liblzma: Added crc32_clmul to crc32_fast.c.

1 files changed, 240 insertions, 7 deletions

diff --git a/src/liblzma/check/crc32_fast.c b/src/liblzma/check/crc32_fast.c
index fa53e99b..5229a4bd 100644
--- a/src/liblzma/check/crc32_fast.c
+++ b/src/liblzma/check/crc32_fast.c
@@ -3,13 +3,28 @@
 /// \file       crc32.c
 /// \brief      CRC32 calculation
 ///
-/// Calculate the CRC32 using the slice-by-eight algorithm.
+/// There are two methods in this file.
+/// crc32_generic uses the slice-by-eight algorithm.
 /// It is explained in this document:
 /// http://www.intel.com/technology/comms/perfnet/download/CRC_generators.pdf
 /// The code in this file is not the same as in Intel's paper, but
 /// the basic principle is identical.
+///
+/// crc32_clmul uses 32/64-bit x86 SSSE3, SSE4.1, and CLMUL instructions.
+/// It was derived from
+/// https://www.researchgate.net/publication/263424619_Fast_CRC_computation
+/// and the public domain code from https://github.com/rawrunprotected/crc
+/// (URLs were checked on 2023-09-29).
+///
+/// FIXME: Builds for 32-bit x86 use crc32_x86.S by default instead
+/// of this file and thus CLMUL version isn't available on 32-bit x86
+/// unless configured with --disable-assembler. Even then the lookup table
+/// isn't omitted in crc32_table.c since it doesn't know that assembly
+/// code has been disabled.
 //
-//  Author:     Lasse Collin
+//  Authors:    Lasse Collin
+//              Ilya Kurdyukov
+//              Hans Jansen
 //
 //  This file has been put into the public domain.
 //  You can do whatever you want with this file.
@@ -19,12 +34,14 @@
 #include "check.h"
 #include "crc_common.h"
 
+///////////////////
+// Generic CRC32 //
+///////////////////
+#ifdef CRC_GENERIC
 
-// If you make any changes, do some benchmarking! Seemingly unrelated
-// changes can very easily ruin the performance (and very probably is
-// very compiler dependent).
-extern LZMA_API(uint32_t)
-lzma_crc32(const uint8_t *buf, size_t size, uint32_t crc)
+
+static uint32_t
+crc32_generic(const uint8_t *buf, size_t size, uint32_t crc)
 {
 	crc = ~crc;
 
@@ -80,3 +97,219 @@ lzma_crc32(const uint8_t *buf, size_t size, uint32_t crc)
 
