AppendLiteral("/execute-read");
        }
        if (protect & PAGE_EXECUTE_READWRITE) {
          path.AppendLiteral("/execute-readwrite");
        }
        if (protect & PAGE_EXECUTE_WRITECOPY) {
          path.AppendLiteral("/execute-writecopy");
        }
        if (protect & PAGE_NOACCESS) {
          path.AppendLiteral("/noaccess");
        }
        if (protect & PAGE_READONLY) {
          path.AppendLiteral("/readonly");
        }
        if (protect & PAGE_READWRITE) {
          path.AppendLiteral("/readwrite");
        }
        if (protect & PAGE_WRITECOPY) {
          path.AppendLiteral("/writecopy");
        }

        // Modifiers. At most one of these should be set.
        if (protect & PAGE_GUARD) {
          path.AppendLiteral("+guard");
        }
        if (protect & PAGE_NOCACHE) {
          path.AppendLiteral("+nocache");
        }
        if (protect & PAGE_WRITECOMBINE) {
          path.AppendLiteral("+writecombine");
        }

        // Annotate likely stack segments, too.
        if (entry->mKind.mIsStack) {
          path.AppendLiteral("+stack");
        }
      }

      // Append the segment count.
      path.AppendPrintf("(segments=%u)", entry->mCount);

      aHandleReport->Callback(""_ns, path, KIND_OTHER, UNITS_BYTES,
                              entry->mSize, "From MEMORY_BASIC_INFORMATION."_ns,
                              aData);
    }

    return NS_OK;
  }
};
NS_IMPL_ISUPPORTS(WindowsAddressSpaceReporter, nsIMemoryReporter)

#endif  // XP_<PLATFORM>

#ifdef HAVE_VSIZE_MAX_CONTIGUOUS_REPORTER
class VsizeMaxContiguousReporter final : public nsIMemoryReporter {
  ~VsizeMaxContiguousReporter() {}

public:
  NS_DECL_ISUPPORTS

  NS_IMETHOD CollectReports(nsIHandleReportCallback* aHandleReport,
                            nsISupports* aData, bool aAnonymize) override {
    int64_t amount;
    if (NS_SUCCEEDED(VsizeMaxContiguousDistinguishedAmount(&amount))) {
      MOZ_COLLECT_REPORT(
          "vsize-max-contiguous", KIND_OTHER, UNITS_BYTES, amount,
          "Size of the maximum contiguous block of available virtual memory.");
    }
    return NS_OK;
  }
};
NS_IMPL_ISUPPORTS(VsizeMaxContiguousReporter, nsIMemoryReporter)
#endif

#ifdef HAVE_PRIVATE_REPORTER
class PrivateReporter final : public nsIMemoryReporter {
  ~PrivateReporter() {}

public:
  NS_DECL_ISUPPORTS

  NS_IMETHOD CollectReports(nsIHandleReportCallback* aHandleReport,
                            nsISupports* aData, bool aAnonymize) override {
    int64_t amount;
    if (NS_SUCCEEDED(PrivateDistinguishedAmount(&amount))) {
      // clang-format off
      MOZ_COLLECT_REPORT(
        "private", KIND_OTHER, UNITS_BYTES, amount,
"Memory that cannot be shared with other processes, including memory that is "
"committed and marked MEM_PRIVATE, data that is not mapped, and executable "
"pages that have been written to.");
      // clang-format on
    }
    return NS_OK;
  }
};
NS_IMPL_ISUPPORTS(PrivateReporter, nsIMemoryReporter)
#endif

#ifdef HAVE_VSIZE_AND_RESIDENT_REPORTERS
class VsizeReporter final : public nsIMemoryReporter {
  ~VsizeReporter() = default;

public:
  NS_DECL_ISUPPORTS

  NS_IMETHOD CollectReports(nsIHandleReportCallback* aHandleReport,
                            nsISupports* aData, bool aAnonymize) override {
    int64_t amount;
    if (NS_SUCCEEDED(VsizeDistinguishedAmount(&amount))) {
      // clang-format off
      MOZ_COLLECT_REPORT(
        "vsize", KIND_OTHER, UNITS_BYTES, amount,
"Memory mapped by the process, including code and data segments, the heap, "
"thread stacks, memory explicitly mapped by the process via mmap and similar "
"operations, and memory shared with other processes. This is the vsize figure "
"as reported by 'top' and 'ps'. This figure is of limited use on Mac, where "
"processes share huge amounts of memory with one another. But even on other "
"operating systems, 'resident' is a much better measure of the memory "
"resources used by the process.");
      // clang-format on
    }
    return NS_OK;
  }
};
NS_IMPL_ISUPPORTS(VsizeReporter, nsIMemoryReporter)

class ResidentReporter final : public nsIMemoryReporter {
  ~ResidentReporter() = default;

public:
  NS_DECL_ISUPPORTS

  NS_IMETHOD CollectReports(nsIHandleReportCallback* aHandleReport,
                            nsISupports* aData, bool aAnonymize) override {
    int64_t amount;
    if (NS_SUCCEEDED(ResidentDistinguishedAmount(&amount))) {
      // clang-format off
      MOZ_COLLECT_REPORT(
        "resident", KIND_OTHER, UNITS_BYTES, amount,
"Memory mapped by the process that is present in physical memory, also known "
"as the resident set size (RSS). This is the best single figure to use when "
"considering the memory resources used by the process, but it depends both on "
"other processes being run and details of the OS kernel and so is best used "
"for comparing the memory usage of a single process at different points in "
"time.");
      // clang-format on
    }
    return NS_OK;
  }
};
NS_IMPL_ISUPPORTS(ResidentReporter, nsIMemoryReporter)

