From 9224ce3d4d1ea0428a70e75134998e08aa45b1e6 Mon Sep 17 00:00:00 2001 From: David Kalnischkies Date: Mon, 30 Mar 2015 20:47:13 +0200 Subject: [PATCH] calculate only expected hashes in methods Methods get told which hashes are expected by the acquire system, which means we can use this list to restrict what we calculate in the methods as any extra we are calculating is wasted effort as we can't compare it with anything anyway. Adding support for a new hash algorithm is therefore 'free' now and if a algorithm is no longer provided in a repository for a file, we automatically stop calculating it. In practice this results in a speed-up in Debian as we don't have SHA512 here (so far), so we practically stop calculating it. --- apt-pkg/contrib/hashes.cc | 63 +++++++++++++++++++++++++++--------- apt-pkg/contrib/hashes.h | 21 +++++++++--- ftparchive/cachedb.cc | 4 +-- ftparchive/writer.cc | 4 +-- methods/cdrom.cc | 2 +- methods/copy.cc | 10 +++--- methods/file.cc | 2 +- methods/ftp.cc | 2 +- methods/gzip.cc | 2 +- methods/http.cc | 14 ++++---- methods/http.h | 2 +- methods/https.cc | 2 +- methods/https.h | 2 +- methods/rred.cc | 2 +- methods/rsh.cc | 2 +- methods/server.cc | 2 +- methods/server.h | 2 +- test/libapt/hashsums_test.cc | 24 ++++++++++++++ 18 files changed, 116 insertions(+), 46 deletions(-) diff --git a/apt-pkg/contrib/hashes.cc b/apt-pkg/contrib/hashes.cc index 6e7080bc9..953465091 100644 --- a/apt-pkg/contrib/hashes.cc +++ b/apt-pkg/contrib/hashes.cc @@ -250,28 +250,34 @@ bool HashStringList::operator!=(HashStringList const &other) const class PrivateHashes { public: unsigned long long FileSize; + unsigned int CalcHashes; - PrivateHashes() : FileSize(0) {} + PrivateHashes(unsigned int const CalcHashes) : FileSize(0), CalcHashes(CalcHashes) {} }; /*}}}*/ // Hashes::Add* - Add the contents of data or FD /*{{{*/ -bool Hashes::Add(const unsigned char * const Data,unsigned long long const Size, unsigned int const Hashes) +bool Hashes::Add(const unsigned char * const Data, unsigned long long const Size) { bool Res = true; APT_IGNORE_DEPRECATED_PUSH - if ((Hashes & MD5SUM) == MD5SUM) + if ((d->CalcHashes & MD5SUM) == MD5SUM) Res &= MD5.Add(Data, Size); - if ((Hashes & SHA1SUM) == SHA1SUM) + if ((d->CalcHashes & SHA1SUM) == SHA1SUM) Res &= SHA1.Add(Data, Size); - if ((Hashes & SHA256SUM) == SHA256SUM) + if ((d->CalcHashes & SHA256SUM) == SHA256SUM) Res &= SHA256.Add(Data, Size); - if ((Hashes & SHA512SUM) == SHA512SUM) + if ((d->CalcHashes & SHA512SUM) == SHA512SUM) Res &= SHA512.Add(Data, Size); APT_IGNORE_DEPRECATED_POP d->FileSize += Size; return Res; } -bool Hashes::AddFD(int const Fd,unsigned long long Size, unsigned int const Hashes) +bool Hashes::Add(const unsigned char * const Data, unsigned long long const Size, unsigned int const Hashes) +{ + d->CalcHashes = Hashes; + return Add(Data, Size); +} +bool Hashes::AddFD(int const Fd,unsigned long long Size) { unsigned char Buf[64*64]; bool const ToEOF = (Size == UntilEOF); @@ -285,12 +291,17 @@ bool Hashes::AddFD(int const Fd,unsigned long long Size, unsigned int const Hash if (ToEOF && Res == 0) // EOF break; Size -= Res; - if (Add(Buf, Res, Hashes) == false) + if (Add(Buf, Res) == false) return false; } return true; } -bool Hashes::AddFD(FileFd &Fd,unsigned long long Size, unsigned int const Hashes) +bool Hashes::AddFD(int const Fd,unsigned long long Size, unsigned int const Hashes) +{ + d->CalcHashes = Hashes; + return AddFD(Fd, Size); +} +bool