 	return ~crc;
 }
+#endif
+
+
+/////////////////////
+// x86 CLMUL CRC32 //
+/////////////////////
+
+#ifdef CRC_CLMUL
+
+#include <immintrin.h>
+
+
+/*
+// These functions were used to generate the constants
+// at the top of crc32_clmul().
+static uint64_t
+calc_lo(uint64_t p, uint64_t a, int n)
+{
+	uint64_t b = 0; int i;
+	for (i = 0; i < n; i++) {
+		b = b >> 1 | (a & 1) << (n - 1);
+		a = (a >> 1) ^ ((0 - (a & 1)) & p);
+	}
+	return b;
+}
+
+// same as ~crc(&a, sizeof(a), ~0)
+static uint64_t
+calc_hi(uint64_t p, uint64_t a, int n)
+{
+	int i;
+	for (i = 0; i < n; i++)
+		a = (a >> 1) ^ ((0 - (a & 1)) & p);
+	return a;
+}
+*/
+
+
+// MSVC (VS2015 - VS2022) produces bad 32-bit x86 code from the CLMUL CRC
+// code when optimizations are enabled (release build). According to the bug
+// report, the ebx register is corrupted and the calculated result is wrong.
+// Trying to workaround the problem with "__asm mov ebx, ebx" didn't help.
+// The following pragma works and performance is still good. x86-64 builds
+// aren't affected by this problem.
+//
+// NOTE: Another pragma after the function restores the optimizations.
+// If the #if condition here is updated, the other one must be updated too.
+#if defined(_MSC_VER) && !defined(__INTEL_COMPILER) && !defined(__clang__) \
+		&& defined(_M_IX86)
+#	pragma optimize("g", off)
+#endif
+
+// EDG-based compilers (Intel's classic compiler and compiler for E2K) can
+// define __GNUC__ but the attribute must not be used with them.
+// The new Clang-based ICX needs the attribute.
+//
+// NOTE: Build systems check for this too, keep them in sync with this.
+#if (defined(__GNUC__) || defined(__clang__)) && !defined(__EDG__)
+__attribute__((__target__("ssse3,sse4.1,pclmul")))
+#endif
+static uint32_t
+crc32_clmul(const uint8_t *buf, size_t size, uint32_t crc)
+{
+	// The prototypes of the intrinsics use signed types while most of
+	// the values are treated as unsigned here. These warnings in this
+	// function have been checked and found to be harmless so silence them.
+#if TUKLIB_GNUC_REQ(4, 6) || defined(__clang__)
+#	pragma GCC diagnostic push
+#	pragma GCC diagnostic ignored "-Wsign-conversion"
+#	pragma GCC diagnostic ignored "-Wconversion"
+#endif
+
+#ifndef CRC_USE_GENERIC_FOR_SMALL_INPUTS
+	// The code assumes that there is at least one byte of input.
+	if (size == 0)
+		return crc;
+#endif
+
+	// uint32_t poly = 0xedb88320;
+	uint64_t p = 0x1db710640; // p << 1
+	uint64_t mu = 0x1f7011641; // calc_lo(p, p, 32) << 1 | 1
+	uint64_t k5 = 0x163cd6124; // calc_hi(p, p, 32) << 1
+	uint64_t k4 = 0x0ccaa009e; // calc_hi(p, p, 64) << 1
+	uint64_t k3 = 0x1751997d0; // calc_hi(p, p, 128) << 1
+
+	__m128i vfold4 = _mm_set_epi64x(mu, p);
+	__m128i vfold8 = _mm_set_epi64x(0, k5);
+	__m128i vfold16 = _mm_set_epi64x(k4, k3);
+
+	__m128i v0, v1, v2;
+
+	crc_simd_body(buf,  size, &v0, &v1, vfold16, _mm_cvtsi32_si128(~crc));
+
+	v1 = _mm_xor_si128(
+			_mm_clmulepi64_si128(v0, vfold16, 0x10), v1); // xxx0
+	v2 = _mm_shuffle_epi32(v1, 0xe7); // 0xx0
+	v0 = _mm_slli_epi64(v1, 32);  // [0]
+	v0 = _mm_clmulepi64_si128(v0, vfold8, 0x00);
+	v0 = _mm_xor_si128(v0, v2);   // [1] [2]
+	v2 = _mm_clmulepi64_si128(v0, vfold4, 0x10);
+	v2 = _mm_clmulepi64_si128(v2, vfold4, 0x00);
+	v0 = _mm_xor_si128(v0, v2);   // [2]
+	return ~_mm_extract_epi32(v0, 2);
+
+#if TUKLIB_GNUC_REQ(4, 6) || defined(__clang__)
+#	pragma GCC diagnostic pop
+#endif
+}