#endif  // HAVE_VSIZE_AND_RESIDENT_REPORTERS

#ifdef HAVE_RESIDENT_UNIQUE_REPORTER
class ResidentUniqueReporter final : public nsIMemoryReporter {
  ~ResidentUniqueReporter() = default;

public:
  NS_DECL_ISUPPORTS

  NS_IMETHOD CollectReports(nsIHandleReportCallback* aHandleReport,
                            nsISupports* aData, bool aAnonymize) override {
    int64_t amount = 0;
    // clang-format off
    if (NS_SUCCEEDED(ResidentUniqueDistinguishedAmount(&amount))) {
      MOZ_COLLECT_REPORT(
        "resident-unique", KIND_OTHER, UNITS_BYTES, amount,
"Memory mapped by the process that is present in physical memory and not "
"shared with any other processes. This is also known as the process's unique "
"set size (USS). This is the amount of RAM we'd expect to be freed if we "
"closed this process.");
    }
#ifdef XP_MACOSX
    if (NS_SUCCEEDED(PhysicalFootprintAmount(&amount))) {
      MOZ_COLLECT_REPORT(
        "resident-phys-footprint", KIND_OTHER, UNITS_BYTES, amount,
"Memory footprint reported by MacOS's task_info API's phys_footprint field. "
"This matches the memory column in Activity Monitor.");
    }
#endif
    // clang-format on
    return NS_OK;
  }
};
NS_IMPL_ISUPPORTS(ResidentUniqueReporter, nsIMemoryReporter)

#endif  // HAVE_RESIDENT_UNIQUE_REPORTER

#ifdef HAVE_SYSTEM_HEAP_REPORTER

class SystemHeapReporter final : public nsIMemoryReporter {
  ~SystemHeapReporter() = default;

public:
  NS_DECL_ISUPPORTS

  NS_IMETHOD CollectReports(nsIHandleReportCallback* aHandleReport,
                            nsISupports* aData, bool aAnonymize) override {
    int64_t amount;
    if (NS_SUCCEEDED(SystemHeapSize(&amount))) {
      // clang-format off
      MOZ_COLLECT_REPORT(
        "system-heap-allocated", KIND_OTHER, UNITS_BYTES, amount,
"Memory used by the system allocator that is currently allocated to the "
"application. This is distinct from the jemalloc heap that Firefox uses for "
"most or all of its heap allocations. Ideally this number is zero, but "
"on some platforms we cannot force every heap allocation through jemalloc.");
      // clang-format on
    }
    return NS_OK;
  }
};
NS_IMPL_ISUPPORTS(SystemHeapReporter, nsIMemoryReporter)
#endif  // HAVE_SYSTEM_HEAP_REPORTER

#ifdef XP_UNIX

#  include <sys/resource.h>

#  define HAVE_RESIDENT_PEAK_REPORTER 1

[[nodiscard]] static nsresult ResidentPeakDistinguishedAmount(int64_t* aN) {
  struct rusage usage;
  if (0 == getrusage(RUSAGE_SELF, &usage)) {
    // The units for ru_maxrrs:
    // - Mac: bytes
    // - Solaris: pages? But some sources it actually always returns 0, so
    //   check for that
    // - Linux, {Net/Open/Free}BSD, DragonFly: KiB
#  ifdef XP_MACOSX
    *aN = usage.ru_maxrss;
#  elif defined(SOLARIS)
    *aN = usage.ru_maxrss * getpagesize();
#  else
    *aN = usage.ru_maxrss * 1024;
#  endif
    if (*aN > 0) {
      return NS_OK;
    }
  }
  return NS_ERROR_FAILURE;
}

class ResidentPeakReporter final : public nsIMemoryReporter {
  ~ResidentPeakReporter() = default;

public:
  NS_DECL_ISUPPORTS

  NS_IMETHOD CollectReports(nsIHandleReportCallback* aHandleReport,
                            nsISupports* aData, bool aAnonymize) override {
    int64_t amount = 0;
    if (NS_SUCCEEDED(ResidentPeakDistinguishedAmount(&amount))) {
      MOZ_COLLECT_REPORT(
          "resident-peak", KIND_OTHER, UNITS_BYTES, amount,
          "The peak 'resident' value for the lifetime of the process.");
    }
    return NS_OK;
  }
};
NS_IMPL_ISUPPORTS(ResidentPeakReporter, nsIMemoryReporter)

#  define HAVE_PAGE_FAULT_REPORTERS 1

class PageFaultsSoftReporter final : public nsIMemoryReporter {
  ~PageFaultsSoftReporter() = default;

public:
  NS_DECL_ISUPPORTS

  NS_IMETHOD CollectReports(nsIHandleReportCallback* aHandleReport,
                            nsISupports* aData, bool aAnonymize) override {
    struct rusage usage;
    int err = getrusage(RUSAGE_SELF, &usage);
    if (err == 0) {
      int64_t amount = usage.ru_minflt;
      // clang-format off
      MOZ_COLLECT_REPORT(
        "page-faults-soft", KIND_OTHER, UNITS_COUNT_CUMULATIVE, amount,
"The number of soft page faults (also known as 'minor page faults') that "
"have occurred since the process started. A soft page fault occurs when the "
"process tries to access a page which is present in physical memory but is "
"not mapped into the process's address space. For instance, a process might "
"observe soft page faults when it loads a shared library which is already "
"present in physical memory. A process may experience many thousands of soft "
"page faults even when the machine has plenty of available physical memory, "
"and because the OS services a soft page fault without accessing the disk, "
"they impact performance much less than hard page faults.");
      // clang-format on
    }
    return NS_OK;
  }
};
NS_IMPL_ISUPPORTS(PageFaultsSoftReporter, nsIMemoryReporter)

