vendor: update buildkit to 3e38a2d
Signed-off-by: CrazyMax <crazy-max@users.noreply.github.com>pull/1041/head
CrazyMax
3 years ago
parent
c3db06cda0
commit
a49ad031a5
@ -1,65 +1,23 @@
|
||||
package system // import "github.com/docker/docker/pkg/system"
|
||||
|
||||
import (
|
||||
"unsafe"
|
||||
|
||||
"github.com/sirupsen/logrus"
|
||||
"golang.org/x/sys/windows"
|
||||
)
|
||||
import "golang.org/x/sys/windows"
|
||||
|
||||
const (
|
||||
// Deprecated: use github.com/docker/pkg/idtools.SeTakeOwnershipPrivilege
|
||||
SeTakeOwnershipPrivilege = "SeTakeOwnershipPrivilege"
|
||||
)
|
||||
|
||||
const (
|
||||
// Deprecated: use github.com/docker/pkg/idtools.ContainerAdministratorSidString
|
||||
ContainerAdministratorSidString = "S-1-5-93-2-1"
|
||||
// Deprecated: use github.com/docker/pkg/idtools.ContainerUserSidString
|
||||
ContainerUserSidString = "S-1-5-93-2-2"
|
||||
)
|
||||
|
||||
var (
|
||||
ntuserApiset = windows.NewLazyDLL("ext-ms-win-ntuser-window-l1-1-0")
|
||||
procGetVersionExW = modkernel32.NewProc("GetVersionExW")
|
||||
)
|
||||
|
||||
// https://docs.microsoft.com/en-us/windows/win32/api/winnt/ns-winnt-osversioninfoexa
|
||||
// TODO: use golang.org/x/sys/windows.OsVersionInfoEx (needs OSVersionInfoSize to be exported)
|
||||
type osVersionInfoEx struct {
|
||||
OSVersionInfoSize uint32
|
||||
MajorVersion uint32
|
||||
MinorVersion uint32
|
||||
BuildNumber uint32
|
||||
PlatformID uint32
|
||||
CSDVersion [128]uint16
|
||||
ServicePackMajor uint16
|
||||
ServicePackMinor uint16
|
||||
SuiteMask uint16
|
||||
ProductType byte
|
||||
Reserve byte
|
||||
}
|
||||
// VER_NT_WORKSTATION, see https://docs.microsoft.com/en-us/windows/win32/api/winnt/ns-winnt-osversioninfoexa
|
||||
const verNTWorkstation = 0x00000001 // VER_NT_WORKSTATION
|
||||
|
||||
// IsWindowsClient returns true if the SKU is client. It returns false on
|
||||
// Windows server, or if an error occurred when making the GetVersionExW
|
||||
// syscall.
|
||||
func IsWindowsClient() bool {
|
||||
osviex := &osVersionInfoEx{OSVersionInfoSize: 284}
|
||||
r1, _, err := procGetVersionExW.Call(uintptr(unsafe.Pointer(osviex)))
|
||||
if r1 == 0 {
|
||||
logrus.WithError(err).Warn("GetVersionExW failed - assuming server SKU")
|
||||
return false
|
||||
}
|
||||
// VER_NT_WORKSTATION, see https://docs.microsoft.com/en-us/windows/win32/api/winnt/ns-winnt-osversioninfoexa
|
||||
const verNTWorkstation = 0x00000001 // VER_NT_WORKSTATION
|
||||
return osviex.ProductType == verNTWorkstation
|
||||
}
|
||||
|
||||
// HasWin32KSupport determines whether containers that depend on win32k can
|
||||
// run on this machine. Win32k is the driver used to implement windowing.
|
||||
func HasWin32KSupport() bool {
|
||||
// For now, check for ntuser API support on the host. In the future, a host
|
||||
// may support win32k in containers even if the host does not support ntuser
|
||||
// APIs.
|
||||
return ntuserApiset.Load() == nil
|
||||
ver := windows.RtlGetVersion()
|
||||
return ver != nil && ver.ProductType == verNTWorkstation
|
||||
}
|
||||
|
||||
@ -0,0 +1,5 @@
|
||||
package huff0
|
||||
|
||||
//go:generate go run generate.go
|
||||
//go:generate asmfmt -w decompress_amd64.s
|
||||
//go:generate asmfmt -w decompress_8b_amd64.s
|
||||
@ -0,0 +1,488 @@
|
||||
// +build !appengine
|
||||
// +build gc
|
||||
// +build !noasm
|
||||
|
||||
#include "textflag.h"
|
||||
#include "funcdata.h"
|
||||
#include "go_asm.h"
|
||||
|
||||
#define bufoff 256 // see decompress.go, we're using [4][256]byte table
|
||||
|
||||
// func decompress4x_main_loop_x86(pbr0, pbr1, pbr2, pbr3 *bitReaderShifted,
|
||||
// peekBits uint8, buf *byte, tbl *dEntrySingle) (int, bool)
|
||||
TEXT ·decompress4x_8b_loop_x86(SB), NOSPLIT, $8
|
||||
#define off R8
|
||||
#define buffer DI
|
||||
#define table SI
|
||||
|
||||
#define br_bits_read R9
|
||||
#define br_value R10
|
||||
#define br_offset R11
|
||||
#define peek_bits R12
|
||||
#define exhausted DX
|
||||
|
||||
#define br0 R13
|
||||
#define br1 R14
|
||||
#define br2 R15
|
||||
#define br3 BP
|
||||
|
||||
MOVQ BP, 0(SP)
|
||||
|
||||
XORQ exhausted, exhausted // exhausted = false
|
||||
XORQ off, off // off = 0
|
||||
|
||||
MOVBQZX peekBits+32(FP), peek_bits
|
||||
MOVQ buf+40(FP), buffer
|
||||
MOVQ tbl+48(FP), table
|
||||
|
||||
MOVQ pbr0+0(FP), br0
|
||||
MOVQ pbr1+8(FP), br1
|
||||
MOVQ pbr2+16(FP), br2
|
||||
MOVQ pbr3+24(FP), br3
|
||||
|
||||
main_loop:
|
||||
|
||||
// const stream = 0
|
||||
// br0.fillFast()
|
||||
MOVBQZX bitReaderShifted_bitsRead(br0), br_bits_read
|
||||
MOVQ bitReaderShifted_value(br0), br_value
|
||||
MOVQ bitReaderShifted_off(br0), br_offset
|
||||
|
||||
// if b.bitsRead >= 32 {
|
||||
CMPQ br_bits_read, $32
|
||||
JB skip_fill0
|
||||
|
||||
SUBQ $32, br_bits_read // b.bitsRead -= 32
|
||||
SUBQ $4, br_offset // b.off -= 4
|
||||
|
||||
// v := b.in[b.off-4 : b.off]
|
||||
// v = v[:4]
|
||||
// low := (uint32(v[0])) | (uint32(v[1]) << 8) | (uint32(v[2]) << 16) | (uint32(v[3]) << 24)
|
||||
MOVQ bitReaderShifted_in(br0), AX
|
||||
MOVL 0(br_offset)(AX*1), AX // AX = uint32(b.in[b.off:b.off+4])
|
||||
|
||||
// b.value |= uint64(low) << (b.bitsRead & 63)
|
||||
MOVQ br_bits_read, CX
|
||||
SHLQ CL, AX
|
||||
ORQ AX, br_value
|
||||
|
||||
// exhausted = exhausted || (br0.off < 4)
|
||||
CMPQ br_offset, $4
|
||||
SETLT DL
|
||||
ORB DL, DH
|
||||
|
||||
// }
|
||||
skip_fill0:
|
||||
|
||||
// val0 := br0.peekTopBits(peekBits)
|
||||
MOVQ br_value, AX
|
||||
MOVQ peek_bits, CX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
// v0 := table[val0&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v0
|
||||
|
||||
// br0.advance(uint8(v0.entry))
|
||||
MOVB AH, BL // BL = uint8(v0.entry >> 8)
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CL, br_value // value <<= n
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// val1 := br0.peekTopBits(peekBits)
|
||||
MOVQ peek_bits, CX
|
||||
MOVQ br_value, AX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
// v1 := table[val1&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v1
|
||||
|
||||
// br0.advance(uint8(v1.entry))
|
||||
MOVB AH, BH // BH = uint8(v1.entry >> 8)
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CX, br_value // value <<= n
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// these two writes get coalesced
|
||||
// buf[stream][off] = uint8(v0.entry >> 8)
|
||||
// buf[stream][off+1] = uint8(v1.entry >> 8)
|
||||
MOVW BX, 0(buffer)(off*1)
|
||||
|
||||
// SECOND PART:
|
||||
// val2 := br0.peekTopBits(peekBits)
|
||||
MOVQ br_value, AX
|
||||
MOVQ peek_bits, CX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
// v2 := table[val0&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v0
|
||||
|
||||
// br0.advance(uint8(v0.entry))
|
||||
MOVB AH, BL // BL = uint8(v0.entry >> 8)
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CL, br_value // value <<= n
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// val3 := br0.peekTopBits(peekBits)
|
||||
MOVQ peek_bits, CX
|
||||
MOVQ br_value, AX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
// v3 := table[val1&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v1
|
||||
|
||||
// br0.advance(uint8(v1.entry))
|
||||
MOVB AH, BH // BH = uint8(v1.entry >> 8)
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CX, br_value // value <<= n
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// these two writes get coalesced
|
||||
// buf[stream][off+2] = uint8(v2.entry >> 8)
|
||||
// buf[stream][off+3] = uint8(v3.entry >> 8)
|
||||
MOVW BX, 0+2(buffer)(off*1)
|
||||
|
||||
// update the bitrader reader structure
|
||||
MOVB br_bits_read, bitReaderShifted_bitsRead(br0)
|
||||
MOVQ br_value, bitReaderShifted_value(br0)
|
||||
MOVQ br_offset, bitReaderShifted_off(br0)
|
||||
|
||||
// const stream = 1
|
||||
// br1.fillFast()
|
||||
MOVBQZX bitReaderShifted_bitsRead(br1), br_bits_read
|
||||
MOVQ bitReaderShifted_value(br1), br_value
|
||||
MOVQ bitReaderShifted_off(br1), br_offset
|
||||
|
||||
// if b.bitsRead >= 32 {
|
||||
CMPQ br_bits_read, $32
|
||||
JB skip_fill1
|
||||
|
||||
SUBQ $32, br_bits_read // b.bitsRead -= 32
|
||||
SUBQ $4, br_offset // b.off -= 4
|
||||
|
||||
// v := b.in[b.off-4 : b.off]
|
||||
// v = v[:4]
|
||||
// low := (uint32(v[0])) | (uint32(v[1]) << 8) | (uint32(v[2]) << 16) | (uint32(v[3]) << 24)
|
||||
MOVQ bitReaderShifted_in(br1), AX
|
||||
MOVL 0(br_offset)(AX*1), AX // AX = uint32(b.in[b.off:b.off+4])
|
||||
|
||||
// b.value |= uint64(low) << (b.bitsRead & 63)
|
||||
MOVQ br_bits_read, CX
|
||||
SHLQ CL, AX
|
||||
ORQ AX, br_value
|
||||
|
||||
// exhausted = exhausted || (br1.off < 4)
|
||||
CMPQ br_offset, $4
|
||||
SETLT DL
|
||||
ORB DL, DH
|
||||
|
||||
// }
|
||||
skip_fill1:
|
||||
|
||||
// val0 := br1.peekTopBits(peekBits)
|
||||
MOVQ br_value, AX
|
||||
MOVQ peek_bits, CX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
// v0 := table[val0&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v0