+#if defined(_MSC_VER) && !defined(__INTEL_COMPILER) && !defined(__clang__) \
+		&& defined(_M_IX86)
+#	pragma optimize("", on)
+#endif
+#endif
+
+#if defined(CRC_GENERIC) && defined(CRC_CLMUL)
+typedef uint32_t (*crc32_func_type)(
+		const uint8_t *buf, size_t size, uint32_t crc);
+
+// Clang 16.0.0 and older has a bug where it marks the ifunc resolver
+// function as unused since it is static and never used outside of
+// __attribute__((__ifunc__())).
+#if defined(HAVE_FUNC_ATTRIBUTE_IFUNC) && defined(__clang__)
+#	pragma GCC diagnostic push
+#	pragma GCC diagnostic ignored "-Wunused-function"
+#endif
+
+static crc32_func_type
+crc32_resolve(void)
+{
+	return is_clmul_supported() ? &crc32_clmul : &crc32_generic;
+}
+
+#if defined(HAVE_FUNC_ATTRIBUTE_IFUNC) && defined(__clang__)
+#	pragma GCC diagnostic pop
+#endif
+
+#ifndef HAVE_FUNC_ATTRIBUTE_IFUNC
+
+#ifdef HAVE_FUNC_ATTRIBUTE_CONSTRUCTOR
+#	define CRC32_SET_FUNC_ATTR __attribute__((__constructor__))
+static crc32_func_type crc32_func;
+#else
+#	define CRC32_SET_FUNC_ATTR
+static uint32_t crc32_dispatch(const uint8_t *buf, size_t size, uint32_t crc);
+static crc32_func_type crc32_func = &crc32_dispatch;
+#endif
+
+CRC32_SET_FUNC_ATTR
+static void
+crc32_set_func(void)
+{
+	crc32_func = crc32_resolve();
+	return;
+}
+
+#ifndef HAVE_FUNC_ATTRIBUTE_CONSTRUCTOR
+static uint32_t
+crc32_dispatch(const uint8_t *buf, size_t size, uint32_t crc)
+{
+	// When __attribute__((__ifunc__(...))) and
+	// __attribute__((__constructor__)) isn't supported, set the
+	// function pointer without any locking. If multiple threads run
+	// the detection code in parallel, they will all end up setting
+	// the pointer to the same value. This avoids the use of
+	// mythread_once() on every call to lzma_crc32() but this likely
+	// isn't strictly standards compliant. Let's change it if it breaks.
+	crc32_set_func();
+	return crc32_func(buf, size, crc);
+}
+
+#endif
+#endif
+#endif
+
+
+#if defined(CRC_GENERIC) && defined(CRC_CLMUL) \
+		&& defined(HAVE_FUNC_ATTRIBUTE_IFUNC)
+extern LZMA_API(uint32_t)
+lzma_crc32(const uint8_t *buf, size_t size, uint32_t crc)
+		__attribute__((__ifunc__("crc32_resolve")));
+#else
+extern LZMA_API(uint32_t)
+lzma_crc32(const uint8_t *buf, size_t size, uint32_t crc)
+{
+#if defined(CRC_GENERIC) && defined(CRC_CLMUL)
+	// If CLMUL is available, it is the best for non-tiny inputs,
+	// being over twice as fast as the generic slice-by-four version.
+	// However, for size <= 16 it's different. In the extreme case
+	// of size == 1 the generic version can be five times faster.
+	// At size >= 8 the CLMUL starts to become reasonable. It
+	// varies depending on the alignment of buf too.
+	//
+	// The above doesn't include the overhead of mythread_once().
+	// At least on x86-64 GNU/Linux, pthread_once() is very fast but
+	// it still makes lzma_crc32(buf, 1, crc) 50-100 % slower. When
+	// size reaches 12-16 bytes the overhead becomes negligible.
+	//
+	// So using the generic version for size <= 16 may give better
+	// performance with tiny inputs but if such inputs happen rarely
+	// it's not so obvious because then the lookup table of the
+	// generic version may not be in the processor cache.
+#ifdef CRC_USE_GENERIC_FOR_SMALL_INPUTS
+	if (size <= 16)
+		return crc32_generic(buf, size, crc);
+#endif
+
+	return crc32_func(buf, size, crc);
+
+#elif defined(CRC_CLMUL)
+	return crc32_clmul(buf, size, crc);
+
+#else
+	return crc32_generic(buf, size, crc);
+#endif
+}
+#endif