[[nodiscard]] static nsresult PageFaultsHardDistinguishedAmount(
    int64_t* aAmount) {
  struct rusage usage;
  int err = getrusage(RUSAGE_SELF, &usage);
  if (err != 0) {
    return NS_ERROR_FAILURE;
  }
  *aAmount = usage.ru_majflt;
  return NS_OK;
}

class PageFaultsHardReporter final : public nsIMemoryReporter {
  ~PageFaultsHardReporter() = default;

public:
  NS_DECL_ISUPPORTS

  NS_IMETHOD CollectReports(nsIHandleReportCallback* aHandleReport,
                            nsISupports* aData, bool aAnonymize) override {
    int64_t amount = 0;
    if (NS_SUCCEEDED(PageFaultsHardDistinguishedAmount(&amount))) {
      // clang-format off
      MOZ_COLLECT_REPORT(
        "page-faults-hard", KIND_OTHER, UNITS_COUNT_CUMULATIVE, amount,
"The number of hard page faults (also known as 'major page faults') that have "
"occurred since the process started. A hard page fault occurs when a process "
"tries to access a page which is not present in physical memory. The "
"operating system must access the disk in order to fulfill a hard page fault. "
"When memory is plentiful, you should see very few hard page faults. But if "
"the process tries to use more memory than your machine has available, you "
"may see many thousands of hard page faults. Because accessing the disk is up "
"to a million times slower than accessing RAM, the program may run very "
"slowly when it is experiencing more than 100 or so hard page faults a "
"second.");
      // clang-format on
    }
    return NS_OK;
  }
};
NS_IMPL_ISUPPORTS(PageFaultsHardReporter, nsIMemoryReporter)

#endif  // XP_UNIX

/**
** memory reporter implementation for jemalloc and OSX malloc,
** to obtain info on total memory in use (that we know about,
** at least -- on OSX, there are sometimes other zones in use).
**/

#ifdef HAVE_JEMALLOC_STATS

static size_t HeapOverhead(const jemalloc_stats_t& aStats) {
  return aStats.waste + aStats.bookkeeping + aStats.pages_dirty +
         aStats.bin_unused;
}

// This has UNITS_PERCENTAGE, so it is multiplied by 100x *again* on top of the
// 100x for the percentage.

// static
int64_t nsMemoryReporterManager::HeapOverheadFraction(
    const jemalloc_stats_t& aStats) {
  size_t heapOverhead = HeapOverhead(aStats);
  size_t heapCommitted = aStats.allocated + heapOverhead;
  return int64_t(10000 * (heapOverhead / (double)heapCommitted));
}

class JemallocHeapReporter final : public nsIMemoryReporter {
  ~JemallocHeapReporter() = default;

public:
  NS_DECL_ISUPPORTS

  NS_IMETHOD CollectReports(nsIHandleReportCallback* aHandleReport,
                            nsISupports* aData, bool aAnonymize) override {
    jemalloc_stats_t stats;
    const size_t num_bins = jemalloc_stats_num_bins();
    nsTArray<jemalloc_bin_stats_t> bin_stats(num_bins);
    bin_stats.SetLength(num_bins);
    jemalloc_stats(&stats, bin_stats.Elements());

    // clang-format off
    MOZ_COLLECT_REPORT(
      "heap/committed/allocated", KIND_OTHER, UNITS_BYTES, stats.allocated,
"Memory mapped by the heap allocator that is currently allocated to the "
"application. This may exceed the amount of memory requested by the "
"application because the allocator regularly rounds up request sizes. (The "
"exact amount requested is not recorded.)");

    MOZ_COLLECT_REPORT(
      "heap-allocated", KIND_OTHER, UNITS_BYTES, stats.allocated,
"The same as 'heap/committed/allocated'.");

    // We mark this and the other heap/committed/overhead reporters as KIND_NONHEAP
    // because KIND_HEAP memory means "counted in heap-allocated", which
    // this is not.
    for (auto& bin : bin_stats) {
      MOZ_ASSERT(bin.size);
      nsPrintfCString path("heap/committed/bin-unused/bin-%zu",
          bin.size);
      aHandleReport->Callback(EmptyCString(), path, KIND_NONHEAP, UNITS_BYTES,
        bin.bytes_unused,
        nsLiteralCString(
          "Unused bytes in all runs of all bins for this size class"),
        aData);
    }

    if (stats.waste > 0) {
      MOZ_COLLECT_REPORT(
        "heap/committed/waste", KIND_NONHEAP, UNITS_BYTES,
        stats.waste,
"Committed bytes which do not correspond to an active allocation and which the "
"allocator is not intentionally keeping alive (i.e., not "
"'heap/{bookkeeping,unused-pages,bin-unused}').");
    }

    MOZ_COLLECT_REPORT(
      "heap/committed/bookkeeping", KIND_NONHEAP, UNITS_BYTES,
      stats.bookkeeping,
"Committed bytes which the heap allocator uses for internal data structures.");