Hashes::AddFD(FileFd &Fd,unsigned long long Size) { unsigned char Buf[64*64]; bool const ToEOF = (Size == 0); @@ -309,20 +320,29 @@ bool Hashes::AddFD(FileFd &Fd,unsigned long long Size, unsigned int const Hashes else if (a == 0) // EOF break; Size -= a; - if (Add(Buf, a, Hashes) == false) + if (Add(Buf, a) == false) return false; } return true; +} +bool Hashes::AddFD(FileFd &Fd,unsigned long long Size, unsigned int const Hashes) +{ + d->CalcHashes = Hashes; + return AddFD(Fd, Size); } /*}}}*/ HashStringList Hashes::GetHashStringList() { HashStringList hashes; APT_IGNORE_DEPRECATED_PUSH - hashes.push_back(HashString("MD5Sum", MD5.Result().Value())); - hashes.push_back(HashString("SHA1", SHA1.Result().Value())); - hashes.push_back(HashString("SHA256", SHA256.Result().Value())); - hashes.push_back(HashString("SHA512", SHA512.Result().Value())); + if ((d->CalcHashes & MD5SUM) == MD5SUM) + hashes.push_back(HashString("MD5Sum", MD5.Result().Value())); + if ((d->CalcHashes & SHA1SUM) == SHA1SUM) + hashes.push_back(HashString("SHA1", SHA1.Result().Value())); + if ((d->CalcHashes & SHA256SUM) == SHA256SUM) + hashes.push_back(HashString("SHA256", SHA256.Result().Value())); + if ((d->CalcHashes & SHA512SUM) == SHA512SUM) + hashes.push_back(HashString("SHA512", SHA512.Result().Value())); APT_IGNORE_DEPRECATED_POP std::string SizeStr; strprintf(SizeStr, "%llu", d->FileSize); @@ -330,6 +350,19 @@ APT_IGNORE_DEPRECATED_POP return hashes; } APT_IGNORE_DEPRECATED_PUSH -Hashes::Hashes() { d = new PrivateHashes(); } +Hashes::Hashes() { d = new PrivateHashes(~0); } +Hashes::Hashes(unsigned int const Hashes) { d = new PrivateHashes(Hashes); } +Hashes::Hashes(HashStringList const &Hashes) { + unsigned int calcHashes = Hashes.usable() ? 0 : ~0; + if (Hashes.find("MD5Sum") != NULL) + calcHashes |= MD5SUM; + if (Hashes.find("SHA1") != NULL) + calcHashes |= SHA1SUM; + if (Hashes.find("SHA256") != NULL) + calcHashes |= SHA256SUM; + if (Hashes.find("SHA512") != NULL) + calcHashes |= SHA512SUM; + d = new PrivateHashes(calcHashes); +} Hashes::~Hashes() { delete d; } APT_IGNORE_DEPRECATED_POP diff --git a/apt-pkg/contrib/hashes.h b/apt-pkg/contrib/hashes.h index 154862457..ac13c8ace 100644 --- a/apt-pkg/contrib/hashes.h +++ b/apt-pkg/contrib/hashes.h @@ -178,7 +178,8 @@ class Hashes static const int UntilEOF = 0; - bool Add(const unsigned char * const Data, unsigned long long const Size, unsigned int const Hashes = ~0); + bool Add(const unsigned char * const Data, unsigned long long const Size); + APT_DEPRECATED bool Add(const unsigned char * const Data, unsigned long long const Size, unsigned int const Hashes); inline bool Add(const char * const Data) {return Add((unsigned char const * const)Data,strlen(Data));}; inline bool Add(const unsigned char * const Beg,const unsigned char * const End) @@ -186,13 +187,24 @@ class Hashes enum SupportedHashes { MD5SUM = (1 << 0), SHA1SUM = (1 << 1), SHA256SUM = (1 << 2), SHA512SUM = (1 << 3) }; - bool AddFD(int const Fd,unsigned long long Size = 0, unsigned int const Hashes = ~0); - bool AddFD(FileFd &Fd,unsigned long long Size = 0, unsigned int const Hashes = ~0); + bool AddFD(int const Fd,unsigned long long Size = 0); + APT_DEPRECATED bool AddFD(int const Fd,unsigned long long Size, unsigned int const Hashes); + bool AddFD(FileFd &Fd,unsigned long long Size = 0); + APT_DEPRECATED bool AddFD(FileFd &Fd,unsigned long long Size, unsigned int const Hashes); HashStringList