Optimize utils and file parsing to prevent timeouts in exotic files (#…

…772) * utils: optimize replaceNonprintableChars() It turns out it can be used quite often in fileformat. * utils/system: replace custom sleep() with std::this_thread::sleep_for() * utils/test.h: remove, not used anywhere * utils/file_i: break long lines * utils/dynamic_buffer: break long lines * utils/const: remove, not used anywhere * utils/alignment: break long lines * utils/byte_value_storage: break long lines * utils/conversion: replace toString(0 with std::to_string() * utils/conversion: remove unused functions * utils/conversion: reduce usage of numToStr() * progress * utils/conversion: refactor hex-string conversions * fileformat: optimize PeFormat::scanForSectionAnomalies() * pelib/ImportDirectory: add guard against infinite cycle * capstone2llvmirtool: fix compilation, add std:: namespace
avast · May 23, 2020 · 64769c5 · 64769c5
1 parent 25cffa4
commit 64769c5
Show file tree

Hide file tree

Showing 95 changed files with 1,071 additions and 1,109 deletions.
diff --git a/include/retdec/fileformat/file_format/pe/pe_format.h b/include/retdec/fileformat/file_format/pe/pe_format.h
@@ -116,7 +116,7 @@ class PeFormat : public FileFormat
 		/// @}
 		/// @name Auxiliary scanning methods
 		/// @{
-		void scanForSectionAnomalies();
+		void scanForSectionAnomalies(unsigned anamaliesLimit = 1000);
 		void scanForResourceAnomalies();
 		void scanForImportAnomalies();
 		void scanForExportAnomalies();

diff --git a/include/retdec/pelib/ImportDirectory.h b/include/retdec/pelib/ImportDirectory.h
@@ -13,6 +13,8 @@
 #ifndef IMPORTDIRECTORY_H
 #define IMPORTDIRECTORY_H
 
+#include <set>
+
 #include "retdec/pelib/PeLibAux.h"
 #include "retdec/pelib/PeHeader.h"
 
@@ -699,6 +701,7 @@ namespace PeLib
 		}
 
 		// FirstThunk - IAT
+		std::set<dword> seenOffsets;
 		for (unsigned int i=0;i<vOldIidCurr.size();i++)
 		{
 			bool hasValidIlt = hasValidOriginalFirstThunk(vOldIidCurr[i].impdesc, peHeader);
@@ -708,7 +711,11 @@ namespace PeLib
 			{
 				return ERROR_INVALID_FILE;
 			}
-
+			if (seenOffsets.count(uiVaoft))
+			{
+				continue;
+			}
+			seenOffsets.insert(uiVaoft);
 			PELIB_THUNK_DATA<bits> tdCurr;
 
 			inStream_w.clear();

diff --git a/include/retdec/utils/alignment.h b/include/retdec/utils/alignment.h
@@ -12,7 +12,11 @@
 namespace retdec {
 namespace utils {
 
-bool isAligned(std::uint64_t value, std::uint64_t alignment, std::uint64_t& remainder);
+bool isAligned(
+		std::uint64_t value,
+		std::uint64_t alignment,
+		std::uint64_t& remainder);
+
 std::uint64_t alignDown(std::uint64_t value, std::uint64_t alignment);
 std::uint64_t alignUp(std::uint64_t value, std::uint64_t alignment);
 

diff --git a/include/retdec/utils/byte_value_storage.h b/include/retdec/utils/byte_value_storage.h
@@ -38,11 +38,24 @@ class ByteValueStorage
 	virtual std::size_t getNumberOfNibblesInByte() const = 0;
 	virtual bool hasMixedEndianForDouble() const = 0;
 
-	virtual bool getXByte(std::uint64_t address, std::uint64_t x, std::uint64_t& res, Endianness e = Endianness::UNKNOWN) const = 0;
-	virtual bool getXBytes(std::uint64_t address, std::uint64_t x, std::vector<std::uint8_t>& res) const = 0;
-
-	virtual bool setXByte(std::uint64_t address, std::uint64_t x, std::uint64_t val, Endianness e = Endianness::UNKNOWN) = 0;
-	virtual bool setXBytes(std::uint64_t address, const std::vector<std::uint8_t>& val) = 0;
+	virtual bool getXByte(
+			std::uint64_t address,
+			std::uint64_t x,
+			std::uint64_t& res,
+			Endianness e = Endianness::UNKNOWN) const = 0;
+	virtual bool getXBytes(
+			std::uint64_t address,
+			std::uint64_t x,
+			std::vector<std::uint8_t>& res) const = 0;
+
+	virtual bool setXByte(
+			std::uint64_t address,
+			std::uint64_t x,
+			std::uint64_t val,
+			Endianness e = Endianness::UNKNOWN) = 0;
+	virtual bool setXBytes(
+			std::uint64_t address,
+			const std::vector<std::uint8_t>& val) = 0;
 
 	Endianness getInverseEndianness() const;
 	bool isLittleEndian() const;
@@ -56,51 +69,168 @@ class ByteValueStorage
 	bool bitsToBig(std::vector<unsigned char>& values) const;
 	bool bitsToLittle(std::vector<unsigned char>& values) const;
 
