go/src/cmd/link/internal/ld/decodesym.go

370 lines
10 KiB
Go
Raw Normal View History

// Copyright 2012 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 ld
import (
cmd/link: prune unused methods Today the linker keeps all methods of reachable types. This is necessary if a program uses reflect.Value.Call. But while use of reflection is widespread in Go for encoders and decoders, using it to call a method is rare. This CL looks for the use of reflect.Value.Call in a program, and if it is absent, adopts a (reasonably conservative) method pruning strategy as part of dead code elimination. Any method that is directly called is kept, and any method that matches a used interface's method signature is kept. Whether or not a method body is kept is determined by the relocation from its receiver's *rtype to its *rtype. A small change in the compiler marks these relocations as R_METHOD so they can be easily collected and manipulated by the linker. As a bonus, this technique removes the text segment of methods that have been inlined. Looking at the output of building cmd/objdump with -ldflags=-v=2 shows that inlined methods like runtime.(*traceAllocBlockPtr).ptr are removed from the program. Relatively little work is necessary to do this. Linking two examples, jujud and cmd/objdump show no more than +2% link time. Binaries that do not use reflect.Call.Value drop 4 - 20% in size: addr2line: -793KB (18%) asm: -346KB (8%) cgo: -490KB (10%) compile: -564KB (4%) dist: -736KB (17%) fix: -404KB (12%) link: -328KB (7%) nm: -827KB (19%) objdump: -712KB (16%) pack: -327KB (14%) yacc: -350KB (10%) Binaries that do use reflect.Call.Value see a modest size decrease of 2 - 6% thanks to pruning of unexported methods: api: -151KB (3%) cover: -222KB (4%) doc: -106KB (2.5%) pprof: -314KB (3%) trace: -357KB (4%) vet: -187KB (2.7%) jujud: -4.4MB (5.8%) cmd/go: -384KB (3.4%) The trivial Hello example program goes from 2MB to 1.68MB: package main import "fmt" func main() { fmt.Println("Hello, 世界") } Method pruning also helps when building small binaries with "-ldflags=-s -w". The above program goes from 1.43MB to 1.2MB. Unfortunately the linker can only tell if reflect.Value.Call has been statically linked, not if it is dynamically used. And while use is rare, it is linked into a very common standard library package, text/template. The result is programs like cmd/go, which don't use reflect.Value.Call, see limited benefit from this CL. If binary size is important enough it may be possible to address this in future work. For #6853. Change-Id: Iabe90e210e813b08c3f8fd605f841f0458973396 Reviewed-on: https://go-review.googlesource.com/20483 Reviewed-by: Russ Cox <rsc@golang.org>
2016-03-07 23:45:04 -05:00
"bytes"
"cmd/internal/obj"
"cmd/internal/sys"
"debug/elf"
cmd/link: prune unused methods Today the linker keeps all methods of reachable types. This is necessary if a program uses reflect.Value.Call. But while use of reflection is widespread in Go for encoders and decoders, using it to call a method is rare. This CL looks for the use of reflect.Value.Call in a program, and if it is absent, adopts a (reasonably conservative) method pruning strategy as part of dead code elimination. Any method that is directly called is kept, and any method that matches a used interface's method signature is kept. Whether or not a method body is kept is determined by the relocation from its receiver's *rtype to its *rtype. A small change in the compiler marks these relocations as R_METHOD so they can be easily collected and manipulated by the linker. As a bonus, this technique removes the text segment of methods that have been inlined. Looking at the output of building cmd/objdump with -ldflags=-v=2 shows that inlined methods like runtime.(*traceAllocBlockPtr).ptr are removed from the program. Relatively little work is necessary to do this. Linking two examples, jujud and cmd/objdump show no more than +2% link time. Binaries that do not use reflect.Call.Value drop 4 - 20% in size: addr2line: -793KB (18%) asm: -346KB (8%) cgo: -490KB (10%) compile: -564KB (4%) dist: -736KB (17%) fix: -404KB (12%) link: -328KB (7%) nm: -827KB (19%) objdump: -712KB (16%) pack: -327KB (14%) yacc: -350KB (10%) Binaries that do use reflect.Call.Value see a modest size decrease of 2 - 6% thanks to pruning of unexported methods: api: -151KB (3%) cover: -222KB (4%) doc: -106KB (2.5%) pprof: -314KB (3%) trace: -357KB (4%) vet: -187KB (2.7%) jujud: -4.4MB (5.8%) cmd/go: -384KB (3.4%) The trivial Hello example program goes from 2MB to 1.68MB: package main import "fmt" func main() { fmt.Println("Hello, 世界") } Method pruning also helps when building small binaries with "-ldflags=-s -w". The above program goes from 1.43MB to 1.2MB. Unfortunately the linker can only tell if reflect.Value.Call has been statically linked, not if it is dynamically used. And while use is rare, it is linked into a very common standard library package, text/template. The result is programs like cmd/go, which don't use reflect.Value.Call, see limited benefit from this CL. If binary size is important enough it may be possible to address this in future work. For #6853. Change-Id: Iabe90e210e813b08c3f8fd605f841f0458973396 Reviewed-on: https://go-review.googlesource.com/20483 Reviewed-by: Russ Cox <rsc@golang.org>
2016-03-07 23:45:04 -05:00
"fmt"
)
// Decoding the type.* symbols. This has to be in sync with
// ../../runtime/type.go, or more specifically, with what
// ../gc/reflect.c stuffs in these.
// tflag is documented in reflect/type.go.
//
// tflag values must be kept in sync with copies in:
// cmd/compile/internal/gc/reflect.go
// cmd/link/internal/ld/decodesym.go
// reflect/type.go
// runtime/type.go
const (
tflagUncommon = 1 << 0
tflagExtraStar = 1 << 1
)
func decode_reloc(s *LSym, off int32) *Reloc {
for i := range s.R {
if s.R[i].Off == off {
return &s.R[i]
}
}
return nil
}
func decode_reloc_sym(s *LSym, off int32) *LSym {
r := decode_reloc(s, off)
if r == nil {
return nil
}
return r.Sym
}
func decode_inuxi(p []byte, sz int) uint64 {
switch sz {
case 2:
return uint64(Ctxt.Arch.ByteOrder.Uint16(p))
case 4:
return uint64(Ctxt.Arch.ByteOrder.Uint32(p))
case 8:
return Ctxt.Arch.ByteOrder.Uint64(p)
default:
Exitf("dwarf: decode inuxi %d", sz)
panic("unreachable")
}
}
func commonsize() int { return 4*SysArch.PtrSize + 8 + 8 } // runtime._type
func structfieldSize() int { return 3 * SysArch.PtrSize } // runtime.structfield
func uncommonSize() int { return 4 + 2 + 2 } // runtime.uncommontype
// Type.commonType.kind
func decodetype_kind(s *LSym) uint8 {
return s.P[2*SysArch.PtrSize+7] & obj.KindMask // 0x13 / 0x1f
}
// Type.commonType.kind
func decodetype_usegcprog(s *LSym) uint8 {
return s.P[2*SysArch.PtrSize+7] & obj.KindGCProg // 0x13 / 0x1f
}
// Type.commonType.size
func decodetype_size(s *LSym) int64 {
return int64(decode_inuxi(s.P, SysArch.PtrSize)) // 0x8 / 0x10
}
