зеркало из https://github.com/mozilla/gecko-dev.git
Bug 1247835 (part 1) - Use binary search instead of a hash table in nsEffectiveTLDService. r=jduell.
This saves 128 KiB of memory per process. It requires putting the eTLD entries in sorted order. --HG-- extra : rebase_source : 96678fcd91089a5151760dcced77e459bcbd4821
This commit is contained in:
Nicholas Nethercote
Родитель
2e0403bcfe
Коммит
9d2a708f22
|
|
@ -30,13 +30,13 @@ NS_IMPL_ISUPPORTS(nsEffectiveTLDService, nsIEffectiveTLDService,
|
|||
#define ETLD_STR_NUM(line) ETLD_STR_NUM_1(line)
|
||||
#define ETLD_ENTRY_OFFSET(name) offsetof(struct etld_string_list, ETLD_STR_NUM(__LINE__))
|
||||
|
||||
const ETLDEntry nsDomainEntry::entries[] = {
|
||||
const ETLDEntry ETLDEntry::entries[] = {
|
||||
#define ETLD_ENTRY(name, ex, wild) { ETLD_ENTRY_OFFSET(name), ex, wild },
|
||||
#include "etld_data.inc"
|
||||
#undef ETLD_ENTRY
|
||||
};
|
||||
|
||||
const union nsDomainEntry::etld_strings nsDomainEntry::strings = {
|
||||
const union ETLDEntry::etld_strings ETLDEntry::strings = {
|
||||
{
|
||||
#define ETLD_ENTRY(name, ex, wild) name,
|
||||
#include "etld_data.inc"
|
||||
|
|
@ -44,10 +44,21 @@ const union nsDomainEntry::etld_strings nsDomainEntry::strings = {
|
|||
}
|
||||
};
|
||||
|
||||
/* static */ const ETLDEntry*
|
||||
ETLDEntry::GetEntry(const char* aDomain)
|
||||
{
|
||||
size_t i;
|
||||
if (BinarySearchIf(entries, 0, ArrayLength(ETLDEntry::entries),
|
||||
Cmp(aDomain), &i)) {
|
||||
return &entries[i];
|
||||
}
|
||||
return nullptr;
|
||||
}
|
||||
|
||||
// Dummy function to statically ensure that our indices don't overflow
|
||||
// the storage provided for them.
|
||||
void
|
||||
nsDomainEntry::FuncForStaticAsserts(void)
|
||||
ETLDEntry::FuncForStaticAsserts(void)
|
||||
{
|
||||
#define ETLD_ENTRY(name, ex, wild) \
|
||||
static_assert(ETLD_ENTRY_OFFSET(name) < (1 << ETLD_ENTRY_N_INDEX_BITS), \
|
||||
|
|
@ -65,33 +76,33 @@ nsDomainEntry::FuncForStaticAsserts(void)
|
|||
static nsEffectiveTLDService *gService = nullptr;
|
||||
|
||||
nsEffectiveTLDService::nsEffectiveTLDService()
|
||||
: mHash(ArrayLength(nsDomainEntry::entries))
|
||||
{
|
||||
}
|
||||
|
||||
nsresult
|
||||
nsEffectiveTLDService::Init()
|
||||
{
|
||||
const ETLDEntry *entries = nsDomainEntry::entries;
|
||||
|
||||
nsresult rv;
|
||||
mIDNService = do_GetService(NS_IDNSERVICE_CONTRACTID, &rv);
|
||||
if (NS_FAILED(rv)) return rv;
|
||||
|
||||
// Initialize eTLD hash from static array
|
||||
for (uint32_t i = 0; i < ArrayLength(nsDomainEntry::entries); i++) {
|
||||
const char *domain = nsDomainEntry::GetEffectiveTLDName(entries[i].strtab_index);
|
||||
#ifdef DEBUG
|
||||
// Sanity-check the eTLD entries.
|
||||
for (uint32_t i = 0; i < ArrayLength(ETLDEntry::entries); i++) {
|
||||
const char* domain = ETLDEntry::entries[i].GetEffectiveTLDName();
|
||||
nsDependentCString name(domain);
|
||||
nsAutoCString normalizedName(domain);
|
||||
NS_ASSERTION(NS_SUCCEEDED(NormalizeHostname(normalizedName)),
|
||||
"normalization failure!");
|
||||
NS_ASSERTION(name.Equals(normalizedName), "domain not normalized!");
|
||||
#endif
|
||||
nsDomainEntry *entry = mHash.PutEntry(domain);
|
||||
NS_ENSURE_TRUE(entry, NS_ERROR_OUT_OF_MEMORY);
|
||||
entry->SetData(entries[i]);
|
||||
MOZ_ASSERT(NS_SUCCEEDED(NormalizeHostname(normalizedName)),
|
||||
"normalization failure!");
|
||||
MOZ_ASSERT(name.Equals(normalizedName), "domain not normalized!");
|
||||
|
||||
// Domains must be in sorted order for binary search to work.
