diff options
author | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-28 13:16:40 +0000 |
---|---|---|
committer | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-28 13:16:40 +0000 |
commit | 47ab3d4a42e9ab51c465c4322d2ec233f6324e6b (patch) | |
tree | a61a0ffd83f4a3def4b36e5c8e99630c559aa723 /src/runtime/internal/sys | |
parent | Initial commit. (diff) | |
download | golang-1.18-47ab3d4a42e9ab51c465c4322d2ec233f6324e6b.tar.xz golang-1.18-47ab3d4a42e9ab51c465c4322d2ec233f6324e6b.zip |
Adding upstream version 1.18.10.upstream/1.18.10upstream
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'src/runtime/internal/sys')
-rw-r--r-- | src/runtime/internal/sys/consts.go | 34 | ||||
-rw-r--r-- | src/runtime/internal/sys/intrinsics.go | 91 | ||||
-rw-r--r-- | src/runtime/internal/sys/intrinsics_386.s | 58 | ||||
-rw-r--r-- | src/runtime/internal/sys/intrinsics_common.go | 158 | ||||
-rw-r--r-- | src/runtime/internal/sys/intrinsics_stubs.go | 13 | ||||
-rw-r--r-- | src/runtime/internal/sys/intrinsics_test.go | 38 | ||||
-rw-r--r-- | src/runtime/internal/sys/sys.go | 7 |
7 files changed, 399 insertions, 0 deletions
diff --git a/src/runtime/internal/sys/consts.go b/src/runtime/internal/sys/consts.go new file mode 100644 index 0000000..fffcf81 --- /dev/null +++ b/src/runtime/internal/sys/consts.go @@ -0,0 +1,34 @@ +// Copyright 2014 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package sys + +import ( + "internal/goarch" + "internal/goos" +) + +// AIX requires a larger stack for syscalls. +const StackGuardMultiplier = StackGuardMultiplierDefault*(1-goos.IsAix) + 2*goos.IsAix + +// DefaultPhysPageSize is the default physical page size. +const DefaultPhysPageSize = goarch.DefaultPhysPageSize + +// PCQuantum is the minimal unit for a program counter (1 on x86, 4 on most other systems). +// The various PC tables record PC deltas pre-divided by PCQuantum. +const PCQuantum = goarch.PCQuantum + +// Int64Align is the required alignment for a 64-bit integer (4 on 32-bit systems, 8 on 64-bit). +const Int64Align = goarch.PtrSize + +// MinFrameSize is the size of the system-reserved words at the bottom +// of a frame (just above the architectural stack pointer). +// It is zero on x86 and PtrSize on most non-x86 (LR-based) systems. +// On PowerPC it is larger, to cover three more reserved words: +// the compiler word, the link editor word, and the TOC save word. +const MinFrameSize = goarch.MinFrameSize + +// StackAlign is the required alignment of the SP register. +// The stack must be at least word aligned, but some architectures require more. +const StackAlign = goarch.StackAlign diff --git a/src/runtime/internal/sys/intrinsics.go b/src/runtime/internal/sys/intrinsics.go new file mode 100644 index 0000000..5af4901 --- /dev/null +++ b/src/runtime/internal/sys/intrinsics.go @@ -0,0 +1,91 @@ +// Copyright 2016 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +//go:build !386 + +// TODO finish intrinsifying 386, deadcode the assembly, remove build tags, merge w/ intrinsics_common +// TODO replace all uses of CtzXX with TrailingZerosXX; they are the same. + +package sys + +// Using techniques from http://supertech.csail.mit.edu/papers/debruijn.pdf + +const deBruijn64ctz = 0x0218a392cd3d5dbf + +var deBruijnIdx64ctz = [64]byte{ + 0, 1, 2, 7, 3, 13, 8, 19, + 4, 25, 14, 28, 9, 34, 20, 40, + 5, 17, 26, 38, 15, 46, 29, 48, + 10, 31, 35, 54, 21, 50, 41, 57, + 63, 6, 12, 18, 24, 27, 33, 39, + 16, 37, 45, 47, 30, 53, 49, 56, + 62, 11, 23, 32, 36, 44, 52, 55, + 61, 22, 43, 51, 60, 42, 59, 58, +} + +const deBruijn32ctz = 0x04653adf + +var deBruijnIdx32ctz = [32]byte{ + 0, 1, 2, 6, 3, 11, 7, 16, + 4, 14, 12, 21, 8, 23, 17, 26, + 31, 5, 10, 15, 13, 20, 22, 25, + 30, 9, 19, 24, 29, 18, 28, 27, +} + +// Ctz64 counts trailing (low-order) zeroes, +// and if all are zero, then 64. +func Ctz64(x uint64) int { + x &= -x // isolate low-order bit + y := x * deBruijn64ctz >> 58 // extract part of