runtime: reintroduce ``dead'' space during GC scan Reintroduce an optimization discarded during the initial conversion from 4-bit heap bitmaps to 2-bit heap bitmaps: when we reach the place in the bitmap where there are no more pointers, mark that position for the GC so that it can avoid scanning past that place. During heapBitsSetType we can also avoid initializing heap bitmap beyond that location, which gives a bit of a win compared to Go 1.4. This particular optimization (not initializing the heap bitmap) may not last: we might change typedmemmove to use the heap bitmap, in which case it would all need to be initialized. The early stop in the GC scan will stay no matter what. Compared to Go 1.4 (github.com/rsc/go, branch go14bench): name old mean new mean delta SetTypeNode64 80.7ns × (1.00,1.01) 57.4ns × (1.00,1.01) -28.83% (p=0.000) SetTypeNode64Dead 80.5ns × (1.00,1.01) 13.1ns × (0.99,1.02) -83.77% (p=0.000) SetTypeNode64Slice 2.16µs × (1.00,1.01) 1.54µs × (1.00,1.01) -28.75% (p=0.000) SetTypeNode64DeadSlice 2.16µs × (1.00,1.01) 1.52µs × (1.00,1.00) -29.74% (p=0.000) Compared to previous CL: name old mean new mean delta SetTypeNode64 56.7ns × (1.00,1.00) 57.4ns × (1.00,1.01) +1.19% (p=0.000) SetTypeNode64Dead 57.2ns × (1.00,1.00) 13.1ns × (0.99,1.02) -77.15% (p=0.000) SetTypeNode64Slice 1.56µs × (1.00,1.01) 1.54µs × (1.00,1.01) -0.89% (p=0.000) SetTypeNode64DeadSlice 1.55µs × (1.00,1.01) 1.52µs × (1.00,1.00) -2.23% (p=0.000) This is the last CL in the sequence converting from the 4-bit heap to the 2-bit heap, with all the same optimizations reenabled. Compared to before that process began (compared to CL 9701 patch set 1): name old mean new mean delta BinaryTree17 5.87s × (0.94,1.09) 5.91s × (0.96,1.06) ~ (p=0.578) Fannkuch11 4.32s × (1.00,1.00) 4.32s × (1.00,1.00) ~ (p=0.474) FmtFprintfEmpty 89.1ns × (0.95,1.16) 89.0ns × (0.93,1.10) ~ (p=0.942) FmtFprintfString 283ns × (0.98,1.02) 298ns × (0.98,1.06) +5.33% (p=0.000) FmtFprintfInt 284ns × (0.98,1.04) 286ns × (0.98,1.03) ~ (p=0.208) FmtFprintfIntInt 486ns × (0.98,1.03) 498ns × (0.97,1.06) +2.48% (p=0.000) FmtFprintfPrefixedInt 400ns × (0.99,1.02) 408ns × (0.98,1.02) +2.23% (p=0.000) FmtFprintfFloat 566ns × (0.99,1.01) 587ns × (0.98,1.01) +3.69% (p=0.000) FmtManyArgs 1.91µs × (0.99,1.02) 1.94µs × (0.99,1.02) +1.81% (p=0.000) GobDecode 15.5ms × (0.98,1.05) 15.8ms × (0.98,1.03) +1.94% (p=0.002) GobEncode 11.9ms × (0.97,1.03) 12.0ms × (0.96,1.09) ~ (p=0.263) Gzip 648ms × (0.99,1.01) 648ms × (0.99,1.01) ~ (p=0.992) Gunzip 143ms × (1.00,1.00) 143ms × (1.00,1.01) ~ (p=0.585) HTTPClientServer 89.2µs × (0.99,1.02) 90.3µs × (0.98,1.01) +1.24% (p=0.000) JSONEncode 32.3ms × (0.97,1.06) 31.6ms × (0.99,1.01) -2.29% (p=0.000) JSONDecode 106ms × (0.99,1.01) 107ms × (1.00,1.01) +0.62% (p=0.000) Mandelbrot200 6.02ms × (1.00,1.00) 6.03ms × (1.00,1.01) ~ (p=0.250) GoParse 6.57ms × (0.97,1.06) 6.53ms × (0.99,1.03) ~ (p=0.243) RegexpMatchEasy0_32 162ns × (1.00,1.00) 161ns × (1.00,1.01) -0.80% (p=0.000) RegexpMatchEasy0_1K 561ns × (0.99,1.02) 541ns × (0.99,1.01) -3.67% (p=0.000) RegexpMatchEasy1_32 145ns × (0.95,1.04) 138ns × (1.00,1.00) -5.04% (p=0.000) RegexpMatchEasy1_1K 864ns × (0.99,1.04) 887ns × (0.99,1.01) +2.57% (p=0.000) RegexpMatchMedium_32 255ns × (0.99,1.04) 253ns × (0.99,1.01) -1.05% (p=0.012) RegexpMatchMedium_1K 73.9µs × (0.98,1.04) 72.8µs × (1.00,1.00) -1.51% (p=0.005) RegexpMatchHard_32 3.92µs × (0.98,1.04) 3.85µs × (1.00,1.01) -1.88% (p=0.002) RegexpMatchHard_1K 120µs × (0.98,1.04) 117µs × (1.00,1.01) -2.02% (p=0.001) Revcomp 936ms × (0.95,1.08) 922ms × (0.97,1.08) ~ (p=0.234) Template 130ms × (0.98,1.04) 126ms × (0.99,1.01) -2.99% (p=0.000) TimeParse 638ns × (0.98,1.05) 628ns × (0.99,1.01) -1.54% (p=0.004) TimeFormat 674ns × (0.99,1.01) 668ns × (0.99,1.01) -0.80% (p=0.001) The slowdown of the first few benchmarks seems to be due to the new atomic operations for certain small size allocations. But the larger benchmarks mostly improve, probably due to the decreased memory pressure from having half as much heap bitmap. CL 9706, which removes the (never used anymore) wbshadow mode, gets back what is lost in the early microbenchmarks. Change-Id: I37423a209e8ec2a2e92538b45cac5422a6acd32d Reviewed-on: https://go-review.googlesource.com/9705 Reviewed-by: Rick Hudson <rlh@golang.org>
2015-05-04 22:53:54 -04:00
// Type.commonType.ptrdata
func decodetype_ptrdata(s *LSym) int64 {
return int64(decode_inuxi(s.P[SysArch.PtrSize:], SysArch.PtrSize)) // 0x8 / 0x10
runtime: reintroduce ``dead'' space during GC scan Reintroduce an optimization discarded during the initial conversion from 4-bit heap bitmaps to 2-bit heap bitmaps: when we reach the place in the bitmap where there are no more pointers, mark that position for the GC so that it can avoid scanning past that place. During heapBitsSetType we can also avoid initializing heap bitmap beyond that location, which gives a bit of a win compared to Go 1.4. This particular optimization (not initializing the heap bitmap) may not last: we might change typedmemmove to use the heap bitmap, in which case it would all need to be initialized. The early stop in the GC scan will stay no matter what. Compared to Go 1.4 (github.com/rsc/go, branch go14bench): name old mean new mean delta SetTypeNode64 80.7ns × (1.00,1.01) 57.4ns × (1.00,1.01) -28.83% (p=0.000) SetTypeNode64Dead 80.5ns × (1.00,1.01) 13.1ns × (0.99,1.02) -83.77% (p=0.000) SetTypeNode64Slice 2.16µs × (1.00,1.01) 1.54µs × (1.00,1.01) -28.75% (p=0.000) SetTypeNode64DeadSlice 2.16µs × (1.00,1.01) 1.52µs × (1.00,1.00) -29.74% (p=0.000) Compared to previous CL: name old mean new mean delta SetTypeNode64 56.7ns × (1.00,1.00) 57.4ns × (1.00,1.01) +1.19% (p=0.000) SetTypeNode64Dead 57.2ns × (1.00,1.00) 13.1ns × (0.99,1.02) -77.15% (p=0.000) SetTypeNode64Slice 1.56µs × (1.00,1.01) 1.54µs × (1.00,1.01) -0.89% (p=0.000) SetTypeNode64DeadSlice 1.55µs × (1.00,1.01) 1.52µs × (1.00,1.00) -2.23% (p=0.000) This is the last CL in the sequence converting from the 4-bit heap to the 2-bit heap, with all the same optimizations reenabled. Compared to before that process began (compared to CL 9701 patch set 1): name old mean new mean delta BinaryTree17 5.87s × (0.94,1.09) 5.91s × (0.96,1.06) ~ (p=0.578) Fannkuch11 4.32s × (1.00,1.00) 4.32s × (1.00,1.00) ~ (p=0.474) FmtFprintfEmpty 89.1ns × (0.95,1.16) 89.0ns × (0.93,1.10) ~ (p=0.942) FmtFprintfString 283ns × (0.98,1.02) 298ns × (0.98,1.06) +5.33% (p=0.000) FmtFprintfInt 284ns × (0.98,1.04) 286ns × (0.98,1.03) ~ (p=0.208) FmtFprintfIntInt 486ns × (0.98,1.03) 498ns × (0.97,1.06) +2.48% (p=0.000) FmtFprintfPrefixedInt 400ns × (0.99,1.02) 408ns × (0.98,1.02) +2.23% (p=0.000) FmtFprintfFloat 566ns × (0.99,1.01) 587ns × (0.98,1.01) +3.69% (p=0.000) FmtManyArgs 1.91µs × (0.99,1.02) 1.94µs × (0.99,1.02) +1.81% (p=0.000) GobDecode 15.5ms × (0.98,1.05) 15.8ms × (0.98,1.03) +1.94% (p=0.002) GobEncode 11.9ms × (0.97,1.03) 12.0ms × (0.96,1.09) ~ (p=0.263) Gzip 648ms × (0.99,1.01) 648ms × (0.99,1.01) ~ (p=0.992) Gunzip 143ms × (1.00,1.00) 143ms × (1.00,1.01) ~ (p=0.585) HTTPClientServer 89.2µs × (0.99,1.02) 90.3µs × (0.98,1.01) +1.24% (p=0.000) JSONEncode 32.3ms × (0.97,1.06) 31.6ms × (0.99,1.01) -2.29% (p=0.000) JSONDecode 106ms × (0.99,1.01) 107ms × (1.00,1.01) +0.62% (p=0.000) Mandelbrot200 6.02ms × (1.00,1.00) 6.03ms × (1.00,1.01) ~ (p=0.250) GoParse 6.57ms × (0.97,1.06) 6.53ms × (0.99,1.03) ~ (p=0.243) RegexpMatchEasy0_32 162ns × (1.00,1.00) 161ns × (1.00,1.01) -0.80% (p=0.000) RegexpMatchEasy0_1K 561ns × (0.99,1.02) 541ns × (0.99,1.01) -3.67% (p=0.000) RegexpMatchEasy1_32 145ns × (0.95,1.04) 138ns × (1.00,1.00) -5.04% (p=0.000) RegexpMatchEasy1_1K 864ns × (0.99,1.04) 887ns × (0.99,1.01) +2.57% (p=0.000) RegexpMatchMedium_32 255ns × (0.99,1.04) 253ns × (0.99,1.01) -1.05% (p=0.012) RegexpMatchMedium_1K 73.9µs × (0.98,1.04) 72.8µs × (1.00,1.00) -1.51% (p=0.005) RegexpMatchHard_32 3.92µs × (0.98,1.04) 3.85µs × (1.00,1.01) -1.88% (p=0.002) RegexpMatchHard_1K 120µs × (0.98,1.04) 117µs × (1.00,1.01) -2.02% (p=0.001) Revcomp 936ms × (0.95,1.08) 922ms × (0.97,1.08) ~ (p=0.234) Template 130ms × (0.98,1.04) 126ms × (0.99,1.01) -2.99% (p=0.000) TimeParse 638ns × (0.98,1.05) 628ns × (0.99,1.01) -1.54% (p=0.004) TimeFormat 674ns × (0.99,1.01) 668ns × (0.99,1.01) -0.80% (p=0.001) The slowdown of the first few benchmarks seems to be due to the new atomic operations for certain small size allocations. But the larger benchmarks mostly improve, probably due to the decreased memory pressure from having half as much heap bitmap. CL 9706, which removes the (never used anymore) wbshadow mode, gets back what is lost in the early microbenchmarks. Change-Id: I37423a209e8ec2a2e92538b45cac5422a6acd32d Reviewed-on: https://go-review.googlesource.com/9705 Reviewed-by: Rick Hudson <rlh@golang.org>
2015-05-04 22:53:54 -04:00
}
// Type.commonType.tflag
func decodetype_hasUncommon(s *LSym) bool {
return s.P[2*SysArch.PtrSize+4]&tflagUncommon != 0
}
// Find the elf.Section of a given shared library that contains a given address.