    MOZ_COLLECT_REPORT(
      "heap/committed/unused-pages/dirty", KIND_NONHEAP, UNITS_BYTES,
      stats.pages_dirty,
"Memory which the allocator could return to the operating system, but hasn't. "
"The allocator keeps this memory around as an optimization, so it doesn't "
"have to ask the OS the next time it needs to fulfill a request. This value "
"is typically not larger than a few megabytes.");

    MOZ_COLLECT_REPORT(
      "heap/decommitted/unused-pages/fresh", KIND_OTHER, UNITS_BYTES, stats.pages_fresh,
"Amount of memory currently mapped but has never been used.");
    // A duplicate entry in the decommitted part of the tree.
    MOZ_COLLECT_REPORT(
      "decommitted/heap/unused-pages/fresh", KIND_OTHER, UNITS_BYTES, stats.pages_fresh,
"Amount of memory currently mapped but has never been used.");

// On MacOS madvised memory is still counted in the resident set until the OS
// actually decommits it.
#ifdef XP_MACOSX
#define MADVISED_GROUP "committed"
#else
#define MADVISED_GROUP "decommitted"
#endif
    MOZ_COLLECT_REPORT(
      "heap/" MADVISED_GROUP "/unused-pages/madvised", KIND_OTHER, UNITS_BYTES,
      stats.pages_madvised,
"Amount of memory currently mapped, not used and that the OS should remove "
"from the application's resident set.");
    // A duplicate entry in the decommitted part of the tree.
    MOZ_COLLECT_REPORT(
      "decommitted/heap/unused-pages/madvised", KIND_OTHER, UNITS_BYTES, stats.pages_madvised,
"Amount of memory currently mapped, not used and that the OS should remove "
"from the application's resident set.");

    {
      size_t decommitted = stats.mapped - stats.allocated - stats.waste - stats.pages_dirty - stats.pages_fresh - stats.bookkeeping - stats.bin_unused;
      MOZ_COLLECT_REPORT(
        "heap/decommitted/unmapped", KIND_OTHER, UNITS_BYTES, decommitted,
  "Amount of memory currently mapped but not committed, "
  "neither in physical memory nor paged to disk.");
      MOZ_COLLECT_REPORT(
        "decommitted/heap/decommitted", KIND_OTHER, UNITS_BYTES, decommitted,
  "Amount of memory currently mapped but not committed, "
  "neither in physical memory nor paged to disk.");
    }
    MOZ_COLLECT_REPORT(
      "heap-chunksize", KIND_OTHER, UNITS_BYTES, stats.chunksize,
      "Size of chunks.");

#ifdef MOZ_PHC
    mozilla::phc::MemoryUsage usage;
    mozilla::phc::PHCMemoryUsage(usage);

    MOZ_COLLECT_REPORT(
      "explicit/phc/metadata", KIND_NONHEAP, UNITS_BYTES,
      usage.mMetadataBytes,
"Memory used by PHC to store stacks and other metadata for each allocation");
    MOZ_COLLECT_REPORT(
      "explicit/phc/fragmentation", KIND_NONHEAP, UNITS_BYTES,
      usage.mFragmentationBytes,
"The amount of memory lost due to rounding up allocations to the next page "
"size. "
"This is also known as 'internal fragmentation'. "
"Note that all allocators have some internal fragmentation, there may still "
"be some internal fragmentation without PHC.");
#endif

    // clang-format on

    return NS_OK;
  }
};
NS_IMPL_ISUPPORTS(JemallocHeapReporter, nsIMemoryReporter)

#endif  // HAVE_JEMALLOC_STATS

// Why is this here?  At first glance, you'd think it could be defined and
// registered with nsMemoryReporterManager entirely within nsAtomTable.cpp.
// However, the obvious time to register it is when the table is initialized,
// and that happens before XPCOM components are initialized, which means the
// RegisterStrongMemoryReporter call fails.  So instead we do it here.
class AtomTablesReporter final : public nsIMemoryReporter {
  MOZ_DEFINE_MALLOC_SIZE_OF(MallocSizeOf)

  ~AtomTablesReporter() = default;

public:
  NS_DECL_ISUPPORTS

  NS_IMETHOD CollectReports(nsIHandleReportCallback* aHandleReport,
                            nsISupports* aData, bool aAnonymize) override {
    AtomsSizes sizes;
    NS_AddSizeOfAtoms(MallocSizeOf, sizes);

    MOZ_COLLECT_REPORT("explicit/atoms/table", KIND_HEAP, UNITS_BYTES,
                       sizes.mTable, "Memory used by the atom table.");

    MOZ_COLLECT_REPORT(
        "explicit/atoms/dynamic-objects-and-chars", KIND_HEAP, UNITS_BYTES,
        sizes.mDynamicAtoms,
        "Memory used by dynamic atom objects and chars (which are stored "
        "at the end of each atom object).");

    return NS_OK;
  }
};
NS_IMPL_ISUPPORTS(AtomTablesReporter, nsIMemoryReporter)

class ThreadsReporter final : public nsIMemoryReporter {
  MOZ_DEFINE_MALLOC_SIZE_OF(MallocSizeOf)
  ~ThreadsReporter() = default;

public:
  NS_DECL_ISUPPORTS

  NS_IMETHOD CollectReports(nsIHandleReportCallback* aHandleReport,
                            nsISupports* aData, bool aAnonymize) override {
#ifdef XP_LINUX
    nsTArray<MemoryMapping> mappings(1024);
    MOZ_TRY(GetMemoryMappings(mappings));
#endif

