Async-Interrupt-1.26/0000755000000000000000000000000013651541474013221 5ustar rootrootAsync-Interrupt-1.26/Makefile.PL0000644000000000000000000000111113456573507015172 0ustar rootrootuse ExtUtils::MakeMaker; use Canary::Stability Async::Interrupt => 1, 5.008; # apparently PL_sighandlerp was introduced with 5.008 - correct me if wrong WriteMakefile( dist => { PREOP => 'pod2text Interrupt.pm | tee README >$(DISTVNAME)/README; chmod -R u=rwX,go=rX . ;', COMPRESS => 'gzip -9v', SUFFIX => '.gz', }, PREREQ_PM => { common::sense => 0, }, CONFIGURE_REQUIRES => { ExtUtils::MakeMaker => 6.52, Canary::Stability => 0 }, NAME => "Async::Interrupt", VERSION_FROM => "Interrupt.pm", ); Async-Interrupt-1.26/ecb.h0000644000000000000000000012534313613114072014117 0ustar rootroot/* * libecb - http://software.schmorp.de/pkg/libecb * * Copyright (©) 2009-2015,2018-2020 Marc Alexander Lehmann * Copyright (©) 2011 Emanuele Giaquinta * All rights reserved. * * Redistribution and use in source and binary forms, with or without modifica- * tion, are permitted provided that the following conditions are met: * * 1. Redistributions of source code must retain the above copyright notice, * this list of conditions and the following disclaimer. * * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR IMPLIED * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MER- * CHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO * EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPE- * CIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTH- * ERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED * OF THE POSSIBILITY OF SUCH DAMAGE. * * Alternatively, the contents of this file may be used under the terms of * the GNU General Public License ("GPL") version 2 or any later version, * in which case the provisions of the GPL are applicable instead of * the above. If you wish to allow the use of your version of this file * only under the terms of the GPL and not to allow others to use your * version of this file under the BSD license, indicate your decision * by deleting the provisions above and replace them with the notice * and other provisions required by the GPL. If you do not delete the * provisions above, a recipient may use your version of this file under * either the BSD or the GPL. */ #ifndef ECB_H #define ECB_H /* 16 bits major, 16 bits minor */ #define ECB_VERSION 0x00010008 #include /* for memcpy */ #if defined (_WIN32) && !defined (__MINGW32__) typedef signed char int8_t; typedef unsigned char uint8_t; typedef signed char int_fast8_t; typedef unsigned char uint_fast8_t; typedef signed short int16_t; typedef unsigned short uint16_t; typedef signed int int_fast16_t; typedef unsigned int uint_fast16_t; typedef signed int int32_t; typedef unsigned int uint32_t; typedef signed int int_fast32_t; typedef unsigned int uint_fast32_t; #if __GNUC__ typedef signed long long int64_t; typedef unsigned long long uint64_t; #else /* _MSC_VER || __BORLANDC__ */ typedef signed __int64 int64_t; typedef unsigned __int64 uint64_t; #endif typedef int64_t int_fast64_t; typedef uint64_t uint_fast64_t; #ifdef _WIN64 #define ECB_PTRSIZE 8 typedef uint64_t uintptr_t; typedef int64_t intptr_t; #else #define ECB_PTRSIZE 4 typedef uint32_t uintptr_t; typedef int32_t intptr_t; #endif #else #include #if (defined INTPTR_MAX ? INTPTR_MAX : ULONG_MAX) > 0xffffffffU #define ECB_PTRSIZE 8 #else #define ECB_PTRSIZE 4 #endif #endif #define ECB_GCC_AMD64 (__amd64 || __amd64__ || __x86_64 || __x86_64__) #define ECB_MSVC_AMD64 (_M_AMD64 || _M_X64) #ifndef ECB_OPTIMIZE_SIZE #if __OPTIMIZE_SIZE__ #define ECB_OPTIMIZE_SIZE 1 #else #define ECB_OPTIMIZE_SIZE 0 #endif #endif /* work around x32 idiocy by defining proper macros */ #if ECB_GCC_AMD64 || ECB_MSVC_AMD64 #if _ILP32 #define ECB_AMD64_X32 1 #else #define ECB_AMD64 1 #endif #endif /* many compilers define _GNUC_ to some versions but then only implement * what their idiot authors think are the "more important" extensions, * causing enormous grief in return for some better fake benchmark numbers. * or so. * we try to detect these and simply assume they are not gcc - if they have * an issue with that they should have done it right in the first place. */ #if !defined __GNUC_MINOR__ || defined __INTEL_COMPILER || defined __SUNPRO_C || defined __SUNPRO_CC || defined __llvm__ || defined __clang__ #define ECB_GCC_VERSION(major,minor) 0 #else #define ECB_GCC_VERSION(major,minor) (__GNUC__ > (major) || (__GNUC__ == (major) && __GNUC_MINOR__ >= (minor))) #endif #define ECB_CLANG_VERSION(major,minor) (__clang_major__ > (major) || (__clang_major__ == (major) && __clang_minor__ >= (minor))) #if __clang__ && defined __has_builtin #define ECB_CLANG_BUILTIN(x) __has_builtin (x) #else #define ECB_CLANG_BUILTIN(x) 0 #endif #if __clang__ && defined __has_extension #define ECB_CLANG_EXTENSION(x) __has_extension (x) #else #define ECB_CLANG_EXTENSION(x) 0 #endif #define ECB_CPP (__cplusplus+0) #define ECB_CPP11 (__cplusplus >= 201103L) #define ECB_CPP14 (__cplusplus >= 201402L) #define ECB_CPP17 (__cplusplus >= 201703L) #if ECB_CPP #define ECB_C 0 #define ECB_STDC_VERSION 0 #else #define ECB_C 1 #define ECB_STDC_VERSION __STDC_VERSION__ #endif #define ECB_C99 (ECB_STDC_VERSION >= 199901L) #define ECB_C11 (ECB_STDC_VERSION >= 201112L) #define ECB_C17 (ECB_STDC_VERSION >= 201710L) #if ECB_CPP #define ECB_EXTERN_C extern "C" #define ECB_EXTERN_C_BEG ECB_EXTERN_C { #define ECB_EXTERN_C_END } #else #define ECB_EXTERN_C extern #define ECB_EXTERN_C_BEG #define ECB_EXTERN_C_END #endif /*****************************************************************************/ /* ECB_NO_THREADS - ecb is not used by multiple threads, ever */ /* ECB_NO_SMP - ecb might be used in multiple threads, but only on a single cpu */ #if ECB_NO_THREADS #define ECB_NO_SMP 1 #endif #if ECB_NO_SMP #define ECB_MEMORY_FENCE do { } while (0) #endif /* http://www-01.ibm.com/support/knowledgecenter/SSGH3R_13.1.0/com.ibm.xlcpp131.aix.doc/compiler_ref/compiler_builtins.html */ #if __xlC__ && ECB_CPP #include #endif #if 1400 <= _MSC_VER #include /* fence functions _ReadBarrier, also bit search functions _BitScanReverse */ #endif #ifndef ECB_MEMORY_FENCE #if ECB_GCC_VERSION(2,5) || defined __INTEL_COMPILER || (__llvm__ && __GNUC__) || __SUNPRO_C >= 0x5110 || __SUNPRO_CC >= 0x5110 #define ECB_MEMORY_FENCE_RELAXED __asm__ __volatile__ ("" : : : "memory") #if __i386 || __i386__ #define ECB_MEMORY_FENCE __asm__ __volatile__ ("lock; orb $0, -1(%%esp)" : : : "memory") #define ECB_MEMORY_FENCE_ACQUIRE __asm__ __volatile__ ("" : : : "memory") #define ECB_MEMORY_FENCE_RELEASE __asm__ __volatile__ ("" : : : "memory") #elif ECB_GCC_AMD64 #define ECB_MEMORY_FENCE __asm__ __volatile__ ("mfence" : : : "memory") #define ECB_MEMORY_FENCE_ACQUIRE __asm__ __volatile__ ("" : : : "memory") #define ECB_MEMORY_FENCE_RELEASE __asm__ __volatile__ ("" : : : "memory") #elif __powerpc__ || __ppc__ || __powerpc64__ || __ppc64__ #define ECB_MEMORY_FENCE __asm__ __volatile__ ("sync" : : : "memory") #elif defined __ARM_ARCH_2__ \ || defined __ARM_ARCH_3__ || defined __ARM_ARCH_3M__ \ || defined __ARM_ARCH_4__ || defined __ARM_ARCH_4T__ \ || defined __ARM_ARCH_5__ || defined __ARM_ARCH_5E__ \ || defined __ARM_ARCH_5T__ || defined __ARM_ARCH_5TE__ \ || defined __ARM_ARCH_5TEJ__ /* should not need any, unless running old code on newer cpu - arm doesn't support that */ #elif defined __ARM_ARCH_6__ || defined __ARM_ARCH_6J__ \ || defined __ARM_ARCH_6K__ || defined __ARM_ARCH_6ZK__ \ || defined __ARM_ARCH_6T2__ #define ECB_MEMORY_FENCE __asm__ __volatile__ ("mcr p15,0,%0,c7,c10,5" : : "r" (0) : "memory") #elif defined __ARM_ARCH_7__ || defined __ARM_ARCH_7A__ \ || defined __ARM_ARCH_7R__ || defined __ARM_ARCH_7M__ #define ECB_MEMORY_FENCE __asm__ __volatile__ ("dmb" : : : "memory") #elif __aarch64__ #define ECB_MEMORY_FENCE __asm__ __volatile__ ("dmb ish" : : : "memory") #elif (__sparc || __sparc__) && !