func findShlibSection(path string, addr uint64) *elf.Section {
for _, shlib := range Ctxt.Shlibs {
if shlib.Path == path {
for _, sect := range shlib.File.Sections {
if sect.Addr <= addr && addr <= sect.Addr+sect.Size {
return sect
}
}
}
}
return nil
}
// Type.commonType.gc
func decodetype_gcprog(s *LSym) []byte {
if s.Type == obj.SDYNIMPORT {
addr := decodetype_gcprog_shlib(s)
sect := findShlibSection(s.File, addr)
if sect != nil {
// A gcprog is a 4-byte uint32 indicating length, followed by
// the actual program.
progsize := make([]byte, 4)
sect.ReadAt(progsize, int64(addr-sect.Addr))
progbytes := make([]byte, Ctxt.Arch.ByteOrder.Uint32(progsize))
sect.ReadAt(progbytes, int64(addr-sect.Addr+4))
return append(progsize, progbytes...)
}
Exitf("cannot find gcprog for %s", s.Name)
return nil
}
return decode_reloc_sym(s, 2*int32(SysArch.PtrSize)+8+1*int32(SysArch.PtrSize)).P
}
func decodetype_gcprog_shlib(s *LSym) uint64 {
if SysArch.Family == sys.ARM64 {
for _, shlib := range Ctxt.Shlibs {
if shlib.Path == s.File {
return shlib.gcdata_addresses[s]
}
}
return 0
}
return decode_inuxi(s.P[2*int32(SysArch.PtrSize)+8+1*int32(SysArch.PtrSize):], SysArch.PtrSize)
}
func decodetype_gcmask(s *LSym) []byte {
if s.Type == obj.SDYNIMPORT {
addr := decodetype_gcprog_shlib(s)
ptrdata := decodetype_ptrdata(s)
sect := findShlibSection(s.File, addr)
if sect != nil {
r := make([]byte, ptrdata/int64(SysArch.PtrSize))
sect.ReadAt(r, int64(addr-sect.Addr))
return r
}
Exitf("cannot find gcmask for %s", s.Name)
return nil
}
mask := decode_reloc_sym(s, 2*int32(SysArch.PtrSize)+8+1*int32(SysArch.PtrSize))
return mask.P
}
// Type.ArrayType.elem and Type.SliceType.Elem
func decodetype_arrayelem(s *LSym) *LSym {
return decode_reloc_sym(s, int32(commonsize())) // 0x1c / 0x30
}
func decodetype_arraylen(s *LSym) int64 {
return int64(decode_inuxi(s.P[commonsize()+2*SysArch.PtrSize:], SysArch.PtrSize))
}
// Type.PtrType.elem
func decodetype_ptrelem(s *LSym) *LSym {
return decode_reloc_sym(s, int32(commonsize())) // 0x1c / 0x30
}
// Type.MapType.key, elem
func decodetype_mapkey(s *LSym) *LSym {
return decode_reloc_sym(s, int32(commonsize())) // 0x1c / 0x30
}
func decodetype_mapvalue(s *LSym) *LSym {
return decode_reloc_sym(s, int32(commonsize())+int32(SysArch.PtrSize)) // 0x20 / 0x38
}
// Type.ChanType.elem
func decodetype_chanelem(s *LSym) *LSym {
return decode_reloc_sym(s, int32(commonsize())) // 0x1c / 0x30
}
// Type.FuncType.dotdotdot
func decodetype_funcdotdotdot(s *LSym) bool {
return uint16(decode_inuxi(s.P[commonsize()+2:], 2))&(1<<15) != 0
}
// Type.FuncType.inCount
func decodetype_funcincount(s *LSym) int {
return int(decode_inuxi(s.P[commonsize():], 2))
}
func decodetype_funcoutcount(s *LSym) int {
return int(uint16(decode_inuxi(s.P[commonsize()+2:], 2)) & (1<<15 - 1))
}
func decodetype_funcintype(s *LSym, i int) *LSym {
uadd := commonsize() + 4
if SysArch.PtrSize == 8 {
uadd += 4
}
if decodetype_hasUncommon(s) {
uadd += uncommonSize()
}
return decode_reloc_sym(s, int32(uadd+i*SysArch.PtrSize))
}
func decodetype_funcouttype(s *LSym, i int) *LSym {
return decodetype_funcintype(s, i+decodetype_funcincount(s))
}
// Type.StructType.fields.Slice::length
func decodetype_structfieldcount(s *LSym) int {
return int(decode_inuxi(s.P[commonsize()+2*SysArch.PtrSize:], SysArch.IntSize))
}
func decodetype_structfieldarrayoff(s *LSym, i int) int {
off := commonsize() + 2*SysArch.PtrSize + 2*SysArch.IntSize
if decodetype_hasUncommon(s) {
off += uncommonSize()
}
off += i * structfieldSize()
return off
}
// decodetype_str returns the contents of an rtype's str field (a nameOff).
func decodetype_str(s *LSym) string {
str := decodetype_name(s, 4*SysArch.PtrSize+8)
if s.P[2*SysArch.PtrSize+4]&tflagExtraStar != 0 {
return str[1:]
}
return str
}
// decodetype_name decodes the name from a reflect.name.