|
||||
|
||||
// br1.advance(uint8(v0.entry))
|
||||
MOVB AH, BL // BL = uint8(v0.entry >> 8)
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CL, br_value // value <<= n
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// val1 := br1.peekTopBits(peekBits)
|
||||
MOVQ peek_bits, CX
|
||||
MOVQ br_value, AX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
// v1 := table[val1&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v1
|
||||
|
||||
// br1.advance(uint8(v1.entry))
|
||||
MOVB AH, BH // BH = uint8(v1.entry >> 8)
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CX, br_value // value <<= n
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// these two writes get coalesced
|
||||
// buf[stream][off] = uint8(v0.entry >> 8)
|
||||
// buf[stream][off+1] = uint8(v1.entry >> 8)
|
||||
MOVW BX, 256(buffer)(off*1)
|
||||
|
||||
// SECOND PART:
|
||||
// val2 := br1.peekTopBits(peekBits)
|
||||
MOVQ br_value, AX
|
||||
MOVQ peek_bits, CX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
// v2 := table[val0&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v0
|
||||
|
||||
// br1.advance(uint8(v0.entry))
|
||||
MOVB AH, BL // BL = uint8(v0.entry >> 8)
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CL, br_value // value <<= n
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// val3 := br1.peekTopBits(peekBits)
|
||||
MOVQ peek_bits, CX
|
||||
MOVQ br_value, AX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
// v3 := table[val1&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v1
|
||||
|
||||
// br1.advance(uint8(v1.entry))
|
||||
MOVB AH, BH // BH = uint8(v1.entry >> 8)
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CX, br_value // value <<= n
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// these two writes get coalesced
|
||||
// buf[stream][off+2] = uint8(v2.entry >> 8)
|
||||
// buf[stream][off+3] = uint8(v3.entry >> 8)
|
||||
MOVW BX, 256+2(buffer)(off*1)
|
||||
|
||||
// update the bitrader reader structure
|
||||
MOVB br_bits_read, bitReaderShifted_bitsRead(br1)
|
||||
MOVQ br_value, bitReaderShifted_value(br1)
|
||||
MOVQ br_offset, bitReaderShifted_off(br1)
|
||||
|
||||
// const stream = 2
|
||||
// br2.fillFast()
|
||||
MOVBQZX bitReaderShifted_bitsRead(br2), br_bits_read
|
||||
MOVQ bitReaderShifted_value(br2), br_value
|
||||
MOVQ bitReaderShifted_off(br2), br_offset
|
||||
|
||||
// if b.bitsRead >= 32 {
|
||||
CMPQ br_bits_read, $32
|
||||
JB skip_fill2
|
||||
|
||||
SUBQ $32, br_bits_read // b.bitsRead -= 32
|
||||
SUBQ $4, br_offset // b.off -= 4
|
||||
|
||||
// v := b.in[b.off-4 : b.off]
|
||||
// v = v[:4]
|
||||
// low := (uint32(v[0])) | (uint32(v[1]) << 8) | (uint32(v[2]) << 16) | (uint32(v[3]) << 24)
|
||||
MOVQ bitReaderShifted_in(br2), AX
|
||||
MOVL 0(br_offset)(AX*1), AX // AX = uint32(b.in[b.off:b.off+4])
|
||||
|
||||
// b.value |= uint64(low) << (b.bitsRead & 63)
|
||||
MOVQ br_bits_read, CX
|
||||
SHLQ CL, AX
|
||||
ORQ AX, br_value
|
||||
|
||||
// exhausted = exhausted || (br2.off < 4)
|
||||
CMPQ br_offset, $4
|
||||
SETLT DL
|
||||
ORB DL, DH
|
||||
|
||||
// }
|
||||
skip_fill2:
|
||||
|
||||
// val0 := br2.peekTopBits(peekBits)
|
||||
MOVQ br_value, AX
|
||||
MOVQ peek_bits, CX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
// v0 := table[val0&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v0
|
||||
|
||||
// br2.advance(uint8(v0.entry))
|
||||
MOVB AH, BL // BL = uint8(v0.entry >> 8)
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CL, br_value // value <<= n
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// val1 := br2.peekTopBits(peekBits)
|
||||
MOVQ peek_bits, CX
|
||||
MOVQ br_value, AX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
// v1 := table[val1&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v1
|
||||
|
||||
// br2.advance(uint8(v1.entry))
|
||||
MOVB AH, BH // BH = uint8(v1.entry >> 8)
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CX, br_value // value <<= n
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// these two writes get coalesced
|
||||
// buf[stream][off] = uint8(v0.entry >> 8)
|
||||
// buf[stream][off+1] = uint8(v1.entry >> 8)
|
||||
MOVW BX, 512(buffer)(off*1)
|
||||
|
||||
// SECOND PART:
|
||||
// val2 := br2.peekTopBits(peekBits)
|
||||
MOVQ br_value, AX
|
||||
MOVQ peek_bits, CX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
// v2 := table[val0&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v0
|
||||
|
||||
// br2.advance(uint8(v0.entry))
|
||||
MOVB AH, BL // BL = uint8(v0.entry >> 8)
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CL, br_value // value <<= n
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// val3 := br2.peekTopBits(peekBits)
|
||||
MOVQ peek_bits, CX
|
||||
MOVQ br_value, AX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
// v3 := table[val1&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v1
|
||||
|
||||
// br2.advance(uint8(v1.entry))
|
||||
MOVB AH, BH // BH = uint8(v1.entry >> 8)
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CX, br_value // value <<= n
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// these two writes get coalesced
|
||||
// buf[stream][off+2] = uint8(v2.entry >> 8)
|
||||
// buf[stream][off+3] = uint8(v3.entry >> 8)
|
||||
MOVW BX, 512+2(buffer)(off*1)
|
||||
|
||||
// update the bitrader reader structure
|
||||
MOVB br_bits_read, bitReaderShifted_bitsRead(br2)
|
||||
MOVQ br_value, bitReaderShifted_value(br2)
|
||||
MOVQ br_offset, bitReaderShifted_off(br2)
|
||||
|
||||
// const stream = 3
|
||||
// br3.fillFast()
|
||||
MOVBQZX bitReaderShifted_bitsRead(br3), br_bits_read
|
||||
MOVQ bitReaderShifted_value(br3), br_value
|
||||
MOVQ bitReaderShifted_off(br3), br_offset
|
||||
|
||||
// if b.bitsRead >= 32 {
|
||||
CMPQ br_bits_read, $32
|
||||
JB skip_fill3
|
||||
|
||||
SUBQ $32, br_bits_read // b.bitsRead -= 32
|
||||
SUBQ $4, br_offset // b.off -= 4
|
||||
|
||||
// v := b.in[b.off-4 : b.off]
|
||||
// v = v[:4]
|
||||
// low := (uint32(v[0])) | (uint32(v[1]) << 8) | (uint32(v[2]) << 16) | (uint32(v[3]) << 24)
|
||||
MOVQ bitReaderShifted_in(br3), AX
|
||||
MOVL 0(br_offset)(AX*1), AX // AX = uint32(b.in[b.off:b.off+4])
|
||||
|
||||
// b.value |= uint64(low) << (b.bitsRead & 63)
|
||||
MOVQ br_bits_read, CX
|
||||
SHLQ CL, AX
|
||||
ORQ AX, br_value
|
||||
|
||||
// exhausted = exhausted || (br3.off < 4)
|
||||
CMPQ br_offset, $4
|
||||
SETLT DL
|
||||
ORB DL, DH
|
||||
|
||||
// }
|
||||
skip_fill3:
|
||||
|
||||
// val0 := br3.peekTopBits(peekBits)
|
||||
MOVQ br_value, AX
|
||||
MOVQ peek_bits, CX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
// v0 := table[val0&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v0
|
||||
|
||||
// br3.advance(uint8(v0.entry))
|
||||
MOVB AH, BL // BL = uint8(v0.entry >> 8)
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CL, br_value // value <<= n
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// val1 := br3.peekTopBits(peekBits)
|
||||
MOVQ peek_bits, CX
|
||||
MOVQ br_value, AX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
// v1 := table[val1&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v1
|
||||
|
||||
// br3.advance(uint8(v1.entry))
|
||||
MOVB AH, BH // BH = uint8(v1.entry >> 8)
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CX, br_value // value <<= n
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// these two writes get coalesced
|
||||
// buf[stream][off] = uint8(v0.entry >> 8)
|
||||
// buf[stream][off+1] = uint8(v1.entry >> 8)
|
||||
MOVW BX, 768(buffer)(off*1)
|
||||
|
||||
// SECOND PART:
|
||||
// val2 := br3.peekTopBits(peekBits)
|
||||
MOVQ br_value, AX
|
||||
MOVQ peek_bits, CX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
// v2 := table[val0&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v0
|
||||
|
||||
// br3.advance(uint8(v0.entry))
|
||||
MOVB AH, BL // BL = uint8(v0.entry >> 8)
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CL, br_value // value <<= n
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// val3 := br3.peekTopBits(peekBits)
|
||||
MOVQ peek_bits, CX
|
||||
MOVQ br_value, AX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
// v3 := table[val1&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v1
|
||||
|
||||
// br3.advance(uint8(v1.entry))
|
||||
MOVB AH, BH // BH = uint8(v1.entry >> 8)
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CX, br_value // value <<= n
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// these two writes get coalesced
|
||||
// buf[stream][off+2] = uint8(v2.entry >> 8)
|
||||
// buf[stream][off+3] = uint8(v3.entry >> 8)
|
||||
MOVW BX, 768+2(buffer)(off*1)
|
||||
|
||||
// update the bitrader reader structure
|
||||
MOVB br_bits_read, bitReaderShifted_bitsRead(br3)
|
||||
MOVQ br_value, bitReaderShifted_value(br3)
|
||||
MOVQ br_offset, bitReaderShifted_off(br3)
|
||||
|
||||
ADDQ $4, off // off += 2
|
||||
|
||||
TESTB DH, DH // any br[i].ofs < 4?