    // Enumerating over active threads requires holding a lock, so we collect
    // info on all threads, and then call our reporter callbacks after releasing
    // the lock.
    struct ThreadData {
      nsCString mName;
      uint32_t mThreadId;
      size_t mPrivateSize;
    };
    AutoTArray<ThreadData, 32> threads;

    size_t eventQueueSizes = 0;
    size_t wrapperSizes = 0;
    size_t threadCount = 0;

    {
      nsThreadManager& tm = nsThreadManager::get();
      OffTheBooksMutexAutoLock lock(tm.ThreadListMutex());
      for (auto* thread : tm.ThreadList()) {
        threadCount++;
        eventQueueSizes += thread->SizeOfEventQueues(MallocSizeOf);
        wrapperSizes += thread->ShallowSizeOfIncludingThis(MallocSizeOf);

        if (!thread->StackBase()) {
          continue;
        }

#if defined(XP_LINUX)
        int idx = mappings.BinaryIndexOf(thread->StackBase());
        if (idx < 0) {
          continue;
        }
        // Referenced() is the combined size of all pages in the region which
        // have ever been touched, and are therefore consuming memory. For stack
        // regions, these pages are guaranteed to be un-shared unless we fork
        // after creating threads (which we don't).
        size_t privateSize = mappings[idx].Referenced();

        // On Linux, we have to be very careful matching memory regions to
        // thread stacks.
        //
        // To begin with, the kernel only reports VM stats for regions of all
        // adjacent pages with the same flags, protection, and backing file.
        // There's no way to get finer-grained usage information for a subset of
        // those pages.
        //
        // Stack segments always have a guard page at the bottom of the stack
        // (assuming we only support stacks that grow down), so there's no
        // danger of them being merged with other stack regions. At the top,
        // there's no protection page, and no way to allocate one without using
        // pthreads directly and allocating our own stacks. So we get around the
        // problem by adding an extra VM flag (NOHUGEPAGES) to our stack region,
        // which we don't expect to be set on any heap regions. But this is not
        // fool-proof.
        //
        // A second kink is that different C libraries (and different versions
        // thereof) report stack base locations and sizes differently with
        // regard to the guard page. For the libraries that include the guard
        // page in the stack size base pointer, we need to adjust those values
        // to compensate. But it's possible that our logic will get out of sync
        // with library changes, or someone will compile with an unexpected
        // library.
        //
        //
        // The upshot of all of this is that there may be configurations that
        // our special cases don't cover. And if there are, we want to know
        // about it. So assert that total size of the memory region we're
        // reporting actually matches the allocated size of the thread stack.
#  ifndef ANDROID
        MOZ_ASSERT(mappings[idx].Size() == thread->StackSize(),
                   "Mapping region size doesn't match stack allocation size");
#  endif
#elif defined(XP_WIN)
        auto memInfo =
            MemoryInfo::Get(thread->StackBase(), thread->StackSize());
        size_t privateSize = memInfo.Committed();
#else
        size_t privateSize = thread->StackSize();
        MOZ_ASSERT_UNREACHABLE(
            "Shouldn't have stack base pointer on this "
            "platform");
#endif

        nsCString threadName;
        thread->GetThreadName(threadName);
        threads.AppendElement(ThreadData{
            std::move(threadName),
            thread->ThreadId(),
            // On Linux, it's possible (but unlikely) that our stack region will
            // have been merged with adjacent heap regions, in which case we'll
            // get combined size information for both. So we take the minimum of
            // the reported private size and the requested stack size to avoid
            // the possible of majorly over-reporting in that case.
            std::min(privateSize, thread->StackSize()),
        });
      }
    }

    for (auto& thread : threads) {
      nsPrintfCString path("explicit/threads/stacks/%s (tid=%u)",
                           thread.mName.get(), thread.mThreadId);

      aHandleReport->Callback(
          ""_ns, path, KIND_NONHEAP, UNITS_BYTES, thread.mPrivateSize,
          nsLiteralCString("The sizes of thread stacks which have been "
                           "committed to memory."),
          aData);
    }

    MOZ_COLLECT_REPORT("explicit/threads/overhead/event-queues", KIND_HEAP,
                       UNITS_BYTES, eventQueueSizes,
                       "The sizes of nsThread event queues and observers.");

    MOZ_COLLECT_REPORT("explicit/threads/overhead/wrappers", KIND_HEAP,
                       UNITS_BYTES, wrapperSizes,
                       "The sizes of nsThread/PRThread wrappers.");

#if defined(XP_WIN)
    // Each thread on Windows has a fixed kernel overhead. For 32 bit Windows,
    // that's 12K. For 64 bit, it's 24K.
    //
    // See
    // https://blogs.technet.microsoft.com/markrussinovich/2009/07/05/pushing-the-limits-of-windows-processes-and-threads/
    constexpr size_t kKernelSize = (sizeof(void*) == 8 ? 24 : 12) * 1024;
#elif defined(XP_LINUX)
    // On Linux, kernel stacks are usually 8K. However, on x86, they are
    // allocated virtually, and start out at 4K. They may grow to 8K, but we
    // have no way of knowing which ones do, so all we can do is guess.
#  if defined(__x86_64__) || defined(__i386__)
    constexpr size_t kKernelSize = 4 * 1024;
#  else
    constexpr size_t kKernelSize = 8 * 1024;
#  endif
#elif defined(XP_MACOSX)
    // On Darwin, kernel stacks are 16K:
    //
    // https://books.google.com/books?id=K8vUkpOXhN4C&lpg=PA513&dq=mach%20kernel%20thread%20stack%20size&pg=PA513#v=onepage&q=mach%20kernel%20thread%20stack%20size&f=false
    constexpr size_t kKernelSize = 16 * 1024;
#else
    // Elsewhere, just assume that kernel stacks require at least 8K.
    constexpr size_t kKernelSize = 8 * 1024;
#endif