-	bool get1Byte(std::uint64_t address, std::uint64_t& res, Endianness e = Endianness::UNKNOWN) const;
-	bool get2Byte(std::uint64_t address, std::uint64_t& res, Endianness e = Endianness::UNKNOWN) const;
-	bool get4Byte(std::uint64_t address, std::uint64_t& res, Endianness e = Endianness::UNKNOWN) const;
-	bool get8Byte(std::uint64_t address, std::uint64_t& res, Endianness e = Endianness::UNKNOWN) const;
-	bool get10Byte(std::uint64_t address, long double& res) const;
-	bool getWord(std::uint64_t address, std::uint64_t& res, Endianness e = Endianness::UNKNOWN) const;
-	bool getFloat(std::uint64_t address, float& res) const;
-	bool getDouble(std::uint64_t address, double& res) const;
-
-	bool set1Byte(std::uint64_t address, std::uint64_t val, Endianness e = Endianness::UNKNOWN);
-	bool set2Byte(std::uint64_t address, std::uint64_t val, Endianness e = Endianness::UNKNOWN);
-	bool set4Byte(std::uint64_t address, std::uint64_t val, Endianness e = Endianness::UNKNOWN);
-	bool set8Byte(std::uint64_t address, std::uint64_t val, Endianness e = Endianness::UNKNOWN);
-	bool set10Byte(std::uint64_t address, long double val);
-	bool setWord(std::uint64_t address, std::uint64_t val, Endianness e = Endianness::UNKNOWN);
-	bool setFloat(std::uint64_t address, float val);
-	bool setDouble(std::uint64_t address, double val);
-
-	bool getXByteArray(std::uint64_t address, std::uint64_t x, std::vector<std::uint64_t>& res, std::size_t size, Endianness e = Endianness::UNKNOWN) const;
-	bool get1ByteArray(std::uint64_t address, std::vector<std::uint64_t>& res, std::size_t size, Endianness e = Endianness::UNKNOWN) const;
-	bool get2ByteArray(std::uint64_t address, std::vector<std::uint64_t>& res, std::size_t size, Endianness e = Endianness::UNKNOWN) const;
-	bool get4ByteArray(std::uint64_t address, std::vector<std::uint64_t>& res, std::size_t size, Endianness e = Endianness::UNKNOWN) const;
-	bool get8ByteArray(std::uint64_t address, std::vector<std::uint64_t>& res, std::size_t size, Endianness e = Endianness::UNKNOWN) const;
-	bool get10ByteArray(std::uint64_t address, std::vector<long double>& res, std::size_t size) const;
-	bool getWordArray(std::uint64_t address, std::vector<std::uint64_t>& res, std::size_t, Endianness e = Endianness::UNKNOWN) const;
-	bool getFloatArray(std::uint64_t address, std::vector<float>& res, std::size_t size) const;
-	bool getDoubleArray(std::uint64_t address, std::vector<double>& res, std::size_t size) const;
-
-	bool getNTBS(std::uint64_t address, std::string& res, std::size_t size = 0) const;
-	bool getNTWS(std::uint64_t address, std::size_t width, std::vector<std::uint64_t>& res) const;
-	bool getNTWSNice(std::uint64_t address, std::size_t width, std::vector<std::uint64_t>& res) const;
+	bool get1Byte(
+			std::uint64_t address,
+			std::uint64_t& res,
+			Endianness e = Endianness::UNKNOWN) const;
+	bool get2Byte(
+			std::uint64_t address,
+			std::uint64_t& res,
+			Endianness e = Endianness::UNKNOWN) const;
+	bool get4Byte(
+			std::uint64_t address,
+			std::uint64_t& res,
+			Endianness e = Endianness::UNKNOWN) const;
+	bool get8Byte(
+			std::uint64_t address,
+			std::uint64_t& res,
+			Endianness e = Endianness::UNKNOWN) const;
+	bool get10Byte(
+		std::uint64_t address,
+		long double& res) const;
+	bool getWord(
+			std::uint64_t address,
+			std::uint64_t& res,
+			Endianness e = Endianness::UNKNOWN) const;
+	bool getFloat(
+			std::uint64_t address,
+			float& res) const;
+	bool getDouble(
+			std::uint64_t address,
+			double& res) const;
+
+	bool set1Byte(
+			std::uint64_t address,
+			std::uint64_t val,
+			Endianness e = Endianness::UNKNOWN);
+	bool set2Byte(
+			std::uint64_t address,
+			std::uint64_t val,
+			Endianness e = Endianness::UNKNOWN);
+	bool set4Byte(
+			std::uint64_t address,
+			std::uint64_t val,
+			Endianness e = Endianness::UNKNOWN);
+	bool set8Byte(
+			std::uint64_t address,
+			std::uint64_t val,
+			Endianness e = Endianness::UNKNOWN);
+	bool set10Byte(
+			std::uint64_t address,
+			long double val);
+	bool setWord(
+			std::uint64_t address,
+			std::uint64_t val,
+			Endianness e = Endianness::UNKNOWN);
+	bool setFloat(
+			std::uint64_t address,
+			float val);
+	bool setDouble(
+			std::uint64_t address,
+			double val);
+
+	bool getXByteArray(
+			std::uint64_t address,
+			std::uint64_t x,
+			std::vector<std::uint64_t>& res,
+			std::size_t size,
+			Endianness e = Endianness::UNKNOWN) const;
+	bool get1ByteArray(
+			std::uint64_t address,
+			std::vector<std::uint64_t>& res,
+			std::size_t size,
+			Endianness e = Endianness::UNKNOWN) const;
+	bool get2ByteArray(
+			std::uint64_t address,
+			std::vector<std::uint64_t>& res,
+			std::size_t size,
+			Endianness e = Endianness::UNKNOWN) const;
+	bool get4ByteArray(
+			std::uint64_t address,
+			std::vector<std::uint64_t>& res,
+			std::size_t size,
+			Endianness e = Endianness::UNKNOWN) const;
+	bool get8ByteArray(
+			std::uint64_t address,
+			std::vector<std::uint64_t>& res,
+			std::size_t size,
+			Endianness e = Endianness::UNKNOWN) const;
+	bool get10ByteArray(
+			std::uint64_t address,
+			std::vector<long double>& res,
+			std::size_t size) const;
+	bool getWordArray(
+			std::uint64_t address,
+			std::vector<std::uint64_t>& res,
+			std::size_t,
+			Endianness e = Endianness::UNKNOWN) const;
+	bool getFloatArray(
+			std::uint64_t address,
+			std::vector<float>& res, std::size_t size) const;
+	bool getDoubleArray(
+			std::uint64_t address,
+			std::vector<double>& res, std::size_t size) const;
+
+	bool getNTBS(
+			std::uint64_t address,
+			std::string& res,
+			std::size_t size = 0) const;
+	bool getNTWS(
+			std::uint64_t address,
+			std::size_t width,
+			std::vector<std::uint64_t>& res) const;
+	bool getNTWSNice(
+			std::uint64_t address,
+			std::size_t width,
+			std::vector<std::uint64_t>& res) const;
 