|
||||
if (i > 0) {
|
||||
const char* domain0 = ETLDEntry::entries[i - 1].GetEffectiveTLDName();
|
||||
MOZ_ASSERT(strcmp(domain0, domain) < 0, "domains not in sorted order!");
|
||||
}
|
||||
}
|
||||
#endif
|
||||
|
||||
MOZ_ASSERT(!gService);
|
||||
gService = this;
|
||||
|
|
@ -108,6 +119,10 @@ nsEffectiveTLDService::~nsEffectiveTLDService()
|
|||
|
||||
MOZ_DEFINE_MALLOC_SIZE_OF(EffectiveTLDServiceMallocSizeOf)
|
||||
|
||||
// The amount of heap memory measured here is tiny. It used to be bigger when
|
||||
// nsEffectiveTLDService used a separate hash table instead of binary search.
|
||||
// Nonetheless, we keep this code here in anticipation of bug 1083971 which will
|
||||
// change ETLDEntries::entries to a heap-allocated array modifiable at runtime.
|
||||
NS_IMETHODIMP
|
||||
nsEffectiveTLDService::CollectReports(nsIHandleReportCallback* aHandleReport,
|
||||
nsISupports* aData, bool aAnonymize)
|
||||
|
|
@ -122,7 +137,6 @@ size_t
|
|||
nsEffectiveTLDService::SizeOfIncludingThis(mozilla::MallocSizeOf aMallocSizeOf)
|
||||
{
|
||||
size_t n = aMallocSizeOf(this);
|
||||
n += mHash.ShallowSizeOfExcludingThis(aMallocSizeOf);
|
||||
|
||||
// Measurement of the following members may be added later if DMD finds it is
|
||||
// worthwhile:
|
||||
|
|
@ -263,8 +277,8 @@ nsEffectiveTLDService::GetBaseDomainInternal(nsCString &aHostname,
|
|||
if (*currDomain == '.')
|
||||
return NS_ERROR_INVALID_ARG;
|
||||
|
||||
// perform the hash lookup.
|
||||
nsDomainEntry *entry = mHash.GetEntry(currDomain);
|
||||
// Perform the lookup.
|
||||
const ETLDEntry* entry = ETLDEntry::GetEntry(currDomain);
|
||||
if (entry) {
|
||||
if (entry->IsWild() && prevDomain) {
|
||||
// wildcard rules imply an eTLD one level inferior to the match.
|
||||
|
|
|
|||
|
|
@ -9,85 +9,47 @@
|
|||
#include "nsIEffectiveTLDService.h"
|
||||
|
||||
#include "nsIMemoryReporter.h"
|
||||
#include "nsTHashtable.h"
|
||||
#include "nsString.h"
|
||||
#include "nsCOMPtr.h"
|
||||
#include "mozilla/Attributes.h"
|
||||
#include "mozilla/BinarySearch.h"
|
||||
#include "mozilla/MemoryReporting.h"
|
||||
|
||||
class nsIIDNService;
|
||||
|
||||
#define ETLD_ENTRY_N_INDEX_BITS 30
|
||||
|
||||
// struct for static data generated from effective_tld_names.dat
|
||||
struct ETLDEntry {
|
||||
friend class nsEffectiveTLDService;
|
||||
|
||||
public:
|
||||
bool IsNormal() const { return wild || !exception; }
|
||||
bool IsException() const { return exception; }
|
||||
bool IsWild() const { return wild; }
|
||||
|
||||
const char* GetEffectiveTLDName() const
|
||||
{
|
||||
return strings.strtab + strtab_index;
|
||||
}
|
||||
|
||||
static const ETLDEntry* GetEntry(const char* aDomain);
|
||||
|
||||
static const size_t ETLD_ENTRY_N_INDEX_BITS = 30;
|
||||
|
||||
// These fields must be public to allow static construction.
|
||||
uint32_t strtab_index : ETLD_ENTRY_N_INDEX_BITS;
|
||||
uint32_t exception : 1;
|
||||
uint32_t wild : 1;
|
||||
};
|
||||
|
||||
|
||||
// hash entry class
|
||||
class nsDomainEntry : public PLDHashEntryHdr
|
||||
{
|
||||
friend class nsEffectiveTLDService;
|
||||
public:
|
||||
// Hash methods
|
||||
typedef const char* KeyType;
|
||||
typedef const char* KeyTypePointer;
|
||||
|
||||
explicit nsDomainEntry(KeyTypePointer aEntry)
|
||||
{
|
||||
}
|
||||
|
||||
nsDomainEntry(const nsDomainEntry& toCopy)
|
||||
{
|
||||
// if we end up here, things will break. nsTHashtable shouldn't
|
||||
// allow this, since we set ALLOW_MEMMOVE to true.