|
||||
JNZ end
|
||||
|
||||
CMPQ off, $bufoff
|
||||
JL main_loop
|
||||
|
||||
end:
|
||||
MOVQ 0(SP), BP
|
||||
|
||||
MOVB off, ret+56(FP)
|
||||
RET
|
||||
|
||||
#undef off
|
||||
#undef buffer
|
||||
#undef table
|
||||
|
||||
#undef br_bits_read
|
||||
#undef br_value
|
||||
#undef br_offset
|
||||
#undef peek_bits
|
||||
#undef exhausted
|
||||
|
||||
#undef br0
|
||||
#undef br1
|
||||
#undef br2
|
||||
#undef br3
|
||||
@ -0,0 +1,197 @@
|
||||
// +build !appengine
|
||||
// +build gc
|
||||
// +build !noasm
|
||||
|
||||
#include "textflag.h"
|
||||
#include "funcdata.h"
|
||||
#include "go_asm.h"
|
||||
|
||||
|
||||
#define bufoff 256 // see decompress.go, we're using [4][256]byte table
|
||||
|
||||
//func decompress4x_main_loop_x86(pbr0, pbr1, pbr2, pbr3 *bitReaderShifted,
|
||||
// peekBits uint8, buf *byte, tbl *dEntrySingle) (int, bool)
|
||||
TEXT ·decompress4x_8b_loop_x86(SB), NOSPLIT, $8
|
||||
#define off R8
|
||||
#define buffer DI
|
||||
#define table SI
|
||||
|
||||
#define br_bits_read R9
|
||||
#define br_value R10
|
||||
#define br_offset R11
|
||||
#define peek_bits R12
|
||||
#define exhausted DX
|
||||
|
||||
#define br0 R13
|
||||
#define br1 R14
|
||||
#define br2 R15
|
||||
#define br3 BP
|
||||
|
||||
MOVQ BP, 0(SP)
|
||||
|
||||
XORQ exhausted, exhausted // exhausted = false
|
||||
XORQ off, off // off = 0
|
||||
|
||||
MOVBQZX peekBits+32(FP), peek_bits
|
||||
MOVQ buf+40(FP), buffer
|
||||
MOVQ tbl+48(FP), table
|
||||
|
||||
MOVQ pbr0+0(FP), br0
|
||||
MOVQ pbr1+8(FP), br1
|
||||
MOVQ pbr2+16(FP), br2
|
||||
MOVQ pbr3+24(FP), br3
|
||||
|
||||
main_loop:
|
||||
{{ define "decode_2_values_x86" }}
|
||||
// const stream = {{ var "id" }}
|
||||
// br{{ var "id"}}.fillFast()
|
||||
MOVBQZX bitReaderShifted_bitsRead(br{{ var "id" }}), br_bits_read
|
||||
MOVQ bitReaderShifted_value(br{{ var "id" }}), br_value
|
||||
MOVQ bitReaderShifted_off(br{{ var "id" }}), br_offset
|
||||
|
||||
// if b.bitsRead >= 32 {
|
||||
CMPQ br_bits_read, $32
|
||||
JB skip_fill{{ var "id" }}
|
||||
|
||||
SUBQ $32, br_bits_read // b.bitsRead -= 32
|
||||
SUBQ $4, br_offset // b.off -= 4
|
||||
|
||||
// v := b.in[b.off-4 : b.off]
|
||||
// v = v[:4]
|
||||
// low := (uint32(v[0])) | (uint32(v[1]) << 8) | (uint32(v[2]) << 16) | (uint32(v[3]) << 24)
|
||||
MOVQ bitReaderShifted_in(br{{ var "id" }}), AX
|
||||
MOVL 0(br_offset)(AX*1), AX // AX = uint32(b.in[b.off:b.off+4])
|
||||
|
||||
// b.value |= uint64(low) << (b.bitsRead & 63)
|
||||
MOVQ br_bits_read, CX
|
||||
SHLQ CL, AX
|
||||
ORQ AX, br_value
|
||||
|
||||
// exhausted = exhausted || (br{{ var "id"}}.off < 4)
|
||||
CMPQ br_offset, $4
|
||||
SETLT DL
|
||||
ORB DL, DH
|
||||
// }
|
||||
skip_fill{{ var "id" }}:
|
||||
|
||||
// val0 := br{{ var "id"}}.peekTopBits(peekBits)
|
||||
MOVQ br_value, AX
|
||||
MOVQ peek_bits, CX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
// v0 := table[val0&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v0
|
||||
|
||||
// br{{ var "id"}}.advance(uint8(v0.entry))
|
||||
MOVB AH, BL // BL = uint8(v0.entry >> 8)
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CL, br_value // value <<= n
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// val1 := br{{ var "id"}}.peekTopBits(peekBits)
|
||||
MOVQ peek_bits, CX
|
||||
MOVQ br_value, AX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
// v1 := table[val1&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v1
|
||||
|
||||
// br{{ var "id"}}.advance(uint8(v1.entry))
|
||||
MOVB AH, BH // BH = uint8(v1.entry >> 8)
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CX, br_value // value <<= n
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
|
||||
// these two writes get coalesced
|
||||
// buf[stream][off] = uint8(v0.entry >> 8)
|
||||
// buf[stream][off+1] = uint8(v1.entry >> 8)
|
||||
MOVW BX, {{ var "bufofs" }}(buffer)(off*1)
|
||||
|
||||
// SECOND PART:
|
||||
// val2 := br{{ var "id"}}.peekTopBits(peekBits)
|
||||
MOVQ br_value, AX
|
||||
MOVQ peek_bits, CX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
// v2 := table[val0&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v0
|
||||
|
||||
// br{{ var "id"}}.advance(uint8(v0.entry))
|
||||
MOVB AH, BL // BL = uint8(v0.entry >> 8)
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CL, br_value // value <<= n
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// val3 := br{{ var "id"}}.peekTopBits(peekBits)
|
||||
MOVQ peek_bits, CX
|
||||
MOVQ br_value, AX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
// v3 := table[val1&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v1
|
||||
|
||||
// br{{ var "id"}}.advance(uint8(v1.entry))
|
||||
MOVB AH, BH // BH = uint8(v1.entry >> 8)
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CX, br_value // value <<= n
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
|
||||
// these two writes get coalesced
|
||||
// buf[stream][off+2] = uint8(v2.entry >> 8)
|
||||
// buf[stream][off+3] = uint8(v3.entry >> 8)
|
||||
MOVW BX, {{ var "bufofs" }}+2(buffer)(off*1)
|
||||
|
||||
// update the bitrader reader structure
|
||||
MOVB br_bits_read, bitReaderShifted_bitsRead(br{{ var "id" }})
|
||||
MOVQ br_value, bitReaderShifted_value(br{{ var "id" }})
|
||||
MOVQ br_offset, bitReaderShifted_off(br{{ var "id" }})
|
||||
{{ end }}
|
||||
|
||||
{{ set "id" "0" }}
|
||||
{{ set "ofs" "0" }}
|
||||
{{ set "bufofs" "0" }} {{/* id * bufoff */}}
|
||||
{{ template "decode_2_values_x86" . }}
|
||||
|
||||
{{ set "id" "1" }}
|
||||
{{ set "ofs" "8" }}
|
||||
{{ set "bufofs" "256" }}
|
||||
{{ template "decode_2_values_x86" . }}
|
||||
|
||||
{{ set "id" "2" }}
|
||||
{{ set "ofs" "16" }}
|
||||
{{ set "bufofs" "512" }}
|
||||
{{ template "decode_2_values_x86" . }}
|
||||
|
||||
{{ set "id" "3" }}
|
||||
{{ set "ofs" "24" }}
|
||||
{{ set "bufofs" "768" }}
|
||||
{{ template "decode_2_values_x86" . }}
|
||||
|
||||
ADDQ $4, off // off += 2
|
||||
|
||||
TESTB DH, DH // any br[i].ofs < 4?