func decodetype_name(s *LSym, off int) string {
r := decode_reloc(s, int32(off))
if r == nil {
return ""
}
data := r.Sym.P
namelen := int(uint16(data[1])<<8 | uint16(data[2]))
return string(data[3 : 3+namelen])
}
cmd/link: prune unused methods Today the linker keeps all methods of reachable types. This is necessary if a program uses reflect.Value.Call. But while use of reflection is widespread in Go for encoders and decoders, using it to call a method is rare. This CL looks for the use of reflect.Value.Call in a program, and if it is absent, adopts a (reasonably conservative) method pruning strategy as part of dead code elimination. Any method that is directly called is kept, and any method that matches a used interface's method signature is kept. Whether or not a method body is kept is determined by the relocation from its receiver's *rtype to its *rtype. A small change in the compiler marks these relocations as R_METHOD so they can be easily collected and manipulated by the linker. As a bonus, this technique removes the text segment of methods that have been inlined. Looking at the output of building cmd/objdump with -ldflags=-v=2 shows that inlined methods like runtime.(*traceAllocBlockPtr).ptr are removed from the program. Relatively little work is necessary to do this. Linking two examples, jujud and cmd/objdump show no more than +2% link time. Binaries that do not use reflect.Call.Value drop 4 - 20% in size: addr2line: -793KB (18%) asm: -346KB (8%) cgo: -490KB (10%) compile: -564KB (4%) dist: -736KB (17%) fix: -404KB (12%) link: -328KB (7%) nm: -827KB (19%) objdump: -712KB (16%) pack: -327KB (14%) yacc: -350KB (10%) Binaries that do use reflect.Call.Value see a modest size decrease of 2 - 6% thanks to pruning of unexported methods: api: -151KB (3%) cover: -222KB (4%) doc: -106KB (2.5%) pprof: -314KB (3%) trace: -357KB (4%) vet: -187KB (2.7%) jujud: -4.4MB (5.8%) cmd/go: -384KB (3.4%) The trivial Hello example program goes from 2MB to 1.68MB: package main import "fmt" func main() { fmt.Println("Hello, 世界") } Method pruning also helps when building small binaries with "-ldflags=-s -w". The above program goes from 1.43MB to 1.2MB. Unfortunately the linker can only tell if reflect.Value.Call has been statically linked, not if it is dynamically used. And while use is rare, it is linked into a very common standard library package, text/template. The result is programs like cmd/go, which don't use reflect.Value.Call, see limited benefit from this CL. If binary size is important enough it may be possible to address this in future work. For #6853. Change-Id: Iabe90e210e813b08c3f8fd605f841f0458973396 Reviewed-on: https://go-review.googlesource.com/20483 Reviewed-by: Russ Cox <rsc@golang.org>
2016-03-07 23:45:04 -05:00
func decodetype_structfieldname(s *LSym, i int) string {
off := decodetype_structfieldarrayoff(s, i)
return decodetype_name(s, off)
cmd/link: prune unused methods Today the linker keeps all methods of reachable types. This is necessary if a program uses reflect.Value.Call. But while use of reflection is widespread in Go for encoders and decoders, using it to call a method is rare. This CL looks for the use of reflect.Value.Call in a program, and if it is absent, adopts a (reasonably conservative) method pruning strategy as part of dead code elimination. Any method that is directly called is kept, and any method that matches a used interface's method signature is kept. Whether or not a method body is kept is determined by the relocation from its receiver's *rtype to its *rtype. A small change in the compiler marks these relocations as R_METHOD so they can be easily collected and manipulated by the linker. As a bonus, this technique removes the text segment of methods that have been inlined. Looking at the output of building cmd/objdump with -ldflags=-v=2 shows that inlined methods like runtime.(*traceAllocBlockPtr).ptr are removed from the program. Relatively little work is necessary to do this. Linking two examples, jujud and cmd/objdump show no more than +2% link time. Binaries that do not use reflect.Call.Value drop 4 - 20% in size: addr2line: -793KB (18%) asm: -346KB (8%) cgo: -490KB (10%) compile: -564KB (4%) dist: -736KB (17%) fix: -404KB (12%) link: -328KB (7%) nm: -827KB (19%) objdump: -712KB (16%) pack: -327KB (14%) yacc: -350KB (10%) Binaries that do use reflect.Call.Value see a modest size decrease of 2 - 6% thanks to pruning of unexported methods: api: -151KB (3%) cover: -222KB (4%) doc: -106KB (2.5%) pprof: -314KB (3%) trace: -357KB (4%) vet: -187KB (2.7%) jujud: -4.4MB (5.8%) cmd/go: -384KB (3.4%) The trivial Hello example program goes from 2MB to 1.68MB: package main import "fmt" func main() { fmt.Println("Hello, 世界") } Method pruning also helps when building small binaries with "-ldflags=-s -w". The above program goes from 1.43MB to 1.2MB. Unfortunately the linker can only tell if reflect.Value.Call has been statically linked, not if it is dynamically used. And while use is rare, it is linked into a very common standard library package, text/template. The result is programs like cmd/go, which don't use reflect.Value.Call, see limited benefit from this CL. If binary size is important enough it may be possible to address this in future work. For #6853. Change-Id: Iabe90e210e813b08c3f8fd605f841f0458973396 Reviewed-on: https://go-review.googlesource.com/20483 Reviewed-by: Russ Cox <rsc@golang.org>
2016-03-07 23:45:04 -05:00
}
func decodetype_structfieldtype(s *LSym, i int) *LSym {
off := decodetype_structfieldarrayoff(s, i)
return decode_reloc_sym(s, int32(off+SysArch.PtrSize))
}
func decodetype_structfieldoffs(s *LSym, i int) int64 {
off := decodetype_structfieldarrayoff(s, i)
return int64(decode_inuxi(s.P[off+2*SysArch.PtrSize:], SysArch.IntSize))
}
// InterfaceType.methods.length
func decodetype_ifacemethodcount(s *LSym) int64 {
return int64(decode_inuxi(s.P[commonsize()+2*SysArch.PtrSize:], SysArch.IntSize))
}
cmd/link: prune unused methods Today the linker keeps all methods of reachable types. This is necessary if a program uses reflect.Value.Call. But while use of reflection is widespread in Go for encoders and decoders, using it to call a method is rare. This CL looks for the use of reflect.Value.Call in a program, and if it is absent, adopts a (reasonably conservative) method pruning strategy as part of dead code elimination. Any method that is directly called is kept, and any method that matches a used interface's method signature is kept. Whether or not a method body is kept is determined by the relocation from its receiver's *rtype to its *rtype. A small change in the compiler marks these relocations as R_METHOD so they can be easily collected and manipulated by the linker. As a bonus, this technique removes the text segment of methods that have been inlined. Looking at the output of building cmd/objdump with -ldflags=-v=2 shows that inlined methods like runtime.(*traceAllocBlockPtr).ptr are removed from the program. Relatively little work is necessary to do this. Linking two examples, jujud and cmd/objdump show no more than +2% link time. Binaries that do not use reflect.Call.Value drop 4 - 20% in size: addr2line: -793KB (18%) asm: -346KB (8%) cgo: -490KB (10%) compile: -564KB (4%) dist: -736KB (17%) fix: -404KB (12%) link: -328KB (7%) nm: -827KB (19%) objdump: -712KB (16%) pack: -327KB (14%) yacc: -350KB (10%) Binaries that do use reflect.Call.Value see a modest size decrease of 2 - 6% thanks to pruning of unexported methods: api: -151KB (3%) cover: -222KB (4%) doc: -106KB (2.5%) pprof: -314KB (3%) trace: -357KB (4%) vet: -187KB (2.7%) jujud: -4.4MB (5.8%) cmd/go: -384KB (3.4%) The trivial Hello example program goes from 2MB to 1.68MB: package main import "fmt" func main() { fmt.Println("Hello, 世界") } Method pruning also helps when building small binaries with "-ldflags=-s -w". The above program goes from 1.43MB to 1.2MB. Unfortunately the linker can only tell if reflect.Value.Call has been statically linked, not if it is dynamically used. And while use is rare, it is linked into a very common standard library package, text/template. The result is programs like cmd/go, which don't use reflect.Value.Call, see limited benefit from this CL. If binary size is important enough it may be possible to address this in future work. For #6853. Change-Id: Iabe90e210e813b08c3f8fd605f841f0458973396 Reviewed-on: https://go-review.googlesource.com/20483 Reviewed-by: Russ Cox <rsc@golang.org>
2016-03-07 23:45:04 -05:00
// methodsig is a fully qualified typed method signature, like
// "Visit(type.go/ast.Node) (type.go/ast.Visitor)".
type methodsig string
// Matches runtime/typekind.go and reflect.Kind.
const (
kindArray = 17
kindChan = 18
kindFunc = 19
kindInterface = 20
kindMap = 21
kindPtr = 22
kindSlice = 23
kindStruct = 25
kindMask = (1 << 5) - 1
)
// decode_methodsig decodes an array of method signature information.
// Each element of the array is size bytes. The first 4 bytes is a
// nameOff for the method name, and the next 4 bytes is a typeOff for
// the function type.