(__sparc_v8__ || defined __sparcv8) #define ECB_MEMORY_FENCE __asm__ __volatile__ ("membar #LoadStore | #LoadLoad | #StoreStore | #StoreLoad" : : : "memory") #define ECB_MEMORY_FENCE_ACQUIRE __asm__ __volatile__ ("membar #LoadStore | #LoadLoad" : : : "memory") #define ECB_MEMORY_FENCE_RELEASE __asm__ __volatile__ ("membar #LoadStore | #StoreStore") #elif defined __s390__ || defined __s390x__ #define ECB_MEMORY_FENCE __asm__ __volatile__ ("bcr 15,0" : : : "memory") #elif defined __mips__ /* GNU/Linux emulates sync on mips1 architectures, so we force its use */ /* anybody else who still uses mips1 is supposed to send in their version, with detection code. */ #define ECB_MEMORY_FENCE __asm__ __volatile__ (".set mips2; sync; .set mips0" : : : "memory") #elif defined __alpha__ #define ECB_MEMORY_FENCE __asm__ __volatile__ ("mb" : : : "memory") #elif defined __hppa__ #define ECB_MEMORY_FENCE __asm__ __volatile__ ("" : : : "memory") #define ECB_MEMORY_FENCE_RELEASE __asm__ __volatile__ ("") #elif defined __ia64__ #define ECB_MEMORY_FENCE __asm__ __volatile__ ("mf" : : : "memory") #elif defined __m68k__ #define ECB_MEMORY_FENCE __asm__ __volatile__ ("" : : : "memory") #elif defined __m88k__ #define ECB_MEMORY_FENCE __asm__ __volatile__ ("tb1 0,%%r0,128" : : : "memory") #elif defined __sh__ #define ECB_MEMORY_FENCE __asm__ __volatile__ ("" : : : "memory") #endif #endif #endif #ifndef ECB_MEMORY_FENCE #if ECB_GCC_VERSION(4,7) /* see comment below (stdatomic.h) about the C11 memory model. */ #define ECB_MEMORY_FENCE __atomic_thread_fence (__ATOMIC_SEQ_CST) #define ECB_MEMORY_FENCE_ACQUIRE __atomic_thread_fence (__ATOMIC_ACQUIRE) #define ECB_MEMORY_FENCE_RELEASE __atomic_thread_fence (__ATOMIC_RELEASE) #define ECB_MEMORY_FENCE_RELAXED __atomic_thread_fence (__ATOMIC_RELAXED) #elif ECB_CLANG_EXTENSION(c_atomic) /* see comment below (stdatomic.h) about the C11 memory model. */ #define ECB_MEMORY_FENCE __c11_atomic_thread_fence (__ATOMIC_SEQ_CST) #define ECB_MEMORY_FENCE_ACQUIRE __c11_atomic_thread_fence (__ATOMIC_ACQUIRE) #define ECB_MEMORY_FENCE_RELEASE __c11_atomic_thread_fence (__ATOMIC_RELEASE) #define ECB_MEMORY_FENCE_RELAXED __c11_atomic_thread_fence (__ATOMIC_RELAXED) #elif ECB_GCC_VERSION(4,4) || defined __INTEL_COMPILER || defined __clang__ #define ECB_MEMORY_FENCE __sync_synchronize () #elif _MSC_VER >= 1500 /* VC++ 2008 */ /* apparently, microsoft broke all the memory barrier stuff in Visual Studio 2008... */ #pragma intrinsic(_ReadBarrier,_WriteBarrier,_ReadWriteBarrier) #define ECB_MEMORY_FENCE _ReadWriteBarrier (); MemoryBarrier() #define ECB_MEMORY_FENCE_ACQUIRE _ReadWriteBarrier (); MemoryBarrier() /* according to msdn, _ReadBarrier is not a load fence */ #define ECB_MEMORY_FENCE_RELEASE _WriteBarrier (); MemoryBarrier() #elif _MSC_VER >= 1400 /* VC++ 2005 */ #pragma intrinsic(_ReadBarrier,_WriteBarrier,_ReadWriteBarrier) #define ECB_MEMORY_FENCE _ReadWriteBarrier () #define ECB_MEMORY_FENCE_ACQUIRE _ReadWriteBarrier () /* according to msdn, _ReadBarrier is not a load fence */ #define ECB_MEMORY_FENCE_RELEASE _WriteBarrier () #elif defined _WIN32 #include #define ECB_MEMORY_FENCE MemoryBarrier () /* actually just xchg on x86... scary */ #elif __SUNPRO_C >= 0x5110 || __SUNPRO_CC >= 0x5110 #include #define ECB_MEMORY_FENCE __machine_rw_barrier () #define ECB_MEMORY_FENCE_ACQUIRE __machine_acq_barrier () #define ECB_MEMORY_FENCE_RELEASE __machine_rel_barrier () #define ECB_MEMORY_FENCE_RELAXED __compiler_barrier () #elif __xlC__ #define ECB_MEMORY_FENCE __sync () #endif #endif #ifndef ECB_MEMORY_FENCE #if ECB_C11 && !defined __STDC_NO_ATOMICS__ /* we assume that these memory fences work on all variables/all memory accesses, */ /* not just C11 atomics and atomic accesses */ #include #define ECB_MEMORY_FENCE atomic_thread_fence (memory_order_seq_cst) #define ECB_MEMORY_FENCE_ACQUIRE atomic_thread_fence (memory_order_acquire) #define ECB_MEMORY_FENCE_RELEASE atomic_thread_fence (memory_order_release) #endif #endif #ifndef ECB_MEMORY_FENCE #if !ECB_AVOID_PTHREADS /* * if you get undefined symbol references to pthread_mutex_lock, * or failure to find pthread.h, then you should implement * the ECB_MEMORY_FENCE operations for your cpu/compiler * OR provide pthread.h and link against the posix thread library * of your system. */ #include #define ECB_NEEDS_PTHREADS 1 #define ECB_MEMORY_FENCE_NEEDS_PTHREADS 1 static pthread_mutex_t ecb_mf_lock = PTHREAD_MUTEX_INITIALIZER; #define ECB_MEMORY_FENCE do { pthread_mutex_lock (&ecb_mf_lock); pthread_mutex_unlock (&ecb_mf_lock); } while (0) #endif #endif #if !defined ECB_MEMORY_FENCE_ACQUIRE && defined ECB_MEMORY_FENCE #define ECB_MEMORY_FENCE_ACQUIRE ECB_MEMORY_FENCE #endif #if !defined ECB_MEMORY_FENCE_RELEASE && defined ECB_MEMORY_FENCE #define ECB_MEMORY_FENCE_RELEASE ECB_MEMORY_FENCE #endif #if !defined ECB_MEMORY_FENCE_RELAXED && defined ECB_MEMORY_FENCE #define ECB_MEMORY_FENCE_RELAXED ECB_MEMORY_FENCE /* very heavy-handed */ #endif /*****************************************************************************/ #if ECB_CPP #define ecb_inline static inline #elif ECB_GCC_VERSION(2,5) #define ecb_inline static __inline__ #elif ECB_C99 #define ecb_inline static inline #else #define ecb_inline static #endif #if ECB_GCC_VERSION(3,3) #define ecb_restrict __restrict__ #elif ECB_C99 #define ecb_restrict restrict #else #define ecb_restrict #endif typedef int ecb_bool; #define ECB_CONCAT_(a, b) a ## b #define ECB_CONCAT(a, b) ECB_CONCAT_(a, b) #define ECB_STRINGIFY_(a) # a #define ECB_STRINGIFY(a) ECB_STRINGIFY_(a) #define ECB_STRINGIFY_EXPR(expr) ((expr), ECB_STRINGIFY_ (expr)) #define ecb_function_ ecb_inline #if ECB_GCC_VERSION(3,1) || ECB_CLANG_VERSION(2,8) #define ecb_attribute(attrlist) __attribute__ (attrlist) #else #define ecb_attribute(attrlist) #endif #if ECB_GCC_VERSION(3,1) || ECB_CLANG_BUILTIN(__builtin_constant_p) #define ecb_is_constant(expr) __builtin_constant_p (expr) #else /* possible C11 impl for integral types typedef struct ecb_is_constant_struct ecb_is_constant_struct; #define ecb_is_constant(expr) _Generic ((1 ? (struct ecb_is_constant_struct *)0 : (void *)((expr) - (expr)), ecb_is_constant_struct *: 0, default: 1)) */ #define ecb_is_constant(expr) 0 #endif #if ECB_GCC_VERSION(3,1) || ECB_CLANG_BUILTIN(__builtin_expect) #define ecb_expect(expr,value) __builtin_expect ((expr),(value)) #else #define ecb_expect(expr,value) (expr) #endif #if ECB_GCC_VERSION(3,1) || ECB_CLANG_BUILTIN(__builtin_prefetch) #define ecb_prefetch(addr,rw,locality) __builtin_prefetch (addr, rw, locality) #else #define ecb_prefetch(addr,rw,locality) #endif /* no emulation for ecb_decltype */ #if ECB_CPP11 // older implementations might have problems with decltype(x)::type, work around it template struct ecb_decltype_t { typedef T type; }; #define ecb_decltype(x) ecb_decltype_t::type #elif ECB_GCC_VERSION(3,0) || ECB_CLANG_VERSION(2,8) #define ecb_decltype(x) __typeof__ (x) #endif #if _MSC_VER >= 1300 #define ecb_deprecated __declspec (deprecated) #else #define ecb_deprecated ecb_attribute ((__deprecated__)) #endif #if _MSC_VER >= 1500 #define ecb_deprecated_message(msg) __declspec (deprecated (msg)) #elif ECB_GCC_VERSION(4,5) #define ecb_deprecated_message(msg) ecb_attribute ((__deprecated__ (msg)) #else #define ecb_deprecated_message(msg) ecb_deprecated #endif #if _MSC_VER >= 1400 #define ecb_noinline __declspec (noinline) #else #define ecb_noinline ecb_attribute ((__noinline__)) #endif #define ecb_unused ecb_attribute ((__unused__)) #define ecb_const ecb_attribute ((__const__)) #define ecb_pure ecb_attribute ((__pure__)) #if ECB_C11 || __IBMC_NORETURN /* http://www-01.ibm.com/support/knowledgecenter/SSGH3R_13.1.0/com.ibm.xlcpp131.aix.doc/language_ref/noreturn.html */ #define ecb_noreturn _Noreturn #elif ECB_CPP11 #define ecb_noreturn [[noreturn]] #elif _MSC_VER >= 1200 /* http://msdn.microsoft.com/en-us/library/k6ktzx3s.aspx */ #define ecb_noreturn __declspec (noreturn) #else #define ecb_noreturn ecb_attribute ((__noreturn__)) #endif #if ECB_GCC_VERSION(4,3) #define ecb_artificial ecb_attribute ((__artificial__)) #define ecb_hot ecb_attribute ((__hot__)) #define ecb_cold ecb_attribute ((__cold__)) #else #define ecb_artificial #define ecb_hot #define ecb_cold #endif /* put around conditional expressions if you are very sure that the */ /* expression is mostly true or mostly false. note that these return */ /* booleans, not the expression. */ #define ecb_expect_false(expr) ecb_expect (!!