src/runtime/internal/sys/intrinsics_common.go

// 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.

const len8tab = "" +
	"\x00\x01\x02\x02\x03\x03\x03\x03\x04\x04\x04\x04\x04\x04\x04\x04" +
	"\x05\x05\x05\x05\x05\x05\x05\x05\x05\x05\x05\x05\x05\x05\x05\x05" +
	"\x06\x06\x06\x06\x06\x06\x06\x06\x06\x06\x06\x06\x06\x06\x06\x06" +
	"\x06\x06\x06\x06\x06\x06\x06\x06\x06\x06\x06\x06\x06\x06\x06\x06" +
	"\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07" +
	"\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07" +
	"\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07" +
	"\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07\x07" +
	"\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08" +
	"\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08" +
	"\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08" +
	"\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08" +
	"\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08" +
	"\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08" +
	"\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08" +
	"\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08\x08"

// Len64 returns the minimum number of bits required to represent x; the result is 0 for x == 0.
//
// nosplit because this is used in src/runtime/histogram.go, which make run in sensitive contexts.
//
//go:nosplit
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)
}

// 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) }

// 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])
}

// 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
}

// 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) {}