cmd/link: prune unused methods Today the linker keeps all methods of reachable types. This is necessary if a program uses reflect.Value.Call. But while use of reflection is widespread in Go for encoders and decoders, using it to call a method is rare. This CL looks for the use of reflect.Value.Call in a program, and if it is absent, adopts a (reasonably conservative) method pruning strategy as part of dead code elimination. Any method that is directly called is kept, and any method that matches a used interface's method signature is kept. Whether or not a method body is kept is determined by the relocation from its receiver's *rtype to its *rtype. A small change in the compiler marks these relocations as R_METHOD so they can be easily collected and manipulated by the linker. As a bonus, this technique removes the text segment of methods that have been inlined. Looking at the output of building cmd/objdump with -ldflags=-v=2 shows that inlined methods like runtime.(*traceAllocBlockPtr).ptr are removed from the program. Relatively little work is necessary to do this. Linking two examples, jujud and cmd/objdump show no more than +2% link time. Binaries that do not use reflect.Call.Value drop 4 - 20% in size: addr2line: -793KB (18%) asm: -346KB (8%) cgo: -490KB (10%) compile: -564KB (4%) dist: -736KB (17%) fix: -404KB (12%) link: -328KB (7%) nm: -827KB (19%) objdump: -712KB (16%) pack: -327KB (14%) yacc: -350KB (10%) Binaries that do use reflect.Call.Value see a modest size decrease of 2 - 6% thanks to pruning of unexported methods: api: -151KB (3%) cover: -222KB (4%) doc: -106KB (2.5%) pprof: -314KB (3%) trace: -357KB (4%) vet: -187KB (2.7%) jujud: -4.4MB (5.8%) cmd/go: -384KB (3.4%) The trivial Hello example program goes from 2MB to 1.68MB: package main import "fmt" func main() { fmt.Println("Hello, 世界") } Method pruning also helps when building small binaries with "-ldflags=-s -w". The above program goes from 1.43MB to 1.2MB. Unfortunately the linker can only tell if reflect.Value.Call has been statically linked, not if it is dynamically used. And while use is rare, it is linked into a very common standard library package, text/template. The result is programs like cmd/go, which don't use reflect.Value.Call, see limited benefit from this CL. If binary size is important enough it may be possible to address this in future work. For #6853. Change-Id: Iabe90e210e813b08c3f8fd605f841f0458973396 Reviewed-on: https://go-review.googlesource.com/20483 Reviewed-by: Russ Cox <rsc@golang.org>
2016-03-07 23:45:04 -05:00
//
// Conveniently this is the layout of both runtime.method and runtime.imethod.
func decode_methodsig(s *LSym, off, size, count int) []methodsig {
var buf bytes.Buffer
var methods []methodsig
for i := 0; i < count; i++ {
buf.WriteString(decodetype_name(s, off))
mtypSym := decode_reloc_sym(s, int32(off+4))
cmd/link: prune unused methods Today the linker keeps all methods of reachable types. This is necessary if a program uses reflect.Value.Call. But while use of reflection is widespread in Go for encoders and decoders, using it to call a method is rare. This CL looks for the use of reflect.Value.Call in a program, and if it is absent, adopts a (reasonably conservative) method pruning strategy as part of dead code elimination. Any method that is directly called is kept, and any method that matches a used interface's method signature is kept. Whether or not a method body is kept is determined by the relocation from its receiver's *rtype to its *rtype. A small change in the compiler marks these relocations as R_METHOD so they can be easily collected and manipulated by the linker. As a bonus, this technique removes the text segment of methods that have been inlined. Looking at the output of building cmd/objdump with -ldflags=-v=2 shows that inlined methods like runtime.(*traceAllocBlockPtr).ptr are removed from the program. Relatively little work is necessary to do this. Linking two examples, jujud and cmd/objdump show no more than +2% link time. Binaries that do not use reflect.Call.Value drop 4 - 20% in size: addr2line: -793KB (18%) asm: -346KB (8%) cgo: -490KB (10%) compile: -564KB (4%) dist: -736KB (17%) fix: -404KB (12%) link: -328KB (7%) nm: -827KB (19%) objdump: -712KB (16%) pack: -327KB (14%) yacc: -350KB (10%) Binaries that do use reflect.Call.Value see a modest size decrease of 2 - 6% thanks to pruning of unexported methods: api: -151KB (3%) cover: -222KB (4%) doc: -106KB (2.5%) pprof: -314KB (3%) trace: -357KB (4%) vet: -187KB (2.7%) jujud: -4.4MB (5.8%) cmd/go: -384KB (3.4%) The trivial Hello example program goes from 2MB to 1.68MB: package main import "fmt" func main() { fmt.Println("Hello, 世界") } Method pruning also helps when building small binaries with "-ldflags=-s -w". The above program goes from 1.43MB to 1.2MB. Unfortunately the linker can only tell if reflect.Value.Call has been statically linked, not if it is dynamically used. And while use is rare, it is linked into a very common standard library package, text/template. The result is programs like cmd/go, which don't use reflect.Value.Call, see limited benefit from this CL. If binary size is important enough it may be possible to address this in future work. For #6853. Change-Id: Iabe90e210e813b08c3f8fd605f841f0458973396 Reviewed-on: https://go-review.googlesource.com/20483 Reviewed-by: Russ Cox <rsc@golang.org>
2016-03-07 23:45:04 -05:00
buf.WriteRune('(')
inCount := decodetype_funcincount(mtypSym)
for i := 0; i < inCount; i++ {
if i > 0 {
buf.WriteString(", ")
}
buf.WriteString(decodetype_funcintype(mtypSym, i).Name)
}
buf.WriteString(") (")
outCount := decodetype_funcoutcount(mtypSym)
for i := 0; i < outCount; i++ {
if i > 0 {
buf.WriteString(", ")
}
buf.WriteString(decodetype_funcouttype(mtypSym, i).Name)
}
buf.WriteRune(')')
off += size
methods = append(methods, methodsig(buf.String()))
buf.Reset()
}
return methods
}
func decodetype_ifacemethods(s *LSym) []methodsig {
if decodetype_kind(s)&kindMask != kindInterface {
panic(fmt.Sprintf("symbol %q is not an interface", s.Name))
}
r := decode_reloc(s, int32(commonsize()+SysArch.PtrSize))
cmd/link: prune unused methods Today the linker keeps all methods of reachable types. This is necessary if a program uses reflect.Value.Call. But while use of reflection is widespread in Go for encoders and decoders, using it to call a method is rare. This CL looks for the use of reflect.Value.Call in a program, and if it is absent, adopts a (reasonably conservative) method pruning strategy as part of dead code elimination. Any method that is directly called is kept, and any method that matches a used interface's method signature is kept. Whether or not a method body is kept is determined by the relocation from its receiver's *rtype to its *rtype. A small change in the compiler marks these relocations as R_METHOD so they can be easily collected and manipulated by the linker. As a bonus, this technique removes the text segment of methods that have been inlined. Looking at the output of building cmd/objdump with -ldflags=-v=2 shows that inlined methods like runtime.(*traceAllocBlockPtr).ptr are removed from the program. Relatively little work is necessary to do this. Linking two examples, jujud and cmd/objdump show no more than +2% link time. Binaries that do not use reflect.Call.Value drop 4 - 20% in size: addr2line: -793KB (18%) asm: -346KB (8%) cgo: -490KB (10%) compile: -564KB (4%) dist: -736KB (17%) fix: -404KB (12%) link: -328KB (7%) nm: -827KB (19%) objdump: -712KB (16%) pack: -327KB (14%) yacc: -350KB (10%) Binaries that do use reflect.Call.Value see a modest size decrease of 2 - 6% thanks to pruning of unexported methods: api: -151KB (3%) cover: -222KB (4%) doc: -106KB (2.5%) pprof: -314KB (3%) trace: -357KB (4%) vet: -187KB (2.7%) jujud: -4.4MB (5.8%) cmd/go: -384KB (3.4%) The trivial Hello example program goes from 2MB to 1.68MB: package main import "fmt" func main() { fmt.Println("Hello, 世界") } Method pruning also helps when building small binaries with "-ldflags=-s -w". The above program goes from 1.43MB to 1.2MB. Unfortunately the linker can only tell if reflect.Value.Call has been statically linked, not if it is dynamically used. And while use is rare, it is linked into a very common standard library package, text/template. The result is programs like cmd/go, which don't use reflect.Value.Call, see limited benefit from this CL. If binary size is important enough it may be possible to address this in future work. For #6853. Change-Id: Iabe90e210e813b08c3f8fd605f841f0458973396 Reviewed-on: https://go-review.googlesource.com/20483 Reviewed-by: Russ Cox <rsc@golang.org>
2016-03-07 23:45:04 -05:00
if r == nil {
return nil
}
if r.Sym != s {
panic(fmt.Sprintf("imethod slice pointer in %q leads to a different symbol", s.Name))
}
off := int(r.Add) // array of reflect.imethod values
numMethods := int(decodetype_ifacemethodcount(s))
sizeofIMethod := 4 + 4