|
||||
JNZ end
|
||||
|
||||
CMPQ off, $bufoff
|
||||
JL main_loop
|
||||
end:
|
||||
MOVQ 0(SP), BP
|
||||
|
||||
MOVB off, ret+56(FP)
|
||||
RET
|
||||
#undef off
|
||||
#undef buffer
|
||||
#undef table
|
||||
|
||||
#undef br_bits_read
|
||||
#undef br_value
|
||||
#undef br_offset
|
||||
#undef peek_bits
|
||||
#undef exhausted
|
||||
|
||||
#undef br0
|
||||
#undef br1
|
||||
#undef br2
|
||||
#undef br3
|
||||
@ -0,0 +1,181 @@
|
||||
//go:build amd64 && !appengine && !noasm && gc
|
||||
// +build amd64,!appengine,!noasm,gc
|
||||
|
||||
// This file contains the specialisation of Decoder.Decompress4X
|
||||
// that uses an asm implementation of its main loop.
|
||||
package huff0
|
||||
|
||||
import (
|
||||
"errors"
|
||||
"fmt"
|
||||
)
|
||||
|
||||
// decompress4x_main_loop_x86 is an x86 assembler implementation
|
||||
// of Decompress4X when tablelog > 8.
|
||||
// go:noescape
|
||||
func decompress4x_main_loop_x86(pbr0, pbr1, pbr2, pbr3 *bitReaderShifted,
|
||||
peekBits uint8, buf *byte, tbl *dEntrySingle) uint8
|
||||
|
||||
// decompress4x_8b_loop_x86 is an x86 assembler implementation
|
||||
// of Decompress4X when tablelog <= 8 which decodes 4 entries
|
||||
// per loop.
|
||||
// go:noescape
|
||||
func decompress4x_8b_loop_x86(pbr0, pbr1, pbr2, pbr3 *bitReaderShifted,
|
||||
peekBits uint8, buf *byte, tbl *dEntrySingle) uint8
|
||||
|
||||
// fallback8BitSize is the size where using Go version is faster.
|
||||
const fallback8BitSize = 800
|
||||
|
||||
// Decompress4X will decompress a 4X encoded stream.
|
||||
// The length of the supplied input must match the end of a block exactly.
|
||||
// The *capacity* of the dst slice must match the destination size of
|
||||
// the uncompressed data exactly.
|
||||
func (d *Decoder) Decompress4X(dst, src []byte) ([]byte, error) {
|
||||
if len(d.dt.single) == 0 {
|
||||
return nil, errors.New("no table loaded")
|
||||
}
|
||||
if len(src) < 6+(4*1) {
|
||||
return nil, errors.New("input too small")
|
||||
}
|
||||
|
||||
use8BitTables := d.actualTableLog <= 8
|
||||
if cap(dst) < fallback8BitSize && use8BitTables {
|
||||
return d.decompress4X8bit(dst, src)
|
||||
}
|
||||
var br [4]bitReaderShifted
|
||||
// Decode "jump table"
|
||||
start := 6
|
||||
for i := 0; i < 3; i++ {
|
||||
length := int(src[i*2]) | (int(src[i*2+1]) << 8)
|
||||
if start+length >= len(src) {
|
||||
return nil, errors.New("truncated input (or invalid offset)")
|
||||
}
|
||||
err := br[i].init(src[start : start+length])
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
start += length
|
||||
}
|
||||
err := br[3].init(src[start:])
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
|
||||
// destination, offset to match first output
|
||||
dstSize := cap(dst)
|
||||
dst = dst[:dstSize]
|
||||
out := dst
|
||||
dstEvery := (dstSize + 3) / 4
|
||||
|
||||
const tlSize = 1 << tableLogMax
|
||||
const tlMask = tlSize - 1
|
||||
single := d.dt.single[:tlSize]
|
||||
|
||||
// Use temp table to avoid bound checks/append penalty.
|
||||
buf := d.buffer()
|
||||
var off uint8
|
||||
var decoded int
|
||||
|
||||
const debug = false
|
||||
|
||||
// see: bitReaderShifted.peekBitsFast()
|
||||
peekBits := uint8((64 - d.actualTableLog) & 63)
|
||||
|
||||
// Decode 2 values from each decoder/loop.
|
||||
const bufoff = 256
|
||||
for {
|
||||
if br[0].off < 4 || br[1].off < 4 || br[2].off < 4 || br[3].off < 4 {
|
||||
break
|
||||
}
|
||||
|
||||
if use8BitTables {
|
||||
off = decompress4x_8b_loop_x86(&br[0], &br[1], &br[2], &br[3], peekBits, &buf[0][0], &single[0])
|
||||
} else {
|
||||
off = decompress4x_main_loop_x86(&br[0], &br[1], &br[2], &br[3], peekBits, &buf[0][0], &single[0])
|
||||
}
|
||||
if debug {
|
||||
fmt.Print("DEBUG: ")
|
||||
fmt.Printf("off=%d,", off)
|
||||
for i := 0; i < 4; i++ {
|
||||
fmt.Printf(" br[%d]={bitsRead=%d, value=%x, off=%d}",
|
||||
i, br[i].bitsRead, br[i].value, br[i].off)
|
||||
}
|
||||
fmt.Println("")
|
||||
}
|
||||
|
||||
if off != 0 {
|
||||
break
|
||||
}
|
||||
|
||||
if bufoff > dstEvery {
|
||||
d.bufs.Put(buf)
|
||||
return nil, errors.New("corruption detected: stream overrun 1")
|
||||
}
|
||||
copy(out, buf[0][:])
|
||||
copy(out[dstEvery:], buf[1][:])
|
||||
copy(out[dstEvery*2:], buf[2][:])
|
||||
copy(out[dstEvery*3:], buf[3][:])
|
||||
out = out[bufoff:]
|
||||
decoded += bufoff * 4
|
||||
// There must at least be 3 buffers left.
|
||||
if len(out) < dstEvery*3 {
|
||||
d.bufs.Put(buf)
|
||||
return nil, errors.New("corruption detected: stream overrun 2")
|
||||
}
|
||||
}
|
||||
if off > 0 {
|
||||
ioff := int(off)
|
||||
if len(out) < dstEvery*3+ioff {
|
||||
d.bufs.Put(buf)
|
||||
return nil, errors.New("corruption detected: stream overrun 3")
|
||||
}
|
||||
copy(out, buf[0][:off])
|
||||
copy(out[dstEvery:], buf[1][:off])
|
||||
copy(out[dstEvery*2:], buf[2][:off])
|
||||
copy(out[dstEvery*3:], buf[3][:off])
|
||||
decoded += int(off) * 4
|
||||
out = out[off:]
|
||||
}
|
||||
|
||||
// Decode remaining.
|
||||
remainBytes := dstEvery - (decoded / 4)
|
||||
for i := range br {
|
||||
offset := dstEvery * i
|
||||
endsAt := offset + remainBytes
|
||||
if endsAt > len(out) {
|
||||
endsAt = len(out)
|
||||
}
|
||||
br := &br[i]
|
||||
bitsLeft := br.remaining()
|
||||
for bitsLeft > 0 {
|
||||
br.fill()
|
||||
if offset >= endsAt {
|
||||
d.bufs.Put(buf)
|
||||
return nil, errors.New("corruption detected: stream overrun 4")
|
||||
}
|
||||
|
||||
// Read value and increment offset.