deBruijn sequence + i := int(deBruijnIdx64ctz[y]) // convert to bit index + z := int((x - 1) >> 57 & 64) // adjustment if zero + return i + z +} + +// Ctz32 counts trailing (low-order) zeroes, +// and if all are zero, then 32. +func Ctz32(x uint32) int { + x &= -x // isolate low-order bit + y := x * deBruijn32ctz >> 27 // extract part of deBruijn sequence + i := int(deBruijnIdx32ctz[y]) // convert to bit index + z := int((x - 1) >> 26 & 32) // adjustment if zero + return i + z +} + +// Ctz8 returns the number of trailing zero bits in x; the result is 8 for x == 0. +func Ctz8(x uint8) int { + return int(ntz8tab[x]) +} + +// Bswap64 returns its input with byte order reversed +// 0x0102030405060708 -> 0x0807060504030201 +func Bswap64(x uint64) uint64 { + c8 := uint64(0x00ff00ff00ff00ff) + a := x >> 8 & c8 + b := (x & c8) << 8 + x = a | b + c16 := uint64(0x0000ffff0000ffff) + a = x >> 16 & c16 + b = (x & c16) << 16 + x = a | b + c32 := uint64(0x00000000ffffffff) + a = x >> 32 & c32 + b = (x & c32) << 32 + x = a | b + return x +} + +// Bswap32 returns its input with byte order reversed +// 0x01020304 -> 0x04030201 +func Bswap32(x uint32) uint32 { + c8 := uint32(0x00ff00ff) + a := x >> 8 & c8 + b := (x & c8) << 8 + x = a | b + c16 := uint32(0x0000ffff) + a = x >> 16 & c16 + b = (x & c16) << 16 + x = a | b + return x +} diff --git a/src/runtime/internal/sys/intrinsics_386.s b/src/runtime/internal/sys/intrinsics_386.s new file mode 100644 index 0000000..784b246 --- /dev/null +++ b/src/runtime/internal/sys/intrinsics_386.s @@ -0,0 +1,58 @@ +// Copyright 2016 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +#include "textflag.h" + +TEXT runtime∕internal∕sys·Ctz64(SB), NOSPLIT, $0-12 + // Try low 32 bits. + MOVL x_lo+0(FP), AX + BSFL AX, AX + JZ tryhigh + MOVL AX, ret+8(FP) + RET + +tryhigh: + // Try high 32 bits. + MOVL x_hi+4(FP), AX + BSFL AX, AX + JZ none + ADDL $32, AX + MOVL AX, ret+8(FP) + RET + +none: + // No bits are set. + MOVL $64, ret+8(FP) + RET + +TEXT runtime∕internal∕sys·Ctz32(SB), NOSPLIT, $0-8 + MOVL x+0(FP), AX + BSFL AX, AX + JNZ 2(PC) + MOVL $32, AX + MOVL AX, ret+4(FP) + RET + +TEXT runtime∕internal∕sys·Ctz8(SB), NOSPLIT, $0-8 + MOVBLZX x+0(FP), AX + BSFL AX, AX + JNZ 2(PC) + MOVL $8, AX + MOVL AX, ret+4(FP) + RET + +TEXT runtime∕internal∕sys·Bswap64(SB), NOSPLIT, $0-16 + MOVL x_lo+0(FP), AX + MOVL x_hi+4(FP), BX + BSWAPL AX + BSWAPL BX + MOVL BX, ret_lo+8(FP) + MOVL AX, ret_hi+12(FP) + RET + +TEXT runtime∕internal∕sys·Bswap32(SB), NOSPLIT, $0-8 + MOVL x+0(FP), AX + BSWAPL AX + MOVL AX, ret+4(FP) + RET diff --git a/src/runtime/internal/sys/intrinsics_common.go b/src/runtime/internal/sys/intrinsics_common.go new file mode 100644 index 0000000..48d9759 --- /dev/null +++ b/src/runtime/internal/sys/intrinsics_common.go @@ -0,0 +1,158 @@ +// Copyright 2019 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +package sys + +// Copied from math/bits to avoid dependence. + +var len8tab = [256]uint8{ + 0x00, 0x01, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, + 0x05, 0x05, 0x05, 0x05, 0x05, 0x05, 0x05, 0x05, 0x05, 0x05, 0x05, 0x05, 0x05, 0x05, 0x05, 0x05, + 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, + 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, + 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, + 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, + 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, + 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, + 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, + 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, + 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, + 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, + 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, + 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, + 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, + 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, +} + +var ntz8tab = [256]uint8{ + 0x08, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, 0x03, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, + 0x04, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, 0x03, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, + 0x05, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, 0x03, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, + 0x04, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, 0x03, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, + 0x06, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, 0x03, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, + 0x04, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, 0x03, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, + 0x05, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, 0x03, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, + 0x04, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, 0x03, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, + 0x07, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, 0x03, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, + 0x04, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, 0x03, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, + 0x05, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, 0x03, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, + 0x04, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, 0x03, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, + 0x06, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, 0x03, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, + 0x04, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, 0x03, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, + 0x05, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, 0x03, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, + 0x04, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, 0x03, 0x00, 0x01, 0x00, 0x02, 0x00, 0x01, 0x00, +} + +// len64 returns the minimum number of bits required to represent x; the result is 0 for x == 0. +func Len64(x uint64) (n int) { + if x >= 1<<32 { + x >>= 32 + n = 32 + } + if x >= 1<<16 { + x >>= 16 + n += 16 + } + if x >= 1<<8 { + x >>= 8 + n += 8 + } + return n + int(len8tab[x]) +} + +// --- OnesCount --- + +const m0 = 0x5555555555555555 // 01010101 ... +const m1 = 0x3333333333333333 // 00110011 ... +const m2 = 0x0f0f0f0f0f0f0f0f // 00001111 ... + +// OnesCount64 returns the number of one bits ("population count") in x. +func OnesCount64(x uint64) int { + // Implementation: Parallel summing of adjacent bits. + // See "Hacker's Delight", Chap. 5: Counting Bits. + // The following pattern shows the general approach: + // + // x = x>>1&(m0&m) + x&(m0&m) + // x = x>>2&(m1&m) + x&(m1&m) + // x = x>>4&(m2&m) + x&(m2&m) + // x = x>>8&(m3&m) + x&(m3&m) + // x = x>>16&(m4&m) + x&(m4&m) + // x = x>>32&(m5&m) + x&(m5&m) + // return int(x) + // + // Masking (& operations) can be left away when there's no + // danger that a field's sum will carry over into the next + // field: Since the result cannot be > 64, 8 bits is enough + // and we can ignore the masks for the shifts by 8 and up. + // Per "Hacker's Delight", the first line can be simplified + // more, but it saves at best one instruction, so we leave + // it alone for clarity. + const m = 1<<64 - 1 + x = x>>1&(m0&m) + x&(m0&m) + x = x>>2&(m1&m) + x&(m1&m) + x = (x>>4 + x) & (m2 & m) + x += x >> 8 + x += x >> 16 + x += x >> 32 + return int(x) & (1<<7 - 1) +} + +var deBruijn64tab = [64]byte{ + 0, 1, 56, 2, 57, 49, 28, 3, 61, 58, 42, 50, 38, 29, 17, 4, + 62, 47, 59, 36, 45, 43, 51, 22, 53, 39, 33, 30, 24, 18, 12, 5, + 63, 55, 48, 27, 60, 41, 37, 16, 46, 35, 44, 21, 52, 32, 23, 11, + 54, 26, 40, 15, 34, 20, 31, 10, 25, 14, 19, 9, 13, 8, 7, 6, +} + +const deBruijn64 = 0x03f79d71b4ca8b09 + +// TrailingZeros64 returns the number of trailing zero bits in x; the result is 64 for x == 0. +func TrailingZeros64(x