cmd/link: prune unused methods Today the linker keeps all methods of reachable types. This is necessary if a program uses reflect.Value.Call. But while use of reflection is widespread in Go for encoders and decoders, using it to call a method is rare. This CL looks for the use of reflect.Value.Call in a program, and if it is absent, adopts a (reasonably conservative) method pruning strategy as part of dead code elimination. Any method that is directly called is kept, and any method that matches a used interface's method signature is kept. Whether or not a method body is kept is determined by the relocation from its receiver's *rtype to its *rtype. A small change in the compiler marks these relocations as R_METHOD so they can be easily collected and manipulated by the linker. As a bonus, this technique removes the text segment of methods that have been inlined. Looking at the output of building cmd/objdump with -ldflags=-v=2 shows that inlined methods like runtime.(*traceAllocBlockPtr).ptr are removed from the program. Relatively little work is necessary to do this. Linking two examples, jujud and cmd/objdump show no more than +2% link time. Binaries that do not use reflect.Call.Value drop 4 - 20% in size: addr2line: -793KB (18%) asm: -346KB (8%) cgo: -490KB (10%) compile: -564KB (4%) dist: -736KB (17%) fix: -404KB (12%) link: -328KB (7%) nm: -827KB (19%) objdump: -712KB (16%) pack: -327KB (14%) yacc: -350KB (10%) Binaries that do use reflect.Call.Value see a modest size decrease of 2 - 6% thanks to pruning of unexported methods: api: -151KB (3%) cover: -222KB (4%) doc: -106KB (2.5%) pprof: -314KB (3%) trace: -357KB (4%) vet: -187KB (2.7%) jujud: -4.4MB (5.8%) cmd/go: -384KB (3.4%) The trivial Hello example program goes from 2MB to 1.68MB: package main import "fmt" func main() { fmt.Println("Hello, 世界") } Method pruning also helps when building small binaries with "-ldflags=-s -w". The above program goes from 1.43MB to 1.2MB. Unfortunately the linker can only tell if reflect.Value.Call has been statically linked, not if it is dynamically used. And while use is rare, it is linked into a very common standard library package, text/template. The result is programs like cmd/go, which don't use reflect.Value.Call, see limited benefit from this CL. If binary size is important enough it may be possible to address this in future work. For #6853. Change-Id: Iabe90e210e813b08c3f8fd605f841f0458973396 Reviewed-on: https://go-review.googlesource.com/20483 Reviewed-by: Russ Cox <rsc@golang.org>
2016-03-07 23:45:04 -05:00
return decode_methodsig(s, off, sizeofIMethod, numMethods)
}
func decodetype_methods(s *LSym) []methodsig {
if !decodetype_hasUncommon(s) {
panic(fmt.Sprintf("no methods on %q", s.Name))
}
off := commonsize() // reflect.rtype
switch decodetype_kind(s) & kindMask {
case kindStruct: // reflect.structType
off += 2*SysArch.PtrSize + 2*SysArch.IntSize
cmd/link: prune unused methods Today the linker keeps all methods of reachable types. This is necessary if a program uses reflect.Value.Call. But while use of reflection is widespread in Go for encoders and decoders, using it to call a method is rare. This CL looks for the use of reflect.Value.Call in a program, and if it is absent, adopts a (reasonably conservative) method pruning strategy as part of dead code elimination. Any method that is directly called is kept, and any method that matches a used interface's method signature is kept. Whether or not a method body is kept is determined by the relocation from its receiver's *rtype to its *rtype. A small change in the compiler marks these relocations as R_METHOD so they can be easily collected and manipulated by the linker. As a bonus, this technique removes the text segment of methods that have been inlined. Looking at the output of building cmd/objdump with -ldflags=-v=2 shows that inlined methods like runtime.(*traceAllocBlockPtr).ptr are removed from the program. Relatively little work is necessary to do this. Linking two examples, jujud and cmd/objdump show no more than +2% link time. Binaries that do not use reflect.Call.Value drop 4 - 20% in size: addr2line: -793KB (18%) asm: -346KB (8%) cgo: -490KB (10%) compile: -564KB (4%) dist: -736KB (17%) fix: -404KB (12%) link: -328KB (7%) nm: -827KB (19%) objdump: -712KB (16%) pack: -327KB (14%) yacc: -350KB (10%) Binaries that do use reflect.Call.Value see a modest size decrease of 2 - 6% thanks to pruning of unexported methods: api: -151KB (3%) cover: -222KB (4%) doc: -106KB (2.5%) pprof: -314KB (3%) trace: -357KB (4%) vet: -187KB (2.7%) jujud: -4.4MB (5.8%) cmd/go: -384KB (3.4%) The trivial Hello example program goes from 2MB to 1.68MB: package main import "fmt" func main() { fmt.Println("Hello, 世界") } Method pruning also helps when building small binaries with "-ldflags=-s -w". The above program goes from 1.43MB to 1.2MB. Unfortunately the linker can only tell if reflect.Value.Call has been statically linked, not if it is dynamically used. And while use is rare, it is linked into a very common standard library package, text/template. The result is programs like cmd/go, which don't use reflect.Value.Call, see limited benefit from this CL. If binary size is important enough it may be possible to address this in future work. For #6853. Change-Id: Iabe90e210e813b08c3f8fd605f841f0458973396 Reviewed-on: https://go-review.googlesource.com/20483 Reviewed-by: Russ Cox <rsc@golang.org>
2016-03-07 23:45:04 -05:00
case kindPtr: // reflect.ptrType
off += SysArch.PtrSize
cmd/link: prune unused methods Today the linker keeps all methods of reachable types. This is necessary if a program uses reflect.Value.Call. But while use of reflection is widespread in Go for encoders and decoders, using it to call a method is rare. This CL looks for the use of reflect.Value.Call in a program, and if it is absent, adopts a (reasonably conservative) method pruning strategy as part of dead code elimination. Any method that is directly called is kept, and any method that matches a used interface's method signature is kept. Whether or not a method body is kept is determined by the relocation from its receiver's *rtype to its *rtype. A small change in the compiler marks these relocations as R_METHOD so they can be easily collected and manipulated by the linker. As a bonus, this technique removes the text segment of methods that have been inlined. Looking at the output of building cmd/objdump with -ldflags=-v=2 shows that inlined methods like runtime.(*traceAllocBlockPtr).ptr are removed from the program. Relatively little work is necessary to do this. Linking two examples, jujud and cmd/objdump show no more than +2% link time. Binaries that do not use reflect.Call.Value drop 4 - 20% in size: addr2line: -793KB (18%) asm: -346KB (8%) cgo: -490KB (10%) compile: -564KB (4%) dist: -736KB (17%) fix: -404KB (12%) link: -328KB (7%) nm: -827KB (19%) objdump: -712KB (16%) pack: -327KB (14%) yacc: -350KB (10%) Binaries that do use reflect.Call.Value see a modest size decrease of 2 - 6% thanks to pruning of unexported methods: api: -151KB (3%) cover: -222KB (4%) doc: -106KB (2.5%) pprof: -314KB (3%) trace: -357KB (4%) vet: -187KB (2.7%) jujud: -4.4MB (5.8%) cmd/go: -384KB (3.4%) The trivial Hello example program goes from 2MB to 1.68MB: package main import "fmt" func main() { fmt.Println("Hello, 世界") } Method pruning also helps when building small binaries with "-ldflags=-s -w". The above program goes from 1.43MB to 1.2MB. Unfortunately the linker can only tell if reflect.Value.Call has been statically linked, not if it is dynamically used. And while use is rare, it is linked into a very common standard library package, text/template. The result is programs like cmd/go, which don't use reflect.Value.Call, see limited benefit from this CL. If binary size is important enough it may be possible to address this in future work. For #6853. Change-Id: Iabe90e210e813b08c3f8fd605f841f0458973396 Reviewed-on: https://go-review.googlesource.com/20483 Reviewed-by: Russ Cox <rsc@golang.org>
2016-03-07 23:45:04 -05:00
case kindFunc: // reflect.funcType
off += SysArch.PtrSize // 4 bytes, pointer aligned
cmd/link: prune unused methods Today the linker keeps all methods of reachable types. This is necessary if a program uses reflect.Value.Call. But while use of reflection is widespread in Go for encoders and decoders, using it to call a method is rare. This CL looks for the use of reflect.Value.Call in a program, and if it is absent, adopts a (reasonably conservative) method pruning strategy as part of dead code elimination. Any method that is directly called is kept, and any method that matches a used interface's method signature is kept. Whether or not a method body is kept is determined by the relocation from its receiver's *rtype to its *rtype. A small change in the compiler marks these relocations as R_METHOD so they can be easily collected and manipulated by the linker. As a bonus, this technique removes the text segment of methods that have been inlined. Looking at the output of building cmd/objdump with -ldflags=-v=2 shows that inlined methods like runtime.(*traceAllocBlockPtr).ptr are removed from the program. Relatively little work is necessary to do this. Linking two examples, jujud and cmd/objdump show no more than +2% link time. Binaries that do not use reflect.Call.Value drop 4 - 20% in size: addr2line: -793KB (18%) asm: -346KB (8%) cgo: -490KB (10%) compile: -564KB (4%) dist: -736KB (17%) fix: -404KB (12%) link: -328KB (7%) nm: -827KB (19%) objdump: -712KB (16%) pack: -327KB (14%) yacc: -350KB (10%) Binaries that do use reflect.Call.Value see a modest size decrease of 2 - 6% thanks to pruning of unexported methods: api: -151KB (3%) cover: -222KB (4%) doc: -106KB (2.5%) pprof: -314KB (3%) trace: -357KB (4%) vet: -187KB (2.7%) jujud: -4.4MB (5.8%) cmd/go: -384KB (3.4%) The trivial Hello example program goes from 2MB to 1.68MB: package main import "fmt" func main() { fmt.Println("Hello, 世界") } Method pruning also helps when building small binaries with "-ldflags=-s -w". The above program goes from 1.43MB to 1.2MB. Unfortunately the linker can only tell if reflect.Value.Call has been statically linked, not if it is dynamically used. And while use is rare, it is linked into a very common standard library package, text/template. The result is programs like cmd/go, which don't use reflect.Value.Call, see limited benefit from this CL. If binary size is important enough it may be possible to address this in future work. For #6853. Change-Id: Iabe90e210e813b08c3f8fd605f841f0458973396 Reviewed-on: https://go-review.googlesource.com/20483 Reviewed-by: Russ Cox <rsc@golang.org>