|
||||
val := br.peekBitsFast(d.actualTableLog)
|
||||
v := single[val&tlMask].entry
|
||||
nBits := uint8(v)
|
||||
br.advance(nBits)
|
||||
bitsLeft -= uint(nBits)
|
||||
out[offset] = uint8(v >> 8)
|
||||
offset++
|
||||
}
|
||||
if offset != endsAt {
|
||||
d.bufs.Put(buf)
|
||||
return nil, fmt.Errorf("corruption detected: short output block %d, end %d != %d", i, offset, endsAt)
|
||||
}
|
||||
decoded += offset - dstEvery*i
|
||||
err = br.close()
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
}
|
||||
d.bufs.Put(buf)
|
||||
if dstSize != decoded {
|
||||
return nil, errors.New("corruption detected: short output block")
|
||||
}
|
||||
return dst, nil
|
||||
}
|
||||
@ -0,0 +1,506 @@
|
||||
// +build !appengine
|
||||
// +build gc
|
||||
// +build !noasm
|
||||
|
||||
#include "textflag.h"
|
||||
#include "funcdata.h"
|
||||
#include "go_asm.h"
|
||||
|
||||
#ifdef GOAMD64_v4
|
||||
#ifndef GOAMD64_v3
|
||||
#define GOAMD64_v3
|
||||
#endif
|
||||
#endif
|
||||
|
||||
#define bufoff 256 // see decompress.go, we're using [4][256]byte table
|
||||
|
||||
// func decompress4x_main_loop_x86(pbr0, pbr1, pbr2, pbr3 *bitReaderShifted,
|
||||
// peekBits uint8, buf *byte, tbl *dEntrySingle) (int, bool)
|
||||
TEXT ·decompress4x_main_loop_x86(SB), NOSPLIT, $8
|
||||
#define off R8
|
||||
#define buffer DI
|
||||
#define table SI
|
||||
|
||||
#define br_bits_read R9
|
||||
#define br_value R10
|
||||
#define br_offset R11
|
||||
#define peek_bits R12
|
||||
#define exhausted DX
|
||||
|
||||
#define br0 R13
|
||||
#define br1 R14
|
||||
#define br2 R15
|
||||
#define br3 BP
|
||||
|
||||
MOVQ BP, 0(SP)
|
||||
|
||||
XORQ exhausted, exhausted // exhausted = false
|
||||
XORQ off, off // off = 0
|
||||
|
||||
MOVBQZX peekBits+32(FP), peek_bits
|
||||
MOVQ buf+40(FP), buffer
|
||||
MOVQ tbl+48(FP), table
|
||||
|
||||
MOVQ pbr0+0(FP), br0
|
||||
MOVQ pbr1+8(FP), br1
|
||||
MOVQ pbr2+16(FP), br2
|
||||
MOVQ pbr3+24(FP), br3
|
||||
|
||||
main_loop:
|
||||
|
||||
// const stream = 0
|
||||
// br0.fillFast()
|
||||
MOVBQZX bitReaderShifted_bitsRead(br0), br_bits_read
|
||||
MOVQ bitReaderShifted_value(br0), br_value
|
||||
MOVQ bitReaderShifted_off(br0), br_offset
|
||||
|
||||
// We must have at least 2 * max tablelog left
|
||||
CMPQ br_bits_read, $64-22
|
||||
JBE skip_fill0
|
||||
|
||||
SUBQ $32, br_bits_read // b.bitsRead -= 32
|
||||
SUBQ $4, br_offset // b.off -= 4
|
||||
|
||||
// v := b.in[b.off-4 : b.off]
|
||||
// v = v[:4]
|
||||
// low := (uint32(v[0])) | (uint32(v[1]) << 8) | (uint32(v[2]) << 16) | (uint32(v[3]) << 24)
|
||||
MOVQ bitReaderShifted_in(br0), AX
|
||||
|
||||
// b.value |= uint64(low) << (b.bitsRead & 63)
|
||||
#ifdef GOAMD64_v3
|
||||
SHLXQ br_bits_read, 0(br_offset)(AX*1), AX // AX = uint32(b.in[b.off:b.off+4]) << (b.bitsRead & 63)
|
||||
|
||||
#else
|
||||
MOVL 0(br_offset)(AX*1), AX // AX = uint32(b.in[b.off:b.off+4])
|
||||
MOVQ br_bits_read, CX
|
||||
SHLQ CL, AX
|
||||
|
||||
#endif
|
||||
|
||||
ORQ AX, br_value
|
||||
|
||||
// exhausted = exhausted || (br0.off < 4)
|
||||
CMPQ br_offset, $4
|
||||
SETLT DL
|
||||
ORB DL, DH
|
||||
|
||||
// }
|
||||
skip_fill0:
|
||||
|
||||
// val0 := br0.peekTopBits(peekBits)
|
||||
#ifdef GOAMD64_v3
|
||||
SHRXQ peek_bits, br_value, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
#else
|
||||
MOVQ br_value, AX
|
||||
MOVQ peek_bits, CX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
#endif
|
||||
|
||||
// v0 := table[val0&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v0
|
||||
|
||||
// br0.advance(uint8(v0.entry))
|
||||
MOVB AH, BL // BL = uint8(v0.entry >> 8)
|
||||
|
||||
#ifdef GOAMD64_v3
|
||||
MOVBQZX AL, CX
|
||||
SHLXQ AX, br_value, br_value // value <<= n
|
||||
|
||||
#else
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CL, br_value // value <<= n
|
||||
|
||||
#endif
|
||||
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
#ifdef GOAMD64_v3
|
||||
SHRXQ peek_bits, br_value, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
#else
|
||||
// val1 := br0.peekTopBits(peekBits)
|
||||
MOVQ peek_bits, CX
|
||||
MOVQ br_value, AX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
#endif
|
||||
|
||||
// v1 := table[val1&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v1
|
||||
|
||||
// br0.advance(uint8(v1.entry))
|
||||
MOVB AH, BH // BH = uint8(v1.entry >> 8)
|
||||
|
||||
#ifdef GOAMD64_v3
|
||||
MOVBQZX AL, CX
|
||||
SHLXQ AX, br_value, br_value // value <<= n
|
||||
|
||||
#else
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CL, br_value // value <<= n
|
||||
|
||||
#endif
|
||||
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// these two writes get coalesced
|
||||
// buf[stream][off] = uint8(v0.entry >> 8)
|
||||
// buf[stream][off+1] = uint8(v1.entry >> 8)
|
||||
MOVW BX, 0(buffer)(off*1)
|
||||
|
||||
// update the bitrader reader structure
|
||||
MOVB br_bits_read, bitReaderShifted_bitsRead(br0)
|
||||
MOVQ br_value, bitReaderShifted_value(br0)
|
||||
MOVQ br_offset, bitReaderShifted_off(br0)
|
||||
|
||||
// const stream = 1
|
||||
// br1.fillFast()
|
||||
MOVBQZX bitReaderShifted_bitsRead(br1), br_bits_read
|
||||
MOVQ bitReaderShifted_value(br1), br_value
|
||||
MOVQ bitReaderShifted_off(br1), br_offset
|
||||
|
||||
// We must have at least 2 * max tablelog left
|
||||
CMPQ br_bits_read, $64-22
|
||||
JBE skip_fill1
|
||||
|
||||
SUBQ $32, br_bits_read // b.bitsRead -= 32
|
||||
SUBQ $4, br_offset // b.off -= 4
|
||||
|
||||
// v := b.in[b.off-4 : b.off]
|
||||
// v = v[:4]
|
||||
// low := (uint32(v[0])) | (uint32(v[1]) << 8) | (uint32(v[2]) << 16) | (uint32(v[3]) << 24)
|
||||
MOVQ bitReaderShifted_in(br1), AX
|
||||
|
||||
// b.value |= uint64(low) << (b.bitsRead & 63)
|
||||
#ifdef GOAMD64_v3
|
||||
SHLXQ br_bits_read, 0(br_offset)(AX*1), AX // AX = uint32(b.in[b.off:b.off+4]) << (b.bitsRead & 63)
|
||||
|
||||
#else
|
||||
MOVL 0(br_offset)(AX*1), AX // AX = uint32(b.in[b.off:b.off+4])
|
||||
MOVQ br_bits_read, CX
|
||||
SHLQ CL, AX
|
||||
|
||||
#endif
|
||||
|
||||
ORQ AX, br_value
|
||||
|
||||
// exhausted = exhausted || (br1.off < 4)
|
||||
CMPQ br_offset, $4
|
||||
SETLT DL
|
||||
ORB DL, DH
|
||||
|
||||
// }
|
||||
skip_fill1:
|
||||
|
||||
// val0 := br1.peekTopBits(peekBits)
|
||||
#ifdef GOAMD64_v3
|
||||
SHRXQ peek_bits, br_value, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
#else
|
||||
MOVQ br_value, AX
|
||||
MOVQ peek_bits, CX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
#endif
|
||||
|
||||
// v0 := table[val0&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v0
|
||||
|
||||
// br1.advance(uint8(v0.entry))
|
||||
MOVB AH, BL // BL = uint8(v0.entry >> 8)
|
||||
|
||||
#ifdef GOAMD64_v3
|
||||
MOVBQZX AL, CX
|
||||
SHLXQ AX, br_value, br_value // value <<= n
|
||||
|
||||
#else
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CL, br_value // value <<= n
|
||||
|
||||
#endif
|
||||
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
#ifdef GOAMD64_v3
|
||||
SHRXQ peek_bits, br_value, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
#else
|
||||
// val1 := br1.peekTopBits(peekBits)
|
||||
MOVQ peek_bits, CX