GetHashStringList(); APT_IGNORE_DEPRECATED_PUSH + /** create a Hashes object to calculate all supported hashes + * + * If ALL is too much, you can limit which Hashes are calculated + * with the following other constructors which mention explicitly + * which hashes to generate. */ Hashes(); + /** @param Hashes bitflag composed of #SupportedHashes */ + Hashes(unsigned int const Hashes); + /** @param Hashes is a list of hashes */ + Hashes(HashStringList const &Hashes); virtual ~Hashes(); APT_IGNORE_DEPRECATED_POP @@ -208,15 +220,16 @@ APT_IGNORE_DEPRECATED_POP } public: +APT_IGNORE_DEPRECATED_PUSH APT_DEPRECATED bool AddFD(int const Fd, unsigned long long Size, bool const addMD5, bool const addSHA1, bool const addSHA256, bool const addSHA512) { return AddFD(Fd, Size, boolsToFlag(addMD5, addSHA1, addSHA256, addSHA512)); }; - APT_DEPRECATED bool AddFD(FileFd &Fd, unsigned long long Size, bool const addMD5, bool const addSHA1, bool const addSHA256, bool const addSHA512) { return AddFD(Fd, Size, boolsToFlag(addMD5, addSHA1, addSHA256, addSHA512)); }; +APT_IGNORE_DEPRECATED_POP }; #endif diff --git a/ftparchive/cachedb.cc b/ftparchive/cachedb.cc index 1dc268594..cc3527ea4 100644 --- a/ftparchive/cachedb.cc +++ b/ftparchive/cachedb.cc @@ -441,8 +441,8 @@ bool CacheDB::GetHashes(bool const GenOnly, unsigned int const DoHashes) if (OpenFile() == false) return false; - Hashes hashes; - if (Fd->Seek(0) == false || hashes.AddFD(*Fd, CurStat.FileSize, FlHashes) == false) + Hashes hashes(FlHashes); + if (Fd->Seek(0) == false || hashes.AddFD(*Fd, CurStat.FileSize) == false) return false; HashStringList hl = hashes.GetHashStringList(); diff --git a/ftparchive/writer.cc b/ftparchive/writer.cc index db68c21f0..593278590 100644 --- a/ftparchive/writer.cc +++ b/ftparchive/writer.cc @@ -1075,8 +1075,8 @@ bool ReleaseWriter::DoPackage(string FileName) CheckSums[NewFileName].size = fd.Size(); - Hashes hs; - hs.AddFD(fd, 0, DoHashes); + Hashes hs(DoHashes); + hs.AddFD(fd); CheckSums[NewFileName].Hashes = hs.GetHashStringList(); fd.Close(); diff --git a/methods/cdrom.cc b/methods/cdrom.cc index 74e2ecc6b..10cb29f66 100644 --- a/methods/cdrom.cc +++ b/methods/cdrom.cc @@ -266,7 +266,7 @@ bool CDROMMethod::Fetch(FetchItem *Itm) Res.LastModified = Buf.st_mtime; Res.Size = Buf.st_size; - Hashes Hash; + Hashes Hash(Itm->ExpectedHashes); FileFd Fd(Res.Filename, FileFd::ReadOnly); Hash.AddFD(Fd); Res.TakeHashes(Hash); diff --git a/methods/copy.cc b/methods/copy.cc index f3cd84215..a8e289df5 100644 --- a/methods/copy.cc +++ b/methods/copy.cc @@ -28,16 +28,16 @@ class CopyMethod : public pkgAcqMethod { virtual bool Fetch(FetchItem *Itm); - void CalculateHashes(FetchResult &Res); + void CalculateHashes(FetchItem const * const Itm, FetchResult &Res); public: CopyMethod() : pkgAcqMethod("1.0",SingleInstance | SendConfig) {}; }; -void CopyMethod::CalculateHashes(FetchResult &Res) +void CopyMethod::CalculateHashes(FetchItem const * const Itm, FetchResult &Res) { - Hashes Hash; + Hashes Hash(Itm->ExpectedHashes); FileFd::CompressMode CompressMode = FileFd::None; if (_config->FindB("Acquire::GzipIndexes", false) == true) CompressMode = FileFd::Extension; @@ -71,7 +71,7 @@ bool CopyMethod::Fetch(FetchItem *Itm) // just calc the hashes if the source and destination are identical if (File == Itm->DestFile) { - CalculateHashes(Res); + CalculateHashes(Itm, Res); URIDone(Res); return true; } @@ -104,7 +104,7 @@ bool CopyMethod::Fetch(FetchItem *Itm) if (utimes(Res.Filename.c_str(), times) != 0) return _error->Errno("utimes",_("Failed to set modification time")); - CalculateHashes(Res); + CalculateHashes(Itm, Res); URIDone(Res); return true; diff --git a/methods/file.cc b/methods/file.cc index 5d9d7b951..043ab04b8 100644 --- a/methods/file.cc +++ b/methods/file.cc @@ -87,7 +87,7 @@ bool FileMethod::Fetch(FetchItem *Itm) if (Res.Filename.empty() == true) return _error->Error(_("File not found")); - Hashes Hash; + Hashes Hash(Itm->ExpectedHashes); FileFd Fd(Res.Filename, FileFd::ReadOnly); Hash.AddFD(Fd); Res.TakeHashes(Hash); diff --git a/methods/ftp.cc b/methods/ftp.cc index 7764acf6a..92d8573f1 100644 --- a/methods/ftp.cc +++ b/methods/ftp.cc @@ -1064,7 +1064,7 @@ bool FtpMethod::Fetch(FetchItem *Itm) } // Open the file - Hashes Hash; + Hashes Hash(Itm->ExpectedHashes); { FileFd Fd(Itm->DestFile,FileFd::WriteAny); if (_error->PendingError() == true) diff --git a/methods/gzip.cc b/methods/gzip.cc index 387c05f2e..65519633c 100644 --- a/methods/gzip.cc +++ b/methods/gzip.cc @@ -91,7 +91,7 @@ bool GzipMethod::Fetch(FetchItem *Itm) return false; // Read data from source, generate checksums and write - Hashes Hash; + Hashes Hash(Itm->ExpectedHashes); bool Failed = false; while (1) { diff --git a/methods/http.cc b/methods/http.cc index 947002cc6..e4773b0e2 100644 --- a/methods/http.cc +++ b/methods/http.cc @@ -64,8 +64,8 @@ const unsigned int CircleBuf::BW_HZ=10; // CircleBuf::CircleBuf - Circular input buffer /*{{{*/ // --------------------------------------------------------------------- /* */ -CircleBuf::CircleBuf(unsigned long long Size) - : Size(Size), Hash(0), TotalWriten(0) +CircleBuf::CircleBuf(unsigned long long Size) + : Size(Size), Hash(NULL), TotalWriten(0) { Buf = new unsigned char[Size]; Reset(); @@ -84,10 +84,10 @@ void CircleBuf::Reset() TotalWriten = 0; MaxGet = (unsigned long long)-1; OutQueue = string(); - if (Hash != 0) + if (Hash != NULL) { delete Hash; - Hash = new Hashes; + Hash = NULL; } } /*}}}*/ @@ -222,7 +222,7 @@ bool CircleBuf::Write(int Fd) TotalWriten += Res; - if (Hash != 0) + if (Hash != NULL) Hash->Add(Buf + (OutP%Size),Res); OutP += Res; @@ -484,10 +484,10 @@ APT_PURE bool HttpServerState::IsOpen() /*{{{*/ return (ServerFd != -1); } /*}}}*/ -bool HttpServerState::InitHashes(FileFd &File) /*{{{*/ +bool HttpServerState::InitHashes(FileFd &File, HashStringList const &ExpectedHashes)/*{{{*/ { delete In.Hash; - In.Hash = new Hashes; + In.Hash = new Hashes(ExpectedHashes); // Set the expected size and read file for the hashes File.Truncate(StartPos); diff --git a/methods/http.h b/methods/http.h index 40a88a7be..6dc872659 100644 --- a/methods/http.h +++ b/methods/http.h @@ -111,7 +111,7 @@ struct HttpServerState: public ServerState virtual bool Open(); virtual bool IsOpen(); virtual bool Close(); - virtual bool InitHashes(FileFd &File); + virtual bool InitHashes(FileFd &File, HashStringList const &ExpectedHashes); virtual Hashes * GetHashes(); virtual bool Die(FileFd &File); virtual bool Flush(FileFd * const File); diff --git a/methods/https.cc b/methods/https.cc index 70f6a1046..81903b239 100644 --- a/methods/https.cc +++ b/methods/https.cc @@ -443,7 +443,7 @@ bool HttpsMethod::Fetch(FetchItem *Itm) Res.LastModified = resultStat.st_mtime; // take hashes - Hashes Hash; + Hashes Hash(Itm->ExpectedHashes); FileFd Fd(Res.Filename, FileFd::ReadOnly); Hash.AddFD(Fd); Res.TakeHashes(Hash); diff --git a/methods/https.h b/methods/https.h