 protected:
-	bool createValueFromBytes(const std::vector<std::uint8_t>& data, std::uint64_t& value, Endianness endian, std::uint64_t offset = 0, std::uint64_t size = 0) const;
-	bool createBytesFromValue(std::uint64_t data, std::uint64_t x, std::vector<std::uint8_t>& value, Endianness endian) const;
-
-	bool get10ByteImpl(const std::vector<std::uint8_t>& data, long double& res) const;
-	bool getFloatImpl(const std::vector<std::uint8_t>& data, float& res) const;
-	bool getDoubleImpl(const std::vector<std::uint8_t>& data, double& res) const;
-
-	using GetNByteFn = std::function<bool(std::uint64_t, std::uint64_t&, Endianness)>;
-	using GetXByteFn = std::function<bool(std::uint64_t, std::uint64_t, std::uint64_t&, Endianness)>;
-	bool getNTBSImpl(const GetNByteFn& get1ByteFn, std::uint64_t address, std::string& res, std::size_t size) const;
-	bool getNTWSImpl(const GetXByteFn& getXByteFn, std::uint64_t address, std::size_t width, std::vector<std::uint64_t>& res) const;
-	bool getNTWSNiceImpl(const GetXByteFn& getXByteFn, std::uint64_t address, std::size_t width, std::vector<std::uint64_t>& res) const;
+	bool createValueFromBytes(
+			const std::vector<std::uint8_t>& data,
+			std::uint64_t& value,
+			Endianness endian,
+			std::uint64_t offset = 0,
+			std::uint64_t size = 0) const;
+	bool createBytesFromValue(
+			std::uint64_t data,
+			std::uint64_t x,
+			std::vector<std::uint8_t>& value,
+			Endianness endian) const;
+
+	bool get10ByteImpl(
+			const std::vector<std::uint8_t>& data,
+			long double& res) const;
+	bool getFloatImpl(
+			const std::vector<std::uint8_t>& data,
+			float& res) const;
+	bool getDoubleImpl(
+			const std::vector<std::uint8_t>& data,
+			double& res) const;
+
+	using GetNByteFn = std::function<bool(
+			std::uint64_t,
+			std::uint64_t&,
+			Endianness)>;
+	using GetXByteFn = std::function<bool(
+			std::uint64_t,
+			std::uint64_t,
+			std::uint64_t&, Endianness)>;
+
+	bool getNTBSImpl(
+			const GetNByteFn& get1ByteFn,
+			std::uint64_t address,
+			std::string& res,
+			std::size_t size) const;
+	bool getNTWSImpl(
+			const GetXByteFn& getXByteFn,
+			std::uint64_t address,
+			std::size_t width,
+			std::vector<std::uint64_t>& res) const;
+	bool getNTWSNiceImpl(
+			const GetXByteFn& getXByteFn,
+			std::uint64_t address,
+			std::size_t width,
+			std::vector<std::uint64_t>& res) const;
 };
 
 } // namespace utils

diff --git a/include/retdec/utils/const.h b/include/retdec/utils/const.h