|
||||
NS_NOTREACHED("nsDomainEntry copy constructor is forbidden!");
|
||||
}
|
||||
|
||||
~nsDomainEntry()
|
||||
{
|
||||
}
|
||||
|
||||
KeyType GetKey() const
|
||||
{
|
||||
return GetEffectiveTLDName(mData.strtab_index);
|
||||
}
|
||||
|
||||
bool KeyEquals(KeyTypePointer aKey) const
|
||||
{
|
||||
return !strcmp(GetKey(), aKey);
|
||||
}
|
||||
|
||||
static KeyTypePointer KeyToPointer(KeyType aKey)
|
||||
{
|
||||
return aKey;
|
||||
}
|
||||
|
||||
static PLDHashNumber HashKey(KeyTypePointer aKey)
|
||||
{
|
||||
// PLDHashTable::HashStringKey doesn't use the table parameter, so we can
|
||||
// safely pass nullptr
|
||||
return PLDHashTable::HashStringKey(nullptr, aKey);
|
||||
}
|
||||
|
||||
enum { ALLOW_MEMMOVE = true };
|
||||
|
||||
void SetData(ETLDEntry entry) { mData = entry; }
|
||||
|
||||
bool IsNormal() { return mData.wild || !mData.exception; }
|
||||
bool IsException() { return mData.exception; }
|
||||
bool IsWild() { return mData.wild; }
|
||||
|
||||
static const char *GetEffectiveTLDName(size_t idx)
|
||||
{
|
||||
return strings.strtab + idx;
|
||||
}
|
||||
|
||||
private:
|
||||
ETLDEntry mData;
|
||||
struct Cmp {
|
||||
int operator()(const ETLDEntry aEntry) const
|
||||
{
|
||||
return strcmp(mName, aEntry.GetEffectiveTLDName());
|
||||
}
|
||||
explicit Cmp(const char* aName) : mName(aName) {}
|
||||
const char* mName;
|
||||
};
|
||||
|
||||
#define ETLD_STR_NUM_1(line) str##line
|
||||
#define ETLD_STR_NUM(line) ETLD_STR_NUM_1(line)
|
||||
struct etld_string_list {
|
||||
|
|
@ -95,11 +57,17 @@ private:
|
|||
#include "etld_data.inc"
|
||||
#undef ETLD_ENTRY
|
||||
};
|
||||
|
||||
// This static string table is all the eTLD domain names packed together.
|
||||
static const union etld_strings {
|
||||
struct etld_string_list list;
|
||||
char strtab[1];
|
||||
} strings;
|
||||
|
||||
// This is the static entries table. Each entry has an index into the string
|
||||
// table. The entries are in sorted order so that binary search can be used.
|
||||
static const ETLDEntry entries[];
|
||||
|
||||
void FuncForStaticAsserts(void);
|
||||
#undef ETLD_STR_NUM
|
||||
#undef ETLD_STR_NUM1
|
||||
|
|
@ -124,7 +92,6 @@ private:
|
|||
nsresult NormalizeHostname(nsCString &aHostname);
|
||||
~nsEffectiveTLDService();
|
||||
|
||||
nsTHashtable<nsDomainEntry> mHash;
|
||||
nsCOMPtr<nsIIDNService> mIDNService;
|
||||
};
|
||||
|
||||
|
|
|
|||
|
|
@ -18,9 +18,9 @@ http://wiki.mozilla.org/Gecko:Effective_TLD_Service
|
|||
|
||||
def getEffectiveTLDs(path):
|
||||
file = codecs.open(path, "r", "UTF-8")
|
||||
entries = []
|
||||
domains = set()
|
||||
while True:
|
||||
line = file.readline()
|
||||
for line in file:
|
||||
# line always contains a line terminator unless the file is empty
|
||||
if len(line) == 0:
|
||||
raise StopIteration
|
||||
|
|
@ -34,7 +34,12 @@ def getEffectiveTLDs(path):
|
|||
assert domain not in domains, \
|
||||
"repeating domain %s makes no sense" % domain
|
||||
domains.add(domain)
|
||||
yield entry
|
||||
entries.append(entry)
|
||||
|
||||
# Sort the entries so we can use binary search on them.
|
||||
entries.sort(key=EffectiveTLDEntry.domain)
|
||||
|
||||
return entries
|
||||
|
||||
def _normalizeHostname(domain):
|
||||
"""
|
||||
|
|
|
|||
Загрузка…
Ссылка в новой задаче