|
||||
MOVQ br_value, AX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
#endif
|
||||
|
||||
// v1 := table[val1&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v1
|
||||
|
||||
// br1.advance(uint8(v1.entry))
|
||||
MOVB AH, BH // BH = uint8(v1.entry >> 8)
|
||||
|
||||
#ifdef GOAMD64_v3
|
||||
MOVBQZX AL, CX
|
||||
SHLXQ AX, br_value, br_value // value <<= n
|
||||
|
||||
#else
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CL, br_value // value <<= n
|
||||
|
||||
#endif
|
||||
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// these two writes get coalesced
|
||||
// buf[stream][off] = uint8(v0.entry >> 8)
|
||||
// buf[stream][off+1] = uint8(v1.entry >> 8)
|
||||
MOVW BX, 256(buffer)(off*1)
|
||||
|
||||
// update the bitrader reader structure
|
||||
MOVB br_bits_read, bitReaderShifted_bitsRead(br1)
|
||||
MOVQ br_value, bitReaderShifted_value(br1)
|
||||
MOVQ br_offset, bitReaderShifted_off(br1)
|
||||
|
||||
// const stream = 2
|
||||
// br2.fillFast()
|
||||
MOVBQZX bitReaderShifted_bitsRead(br2), br_bits_read
|
||||
MOVQ bitReaderShifted_value(br2), br_value
|
||||
MOVQ bitReaderShifted_off(br2), br_offset
|
||||
|
||||
// We must have at least 2 * max tablelog left
|
||||
CMPQ br_bits_read, $64-22
|
||||
JBE skip_fill2
|
||||
|
||||
SUBQ $32, br_bits_read // b.bitsRead -= 32
|
||||
SUBQ $4, br_offset // b.off -= 4
|
||||
|
||||
// v := b.in[b.off-4 : b.off]
|
||||
// v = v[:4]
|
||||
// low := (uint32(v[0])) | (uint32(v[1]) << 8) | (uint32(v[2]) << 16) | (uint32(v[3]) << 24)
|
||||
MOVQ bitReaderShifted_in(br2), AX
|
||||
|
||||
// b.value |= uint64(low) << (b.bitsRead & 63)
|
||||
#ifdef GOAMD64_v3
|
||||
SHLXQ br_bits_read, 0(br_offset)(AX*1), AX // AX = uint32(b.in[b.off:b.off+4]) << (b.bitsRead & 63)
|
||||
|
||||
#else
|
||||
MOVL 0(br_offset)(AX*1), AX // AX = uint32(b.in[b.off:b.off+4])
|
||||
MOVQ br_bits_read, CX
|
||||
SHLQ CL, AX
|
||||
|
||||
#endif
|
||||
|
||||
ORQ AX, br_value
|
||||
|
||||
// exhausted = exhausted || (br2.off < 4)
|
||||
CMPQ br_offset, $4
|
||||
SETLT DL
|
||||
ORB DL, DH
|
||||
|
||||
// }
|
||||
skip_fill2:
|
||||
|
||||
// val0 := br2.peekTopBits(peekBits)
|
||||
#ifdef GOAMD64_v3
|
||||
SHRXQ peek_bits, br_value, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
#else
|
||||
MOVQ br_value, AX
|
||||
MOVQ peek_bits, CX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
#endif
|
||||
|
||||
// v0 := table[val0&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v0
|
||||
|
||||
// br2.advance(uint8(v0.entry))
|
||||
MOVB AH, BL // BL = uint8(v0.entry >> 8)
|
||||
|
||||
#ifdef GOAMD64_v3
|
||||
MOVBQZX AL, CX
|
||||
SHLXQ AX, br_value, br_value // value <<= n
|
||||
|
||||
#else
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CL, br_value // value <<= n
|
||||
|
||||
#endif
|
||||
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
#ifdef GOAMD64_v3
|
||||
SHRXQ peek_bits, br_value, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
#else
|
||||
// val1 := br2.peekTopBits(peekBits)
|
||||
MOVQ peek_bits, CX
|
||||
MOVQ br_value, AX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
#endif
|
||||
|
||||
// v1 := table[val1&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v1
|
||||
|
||||
// br2.advance(uint8(v1.entry))
|
||||
MOVB AH, BH // BH = uint8(v1.entry >> 8)
|
||||
|
||||
#ifdef GOAMD64_v3
|
||||
MOVBQZX AL, CX
|
||||
SHLXQ AX, br_value, br_value // value <<= n
|
||||
|
||||
#else
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CL, br_value // value <<= n
|
||||
|
||||
#endif
|
||||
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// these two writes get coalesced
|
||||
// buf[stream][off] = uint8(v0.entry >> 8)
|
||||
// buf[stream][off+1] = uint8(v1.entry >> 8)
|
||||
MOVW BX, 512(buffer)(off*1)
|
||||
|
||||
// update the bitrader reader structure
|
||||
MOVB br_bits_read, bitReaderShifted_bitsRead(br2)
|
||||
MOVQ br_value, bitReaderShifted_value(br2)
|
||||
MOVQ br_offset, bitReaderShifted_off(br2)
|
||||
|
||||
// const stream = 3
|
||||
// br3.fillFast()
|
||||
MOVBQZX bitReaderShifted_bitsRead(br3), br_bits_read
|
||||
MOVQ bitReaderShifted_value(br3), br_value
|
||||
MOVQ bitReaderShifted_off(br3), br_offset
|
||||
|
||||
// We must have at least 2 * max tablelog left
|
||||
CMPQ br_bits_read, $64-22
|
||||
JBE skip_fill3
|
||||
|
||||
SUBQ $32, br_bits_read // b.bitsRead -= 32
|
||||
SUBQ $4, br_offset // b.off -= 4
|
||||
|
||||
// v := b.in[b.off-4 : b.off]
|
||||
// v = v[:4]
|
||||
// low := (uint32(v[0])) | (uint32(v[1]) << 8) | (uint32(v[2]) << 16) | (uint32(v[3]) << 24)
|
||||
MOVQ bitReaderShifted_in(br3), AX
|
||||
|
||||
// b.value |= uint64(low) << (b.bitsRead & 63)
|
||||
#ifdef GOAMD64_v3
|
||||
SHLXQ br_bits_read, 0(br_offset)(AX*1), AX // AX = uint32(b.in[b.off:b.off+4]) << (b.bitsRead & 63)
|
||||
|
||||
#else
|
||||
MOVL 0(br_offset)(AX*1), AX // AX = uint32(b.in[b.off:b.off+4])
|
||||
MOVQ br_bits_read, CX
|
||||
SHLQ CL, AX
|
||||
|
||||
#endif
|
||||
|
||||
ORQ AX, br_value
|
||||
|
||||
// exhausted = exhausted || (br3.off < 4)
|
||||
CMPQ br_offset, $4
|
||||
SETLT DL
|
||||
ORB DL, DH
|
||||
|
||||
// }
|
||||
skip_fill3:
|
||||
|
||||
// val0 := br3.peekTopBits(peekBits)
|
||||
#ifdef GOAMD64_v3
|
||||
SHRXQ peek_bits, br_value, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
#else
|
||||
MOVQ br_value, AX
|
||||
MOVQ peek_bits, CX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
#endif
|
||||
|
||||
// v0 := table[val0&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v0
|
||||
|
||||
// br3.advance(uint8(v0.entry))
|
||||
MOVB AH, BL // BL = uint8(v0.entry >> 8)
|
||||
|
||||
#ifdef GOAMD64_v3
|
||||
MOVBQZX AL, CX
|
||||
SHLXQ AX, br_value, br_value // value <<= n
|
||||
|
||||
#else
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CL, br_value // value <<= n
|
||||
|
||||
#endif
|
||||
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
#ifdef GOAMD64_v3
|
||||
SHRXQ peek_bits, br_value, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
#else
|
||||
// val1 := br3.peekTopBits(peekBits)
|
||||
MOVQ peek_bits, CX
|
||||
MOVQ br_value, AX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
|
||||
#endif
|
||||
|
||||
// v1 := table[val1&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v1
|
||||
|
||||
// br3.advance(uint8(v1.entry))
|
||||
MOVB AH, BH // BH = uint8(v1.entry >> 8)
|
||||
|
||||
#ifdef GOAMD64_v3
|
||||
MOVBQZX AL, CX
|
||||
SHLXQ AX, br_value, br_value // value <<= n
|
||||
|
||||
#else
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CL, br_value // value <<= n
|
||||
|
||||
#endif
|
||||
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
// these two writes get coalesced
|
||||
// buf[stream][off] = uint8(v0.entry >> 8)
|
||||
// buf[stream][off+1] = uint8(v1.entry >> 8)
|
||||
MOVW BX, 768(buffer)(off*1)
|
||||
|
||||
// update the bitrader reader structure
|
||||
MOVB br_bits_read, bitReaderShifted_bitsRead(br3)
|
||||
MOVQ br_value, bitReaderShifted_value(br3)
|
||||
MOVQ br_offset, bitReaderShifted_off(br3)
|
||||
|
||||
ADDQ $2, off // off += 2
|
||||
|
||||
TESTB DH, DH // any br[i].ofs < 4?