(expr), 0) #define ecb_expect_true(expr) ecb_expect (!!(expr), 1) /* for compatibility to the rest of the world */ #define ecb_likely(expr) ecb_expect_true (expr) #define ecb_unlikely(expr) ecb_expect_false (expr) /* count trailing zero bits and count # of one bits */ #if ECB_GCC_VERSION(3,4) \ || (ECB_CLANG_BUILTIN(__builtin_clz) && ECB_CLANG_BUILTIN(__builtin_clzll) \ && ECB_CLANG_BUILTIN(__builtin_ctz) && ECB_CLANG_BUILTIN(__builtin_ctzll) \ && ECB_CLANG_BUILTIN(__builtin_popcount)) /* we assume int == 32 bit, long == 32 or 64 bit and long long == 64 bit */ #define ecb_ld32(x) (__builtin_clz (x) ^ 31) #define ecb_ld64(x) (__builtin_clzll (x) ^ 63) #define ecb_ctz32(x) __builtin_ctz (x) #define ecb_ctz64(x) __builtin_ctzll (x) #define ecb_popcount32(x) __builtin_popcount (x) /* no popcountll */ #else ecb_function_ ecb_const int ecb_ctz32 (uint32_t x); ecb_function_ ecb_const int ecb_ctz32 (uint32_t x) { #if 1400 <= _MSC_VER && (_M_IX86 || _M_X64 || _M_IA64 || _M_ARM) unsigned long r; _BitScanForward (&r, x); return (int)r; #else int r = 0; x &= ~x + 1; /* this isolates the lowest bit */ #if ECB_branchless_on_i386 r += !!(x & 0xaaaaaaaa) << 0; r += !!(x & 0xcccccccc) << 1; r += !!(x & 0xf0f0f0f0) << 2; r += !!(x & 0xff00ff00) << 3; r += !!(x & 0xffff0000) << 4; #else if (x & 0xaaaaaaaa) r += 1; if (x & 0xcccccccc) r += 2; if (x & 0xf0f0f0f0) r += 4; if (x & 0xff00ff00) r += 8; if (x & 0xffff0000) r += 16; #endif return r; #endif } ecb_function_ ecb_const int ecb_ctz64 (uint64_t x); ecb_function_ ecb_const int ecb_ctz64 (uint64_t x) { #if 1400 <= _MSC_VER && (_M_X64 || _M_IA64 || _M_ARM) unsigned long r; _BitScanForward64 (&r, x); return (int)r; #else int shift = x & 0xffffffff ? 0 : 32; return ecb_ctz32 (x >> shift) + shift; #endif } ecb_function_ ecb_const int ecb_popcount32 (uint32_t x); ecb_function_ ecb_const int ecb_popcount32 (uint32_t x) { x -= (x >> 1) & 0x55555555; x = ((x >> 2) & 0x33333333) + (x & 0x33333333); x = ((x >> 4) + x) & 0x0f0f0f0f; x *= 0x01010101; return x >> 24; } ecb_function_ ecb_const int ecb_ld32 (uint32_t x); ecb_function_ ecb_const int ecb_ld32 (uint32_t x) { #if 1400 <= _MSC_VER && (_M_IX86 || _M_X64 || _M_IA64 || _M_ARM) unsigned long r; _BitScanReverse (&r, x); return (int)r; #else int r = 0; if (x >> 16) { x >>= 16; r += 16; } if (x >> 8) { x >>= 8; r += 8; } if (x >> 4) { x >>= 4; r += 4; } if (x >> 2) { x >>= 2; r += 2; } if (x >> 1) { r += 1; } return r; #endif } ecb_function_ ecb_const int ecb_ld64 (uint64_t x); ecb_function_ ecb_const int ecb_ld64 (uint64_t x) { #if 1400 <= _MSC_VER && (_M_X64 || _M_IA64 || _M_ARM) unsigned long r; _BitScanReverse64 (&r, x); return (int)r; #else int r = 0; if (x >> 32) { x >>= 32; r += 32; } return r + ecb_ld32 (x); #endif } #endif ecb_function_ ecb_const ecb_bool ecb_is_pot32 (uint32_t x); ecb_function_ ecb_const ecb_bool ecb_is_pot32 (uint32_t x) { return !(x & (x - 1)); } ecb_function_ ecb_const ecb_bool ecb_is_pot64 (uint64_t x); ecb_function_ ecb_const ecb_bool ecb_is_pot64 (uint64_t x) { return !(x & (x - 1)); } ecb_function_ ecb_const uint8_t ecb_bitrev8 (uint8_t x); ecb_function_ ecb_const uint8_t ecb_bitrev8 (uint8_t x) { return ( (x * 0x0802U & 0x22110U) | (x * 0x8020U & 0x88440U)) * 0x10101U >> 16; } ecb_function_ ecb_const uint16_t ecb_bitrev16 (uint16_t x); ecb_function_ ecb_const uint16_t ecb_bitrev16 (uint16_t x) { x = ((x >> 1) & 0x5555) | ((x & 0x5555) << 1); x = ((x >> 2) & 0x3333) | ((x & 0x3333) << 2); x = ((x >> 4) & 0x0f0f) | ((x & 0x0f0f) << 4); x = ( x >> 8 ) | ( x << 8); return x; } ecb_function_ ecb_const uint32_t ecb_bitrev32 (uint32_t x); ecb_function_ ecb_const uint32_t ecb_bitrev32 (uint32_t x) { x = ((x >> 1) & 0x55555555) | ((x & 0x55555555) << 1); x = ((x >> 2) & 0x33333333) | ((x & 0x33333333) << 2); x = ((x >> 4) & 0x0f0f0f0f) | ((x & 0x0f0f0f0f) << 4); x = ((x >> 8) & 0x00ff00ff) | ((x & 0x00ff00ff) << 8); x = ( x >> 16 ) | ( x << 16); return x; } /* popcount64 is only available on 64 bit cpus as gcc builtin */ /* so for this version we are lazy */ ecb_function_ ecb_const int ecb_popcount64 (uint64_t x); ecb_function_ ecb_const int ecb_popcount64 (uint64_t x) { return ecb_popcount32 (x) + ecb_popcount32 (x >> 32); } ecb_inline ecb_const uint8_t ecb_rotl8 (uint8_t x, unsigned int count); ecb_inline ecb_const uint8_t ecb_rotr8 (uint8_t x, unsigned int count); ecb_inline ecb_const uint16_t ecb_rotl16 (uint16_t x, unsigned int count); ecb_inline ecb_const uint16_t ecb_rotr16 (uint16_t x, unsigned int count); ecb_inline ecb_const uint32_t ecb_rotl32 (uint32_t x, unsigned int count); ecb_inline ecb_const uint32_t ecb_rotr32 (uint32_t x, unsigned int count); ecb_inline ecb_const uint64_t ecb_rotl64 (uint64_t x, unsigned int count); ecb_inline ecb_const uint64_t ecb_rotr64 (uint64_t x, unsigned int count); ecb_inline ecb_const uint8_t ecb_rotl8 (uint8_t x, unsigned int count) { return (x >> ( 8 - count)) | (x << count); } ecb_inline ecb_const uint8_t ecb_rotr8 (uint8_t x, unsigned int count) { return (x << ( 8 - count)) | (x >> count); } ecb_inline ecb_const uint16_t ecb_rotl16 (uint16_t x, unsigned int count) { return (x >> (16 - count)) | (x << count); } ecb_inline ecb_const uint16_t ecb_rotr16 (uint16_t x, unsigned int count) { return (x << (16 - count)) | (x >> count); } ecb_inline ecb_const uint32_t ecb_rotl32 (uint32_t x, unsigned int count) { return (x >> (32 - count)) | (x << count); } ecb_inline ecb_const uint32_t ecb_rotr32 (uint32_t x, unsigned int count) { return (x << (32 - count)) | (x >> count); } ecb_inline ecb_const uint64_t ecb_rotl64 (uint64_t x, unsigned int count) { return (x >> (64 - count)) | (x << count); } ecb_inline ecb_const uint64_t ecb_rotr64 (uint64_t x, unsigned int count) { return (x << (64 - count)) | (x >> count); } #if ECB_CPP inline uint8_t ecb_ctz (uint8_t v) { return ecb_ctz32 (v); } inline uint16_t ecb_ctz (uint16_t v) { return ecb_ctz32 (v); } inline uint32_t ecb_ctz (uint32_t v) { return ecb_ctz32 (v); } inline uint64_t ecb_ctz (uint64_t v) { return ecb_ctz64 (v); } inline bool ecb_is_pot (uint8_t v) { return ecb_is_pot32 (v); } inline bool ecb_is_pot (uint16_t v) { return ecb_is_pot32 (v); } inline bool ecb_is_pot (uint32_t v) { return ecb_is_pot32 (v); } inline bool ecb_is_pot (uint64_t v) { return ecb_is_pot64 (v); } inline int ecb_ld (uint8_t v) { return ecb_ld32 (v); } inline int ecb_ld (uint16_t v) { return ecb_ld32 (v); } inline int ecb_ld (uint32_t v) { return ecb_ld32 (v); } inline int ecb_ld (uint64_t v) { return ecb_ld64 (v); } inline int ecb_popcount (uint8_t v) { return ecb_popcount32 (v); } inline int ecb_popcount (uint16_t v) { return ecb_popcount32 (v); } inline int ecb_popcount (uint32_t v) { return ecb_popcount32 (v); } inline int ecb_popcount (uint64_t v) { return ecb_popcount64 (v); } inline