2016-03-07 23:45:04 -05:00
case kindSlice: // reflect.sliceType
off += SysArch.PtrSize
cmd/link: prune unused methods Today the linker keeps all methods of reachable types. This is necessary if a program uses reflect.Value.Call. But while use of reflection is widespread in Go for encoders and decoders, using it to call a method is rare. This CL looks for the use of reflect.Value.Call in a program, and if it is absent, adopts a (reasonably conservative) method pruning strategy as part of dead code elimination. Any method that is directly called is kept, and any method that matches a used interface's method signature is kept. Whether or not a method body is kept is determined by the relocation from its receiver's *rtype to its *rtype. A small change in the compiler marks these relocations as R_METHOD so they can be easily collected and manipulated by the linker. As a bonus, this technique removes the text segment of methods that have been inlined. Looking at the output of building cmd/objdump with -ldflags=-v=2 shows that inlined methods like runtime.(*traceAllocBlockPtr).ptr are removed from the program. Relatively little work is necessary to do this. Linking two examples, jujud and cmd/objdump show no more than +2% link time. Binaries that do not use reflect.Call.Value drop 4 - 20% in size: addr2line: -793KB (18%) asm: -346KB (8%) cgo: -490KB (10%) compile: -564KB (4%) dist: -736KB (17%) fix: -404KB (12%) link: -328KB (7%) nm: -827KB (19%) objdump: -712KB (16%) pack: -327KB (14%) yacc: -350KB (10%) Binaries that do use reflect.Call.Value see a modest size decrease of 2 - 6% thanks to pruning of unexported methods: api: -151KB (3%) cover: -222KB (4%) doc: -106KB (2.5%) pprof: -314KB (3%) trace: -357KB (4%) vet: -187KB (2.7%) jujud: -4.4MB (5.8%) cmd/go: -384KB (3.4%) The trivial Hello example program goes from 2MB to 1.68MB: package main import "fmt" func main() { fmt.Println("Hello, 世界") } Method pruning also helps when building small binaries with "-ldflags=-s -w". The above program goes from 1.43MB to 1.2MB. Unfortunately the linker can only tell if reflect.Value.Call has been statically linked, not if it is dynamically used. And while use is rare, it is linked into a very common standard library package, text/template. The result is programs like cmd/go, which don't use reflect.Value.Call, see limited benefit from this CL. If binary size is important enough it may be possible to address this in future work. For #6853. Change-Id: Iabe90e210e813b08c3f8fd605f841f0458973396 Reviewed-on: https://go-review.googlesource.com/20483 Reviewed-by: Russ Cox <rsc@golang.org>
2016-03-07 23:45:04 -05:00
case kindArray: // reflect.arrayType
off += 3 * SysArch.PtrSize
cmd/link: prune unused methods Today the linker keeps all methods of reachable types. This is necessary if a program uses reflect.Value.Call. But while use of reflection is widespread in Go for encoders and decoders, using it to call a method is rare. This CL looks for the use of reflect.Value.Call in a program, and if it is absent, adopts a (reasonably conservative) method pruning strategy as part of dead code elimination. Any method that is directly called is kept, and any method that matches a used interface's method signature is kept. Whether or not a method body is kept is determined by the relocation from its receiver's *rtype to its *rtype. A small change in the compiler marks these relocations as R_METHOD so they can be easily collected and manipulated by the linker. As a bonus, this technique removes the text segment of methods that have been inlined. Looking at the output of building cmd/objdump with -ldflags=-v=2 shows that inlined methods like runtime.(*traceAllocBlockPtr).ptr are removed from the program. Relatively little work is necessary to do this. Linking two examples, jujud and cmd/objdump show no more than +2% link time. Binaries that do not use reflect.Call.Value drop 4 - 20% in size: addr2line: -793KB (18%) asm: -346KB (8%) cgo: -490KB (10%) compile: -564KB (4%) dist: -736KB (17%) fix: -404KB (12%) link: -328KB (7%) nm: -827KB (19%) objdump: -712KB (16%) pack: -327KB (14%) yacc: -350KB (10%) Binaries that do use reflect.Call.Value see a modest size decrease of 2 - 6% thanks to pruning of unexported methods: api: -151KB (3%) cover: -222KB (4%) doc: -106KB (2.5%) pprof: -314KB (3%) trace: -357KB (4%) vet: -187KB (2.7%) jujud: -4.4MB (5.8%) cmd/go: -384KB (3.4%) The trivial Hello example program goes from 2MB to 1.68MB: package main import "fmt" func main() { fmt.Println("Hello, 世界") } Method pruning also helps when building small binaries with "-ldflags=-s -w". The above program goes from 1.43MB to 1.2MB. Unfortunately the linker can only tell if reflect.Value.Call has been statically linked, not if it is dynamically used. And while use is rare, it is linked into a very common standard library package, text/template. The result is programs like cmd/go, which don't use reflect.Value.Call, see limited benefit from this CL. If binary size is important enough it may be possible to address this in future work. For #6853. Change-Id: Iabe90e210e813b08c3f8fd605f841f0458973396 Reviewed-on: https://go-review.googlesource.com/20483 Reviewed-by: Russ Cox <rsc@golang.org>
2016-03-07 23:45:04 -05:00
case kindChan: // reflect.chanType
off += 2 * SysArch.PtrSize
cmd/link: prune unused methods Today the linker keeps all methods of reachable types. This is necessary if a program uses reflect.Value.Call. But while use of reflection is widespread in Go for encoders and decoders, using it to call a method is rare. This CL looks for the use of reflect.Value.Call in a program, and if it is absent, adopts a (reasonably conservative) method pruning strategy as part of dead code elimination. Any method that is directly called is kept, and any method that matches a used interface's method signature is kept. Whether or not a method body is kept is determined by the relocation from its receiver's *rtype to its *rtype. A small change in the compiler marks these relocations as R_METHOD so they can be easily collected and manipulated by the linker. As a bonus, this technique removes the text segment of methods that have been inlined. Looking at the output of building cmd/objdump with -ldflags=-v=2 shows that inlined methods like runtime.(*traceAllocBlockPtr).ptr are removed from the program. Relatively little work is necessary to do this. Linking two examples, jujud and cmd/objdump show no more than +2% link time. Binaries that do not use reflect.Call.Value drop 4 - 20% in size: addr2line: -793KB (18%) asm: -346KB (8%) cgo: -490KB (10%) compile: -564KB (4%) dist: -736KB (17%) fix: -404KB (12%) link: -328KB (7%) nm: -827KB (19%) objdump: -712KB (16%) pack: -327KB (14%) yacc: -350KB (10%) Binaries that do use reflect.Call.Value see a modest size decrease of 2 - 6% thanks to pruning of unexported methods: api: -151KB (3%) cover: -222KB (4%) doc: -106KB (2.5%) pprof: -314KB (3%) trace: -357KB (4%) vet: -187KB (2.7%) jujud: -4.4MB (5.8%) cmd/go: -384KB (3.4%) The trivial Hello example program goes from 2MB to 1.68MB: package main import "fmt" func main() { fmt.Println("Hello, 世界") } Method pruning also helps when building small binaries with "-ldflags=-s -w". The above program goes from 1.43MB to 1.2MB. Unfortunately the linker can only tell if reflect.Value.Call has been statically linked, not if it is dynamically used. And while use is rare, it is linked into a very common standard library package, text/template. The result is programs like cmd/go, which don't use reflect.Value.Call, see limited benefit from this CL. If binary size is important enough it may be possible to address this in future work. For #6853. Change-Id: Iabe90e210e813b08c3f8fd605f841f0458973396 Reviewed-on: https://go-review.googlesource.com/20483 Reviewed-by: Russ Cox <rsc@golang.org>
2016-03-07 23:45:04 -05:00
case kindMap: // reflect.mapType
off += 4*SysArch.PtrSize + 8