    MOZ_COLLECT_REPORT("explicit/threads/overhead/kernel", KIND_NONHEAP,
                       UNITS_BYTES, threadCount * kKernelSize,
                       "The total kernel overhead for all active threads.");

    return NS_OK;
  }
};
NS_IMPL_ISUPPORTS(ThreadsReporter, nsIMemoryReporter)

#ifdef DEBUG

// Ideally, this would be implemented in BlockingResourceBase.cpp.
// However, this ends up breaking the linking step of various unit tests due
// to adding a new dependency to libdmd for a commonly used feature (mutexes)
// in  DMD  builds. So instead we do it here.
class DeadlockDetectorReporter final : public nsIMemoryReporter {
  MOZ_DEFINE_MALLOC_SIZE_OF(MallocSizeOf)

  ~DeadlockDetectorReporter() = default;

public:
  NS_DECL_ISUPPORTS

  NS_IMETHOD CollectReports(nsIHandleReportCallback* aHandleReport,
                            nsISupports* aData, bool aAnonymize) override {
    MOZ_COLLECT_REPORT(
        "explicit/deadlock-detector", KIND_HEAP, UNITS_BYTES,
        BlockingResourceBase::SizeOfDeadlockDetector(MallocSizeOf),
        "Memory used by the deadlock detector.");

    return NS_OK;
  }
};
NS_IMPL_ISUPPORTS(DeadlockDetectorReporter, nsIMemoryReporter)

#endif

#ifdef MOZ_DMD

namespace mozilla {
namespace dmd {

class DMDReporter final : public nsIMemoryReporter {
public:
  NS_DECL_ISUPPORTS

  NS_IMETHOD CollectReports(nsIHandleReportCallback* aHandleReport,
                            nsISupports* aData, bool aAnonymize) override {
    dmd::Sizes sizes;
    dmd::SizeOf(&sizes);

    MOZ_COLLECT_REPORT(
        "explicit/dmd/stack-traces/used", KIND_HEAP, UNITS_BYTES,
        sizes.mStackTracesUsed,
        "Memory used by stack traces which correspond to at least "
        "one heap block DMD is tracking.");

    MOZ_COLLECT_REPORT(
        "explicit/dmd/stack-traces/unused", KIND_HEAP, UNITS_BYTES,
        sizes.mStackTracesUnused,
        "Memory used by stack traces which don't correspond to any heap "
        "blocks DMD is currently tracking.");

    MOZ_COLLECT_REPORT("explicit/dmd/stack-traces/table", KIND_HEAP,
                       UNITS_BYTES, sizes.mStackTraceTable,
                       "Memory used by DMD's stack trace table.");

    MOZ_COLLECT_REPORT("explicit/dmd/live-block-table", KIND_HEAP, UNITS_BYTES,
                       sizes.mLiveBlockTable,
                       "Memory used by DMD's live block table.");

    MOZ_COLLECT_REPORT("explicit/dmd/dead-block-list", KIND_HEAP, UNITS_BYTES,
                       sizes.mDeadBlockTable,
                       "Memory used by DMD's dead block list.");

    return NS_OK;
  }

private:
  ~DMDReporter() = default;
};
NS_IMPL_ISUPPORTS(DMDReporter, nsIMemoryReporter)

}  // namespace dmd
}  // namespace mozilla

#endif  // MOZ_DMD

#ifdef MOZ_WIDGET_ANDROID
class AndroidMemoryReporter final : public nsIMemoryReporter {
public:
  NS_DECL_ISUPPORTS

  AndroidMemoryReporter() = default;

  NS_IMETHOD
  CollectReports(nsIHandleReportCallback* aHandleReport, nsISupports* aData,
                 bool aAnonymize) override {
    if (!jni::IsAvailable() || jni::GetAPIVersion() < 23) {
      return NS_OK;
    }

    int32_t heap = java::GeckoAppShell::GetMemoryUsage("summary.java-heap"_ns);
    if (heap > 0) {
      MOZ_COLLECT_REPORT("java-heap", KIND_OTHER, UNITS_BYTES, heap * 1024,
                         "The private Java Heap usage");
    }
    return NS_OK;
  }

private:
  ~AndroidMemoryReporter() = default;
};

NS_IMPL_ISUPPORTS(AndroidMemoryReporter, nsIMemoryReporter)
#endif

/**
** nsMemoryReporterManager implementation
**/

NS_IMPL_ISUPPORTS(nsMemoryReporterManager, nsIMemoryReporterManager,
                  nsIMemoryReporter)

NS_IMETHODIMP
nsMemoryReporterManager::Init() {
  if (!NS_IsMainThread()) {
    MOZ_CRASH();
  }