uint8_t ecb_bitrev (uint8_t v) { return ecb_bitrev8 (v); } inline uint16_t ecb_bitrev (uint16_t v) { return ecb_bitrev16 (v); } inline uint32_t ecb_bitrev (uint32_t v) { return ecb_bitrev32 (v); } inline uint8_t ecb_rotl (uint8_t v, unsigned int count) { return ecb_rotl8 (v, count); } inline uint16_t ecb_rotl (uint16_t v, unsigned int count) { return ecb_rotl16 (v, count); } inline uint32_t ecb_rotl (uint32_t v, unsigned int count) { return ecb_rotl32 (v, count); } inline uint64_t ecb_rotl (uint64_t v, unsigned int count) { return ecb_rotl64 (v, count); } inline uint8_t ecb_rotr (uint8_t v, unsigned int count) { return ecb_rotr8 (v, count); } inline uint16_t ecb_rotr (uint16_t v, unsigned int count) { return ecb_rotr16 (v, count); } inline uint32_t ecb_rotr (uint32_t v, unsigned int count) { return ecb_rotr32 (v, count); } inline uint64_t ecb_rotr (uint64_t v, unsigned int count) { return ecb_rotr64 (v, count); } #endif #if ECB_GCC_VERSION(4,3) || (ECB_CLANG_BUILTIN(__builtin_bswap32) && ECB_CLANG_BUILTIN(__builtin_bswap64)) #if ECB_GCC_VERSION(4,8) || ECB_CLANG_BUILTIN(__builtin_bswap16) #define ecb_bswap16(x) __builtin_bswap16 (x) #else #define ecb_bswap16(x) (__builtin_bswap32 (x) >> 16) #endif #define ecb_bswap32(x) __builtin_bswap32 (x) #define ecb_bswap64(x) __builtin_bswap64 (x) #elif _MSC_VER #include #define ecb_bswap16(x) ((uint16_t)_byteswap_ushort ((uint16_t)(x))) #define ecb_bswap32(x) ((uint32_t)_byteswap_ulong ((uint32_t)(x))) #define ecb_bswap64(x) ((uint64_t)_byteswap_uint64 ((uint64_t)(x))) #else ecb_function_ ecb_const uint16_t ecb_bswap16 (uint16_t x); ecb_function_ ecb_const uint16_t ecb_bswap16 (uint16_t x) { return ecb_rotl16 (x, 8); } ecb_function_ ecb_const uint32_t ecb_bswap32 (uint32_t x); ecb_function_ ecb_const uint32_t ecb_bswap32 (uint32_t x) { return (((uint32_t)ecb_bswap16 (x)) << 16) | ecb_bswap16 (x >> 16); } ecb_function_ ecb_const uint64_t ecb_bswap64 (uint64_t x); ecb_function_ ecb_const uint64_t ecb_bswap64 (uint64_t x) { return (((uint64_t)ecb_bswap32 (x)) << 32) | ecb_bswap32 (x >> 32); } #endif #if ECB_GCC_VERSION(4,5) || ECB_CLANG_BUILTIN(__builtin_unreachable) #define ecb_unreachable() __builtin_unreachable () #else /* this seems to work fine, but gcc always emits a warning for it :/ */ ecb_inline ecb_noreturn void ecb_unreachable (void); ecb_inline ecb_noreturn void ecb_unreachable (void) { } #endif /* try to tell the compiler that some condition is definitely true */ #define ecb_assume(cond) if (!(cond)) ecb_unreachable (); else 0 ecb_inline ecb_const uint32_t ecb_byteorder_helper (void); ecb_inline ecb_const uint32_t ecb_byteorder_helper (void) { /* the union code still generates code under pressure in gcc, */ /* but less than using pointers, and always seems to */ /* successfully return a constant. */ /* the reason why we have this horrible preprocessor mess */ /* is to avoid it in all cases, at least on common architectures */ /* or when using a recent enough gcc version (>= 4.6) */ #if (defined __BYTE_ORDER__ && __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__) \ || ((__i386 || __i386__ || _M_IX86 || ECB_GCC_AMD64 || ECB_MSVC_AMD64) && !__VOS__) #define ECB_LITTLE_ENDIAN 1 return 0x44332211; #elif (defined __BYTE_ORDER__ && __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__) \ || ((__AARCH64EB__ || __MIPSEB__ || __ARMEB__) && !__VOS__) #define ECB_BIG_ENDIAN 1 return 0x11223344; #else union { uint8_t c[4]; uint32_t u; } u = { 0x11, 0x22, 0x33, 0x44 }; return u.u; #endif } ecb_inline ecb_const ecb_bool ecb_big_endian (void); ecb_inline ecb_const ecb_bool ecb_big_endian (void) { return ecb_byteorder_helper () == 0x11223344; } ecb_inline ecb_const ecb_bool ecb_little_endian (void); ecb_inline ecb_const ecb_bool ecb_little_endian (void) { return ecb_byteorder_helper () == 0x44332211; } /*****************************************************************************/ /* unaligned load/store */ ecb_inline uint_fast16_t ecb_be_u16_to_host (uint_fast16_t v) { return ecb_little_endian () ? ecb_bswap16 (v) : v; } ecb_inline uint_fast32_t ecb_be_u32_to_host (uint_fast32_t v) { return ecb_little_endian () ? ecb_bswap32 (v) : v; } ecb_inline uint_fast64_t ecb_be_u64_to_host (uint_fast64_t v) { return ecb_little_endian () ? ecb_bswap64 (v) : v; } ecb_inline uint_fast16_t ecb_le_u16_to_host (uint_fast16_t v) { return ecb_big_endian () ? ecb_bswap16 (v) : v; } ecb_inline uint_fast32_t ecb_le_u32_to_host (uint_fast32_t v) { return ecb_big_endian () ? ecb_bswap32 (v) : v; } ecb_inline uint_fast64_t ecb_le_u64_to_host (uint_fast64_t v) { return ecb_big_endian () ? ecb_bswap64 (v) : v; } ecb_inline uint_fast16_t ecb_peek_u16_u (const void *ptr) { uint16_t v; memcpy (&v, ptr, sizeof (v)); return v; } ecb_inline uint_fast32_t ecb_peek_u32_u (const void *ptr) { uint32_t v; memcpy (&v, ptr, sizeof (v)); return v; } ecb_inline uint_fast64_t ecb_peek_u64_u (const void *ptr) { uint64_t v; memcpy (&v, ptr, sizeof (v)); return v; } ecb_inline uint_fast16_t ecb_peek_be_u16_u (const void *ptr) { return ecb_be_u16_to_host (ecb_peek_u16_u (ptr)); } ecb_inline uint_fast32_t ecb_peek_be_u32_u (const void *ptr) { return ecb_be_u32_to_host (ecb_peek_u32_u (ptr)); } ecb_inline uint_fast64_t ecb_peek_be_u64_u (const void *ptr) { return ecb_be_u64_to_host (ecb_peek_u64_u (ptr)); } ecb_inline uint_fast16_t ecb_peek_le_u16_u (const void *ptr) { return ecb_le_u16_to_host (ecb_peek_u16_u (ptr)); } ecb_inline uint_fast32_t ecb_peek_le_u32_u (const void *ptr) { return ecb_le_u32_to_host (ecb_peek_u32_u (ptr)); } ecb_inline uint_fast64_t ecb_peek_le_u64_u (const void *ptr) { return ecb_le_u64_to_host (ecb_peek_u64_u (ptr)); } ecb_inline uint_fast16_t ecb_host_to_be_u16 (uint_fast16_t v) { return ecb_little_endian () ? ecb_bswap16 (v) : v; } ecb_inline uint_fast32_t ecb_host_to_be_u32 (uint_fast32_t v) { return ecb_little_endian () ? ecb_bswap32 (v) : v; } ecb_inline uint_fast64_t ecb_host_to_be_u64 (uint_fast64_t v) { return ecb_little_endian () ? ecb_bswap64 (v) : v; } ecb_inline uint_fast16_t ecb_host_to_le_u16 (uint_fast16_t v) { return ecb_big_endian () ? ecb_bswap16 (v) : v; } ecb_inline uint_fast32_t ecb_host_to_le_u32 (uint_fast32_t v) { return ecb_big_endian () ? ecb_bswap32 (v) : v; } ecb_inline uint_fast64_t ecb_host_to_le_u64 (uint_fast64_t v) { return ecb_big_endian () ? ecb_bswap64 (v) : v; } ecb_inline void ecb_poke_u16_u (void *ptr, uint16_t v) { memcpy (ptr, &v, sizeof (v)); } ecb_inline void ecb_poke_u32_u (void *ptr, uint32_t v) { memcpy (ptr, &v, sizeof (v)); } ecb_inline void ecb_poke_u64_u (void *ptr, uint64_t v) { memcpy (ptr, &v, sizeof (v)); } ecb_inline void ecb_poke_be_u16_u (void *ptr, uint_fast16_t v) { ecb_poke_u16_u (ptr, ecb_host_to_be_u16 (v)); } ecb_inline void ecb_poke_be_u32_u (void *ptr, uint_fast32_t v) { ecb_poke_u32_u (ptr, ecb_host_to_be_u32 (v)); } ecb_inline void ecb_poke_be_u64_u (void *ptr, uint_fast64_t v) { ecb_poke_u64_u (ptr, ecb_host_to_be_u64 (v)); } ecb_inline void ecb_poke_le_u16_u (void *ptr, uint_fast16_t v) { ecb_poke_u16_u (ptr, ecb_host_to_le_u16 (v)); } ecb_inline void ecb_poke_le_u32_u (void *ptr, uint_fast32_t v) { ecb_poke_u32_u (ptr, ecb_host_to_le_u32 (v)); } ecb_inline void ecb_poke_le_u64_u (void *ptr, uint_fast64_t v) { ecb_poke_u64_u (ptr, ecb_host_to_le_u64 (v)); } #if ECB_CPP inline uint8_t ecb_bswap (uint8_t v) { return v; } inline uint16_t ecb_bswap (uint16_t v) { return ecb_bswap16 (v); } inline uint32_t ecb_bswap (uint32_t v) { return ecb_bswap32 (v); } inline uint64_t ecb_bswap (uint64_t v) { return ecb_bswap64 (v); } template inline T ecb_be_to_host (T v) { return ecb_little_endian () ? ecb_bswap (v) : v; } template inline T ecb_le_to_host (T v) { return ecb_big_endian () ? ecb_bswap (v) : v; } template inline T ecb_peek (const void *ptr) { return *(const T *)ptr; } template inline T ecb_peek_be (const void *ptr) { return ecb_be_to_host (ecb_peek (ptr)); } template inline T ecb_peek_le (const void *ptr) { return ecb_le_to_host (ecb_peek (ptr)); } template inline T ecb_peek_u (const void *ptr) { T v; memcpy (&v, ptr, sizeof (v)); return v; } template inline T ecb_peek_be_u (const void *ptr) { return ecb_be_to_host (ecb_peek_u (ptr)); } template inline T ecb_peek_le_u (const void *ptr) { return ecb_le_to_host (ecb_peek_u (ptr)); } template inline T ecb_host_to_be (T v) { return ecb_little_endian () ? ecb_bswap (v) : v; } template inline T ecb_host_to_le (T v) { return ecb_big_endian () ? ecb_bswap (v) : v; } template inline void ecb_poke (void *ptr, T v) { *(T *)ptr = v; } template inline void ecb_poke_be (void *ptr, T v) { return ecb_poke (ptr, ecb_host_to_be (v)); } template inline void ecb_poke_le (void *ptr, T v) { return ecb_poke (ptr, ecb_host_to_le (v)); } template inline void ecb_poke_u (void *ptr, T v) { memcpy (ptr, &v, sizeof (v)); } template inline void ecb_poke_be_u (void *ptr, T v) { return ecb_poke_u (ptr, ecb_host_to_be (v)); } template inline void ecb_poke_le_u (void *ptr, T v) { return ecb_poke_u (ptr, ecb_host_to_le (v)); } #endif /*****************************************************************************/ #if ECB_GCC_VERSION(3,0) || ECB_C99 #define ecb_mod(m,n) ((m) % (n) + ((m) % (n) < 0 ? (n) : 0)) #else #define ecb_mod(m,n) ((m) < 0 ? ((n) - 1 - ((-1 - (m)) % (n))) : ((m) % (n))) #endif #if ECB_CPP template static inline T ecb_div_rd (T val, T div) { return val < 0 ? - ((-val + div - 1) / div) : (val ) / div; } template static inline T ecb_div_ru (T val, T div) { return val < 0 ? - ((-val ) / div) : (val + div - 1) / div; } #else #define ecb_div_rd(val,div) ((val) < 0 ? - ((-(val) + (div) - 1) / (div)) : ((val) ) / (div)) #define ecb_div_ru(val,div) ((val) < 0 ? - ((-(val) ) / (div)) : ((val) + (div) - 1) / (div)) #endif #if ecb_cplusplus_does_not_suck /* does not work for local types (http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2008/n2657.htm) */ template static inline int ecb_array_length (const T (&arr)[N]) { return N; } #else #define ecb_array_length(name) (sizeof (name) / sizeof (name [0])) #endif /*****************************************************************************/ ecb_function_ ecb_const uint32_t ecb_binary16_to_binary32 (uint32_t x); ecb_function_ ecb_const uint32_t ecb_binary16_to_binary32 (uint32_t x) { unsigned int s = (x & 0x8000) << (31 - 15); int e = (x >> 10) & 0x001f; unsigned int m = x & 0x03ff; if (ecb_expect_false (e == 31)) /* infinity or NaN */ e = 255 - (127 - 15); else if (ecb_expect_false (!e)) { if (ecb_expect_true (!m)) /* zero, handled by code below by forcing e to 0 */ e = 0 - (127 - 15); else { /* subnormal, renormalise */ unsigned int s = 10 - ecb_ld32 (m); m = (m << s) & 0x3ff; /* mask implicit bit */ e -= s - 1; } } /* e and m now are normalised, or zero, (or inf or nan) */ e += 127 - 15; return s | (e << 23) | (m << (23 - 10)); } ecb_function_ ecb_const uint16_t ecb_binary32_to_binary16 (uint32_t x); ecb_function_ ecb_const uint16_t ecb_binary32_to_binary16 (uint32_t x) { unsigned int s = (x >> 16) & 0x00008000; /* sign bit, the easy part */ unsigned int e = ((x >> 23) & 0x000000ff) - (127 - 15); /* the desired exponent */ unsigned int m = x & 0x007fffff; x &= 0x7fffffff; /* if it's within range of binary16 normals, use fast path */ if (ecb_expect_true (0x38800000 <= x && x <= 0x477fefff)) { /* mantissa round-to-even */ m += 0x00000fff + ((m >> (23 - 10)) & 1); /* handle overflow */ if (ecb_expect_false (m >= 0x00800000)) { m >>= 1; e += 1; } return s | (e << 10) | (m >> (23 - 10)); } /* handle large numbers and infinity */ if (ecb_expect_true (0x477fefff < x && x <= 0x7f800000)) return s | 0x7c00; /* handle zero, subnormals and small numbers */ if (ecb_expect_true (x < 0x38800000)) { /* zero */ if (ecb_expect_true (!x)) return s; /* handle subnormals */ /* too small, will be zero */ if (e < (14 - 24)) /* might not be sharp, but is good enough */ return s; m |= 0x00800000; /* make implicit bit explicit */ /* very tricky - we need to round to the nearest e (+10) bit value */ { unsigned int bits = 14 - e; unsigned int half = (1 << (bits - 1)) - 1; unsigned int even = (m >> bits) & 1; /* if this overflows, we will end up with a normalised number */ m = (m + half + even) >> bits; } return s | m; } /* handle NaNs, preserve leftmost nan bits, but make sure we don't turn them into infinities */ m >>= 13; return s | 0x7c00 | m | !m; } /*******************************************************************************/ /* floating point stuff, can be disabled by defining ECB_NO_LIBM */ /* basically, everything uses "ieee pure-endian" floating point numbers */ /* the only noteworthy exception is ancient armle, which uses order 43218765 */ #if 0 \ || __i386 || __i386__ \ || ECB_GCC_AMD64 \ || __powerpc__ || __ppc__ || __powerpc64__ || __ppc64__ \ || defined __s390__ || defined __s390x__ \ || defined __mips__ \ || defined __alpha__ \ || defined __hppa__ \ || defined __ia64__ \ || defined __m68k__ \ || defined __m88k__ \ || defined __sh__ \ || defined _M_IX86 || defined ECB_MSVC_AMD64 || defined _M_IA64 \ || (defined __arm__ && (defined __ARM_EABI__ || defined __EABI__ || defined __VFP_FP__ || defined _WIN32_WCE || defined __ANDROID__)) \ || defined __aarch64__ #define ECB_STDFP 1 #else #define ECB_STDFP 0 #endif #ifndef ECB_NO_LIBM #include /* for frexp*, ldexp*, INFINITY, NAN */ /* only the oldest of old doesn't have this one. solaris. */ #ifdef INFINITY #define ECB_INFINITY INFINITY #else #define ECB_INFINITY HUGE_VAL #endif #ifdef NAN #define ECB_NAN NAN #else #define ECB_NAN ECB_INFINITY #endif #if ECB_C99 || _XOPEN_VERSION >= 600 || _POSIX_VERSION >= 200112L #define ecb_ldexpf(x,e) ldexpf ((x), (e)) #define ecb_frexpf(x,e) frexpf ((x), (e)) #else #define ecb_ldexpf(x,e) (float) ldexp ((double) (x), (e)) #define ecb_frexpf(x,e) (float) frexp ((double) (x), (e)) #endif /* convert a float to ieee single/binary32 */ ecb_function_ ecb_const uint32_t ecb_float_to_binary32 (float x); ecb_function_ ecb_const uint32_t ecb_float_to_binary32 (float x) { uint32_t r; #if ECB_STDFP memcpy (&r, &x, 4); #else /* slow emulation, works for anything but -0 */ uint32_t m; int e; if (x == 0e0f ) return 0x00000000U; if (x > +3.40282346638528860e+38f) return 0x7f800000U; if (x < -3.40282346638528860e+38f) return 0xff800000U; if (x != x ) return 0x7fbfffffU; m = ecb_frexpf (x, &e) * 0x1000000U; r = m & 0x80000000U; if (r) m = -m; if (e <= -126) { m &= 0xffffffU; m >>= (-125 - e); e = -126; } r |= (e + 126) << 23; r |= m & 0x7fffffU; #endif return r; } /* converts an ieee single/binary32 to a float */ ecb_function_ ecb_const float ecb_binary32_to_float (uint32_t x); ecb_function_ ecb_const float ecb_binary32_to_float (uint32_t x) { float r; #if ECB_STDFP memcpy (&r, &x, 4); #else /* emulation, only works for normals and subnormals and +0 */ int neg = x >> 31; int e = (x >> 23) & 0xffU; x &= 0x7fffffU; if (e) x |= 0x800000U; else e = 1; /* we distrust ldexpf a bit and do the 2**-24 scaling by an extra multiply */ r = ecb_ldexpf (x * (0.5f / 0x800000U), e - 126); r = neg ? -r : r; #endif return r; } /* convert a double to ieee double/binary64 */ ecb_function_ ecb_const uint64_t ecb_double_to_binary64 (double x); ecb_function_ ecb_const uint64_t ecb_double_to_binary64 (double x) { uint64_t r; #if ECB_STDFP memcpy (&r, &x, 8); #else /* slow emulation, works for anything but -0 */ uint64_t m; int e; if (x == 0e0 ) return 0x0000000000000000U; if (x > +1.79769313486231470e+308) return 0x7ff0000000000000U; if (x < -1.79769313486231470e+308) return 0xfff0000000000000U; if (x != x ) return 0X7ff7ffffffffffffU; m = frexp (x, &e) * 0x20000000000000U; r = m & 0x8000000000000000;; if (r) m = -m; if (e <= -1022) { m &= 0x1fffffffffffffU; m >>= (-1021 - e); e = -1022; } r |= ((uint64_t)(e + 1022)) << 52; r |= m & 0xfffffffffffffU; #endif return r; } /* converts an ieee double/binary64 to a double */ ecb_function_ ecb_const double ecb_binary64_to_double (uint64_t x); ecb_function_ ecb_const double ecb_binary64_to_double (uint64_t x) { double r; #if ECB_STDFP memcpy (&r, &x, 8); #else /* emulation, only works for normals and subnormals and +0 */ int neg = x >> 63; int e = (x >> 52) & 0x7ffU; x &= 0xfffffffffffffU; if (e) x |= 0x10000000000000U; else e = 1; /* we distrust ldexp a bit and do the 2**-53 scaling by an extra multiply */ r = ldexp (x * (0.5 / 0x10000000000000U), e - 1022); r = neg ? -r : r; #endif return r; } /* convert a float to ieee half/binary16 */ ecb_function_ ecb_const uint16_t ecb_float_to_binary16 (float x); ecb_function_ ecb_const uint16_t ecb_float_to_binary16 (float x) { return ecb_binary32_to_binary16 (ecb_float_to_binary32 (x)); } /* convert an ieee half/binary16 to float */ ecb_function_ ecb_const float ecb_binary16_to_float (uint16_t x); ecb_function_ ecb_const float ecb_binary16_to_float (uint16_t x) { return ecb_binary32_to_float (ecb_binary16_to_binary32 (x)); } #endif #endif Async-Interrupt-1.26/COPYING0000644000000000000000000000007610211640730014237 0ustar rootrootThis module is licensed under the same terms as perl itself. Async-Interrupt-1.26/t/0000755000000000000000000000000013651541474013464 5ustar rootrootAsync-Interrupt-1.26/t/04_apipe.t0000644000000000000000000000200011227750114015230 0ustar rootroot#! perl no warnings; print "1..14\n"; $|=1; use Async::Interrupt; my $ai = new Async::Interrupt cb => sub { print "ok $_[0]\n" }; my $fd = $ai->pipe_fileno; print "ok 1\n"; $ai->signal (2); print "ok 3\n"; my ($vr, $vR); vec ($vr, $ai->pipe_fileno, 1) = 1; my $n = select $vR=$vr, undef, undef, 0; print $n == 0 ? "" : "not ", "ok 4 # $n\n"; $ai->block; $ai->signal (7); print "ok 5\n"; my $n = select $vR=$vr, undef, undef, 0; print $n == 1 ? "" : "not ", "ok 6 # $n\n"; $ai->unblock; my $n = select $vR=$vr, undef, undef, 0; print $n == 0 ? "" : "not ", "ok 8 # $n\n"; $ai->signal (9); my $n = select $vR=$vr, undef, undef, 0; print $n == 0 ? "" : "not ", "ok 10 # $n\n"; $ai->pipe_disable; $ai->scope_block; $ai->signal (14); my $n = select $vR=$vr, undef, undef, 0; print $n == 0 ? "" : "not ", "ok 11 # $n\n"; $ai->post_fork; print $fd == $ai->pipe_fileno ? "" : "not ", "ok 12\n"; $ai->post_fork; print $fd == $ai->pipe_fileno ? "" : "not ", "ok 13\n"; undef $ai; # will cause signal to be sent Async-Interrupt-1.26/t/02_pipe.t0000644000000000000000000000223611226205764015105 0ustar rootroot#! perl no warnings; use Socket; my ($pr, $pw); unless (socketpair $pr, $pw, Socket::AF_UNIX (), Socket::SOCK_STREAM (), 0) { print "1..0 # SKIP socketpair failed - broken platform, skipping tests\n"; exit; } print "1..12\n"; $|=1; use Async::Interrupt; # we ignore the requirement to put handles into nonblocking mode # IN THIS TEST only. never do that in real life. my $ai = new Async::Interrupt pipe => [$pr, $pw], cb => sub { print "ok $_[0]\n" }; print "ok 1\n"; $ai->signal (2); print "ok 3\n"; my ($vr, $vR); vec ($vr, fileno $pr, 1) = 1; my $n = select $vR=$vr, undef, undef, 0; print $n == 0 ? "" : "not ", "ok 4 # $n\n"; $ai->block; $ai->signal (7); print "ok 5\n"; my $n = select $vR=$vr, undef, undef, 0; print $n == 1 ? "" : "not ", "ok 6 # $n\n"; $ai->unblock; my $n = select $vR=$vr, undef, undef, 0; print $n == 0 ? "" : "not ", "ok 8 # $n\n"; $ai->signal (9); my $n = select $vR=$vr, undef, undef, 0; print $n == 0 ? "" : "not ", "ok 10 # $n\n"; $ai->pipe_disable; $ai->scope_block; $ai->signal (12); my $n = select $vR=$vr, undef, undef, 0; print $n == 0 ? "" : "not ", "ok 11 # $n\n"; undef $ai; # will cause signal to be sent Async-Interrupt-1.26/t/06_epipe.t0000644000000000000000000000110211373362762015252 0ustar rootroot#! perl no warnings; print "1..6\n"; $|=1; use Async::Interrupt; my $ep = new Async::Interrupt::EventPipe; my $fd = $ep->fileno; print "ok 1\n"; my ($vr, $vR); vec ($vr, $fd, 1) = 1; my $n = select $vR=$vr, undef, undef, 0; print $n == 0 ? "" : "not ", "ok 2 # $n\n"; $ep->signal; my $n = select $vR=$vr, undef, undef, 0; print $n == 1 ? "" : "not ", "ok 3 # $n\n"; $ep->drain; my $n = select $vR=$vr, undef, undef, 0; print $n == 0 ? "" : "not ", "ok 4 # $n\n"; print "ok 5 # ", join " ", $ep->signal_func, "\n"; print "ok 6 # ", join " ", $ep->drain_func, "\n"; Async-Interrupt-1.26/t/01_basic.t0000644000000000000000000000106211223156162015216 0ustar rootrootprint "1..12\n"; $|=1; use Async::Interrupt; my $ai = new Async::Interrupt cb => sub { print "ok $_[0]\n" }; my $ai2 = new Async::Interrupt; print $$ai ? "" : "not ", "ok 1 # $$ai\n"; my ($a, $b) = $ai->signal_func; print $a ? "" : "not ", "ok 2 # $a\n"; print $b ? "" : "not ", "ok 3 # $b\n"; $ai->signal (4); my $ai3 = new Async::Interrupt; print "ok 5\n"; $ai->block; $ai->signal (7); print "ok 6\n"; $ai->unblock; print "ok 8\n"; undef $ai2; print "ok 9\n"; { $ai->scope_block; $ai->signal (11); print "ok 10\n"; } print "ok 12\n"; Async-Interrupt-1.26/t/03_signal.t0000644000000000000000000000076711233451105015423 0ustar rootrootunless (exists $SIG{USR1}) { print "1..0 # SKIP no SIGUSR1 - broken platform, skipping tests\n"; exit; } print "1..9\n"; $|=1; use Async::Interrupt; my $three = 3; my $ai = new Async::Interrupt cb => sub { print "ok ", $three++, "\n" }, signal => "CHLD"; print "ok 1\n"; { $ai->scope_block; $ai->scope_block; kill CHLD => $$; print "ok 2\n"; } kill CHLD, $$; $ai->signal_hysteresis (1); kill CHLD, $$; kill CHLD, $$; kill CHLD, $$; kill CHLD, $$; print "ok 9\n"; Async-Interrupt-1.26/t/05_var.t0000644000000000000000000000063411227754440014744 0ustar rootrootprint "1..7\n"; $|=1; use Async::Interrupt; my $var; my $ai = new Async::Interrupt var => \$var, cb => sub { print $var ? "not " : "", "ok $_[0]\n" }; print $$ai ? "" : "not ", "ok 1 # $$ai\n"; print $ai->c_var ? "" : "not ", "ok 2\n"; $ai->signal (3); print $var == 0 ? "" : "not ", "ok 4\n"; $ai->block; $ai->signal (7); print "ok 5\n"; print $var == 7 ? "" : "not ", "ok 6\n"; $ai->unblock; Async-Interrupt-1.26/t/00_load.t0000644000000000000000000000017611223114451015053 0ustar rootrootBEGIN { $| = 1; print "1..1\n"; } END {print "not ok 1\n" unless $loaded;} use Async::Interrupt; $loaded = 1; print "ok 1\n"; Async-Interrupt-1.26/Interrupt.pm0000644000000000000000000005326713651541472015566 0ustar rootroot=head1 NAME Async::Interrupt - allow C/XS libraries to interrupt perl asynchronously =head1 SYNOPSIS use Async::Interrupt; =head1 DESCRIPTION This module implements a single feature only of interest to advanced perl modules, namely asynchronous interruptions (think "UNIX signals", which are very similar). Sometimes, modules wish to run code asynchronously (in another thread, or from a signal handler), and then signal the perl interpreter on certain events. One common way is to write some data to a pipe and use an event handling toolkit to watch for I/O events. Another way is to send a signal. Those methods are slow, and in the case of a pipe, also not asynchronous - it won't interrupt a running perl interpreter. This module implements asynchronous notifications that enable you to signal running perl code from another thread, asynchronously, and sometimes even without using a single syscall. =head2 USAGE SCENARIOS =over 4 =item Race-free signal handling There seems to be no way to do race-free signal handling in perl: to catch a signal, you have to execute Perl