|
||||
JNZ end
|
||||
|
||||
CMPQ off, $bufoff
|
||||
JL main_loop
|
||||
|
||||
end:
|
||||
MOVQ 0(SP), BP
|
||||
|
||||
MOVB off, ret+56(FP)
|
||||
RET
|
||||
|
||||
#undef off
|
||||
#undef buffer
|
||||
#undef table
|
||||
|
||||
#undef br_bits_read
|
||||
#undef br_value
|
||||
#undef br_offset
|
||||
#undef peek_bits
|
||||
#undef exhausted
|
||||
|
||||
#undef br0
|
||||
#undef br1
|
||||
#undef br2
|
||||
#undef br3
|
||||
@ -0,0 +1,195 @@
|
||||
// +build !appengine
|
||||
// +build gc
|
||||
// +build !noasm
|
||||
|
||||
#include "textflag.h"
|
||||
#include "funcdata.h"
|
||||
#include "go_asm.h"
|
||||
|
||||
#ifdef GOAMD64_v4
|
||||
#ifndef GOAMD64_v3
|
||||
#define GOAMD64_v3
|
||||
#endif
|
||||
#endif
|
||||
|
||||
#define bufoff 256 // see decompress.go, we're using [4][256]byte table
|
||||
|
||||
//func decompress4x_main_loop_x86(pbr0, pbr1, pbr2, pbr3 *bitReaderShifted,
|
||||
// peekBits uint8, buf *byte, tbl *dEntrySingle) (int, bool)
|
||||
TEXT ·decompress4x_main_loop_x86(SB), NOSPLIT, $8
|
||||
#define off R8
|
||||
#define buffer DI
|
||||
#define table SI
|
||||
|
||||
#define br_bits_read R9
|
||||
#define br_value R10
|
||||
#define br_offset R11
|
||||
#define peek_bits R12
|
||||
#define exhausted DX
|
||||
|
||||
#define br0 R13
|
||||
#define br1 R14
|
||||
#define br2 R15
|
||||
#define br3 BP
|
||||
|
||||
MOVQ BP, 0(SP)
|
||||
|
||||
XORQ exhausted, exhausted // exhausted = false
|
||||
XORQ off, off // off = 0
|
||||
|
||||
MOVBQZX peekBits+32(FP), peek_bits
|
||||
MOVQ buf+40(FP), buffer
|
||||
MOVQ tbl+48(FP), table
|
||||
|
||||
MOVQ pbr0+0(FP), br0
|
||||
MOVQ pbr1+8(FP), br1
|
||||
MOVQ pbr2+16(FP), br2
|
||||
MOVQ pbr3+24(FP), br3
|
||||
|
||||
main_loop:
|
||||
{{ define "decode_2_values_x86" }}
|
||||
// const stream = {{ var "id" }}
|
||||
// br{{ var "id"}}.fillFast()
|
||||
MOVBQZX bitReaderShifted_bitsRead(br{{ var "id" }}), br_bits_read
|
||||
MOVQ bitReaderShifted_value(br{{ var "id" }}), br_value
|
||||
MOVQ bitReaderShifted_off(br{{ var "id" }}), br_offset
|
||||
|
||||
// We must have at least 2 * max tablelog left
|
||||
CMPQ br_bits_read, $64-22
|
||||
JBE skip_fill{{ var "id" }}
|
||||
|
||||
SUBQ $32, br_bits_read // b.bitsRead -= 32
|
||||
SUBQ $4, br_offset // b.off -= 4
|
||||
|
||||
// v := b.in[b.off-4 : b.off]
|
||||
// v = v[:4]
|
||||
// low := (uint32(v[0])) | (uint32(v[1]) << 8) | (uint32(v[2]) << 16) | (uint32(v[3]) << 24)
|
||||
MOVQ bitReaderShifted_in(br{{ var "id" }}), AX
|
||||
|
||||
// b.value |= uint64(low) << (b.bitsRead & 63)
|
||||
#ifdef GOAMD64_v3
|
||||
SHLXQ br_bits_read, 0(br_offset)(AX*1), AX // AX = uint32(b.in[b.off:b.off+4]) << (b.bitsRead & 63)
|
||||
#else
|
||||
MOVL 0(br_offset)(AX*1), AX // AX = uint32(b.in[b.off:b.off+4])
|
||||
MOVQ br_bits_read, CX
|
||||
SHLQ CL, AX
|
||||
#endif
|
||||
|
||||
ORQ AX, br_value
|
||||
|
||||
// exhausted = exhausted || (br{{ var "id"}}.off < 4)
|
||||
CMPQ br_offset, $4
|
||||
SETLT DL
|
||||
ORB DL, DH
|
||||
// }
|
||||
skip_fill{{ var "id" }}:
|
||||
|
||||
// val0 := br{{ var "id"}}.peekTopBits(peekBits)
|
||||
#ifdef GOAMD64_v3
|
||||
SHRXQ peek_bits, br_value, AX // AX = (value >> peek_bits) & mask
|
||||
#else
|
||||
MOVQ br_value, AX
|
||||
MOVQ peek_bits, CX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
#endif
|
||||
|
||||
// v0 := table[val0&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v0
|
||||
|
||||
// br{{ var "id"}}.advance(uint8(v0.entry))
|
||||
MOVB AH, BL // BL = uint8(v0.entry >> 8)
|
||||
|
||||
#ifdef GOAMD64_v3
|
||||
MOVBQZX AL, CX
|
||||
SHLXQ AX, br_value, br_value // value <<= n
|
||||
#else
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CL, br_value // value <<= n
|
||||
#endif
|
||||
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
|
||||
#ifdef GOAMD64_v3
|
||||
SHRXQ peek_bits, br_value, AX // AX = (value >> peek_bits) & mask
|
||||
#else
|
||||
// val1 := br{{ var "id"}}.peekTopBits(peekBits)
|
||||
MOVQ peek_bits, CX
|
||||
MOVQ br_value, AX
|
||||
SHRQ CL, AX // AX = (value >> peek_bits) & mask
|
||||
#endif
|
||||
|
||||
// v1 := table[val1&mask]
|
||||
MOVW 0(table)(AX*2), AX // AX - v1
|
||||
|
||||
// br{{ var "id"}}.advance(uint8(v1.entry))
|
||||
MOVB AH, BH // BH = uint8(v1.entry >> 8)
|
||||
|
||||
#ifdef GOAMD64_v3
|
||||
MOVBQZX AL, CX
|
||||
SHLXQ AX, br_value, br_value // value <<= n
|
||||
#else
|
||||
MOVBQZX AL, CX
|
||||
SHLQ CL, br_value // value <<= n
|
||||
#endif
|
||||
|
||||
ADDQ CX, br_bits_read // bits_read += n
|
||||
|
||||
|
||||
// these two writes get coalesced
|
||||
// buf[stream][off] = uint8(v0.entry >> 8)
|
||||
// buf[stream][off+1] = uint8(v1.entry >> 8)
|
||||
MOVW BX, {{ var "bufofs" }}(buffer)(off*1)
|
||||
|
||||
// update the bitrader reader structure
|
||||
MOVB br_bits_read, bitReaderShifted_bitsRead(br{{ var "id" }})
|
||||
MOVQ br_value, bitReaderShifted_value(br{{ var "id" }})
|
||||
MOVQ br_offset, bitReaderShifted_off(br{{ var "id" }})
|
||||
{{ end }}
|
||||
|
||||
{{ set "id" "0" }}
|
||||
{{ set "ofs" "0" }}
|
||||
{{ set "bufofs" "0" }} {{/* id * bufoff */}}
|
||||
{{ template "decode_2_values_x86" . }}
|
||||
|
||||
{{ set "id" "1" }}
|
||||
{{ set "ofs" "8" }}
|
||||
{{ set "bufofs" "256" }}
|
||||
{{ template "decode_2_values_x86" . }}
|
||||
|
||||
{{ set "id" "2" }}
|
||||
{{ set "ofs" "16" }}
|
||||
{{ set "bufofs" "512" }}
|
||||
{{ template "decode_2_values_x86" . }}
|
||||
|
||||
{{ set "id" "3" }}
|
||||
{{ set "ofs" "24" }}
|
||||
{{ set "bufofs" "768" }}
|
||||
{{ template "decode_2_values_x86" . }}
|
||||
|
||||
ADDQ $2, off // off += 2
|
||||
|
||||
TESTB DH, DH // any br[i].ofs < 4?
|
||||
JNZ end
|
||||
|
||||
CMPQ off, $bufoff
|
||||
JL main_loop
|
||||
end:
|
||||
MOVQ 0(SP), BP
|
||||
|
||||
MOVB off, ret+56(FP)
|
||||
RET
|
||||
#undef off
|
||||
#undef buffer
|
||||
#undef table
|
||||
|
||||
#undef br_bits_read
|
||||
#undef br_value
|
||||
#undef br_offset
|
||||
#undef peek_bits
|
||||
#undef exhausted
|
||||
|
||||
#undef br0
|
||||
#undef br1
|
||||
#undef br2
|
||||
#undef br3
|
||||
@ -0,0 +1,193 @@
|
||||
//go:build !amd64 || appengine || !gc || noasm
|
||||
// +build !amd64 appengine !gc noasm
|
||||
|
||||
// This file contains a generic implementation of Decoder.Decompress4X.
|
||||
package huff0
|
||||
|
||||
import (
|
||||
"errors"
|
||||
"fmt"
|
||||
)
|
||||
|
||||
// Decompress4X will decompress a 4X encoded stream.
|
||||
// The length of the supplied input must match the end of a block exactly.
|
||||
// The *capacity* of the dst slice must match the destination size of
|
||||
// the uncompressed data exactly.
|
||||
func (d *Decoder) Decompress4X(dst, src []byte) ([]byte, error) {
|
||||
if len(d.dt.single) == 0 {
|
||||
return nil, errors.New("no table loaded")
|
||||
}
|
||||
if len(src) < 6+(4*1) {
|
||||
return nil, errors.New("input too small")
|
||||
}
|
||||
if use8BitTables && d.actualTableLog <= 8 {
|
||||
return d.decompress4X8bit(dst, src)
|
||||
}
|
||||
|
||||
var br [4]bitReaderShifted
|
||||
// Decode "jump table"
|
||||
start := 6
|
||||
for i := 0; i < 3; i++ {
|
||||
length := int(src[i*2]) | (int(src[i*2+1]) << 8)
|
||||
if start+length >= len(src) {
|
||||
return nil, errors.New("truncated input (or invalid offset)")
|
||||
}
|
||||
err := br[i].init(src[start : start+length])
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
start += length
|
||||
}
|
||||
err := br[3].init(src[start:])
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
|
||||
// destination, offset to match first output
|
||||
dstSize := cap(dst)
|
||||
dst = dst[:dstSize]
|
||||
out := dst
|
||||
dstEvery := (dstSize + 3) / 4
|
||||
|
||||
const tlSize = 1 << tableLogMax
|
||||
const tlMask = tlSize - 1
|
||||
single := d.dt.single[:tlSize]
|
||||
|
||||
// Use temp table to avoid bound checks/append penalty.
|
||||
buf := d.buffer()
|
||||
var off uint8
|
||||
var decoded int
|
||||
|
||||
// Decode 2 values from each decoder/loop.