  // Under normal circumstances this function is only called once. However,
  // we've (infrequently) seen memory report dumps in crash reports that
  // suggest that this function is sometimes called multiple times. That in
  // turn means that multiple reporters of each kind are registered, which
  // leads to duplicated reports of individual measurements such as "resident",
  // "vsize", etc.
  //
  // It's unclear how these multiple calls can occur. The only plausible theory
  // so far is badly-written extensions, because this function is callable from
  // JS code via nsIMemoryReporter.idl.
  //
  // Whatever the cause, it's a bad thing. So we protect against it with the
  // following check.
  static bool isInited = false;
  if (isInited) {
    NS_WARNING("nsMemoryReporterManager::Init() has already been called!");
    return NS_OK;
  }
  isInited = true;

  {
    // Add a series of low-level reporters meant to be executed in order and
    // before any other reporters. These reporters are never released until
    // the manager dies (at process shutdown). Note that this currently only
    // works for reporters expecting to be executed sync.
    //
    // Note that we explicitly handle our self-reporting inside
    // GetReportsForThisProcessExtended, such that we do not need to register
    // ourself to any array/table here.

    mozilla::MutexAutoLock autoLock(mMutex);

#ifdef HAVE_JEMALLOC_STATS
    mStrongEternalReporters->AppendElement(new JemallocHeapReporter());
#endif

#ifdef HAVE_VSIZE_AND_RESIDENT_REPORTERS
    mStrongEternalReporters->AppendElement(new VsizeReporter());
    mStrongEternalReporters->AppendElement(new ResidentReporter());
#endif

#ifdef HAVE_VSIZE_MAX_CONTIGUOUS_REPORTER
    mStrongEternalReporters->AppendElement(new VsizeMaxContiguousReporter());
#endif

#ifdef HAVE_RESIDENT_PEAK_REPORTER
    mStrongEternalReporters->AppendElement(new ResidentPeakReporter());
#endif

#ifdef HAVE_RESIDENT_UNIQUE_REPORTER
    mStrongEternalReporters->AppendElement(new ResidentUniqueReporter());
#endif

#ifdef HAVE_PAGE_FAULT_REPORTERS
    mStrongEternalReporters->AppendElement(new PageFaultsSoftReporter());
    mStrongEternalReporters->AppendElement(new PageFaultsHardReporter());
#endif

#ifdef HAVE_PRIVATE_REPORTER
    mStrongEternalReporters->AppendElement(new PrivateReporter());
#endif

#ifdef HAVE_SYSTEM_HEAP_REPORTER
    mStrongEternalReporters->AppendElement(new SystemHeapReporter());
#endif

    mStrongEternalReporters->AppendElement(new AtomTablesReporter());

    mStrongEternalReporters->AppendElement(new ThreadsReporter());

#ifdef DEBUG
    mStrongEternalReporters->AppendElement(new DeadlockDetectorReporter());
#endif

#ifdef MOZ_GECKO_PROFILER
    // We have to register this here rather than in profiler_init() because
    // profiler_init() runs prior to nsMemoryReporterManager's creation.
    mStrongEternalReporters->AppendElement(new GeckoProfilerReporter());
#endif

#ifdef MOZ_DMD
    mStrongEternalReporters->AppendElement(new mozilla::dmd::DMDReporter());
#endif

#ifdef XP_WIN
    mStrongEternalReporters->AppendElement(new WindowsAddressSpaceReporter());
#endif

#ifdef MOZ_WIDGET_ANDROID
    mStrongEternalReporters->AppendElement(new AndroidMemoryReporter());
#endif
  }  // autoLock(mMutex);

#ifdef XP_UNIX
  nsMemoryInfoDumper::Initialize();
#endif

  return NS_OK;
}

nsMemoryReporterManager::nsMemoryReporterManager()
    : mMutex("nsMemoryReporterManager::mMutex"),
      mIsRegistrationBlocked(false),
      mStrongEternalReporters(new StrongReportersArray()),
      mStrongReporters(new StrongReportersTable()),
      mWeakReporters(new WeakReportersTable()),
      mSavedStrongEternalReporters(nullptr),
      mSavedStrongReporters(nullptr),
      mSavedWeakReporters(nullptr),
      mNextGeneration(1),
      mPendingProcessesState(nullptr),
      mPendingReportersState(nullptr)
#ifdef HAVE_JEMALLOC_STATS
      ,
      mThreadPool(do_GetService(NS_STREAMTRANSPORTSERVICE_CONTRACTID))
#endif
{
}

nsMemoryReporterManager::~nsMemoryReporterManager() {
  NS_ASSERTION(!mSavedStrongEternalReporters,
               "failed to restore eternal reporters");
  NS_ASSERTION(!mSavedStrongReporters, "failed to restore strong reporters");
  NS_ASSERTION(!mSavedWeakReporters, "failed to restore weak reporters");
}

NS_IMETHODIMP
nsMemoryReporterManager::CollectReports(nsIHandleReportCallback* aHandleReport,
                                        nsISupports* aData, bool aAnonymize) {
  size_t n = MallocSizeOf(this);
  {
    mozilla::MutexAutoLock autoLock(mMutex);
    n += mStrongEternalReporters->ShallowSizeOfIncludingThis(MallocSizeOf);
    n += mStrongReporters->ShallowSizeOfIncludingThis(MallocSizeOf);
    n += mWeakReporters->ShallowSizeOfIncludingThis(MallocSizeOf);
    // Note that we do not include the mSaved<X>Reporters here, as during
    // normal operations they are always nullptr and during testing we want to
    // hide the saved variants entirely.
  }