cmd/link: prune unused methods Today the linker keeps all methods of reachable types. This is necessary if a program uses reflect.Value.Call. But while use of reflection is widespread in Go for encoders and decoders, using it to call a method is rare. This CL looks for the use of reflect.Value.Call in a program, and if it is absent, adopts a (reasonably conservative) method pruning strategy as part of dead code elimination. Any method that is directly called is kept, and any method that matches a used interface's method signature is kept. Whether or not a method body is kept is determined by the relocation from its receiver's *rtype to its *rtype. A small change in the compiler marks these relocations as R_METHOD so they can be easily collected and manipulated by the linker. As a bonus, this technique removes the text segment of methods that have been inlined. Looking at the output of building cmd/objdump with -ldflags=-v=2 shows that inlined methods like runtime.(*traceAllocBlockPtr).ptr are removed from the program. Relatively little work is necessary to do this. Linking two examples, jujud and cmd/objdump show no more than +2% link time. Binaries that do not use reflect.Call.Value drop 4 - 20% in size: addr2line: -793KB (18%) asm: -346KB (8%) cgo: -490KB (10%) compile: -564KB (4%) dist: -736KB (17%) fix: -404KB (12%) link: -328KB (7%) nm: -827KB (19%) objdump: -712KB (16%) pack: -327KB (14%) yacc: -350KB (10%) Binaries that do use reflect.Call.Value see a modest size decrease of 2 - 6% thanks to pruning of unexported methods: api: -151KB (3%) cover: -222KB (4%) doc: -106KB (2.5%) pprof: -314KB (3%) trace: -357KB (4%) vet: -187KB (2.7%) jujud: -4.4MB (5.8%) cmd/go: -384KB (3.4%) The trivial Hello example program goes from 2MB to 1.68MB: package main import "fmt" func main() { fmt.Println("Hello, 世界") } Method pruning also helps when building small binaries with "-ldflags=-s -w". The above program goes from 1.43MB to 1.2MB. Unfortunately the linker can only tell if reflect.Value.Call has been statically linked, not if it is dynamically used. And while use is rare, it is linked into a very common standard library package, text/template. The result is programs like cmd/go, which don't use reflect.Value.Call, see limited benefit from this CL. If binary size is important enough it may be possible to address this in future work. For #6853. Change-Id: Iabe90e210e813b08c3f8fd605f841f0458973396 Reviewed-on: https://go-review.googlesource.com/20483 Reviewed-by: Russ Cox <rsc@golang.org>
2016-03-07 23:45:04 -05:00
case kindInterface: // reflect.interfaceType
off += SysArch.PtrSize + 2*SysArch.IntSize
cmd/link: prune unused methods Today the linker keeps all methods of reachable types. This is necessary if a program uses reflect.Value.Call. But while use of reflection is widespread in Go for encoders and decoders, using it to call a method is rare. This CL looks for the use of reflect.Value.Call in a program, and if it is absent, adopts a (reasonably conservative) method pruning strategy as part of dead code elimination. Any method that is directly called is kept, and any method that matches a used interface's method signature is kept. Whether or not a method body is kept is determined by the relocation from its receiver's *rtype to its *rtype. A small change in the compiler marks these relocations as R_METHOD so they can be easily collected and manipulated by the linker. As a bonus, this technique removes the text segment of methods that have been inlined. Looking at the output of building cmd/objdump with -ldflags=-v=2 shows that inlined methods like runtime.(*traceAllocBlockPtr).ptr are removed from the program. Relatively little work is necessary to do this. Linking two examples, jujud and cmd/objdump show no more than +2% link time. Binaries that do not use reflect.Call.Value drop 4 - 20% in size: addr2line: -793KB (18%) asm: -346KB (8%) cgo: -490KB (10%) compile: -564KB (4%) dist: -736KB (17%) fix: -404KB (12%) link: -328KB (7%) nm: -827KB (19%) objdump: -712KB (16%) pack: -327KB (14%) yacc: -350KB (10%) Binaries that do use reflect.Call.Value see a modest size decrease of 2 - 6% thanks to pruning of unexported methods: api: -151KB (3%) cover: -222KB (4%) doc: -106KB (2.5%) pprof: -314KB (3%) trace: -357KB (4%) vet: -187KB (2.7%) jujud: -4.4MB (5.8%) cmd/go: -384KB (3.4%) The trivial Hello example program goes from 2MB to 1.68MB: package main import "fmt" func main() { fmt.Println("Hello, 世界") } Method pruning also helps when building small binaries with "-ldflags=-s -w". The above program goes from 1.43MB to 1.2MB. Unfortunately the linker can only tell if reflect.Value.Call has been statically linked, not if it is dynamically used. And while use is rare, it is linked into a very common standard library package, text/template. The result is programs like cmd/go, which don't use reflect.Value.Call, see limited benefit from this CL. If binary size is important enough it may be possible to address this in future work. For #6853. Change-Id: Iabe90e210e813b08c3f8fd605f841f0458973396 Reviewed-on: https://go-review.googlesource.com/20483 Reviewed-by: Russ Cox <rsc@golang.org>
2016-03-07 23:45:04 -05:00
default:
// just Sizeof(rtype)
}
mcount := int(decode_inuxi(s.P[off+4:], 2))
moff := int(decode_inuxi(s.P[off+4+2:], 2))
off += moff // offset to array of reflect.method values
const sizeofMethod = 4 * 4 // sizeof reflect.method in program
cmd/compile, etc: store method tables as offsets This CL introduces the typeOff type and a lookup method of the same name that can turn a typeOff offset into an *rtype. In a typical Go binary (built with buildmode=exe, pie, c-archive, or c-shared), there is one moduledata and all typeOff values are offsets relative to firstmoduledata.types. This makes computing the pointer cheap in typical programs. With buildmode=shared (and one day, buildmode=plugin) there are multiple modules whose relative offset is determined at runtime. We identify a type in the general case by the pair of the original *rtype that references it and its typeOff value. We determine the module from the original pointer, and then use the typeOff from there to compute the final *rtype. To ensure there is only one *rtype representing each type, the runtime initializes a typemap for each module, using any identical type from an earlier module when resolving that offset. This means that types computed from an offset match the type mapped by the pointer dynamic relocations. A series of followup CLs will replace other *rtype values with typeOff (and name/*string with nameOff). For types created at runtime by reflect, type offsets are treated as global IDs and reference into a reflect offset map kept by the runtime. darwin/amd64: cmd/go: -57KB (0.6%) jujud: -557KB (0.8%) linux/amd64 PIE: cmd/go: -361KB (3.0%) jujud: -3.5MB (4.2%) For #6853. Change-Id: Icf096fd884a0a0cb9f280f46f7a26c70a9006c96 Reviewed-on: https://go-review.googlesource.com/21285 Reviewed-by: Ian Lance Taylor <iant@golang.org> Run-TryBot: David Crawshaw <crawshaw@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org>
2016-03-28 10:32:27 -04:00
return decode_methodsig(s, off, sizeofMethod, mcount)
cmd/link: prune unused methods Today the linker keeps all methods of reachable types. This is necessary if a program uses reflect.Value.Call. But while use of reflection is widespread in Go for encoders and decoders, using it to call a method is rare. This CL looks for the use of reflect.Value.Call in a program, and if it is absent, adopts a (reasonably conservative) method pruning strategy as part of dead code elimination. Any method that is directly called is kept, and any method that matches a used interface's method signature is kept. Whether or not a method body is kept is determined by the relocation from its receiver's *rtype to its *rtype. A small change in the compiler marks these relocations as R_METHOD so they can be easily collected and manipulated by the linker. As a bonus, this technique removes the text segment of methods that have been inlined. Looking at the output of building cmd/objdump with -ldflags=-v=2 shows that inlined methods like runtime.(*traceAllocBlockPtr).ptr are removed from the program. Relatively little work is necessary to do this. Linking two examples, jujud and cmd/objdump show no more than +2% link time. Binaries that do not use reflect.Call.Value drop 4 - 20% in size: addr2line: -793KB (18%) asm: -346KB (8%) cgo: -490KB (10%) compile: -564KB (4%) dist: -736KB (17%) fix: -404KB (12%) link: -328KB (7%) nm: -827KB (19%) objdump: -712KB (16%) pack: -327KB (14%) yacc: -350KB (10%) Binaries that do use reflect.Call.Value see a modest size decrease of 2 - 6% thanks to pruning of unexported methods: api: -151KB (3%) cover: -222KB (4%) doc: -106KB (2.5%) pprof: -314KB (3%) trace: -357KB (4%) vet: -187KB (2.7%) jujud: -4.4MB (5.8%) cmd/go: -384KB (3.4%) The trivial Hello example program goes from 2MB to 1.68MB: package main import "fmt" func main() { fmt.Println("Hello, 世界") } Method pruning also helps when building small binaries with "-ldflags=-s -w". The above program goes from 1.43MB to 1.2MB. Unfortunately the linker can only tell if reflect.Value.Call has been statically linked, not if it is dynamically used. And while use is rare, it is linked into a very common standard library package, text/template. The result is programs like cmd/go, which don't use reflect.Value.Call, see limited benefit from this CL. If binary size is important enough it may be possible to address this in future work. For #6853. Change-Id: Iabe90e210e813b08c3f8fd605f841f0458973396 Reviewed-on: https://go-review.googlesource.com/20483 Reviewed-by: Russ Cox <rsc@golang.org>
2016-03-07 23:45:04 -05:00
}