code, and between entering the interpreter C on, fixing the race completely. This can be used to implement the signal handling in event loops, e.g. L, L, L and so on. =item Background threads want speedy reporting Assume you want very exact timing, and you can spare an extra cpu core for that. Then you can run an extra thread that signals your perl interpreter. This means you can get a very exact timing source while your perl code is number crunching, without even using a syscall to communicate between your threads. For example the deliantra game server uses a variant of this technique to interrupt background processes regularly to send map updates to game clients. Or L uses an interrupt object to wake up perl when new events have arrived. L and L could also use this to speed up result reporting. =item Speedy event loop invocation One could use this module e.g. in L to interrupt a running coro-thread and cause it to enter the event loop. Or one could bind to C and tell some important sockets to send this signal, causing the event loop to be entered to reduce network latency. =back =head2 HOW TO USE You can use this module by creating an C object for each such event source. This object stores a perl and/or a C-level callback that is invoked when the C object gets signalled. It is executed at the next time the perl interpreter is running (i.e. it will interrupt a computation, but not an XS function or a syscall). You can signal the C object either by calling it's C<< ->signal >> method, or, more commonly, by calling a C function. There is also the built-in (POSIX) signal source. The C<< ->signal_func >> returns the address of the C function that is to be called (plus an argument to be used during the call). The signalling function also takes an integer argument in the range SIG_ATOMIC_MIN to SIG_ATOMIC_MAX (guaranteed to allow at least 0..127). Since this kind of interruption is fast, but can only interrupt a I interpreter, there is optional support for signalling a pipe - that means you can also wait for the pipe to become readable (e.g. via L or L). This, of course, incurs the overhead of a C and C syscall. =head1 USAGE EXAMPLES =head2 Implementing race-free signal handling This example uses a single event pipe for all signals, and one Async::Interrupt per signal. This code is actually what the L module uses itself when Async::Interrupt is available. First, create the event pipe and hook it into the event loop $SIGPIPE = new Async::Interrupt::EventPipe; $SIGPIPE_W = AnyEvent->io ( fh => $SIGPIPE->fileno, poll => "r", cb => \&_signal_check, # defined later ); Then, for each signal to hook, create an Async::Interrupt object. The callback just sets a global variable, as we are only interested in synchronous signals (i.e. when the event loop polls), which is why the pipe draining is not done automatically. my $interrupt = new Async::Interrupt cb => sub { undef $SIGNAL_RECEIVED{$signum} }, signal => $signum, pipe => [$SIGPIPE->filenos], pipe_autodrain => 0, ; Finally, the I/O callback for the event pipe handles the signals: sub _signal_check { # drain the pipe first $SIGPIPE->drain; # two loops, just to be sure while (%SIGNAL_RECEIVED) { for (keys %SIGNAL_RECEIVED) { delete $SIGNAL_RECEIVED{$_}; warn "signal $_ received\n"; } } } =head2 Interrupt perl from another thread This example interrupts the Perl interpreter from another thread, via the XS API. This is used by e.g. the L module. On the Perl level, a new loop object (which contains the thread) is created, by first calling some XS constructor, querying the C-level callback function and feeding that as the C into the Async::Interrupt constructor: my $self = XS_thread_constructor; my ($c_func, $c_arg) = _c_func $self; # return the c callback my $asy = new Async::Interrupt c_cb => [$c_func, $c_arg]; Then the newly created Interrupt object is queried for the signaling function that the newly created thread should call, and this is in turn told to the thread object: _attach $self, $asy->signal_func; So to repeat: first the XS object is created, then it is queried for the callback that should be called when the Interrupt object gets signalled. Then the interrupt object is queried for the callback function that the thread should call to signal the Interrupt object, and this callback is then attached to the thread. You have to be careful that your new thread is not signalling before the signal function was configured, for example by starting the background thread only within C<_attach>. That concludes the Perl part. The XS part consists of the actual constructor which creates a thread, which is not relevant for this example, and two functions, C<_c_func>, which returns the Perl-side callback, and C<_attach>, which configures the signalling functioon that is safe toc all from another thread. For simplicity, we will use global variables to store the functions, normally you would somehow attach them to C<$self>. The C simply returns the address of a static function and arranges for the object pointed to by C<$self> to be passed to it, as an integer: void _c_func (SV *loop) PPCODE: EXTEND (SP, 2); PUSHs (sv_2mortal (newSViv (PTR2IV (c_func)))); PUSHs (sv_2mortal (newSViv (SvRV (loop)))); This would be the callback (since it runs in a normal Perl context, it is permissible to manipulate Perl values): static void c_func (pTHX_ void *loop_, int value) { SV *loop_object = (SV *)loop_; ... } And this attaches the signalling callback: static void (*my_sig_func) (void *signal_arg, int value); static void *my_sig_arg; void _attach (SV *loop_, IV sig_func, void *sig_arg) CODE: { my_sig_func = sig_func; my_sig_arg = sig_arg; /* now run the thread */ thread_create (&u->tid, l_run, 0); } And C (the background thread) would eventually call the signaling function: my_sig_func (my_sig_arg, 0); You can have a look at L for an actual example using intra-thread communication, locking and so on. =head1 THE Async::Interrupt CLASS =over 4 =cut package Async::Interrupt; use common::sense; BEGIN { # the next line forces initialisation of internal # signal handling variables, otherwise, PL_sig_pending # etc. might be null pointers. $SIG{KILL} = sub { }; our $VERSION = 1.26; require XSLoader; XSLoader::load ("Async::Interrupt", $VERSION); } our $DIED = sub { warn "$@" }; =item $async = new Async::Interrupt key => value... Creates a new Async::Interrupt object. You may only use async notifications on this object while it exists, so you need to keep a reference to it at all times while it is used. Optional constructor arguments include (normally you would specify at least one of C or C). =over 4 =item cb => $coderef->($value) Registers a perl callback to be invoked whenever the async interrupt is signalled. Note that, since this callback can be invoked at basically any time, it must not modify any well-known global variables such as C<$/> without restoring them again before returning. The exceptions are C<$!