|
||||
const bufoff = 256
|
||||
for {
|
||||
if br[0].off < 4 || br[1].off < 4 || br[2].off < 4 || br[3].off < 4 {
|
||||
break
|
||||
}
|
||||
|
||||
{
|
||||
const stream = 0
|
||||
const stream2 = 1
|
||||
br[stream].fillFast()
|
||||
br[stream2].fillFast()
|
||||
|
||||
val := br[stream].peekBitsFast(d.actualTableLog)
|
||||
val2 := br[stream2].peekBitsFast(d.actualTableLog)
|
||||
v := single[val&tlMask]
|
||||
v2 := single[val2&tlMask]
|
||||
br[stream].advance(uint8(v.entry))
|
||||
br[stream2].advance(uint8(v2.entry))
|
||||
buf[stream][off] = uint8(v.entry >> 8)
|
||||
buf[stream2][off] = uint8(v2.entry >> 8)
|
||||
|
||||
val = br[stream].peekBitsFast(d.actualTableLog)
|
||||
val2 = br[stream2].peekBitsFast(d.actualTableLog)
|
||||
v = single[val&tlMask]
|
||||
v2 = single[val2&tlMask]
|
||||
br[stream].advance(uint8(v.entry))
|
||||
br[stream2].advance(uint8(v2.entry))
|
||||
buf[stream][off+1] = uint8(v.entry >> 8)
|
||||
buf[stream2][off+1] = uint8(v2.entry >> 8)
|
||||
}
|
||||
|
||||
{
|
||||
const stream = 2
|
||||
const stream2 = 3
|
||||
br[stream].fillFast()
|
||||
br[stream2].fillFast()
|
||||
|
||||
val := br[stream].peekBitsFast(d.actualTableLog)
|
||||
val2 := br[stream2].peekBitsFast(d.actualTableLog)
|
||||
v := single[val&tlMask]
|
||||
v2 := single[val2&tlMask]
|
||||
br[stream].advance(uint8(v.entry))
|
||||
br[stream2].advance(uint8(v2.entry))
|
||||
buf[stream][off] = uint8(v.entry >> 8)
|
||||
buf[stream2][off] = uint8(v2.entry >> 8)
|
||||
|
||||
val = br[stream].peekBitsFast(d.actualTableLog)
|
||||
val2 = br[stream2].peekBitsFast(d.actualTableLog)
|
||||
v = single[val&tlMask]
|
||||
v2 = single[val2&tlMask]
|
||||
br[stream].advance(uint8(v.entry))
|
||||
br[stream2].advance(uint8(v2.entry))
|
||||
buf[stream][off+1] = uint8(v.entry >> 8)
|
||||
buf[stream2][off+1] = uint8(v2.entry >> 8)
|
||||
}
|
||||
|
||||
off += 2
|
||||
|
||||
if off == 0 {
|
||||
if bufoff > dstEvery {
|
||||
d.bufs.Put(buf)
|
||||
return nil, errors.New("corruption detected: stream overrun 1")
|
||||
}
|
||||
copy(out, buf[0][:])
|
||||
copy(out[dstEvery:], buf[1][:])
|
||||
copy(out[dstEvery*2:], buf[2][:])
|
||||
copy(out[dstEvery*3:], buf[3][:])
|
||||
out = out[bufoff:]
|
||||
decoded += bufoff * 4
|
||||
// There must at least be 3 buffers left.
|
||||
if len(out) < dstEvery*3 {
|
||||
d.bufs.Put(buf)
|
||||
return nil, errors.New("corruption detected: stream overrun 2")
|
||||
}
|
||||
}
|
||||
}
|
||||
if off > 0 {
|
||||
ioff := int(off)
|
||||
if len(out) < dstEvery*3+ioff {
|
||||
d.bufs.Put(buf)
|
||||
return nil, errors.New("corruption detected: stream overrun 3")
|
||||
}
|
||||
copy(out, buf[0][:off])
|
||||
copy(out[dstEvery:], buf[1][:off])
|
||||
copy(out[dstEvery*2:], buf[2][:off])
|
||||
copy(out[dstEvery*3:], buf[3][:off])
|
||||
decoded += int(off) * 4
|
||||
out = out[off:]
|
||||
}
|
||||
|
||||
// Decode remaining.
|
||||
remainBytes := dstEvery - (decoded / 4)
|
||||
for i := range br {
|
||||
offset := dstEvery * i
|
||||
endsAt := offset + remainBytes
|
||||
if endsAt > len(out) {
|
||||
endsAt = len(out)
|
||||
}
|
||||
br := &br[i]
|
||||
bitsLeft := br.remaining()
|
||||
for bitsLeft > 0 {
|
||||
br.fill()
|
||||
if offset >= endsAt {
|
||||
d.bufs.Put(buf)
|
||||
return nil, errors.New("corruption detected: stream overrun 4")
|
||||
}
|
||||
|
||||
// Read value and increment offset.
|
||||
val := br.peekBitsFast(d.actualTableLog)
|
||||
v := single[val&tlMask].entry
|
||||
nBits := uint8(v)
|
||||
br.advance(nBits)
|
||||
bitsLeft -= uint(nBits)
|
||||
out[offset] = uint8(v >> 8)
|
||||
offset++
|
||||
}
|
||||
if offset != endsAt {
|
||||
d.bufs.Put(buf)
|
||||
return nil, fmt.Errorf("corruption detected: short output block %d, end %d != %d", i, offset, endsAt)
|
||||
}
|
||||
decoded += offset - dstEvery*i
|
||||
err = br.close()
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
}
|
||||
d.bufs.Put(buf)
|
||||
if dstSize != decoded {
|
||||
return nil, errors.New("corruption detected: short output block")
|
||||
}
|
||||
return dst, nil
|
||||
}
|
||||
@ -1,14 +0,0 @@
|
||||
language: go
|
||||
sudo: required
|
||||
dist: trusty
|
||||
|
||||
go:
|
||||
- 1.9
|
||||
- tip
|
||||
|
||||
script:
|
||||
- go test -v ./...
|
||||
|
||||
before_script:
|
||||
- sudo apt-get update
|
||||
- sudo apt-get -y install libsofthsm
|
||||
196
vendor/github.com/miekg/pkcs11/const.go → vendor/github.com/miekg/pkcs11/zconst.go
generated
vendored
196
vendor/github.com/miekg/pkcs11/const.go → vendor/github.com/miekg/pkcs11/zconst.go
generated
vendored
3
vendor/github.com/moby/buildkit/session/sshforward/sshprovider/agentprovider.go
generated
vendored
3
vendor/github.com/moby/buildkit/session/sshforward/sshprovider/agentprovider.go
generated
vendored
@ -1,74 +0,0 @@
|
||||
// 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.
|
||||
|
||||
//go:build go1.13
|
||||
// +build go1.13
|
||||
|
||||
// Package ed25519 implements the Ed25519 signature algorithm. See
|
||||
// https://ed25519.cr.yp.to/.
|
||||
//
|
||||
// These functions are also compatible with the “Ed25519” function defined in
|
||||
// RFC 8032. However, unlike RFC 8032's formulation, this package's private key
|
||||
// representation includes a public key suffix to make multiple signing
|
||||
// operations with the same key more efficient. This package refers to the RFC
|
||||
// 8032 private key as the “seed”.
|
||||
//
|
||||
// Beginning with Go 1.13, the functionality of this package was moved to the
|
||||
// standard library as crypto/ed25519. This package only acts as a compatibility
|
||||
// wrapper.
|
||||
package ed25519
|
||||
|
||||
import (
|
||||
"crypto/ed25519"
|
||||
"io"
|
||||
)
|
||||
|
||||
const (
|
||||
// PublicKeySize is the size, in bytes, of public keys as used in this package.
|
||||
PublicKeySize = 32
|
||||
// PrivateKeySize is the size, in bytes, of private keys as used in this package.
|
||||
PrivateKeySize = 64
|
||||
// SignatureSize is the size, in bytes, of signatures generated and verified by this package.
|
||||
SignatureSize = 64
|
||||
// SeedSize is the size, in bytes, of private key seeds. These are the private key representations used by RFC 8032.
|
||||
SeedSize = 32
|
||||
)
|
||||
|
||||
// PublicKey is the type of Ed25519 public keys.
|
||||
//
|
||||
// This type is an alias for crypto/ed25519's PublicKey type.
|
||||
// See the crypto/ed25519 package for the methods on this type.
|
||||
type PublicKey = ed25519.PublicKey
|
||||
|
||||
// PrivateKey is the type of Ed25519 private keys. It implements crypto.Signer.
|
||||
//
|
||||
// This type is an alias for crypto/ed25519's PrivateKey type.
|
||||
// See the crypto/ed25519 package for the methods on this type.
|
||||
type PrivateKey = ed25519.PrivateKey
|
||||
|
||||
// GenerateKey generates a public/private key pair using entropy from rand.
|
||||
// If rand is nil, crypto/rand.Reader will be used.
|
||||
func GenerateKey(rand io.Reader) (PublicKey, PrivateKey, error) {
|
||||
return ed25519.GenerateKey(rand)
|
||||
}
|
||||
|
||||
// NewKeyFromSeed calculates a private key from a seed. It will panic if
|
||||
// len(seed) is not SeedSize. This function is provided for interoperability
|
||||
// with RFC 8032. RFC 8032's private keys correspond to seeds in this
|
||||
// package.
|
||||
func NewKeyFromSeed(seed []byte) PrivateKey {
|
||||
return ed25519.NewKeyFromSeed(seed)
|
||||
}
|
||||
|
||||
// Sign signs the message with privateKey and returns a signature. It will
|
||||
// panic if len(privateKey) is not PrivateKeySize.
|
||||
func Sign(privateKey PrivateKey, message []byte) []byte {
|
||||
return ed25519.Sign(privateKey, message)
|
||||
}
|
||||
|
||||
// Verify reports whether sig is a valid signature of message by publicKey. It
|
||||
// will panic if len(publicKey) is not PublicKeySize.
|
||||
func Verify(publicKey PublicKey, message, sig []byte) bool {
|
||||
return ed25519.Verify(publicKey, message, sig)
|
||||
}
|
||||
File diff suppressed because it is too large
Load Diff
File diff suppressed because it is too large
Load Diff
Loading…
Reference in New Issue