uint64) int { + if x == 0 { + return 64 + } + // If popcount is fast, replace code below with return popcount(^x & (x - 1)). + // + // x & -x leaves only the right-most bit set in the word. Let k be the + // index of that bit. Since only a single bit is set, the value is two + // to the power of k. Multiplying by a power of two is equivalent to + // left shifting, in this case by k bits. The de Bruijn (64 bit) constant + // is such that all six bit, consecutive substrings are distinct. + // Therefore, if we have a left shifted version of this constant we can + // find by how many bits it was shifted by looking at which six bit + // substring ended up at the top of the word. + // (Knuth, volume 4, section 7.3.1) + return int(deBruijn64tab[(x&-x)*deBruijn64>>(64-6)]) +} + +// LeadingZeros64 returns the number of leading zero bits in x; the result is 64 for x == 0. +func LeadingZeros64(x uint64) int { return 64 - Len64(x) } + +// LeadingZeros8 returns the number of leading zero bits in x; the result is 8 for x == 0. +func LeadingZeros8(x uint8) int { return 8 - Len8(x) } + +// TrailingZeros8 returns the number of trailing zero bits in x; the result is 8 for x == 0. +func TrailingZeros8(x uint8) int { + return int(ntz8tab[x]) +} + +// Len8 returns the minimum number of bits required to represent x; the result is 0 for x == 0. +func Len8(x uint8) int { + return int(len8tab[x]) +} + +// Prefetch prefetches data from memory addr to cache +// +// AMD64: Produce PREFETCHT0 instruction +// +// ARM64: Produce PRFM instruction with PLDL1KEEP option +func Prefetch(addr uintptr) {} + +// PrefetchStreamed prefetches data from memory addr, with a hint that this data is being streamed. +// That is, it is likely to be accessed very soon, but only once. If possible, this will avoid polluting the cache. +// +// AMD64: Produce PREFETCHNTA instruction +// +// ARM64: Produce PRFM instruction with PLDL1STRM option +func PrefetchStreamed(addr uintptr) {} diff --git a/src/runtime/internal/sys/intrinsics_stubs.go b/src/runtime/internal/sys/intrinsics_stubs.go new file mode 100644 index 0000000..a020652 --- /dev/null +++ b/src/runtime/internal/sys/intrinsics_stubs.go @@ -0,0 +1,13 @@ +// Copyright 2016 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +//go:build 386 + +package sys + +func Ctz64(x uint64) int +func Ctz32(x uint32) int +func Ctz8(x uint8) int +func Bswap64(x uint64) uint64 +func Bswap32(x uint32) uint32 diff --git a/src/runtime/internal/sys/intrinsics_test.go b/src/runtime/internal/sys/intrinsics_test.go new file mode 100644 index 0000000..0444183 --- /dev/null +++ b/src/runtime/internal/sys/intrinsics_test.go @@ -0,0 +1,38 @@ +package sys_test + +import ( + "runtime/internal/sys" + "testing" +) + +func TestCtz64(t *testing.T) { + for i := 0; i <= 64; i++ { + x := uint64(5) << uint(i) + if got := sys.Ctz64(x); got != i { + t.Errorf("Ctz64(%d)=%d, want %d", x, got, i) + } + } +} +func TestCtz32(t *testing.T) { + for i := 0; i <= 32; i++ { + x := uint32(5) << uint(i) + if got := sys.Ctz32(x); got != i { + t.Errorf("Ctz32(%d)=%d, want %d", x, got, i) + } + } +} + +func TestBswap64(t *testing.T) { + x := uint64(0x1122334455667788) + y := sys.Bswap64(x) + if y != 0x8877665544332211 { + t.Errorf("Bswap(%x)=%x, want 0x8877665544332211", x, y) + } +} +func TestBswap32(t *testing.T) { + x := uint32(0x11223344) + y := sys.Bswap32(x) + if y != 0x44332211 { + t.Errorf("Bswap(%x)=%x, want 0x44332211", x, y) + } +} diff --git a/src/runtime/internal/sys/sys.go b/src/runtime/internal/sys/sys.go new file mode 100644 index 0000000..694101d --- /dev/null +++ b/src/runtime/internal/sys/sys.go @@ -0,0 +1,7 @@ +// Copyright 2015 The Go Authors. All rights reserved. +// Use of this source code is governed by a BSD-style +// license that can be found in the LICENSE file. + +// package sys contains system- and configuration- and architecture-specific +// constants used by the runtime. +package sys |