> and C<$@>, which are saved and restored by Async::Interrupt. If the callback should throw an exception, then it will be caught, and C<$Async::Interrupt::DIED> will be called with C<$@> containing the exception. The default will simply C about the message and continue. =item c_cb => [$c_func, $c_arg] Registers a C callback the be invoked whenever the async interrupt is signalled. The C callback must have the following prototype: void c_func (pTHX_ void *c_arg, int value); Both C<$c_func> and C<$c_arg> must be specified as integers/IVs, and C<$value> is the C passed to some earlier call to either C<$signal> or the C function. Note that, because the callback can be invoked at almost any time, you have to be careful at saving and restoring global variables that Perl might use (the exception is C, which is saved and restored by Async::Interrupt). The callback itself runs as part of the perl context, so you can call any perl functions and modify any perl data structures (in which case the requirements set out for C apply as well). =item var => $scalar_ref When specified, then the given argument must be a reference to a scalar. The scalar will be set to C<0> initially. Signalling the interrupt object will set it to the passed value, handling the interrupt will reset it to C<0> again. Note that the only thing you are legally allowed to do is to is to check the variable in a boolean or integer context (e.g. comparing it with a string, or printing it, will I it and might cause your program to crash or worse). =item signal => $signame_or_value When this parameter is specified, then the Async::Interrupt will hook the given signal, that is, it will effectively call C<< ->signal (0) >> each time the given signal is caught by the process. Only one async can hook a given signal, and the signal will be restored to defaults when the Async::Interrupt object gets destroyed. =item signal_hysteresis => $boolean Sets the initial signal hysteresis state, see the C method, below. =item pipe => [$fileno_or_fh_for_reading, $fileno_or_fh_for_writing] Specifies two file descriptors (or file handles) that should be signalled whenever the async interrupt is signalled. This means a single octet will be written to it, and before the callback is being invoked, it will be read again. Due to races, it is unlikely but possible that multiple octets are written. It is required that the file handles are both in nonblocking mode. The object will keep a reference to the file handles. This can be used to ensure that async notifications will interrupt event frameworks as well. Note that C will create a suitable signal fd automatically when your program requests one, so you don't have to specify this argument when all you want is an extra file descriptor to watch. If you want to share a single event pipe between multiple Async::Interrupt objects, you can use the C class to manage those. =item pipe_autodrain => $boolean Sets the initial autodrain state, see the C method, below. =back =cut sub new { my ($class, %arg) = @_; my $self = bless \(_alloc $arg{cb}, @{$arg{c_cb}}[0,1], @{$arg{pipe}}[0,1], $arg{signal}, $arg{var}), $class; # urgs, reminds me of Event for my $attr (qw(pipe_autodrain signal_hysteresis)) { $self->$attr ($arg{$attr}) if exists $arg{$attr}; } $self } =item ($signal_func, $signal_arg) = $async->signal_func Returns the address of a function to call asynchronously. The function has the following prototype and needs to be passed the specified C<$signal_arg>, which is a C cast to C: void (*signal_func) (void *signal_arg, int value) An example call would look like: signal_func (signal_arg, 0); The function is safe to call from within signal and thread contexts, at any time. The specified C is passed to both C and Perl callback. C<$value> must be in the valid range for a C, except C<0> (1..127 is portable). If the function is called while the Async::Interrupt object is already signaled but before the callbacks are being executed, then the stored C is either the old or the new one. Due to the asynchronous nature of the code, the C can even be passed to two consecutive invocations of the callback. =item $address = $async->c_var Returns the address (cast to IV) of an C variable. The variable is set to C<0> initially and gets set to the passed value whenever the object gets signalled, and reset to C<0> once the interrupt has been handled. Note that it is often beneficial to just call C to handle any interrupts. Example: call some XS function to store the address, then show C code waiting for it. my_xs_func $async->c_var; static IV *valuep; void my_xs_func (void *addr) CODE: valuep = (IV *)addr; // code in a loop, waiting while (!*valuep) ; // do something =item $async->signal ($value=1) This signals the given async object from Perl code. Semi-obviously, this will instantly trigger the callback invocation (it does not, as the name might imply, do anything with POSIX signals). C<$value> must be in the valid range for a C, except C<0> (1..127 is portable). =item $async->handle Calls the callback if the object is pending. This method does not need to be called normally, as it will be invoked automatically. However, it can be used to force handling of outstanding interrupts while the object is blocked. One reason why one might want to do that is when you want to switch from asynchronous interruptions to synchronous one, using e.g. an event loop. To do that, one would first C<< $async->block >> the interrupt object, then register a read watcher on the C that calls C<< $async->handle >>. This disables asynchronous interruptions, but ensures that interrupts are handled by the event loop. =item $async->signal_hysteresis ($enable) Enables or disables signal hysteresis (default: disabled). If a POSIX signal is used as a signal source for the interrupt object, then enabling signal hysteresis causes Async::Interrupt to reset the signal action to C in the signal handler and restore it just before handling the interruption. When you expect a lot of signals (e.g. when using SIGIO), then enabling signal hysteresis can reduce the number of handler invocations considerably, at the cost of two extra syscalls. Note that setting the signal to C can have unintended side effects when you fork and exec other programs, as often they do not expect signals to be ignored by default. =item $async->block =item $async->unblock Sometimes you need a "critical section" of code that will not be interrupted by an Async::Interrupt. This can be implemented by calling C<< $async->block >> before the critical section, and C<< $async->unblock >> afterwards. Note that there must be exactly one call of C for every previous call to C (i.e. calls can nest). Since ensuring this in the presence of exceptions and threads is usually more difficult than you imagine, I recommend using C<< $async->scoped_block >> instead. =item $async->scope_block This call C<< $async->block >> and installs a handler that is called when the current scope is exited (via an exception, by canceling the Coro thread, by calling last/goto etc.). This is the recommended (and fastest) way to implement critical sections. =item ($block_func, $block_arg) = $async->scope_block_func Returns the address of a function that implements the C functionality. It has the following prototype and needs to be passed the specified C<$block_arg>, which is a C cast to C: void (*block_func) (void *block_arg) An example call would look like: block_func (block_arg); The function is safe to call only from within the toplevel of a perl XS function and will call C and C (in this order!). =item $async->pipe_enable =item $async->pipe_disable Enable/disable signalling the pipe when the interrupt occurs (default is enabled). Writing to a pipe is relatively expensive, so it can be disabled when you know you are not waiting for it (for example, with L you could disable the pipe in a check watcher, and enable it in a prepare watcher). Note that currently, while C is in effect, no attempt to read from the pipe will be done when handling events. This might change as soon as I realize why this is a mistake. =item $fileno = $async->pipe_fileno Returns the reading side of the signalling pipe. If no signalling pipe is currently attached to the object, it will dynamically create one. Note that the only valid operation on this file descriptor is to wait until it is readable. The fd might belong currently to a pipe, a tcp socket, or an eventfd, depending on the platform, and is guaranteed to be C