  MOZ_COLLECT_REPORT("explicit/memory-reporter-manager", KIND_HEAP, UNITS_BYTES,
                     n, "Memory used by the memory reporter infrastructure.");

  return NS_OK;
}

#ifdef DEBUG_CHILD_PROCESS_MEMORY_REPORTING
#  define MEMORY_REPORTING_LOG(format, ...) \
    printf_stderr("++++ MEMORY REPORTING: " format, ##__VA_ARGS__);
#else
#  define MEMORY_REPORTING_LOG(...)
#endif

NS_IMETHODIMP
nsMemoryReporterManager::GetReports(
    nsIHandleReportCallback* aHandleReport, nsISupports* aHandleReportData,
    nsIFinishReportingCallback* aFinishReporting,
    nsISupports* aFinishReportingData, bool aAnonymize) {
  return GetReportsExtended(aHandleReport, aHandleReportData, aFinishReporting,
                            aFinishReportingData, aAnonymize,
                            /* minimize = */ false,
                            /* DMDident = */ u""_ns);
}

NS_IMETHODIMP
nsMemoryReporterManager::GetReportsExtended(
    nsIHandleReportCallback* aHandleReport, nsISupports* aHandleReportData,
    nsIFinishReportingCallback* aFinishReporting,
    nsISupports* aFinishReportingData, bool aAnonymize, bool aMinimize,
    const nsAString& aDMDDumpIdent) {
  nsresult rv;

  // Memory reporters are not necessarily threadsafe, so this function must
  // be called from the main thread.
  if (!NS_IsMainThread()) {
    MOZ_CRASH();
  }

  uint32_t generation = mNextGeneration++;

  if (mPendingProcessesState) {
    // A request is in flight.  Don't start another one.  And don't report
    // an error;  just ignore it, and let the in-flight request finish.
    MEMORY_REPORTING_LOG("GetReports (gen=%u, s->gen=%u): abort\n", generation,
                         mPendingProcessesState->mGeneration);
    return NS_OK;
  }

  MEMORY_REPORTING_LOG("GetReports (gen=%u)\n", generation);

  uint32_t concurrency = Preferences::GetUint("memory.report_concurrency", 1);
  MOZ_ASSERT(concurrency >= 1);
  if (concurrency < 1) {
    concurrency = 1;
  }
  mPendingProcessesState = new PendingProcessesState(
      generation, aAnonymize, aMinimize, concurrency, aHandleReport,
      aHandleReportData, aFinishReporting, aFinishReportingData, aDMDDumpIdent);

  if (aMinimize) {
    nsCOMPtr<nsIRunnable> callback =
        NewRunnableMethod("nsMemoryReporterManager::StartGettingReports", this,
                          &nsMemoryReporterManager::StartGettingReports);
    rv = MinimizeMemoryUsage(callback);
  } else {
    rv = StartGettingReports();
  }
  return rv;
}

// MainThread only
nsresult nsMemoryReporterManager::StartGettingReports() {
  PendingProcessesState* s = mPendingProcessesState;
  nsresult rv;

  // Get reports for this process.
  FILE* parentDMDFile = nullptr;
#ifdef MOZ_DMD
  if (!s->mDMDDumpIdent.IsEmpty()) {
    rv = nsMemoryInfoDumper::OpenDMDFile(s->mDMDDumpIdent, getpid(),
                                         &parentDMDFile);
    if (NS_WARN_IF(NS_FAILED(rv))) {
      // Proceed with the memory report as if DMD were disabled.
      parentDMDFile = nullptr;
    }
  }
#endif

  // This is async.
  GetReportsForThisProcessExtended(
      s->mHandleReport, s->mHandleReportData, s->mAnonymize, parentDMDFile,
      s->mFinishReporting, s->mFinishReportingData);

  nsTArray<dom::ContentParent*> childWeakRefs;
  dom::ContentParent::GetAll(childWeakRefs);
  if (!childWeakRefs.IsEmpty()) {
    // Request memory reports from child processes.  This happens
    // after the parent report so that the parent's main thread will
    // be free to process the child reports, instead of causing them
    // to be buffered and consume (possibly scarce) memory.

    for (size_t i = 0; i < childWeakRefs.Length(); ++i) {
      s->mChildrenPending.AppendElement(childWeakRefs[i]);
    }
  }

  if (gfx::GPUProcessManager* gpu = gfx::GPUProcessManager::Get()) {
    if (RefPtr<MemoryReportingProcess> proc = gpu->GetProcessMemoryReporter()) {
      s->mChildrenPending.AppendElement(proc.forget());
    }
  }

  if (RDDProcessManager* rdd = RDDProcessManager::Get()) {
    if (RefPtr<MemoryReportingProcess> proc = rdd->GetProcessMemoryReporter()) {
--> --------------------

--> maximum size reached

--> --------------------

quality95%

¤ Dauer der Verarbeitung: 0.11 Sekunden (vorverarbeitet) ¤

Wurzel

Suchen

Beweissystem der NASA

Beweissystem Isabelle

NIST Cobol Testsuite

Cephes Mathematical Library

Wiener Entwicklungsmethode

Haftungshinweis

Die Informationen auf dieser Webseite wurden nach bestem Wissen sorgfältig zusammengestellt. Es wird jedoch weder Vollständigkeit, noch Richtigkeit, noch Qualität der bereit gestellten Informationen zugesichert.

Bemerkung:

Die farbliche Syntaxdarstellung ist noch experimentell.