index 4cc48fc34..dc0ff3322 100644 --- a/methods/https.h +++ b/methods/https.h @@ -42,7 +42,7 @@ class HttpsServerState : public ServerState virtual bool Open() { return false; } virtual bool IsOpen() { return false; } virtual bool Close() { return false; } - virtual bool InitHashes(FileFd &/*File*/) { return false; } + virtual bool InitHashes(FileFd &/*File*/, HashStringList const &/*ExpectedHashes*/) { return false; } virtual Hashes * GetHashes() { return NULL; } virtual bool Die(FileFd &/*File*/) { return false; } virtual bool Flush(FileFd * const /*File*/) { return false; } diff --git a/methods/rred.cc b/methods/rred.cc index 774b58a40..554ac99b4 100644 --- a/methods/rred.cc +++ b/methods/rred.cc @@ -581,7 +581,7 @@ class RredMethod : public pkgAcqMethod { FILE *inp = fopen(Path.c_str(), "r"); FILE *out = fopen(Itm->DestFile.c_str(), "w"); - Hashes hash; + Hashes hash(Itm->ExpectedHashes); patch.apply_against_file(out, inp, &hash); diff --git a/methods/rsh.cc b/methods/rsh.cc index 0e949160b..52349c61c 100644 --- a/methods/rsh.cc +++ b/methods/rsh.cc @@ -477,7 +477,7 @@ bool RSHMethod::Fetch(FetchItem *Itm) } // Open the file - Hashes Hash; + Hashes Hash(Itm->ExpectedHashes); { FileFd Fd(Itm->DestFile,FileFd::WriteAny); if (_error->PendingError() == true) diff --git a/methods/server.cc b/methods/server.cc index 91ec824d1..e403f1071 100644 --- a/methods/server.cc +++ b/methods/server.cc @@ -357,7 +357,7 @@ ServerMethod::DealWithHeaders(FetchResult &Res) FailFd = File->Fd(); FailTime = Server->Date; - if (Server->InitHashes(*File) == false) + if (Server->InitHashes(*File, Queue->ExpectedHashes) == false) { _error->Errno("read",_("Problem hashing file")); return ERROR_NOT_FROM_SERVER; diff --git a/methods/server.h b/methods/server.h index b974ec89a..45622dd34 100644 --- a/methods/server.h +++ b/methods/server.h @@ -85,7 +85,7 @@ struct ServerState virtual bool Open() = 0; virtual bool IsOpen() = 0; virtual bool Close() = 0; - virtual bool InitHashes(FileFd &File) = 0; + virtual bool InitHashes(FileFd &File, HashStringList const &ExpectedHashes) = 0; virtual Hashes * GetHashes() = 0; virtual bool Die(FileFd &File) = 0; virtual bool Flush(FileFd * const File) = 0; diff --git a/test/libapt/hashsums_test.cc b/test/libapt/hashsums_test.cc index a19a0befd..edcd8a11a 100644 --- a/test/libapt/hashsums_test.cc +++ b/test/libapt/hashsums_test.cc @@ -193,6 +193,30 @@ TEST(HashSumsTest, FileBased) EXPECT_EQ(FileSize, list.find("Checksum-FileSize")->HashValue()); } fd.Seek(0); + { + Hashes hashes(Hashes::MD5SUM | Hashes::SHA512SUM); + hashes.AddFD(fd); + HashStringList list = hashes.GetHashStringList(); + EXPECT_FALSE(list.empty()); + EXPECT_EQ(3, list.size()); + EXPECT_EQ(md5.Value(), list.find("MD5Sum")->HashValue()); + EXPECT_EQ(NULL, list.find("SHA1")); + EXPECT_EQ(NULL, list.find("SHA256")); + EXPECT_EQ(sha512.Value(), list.find("SHA512")->HashValue()); + EXPECT_EQ(FileSize, list.find("Checksum-FileSize")->HashValue()); + fd.Seek(0); + Hashes hashes2(list); + hashes2.AddFD(fd); + list = hashes2.GetHashStringList(); + EXPECT_FALSE(list.empty()); + EXPECT_EQ(3, list.size()); + EXPECT_EQ(md5.Value(), list.find("MD5Sum")->HashValue()); + EXPECT_EQ(NULL, list.find("SHA1")); + EXPECT_EQ(NULL, list.find("SHA256")); + EXPECT_EQ(sha512.Value(), list.find("SHA512")->HashValue()); + EXPECT_EQ(FileSize, list.find("Checksum-FileSize")->HashValue()); + } + fd.Seek(0); { MD5Summation MD5; MD5.AddFD(fd.Fd()); -- 2.45.2