From 72ee5bad9f9bd8979e14fab02fb07e39c5e9fd8c Mon Sep 17 00:00:00 2001 From: Ian Lance Taylor Date: Fri, 2 Oct 2020 16:03:37 -0700 Subject: cmd/cgo: split gofrontend mangling checks into cmd/internal/pkgpath This is a step toward porting https://golang.org/cl/219817 from the gofrontend repo to the main repo. Note that this also corrects the implementation of the v2 mangling scheme to use ..u and ..U where appropriate. For #37272 Change-Id: I64a1e7ca1c84348efcbf1cf62049eeb05c830ed8 Reviewed-on: https://go-review.googlesource.com/c/go/+/259298 Trust: Ian Lance Taylor Run-TryBot: Ian Lance Taylor TryBot-Result: Go Bot Reviewed-by: Than McIntosh --- src/cmd/internal/pkgpath/pkgpath.go | 114 +++++++++++++++++++++++++++++ src/cmd/internal/pkgpath/pkgpath_test.go | 121 +++++++++++++++++++++++++++++++ 2 files changed, 235 insertions(+) create mode 100644 src/cmd/internal/pkgpath/pkgpath.go create mode 100644 src/cmd/internal/pkgpath/pkgpath_test.go (limited to 'src/cmd/internal/pkgpath') diff --git a/src/cmd/internal/pkgpath/pkgpath.go b/src/cmd/internal/pkgpath/pkgpath.go new file mode 100644 index 0000000000..0b24468be6 --- /dev/null +++ b/src/cmd/internal/pkgpath/pkgpath.go @@ -0,0 +1,114 @@ +// Copyright 2020 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 pkgpath determines the package path used by gccgo/GoLLVM symbols. +// This package is not used for the gc compiler. +package pkgpath + +import ( + "bytes" + "errors" + "fmt" + "io/ioutil" + "os" + "os/exec" + "strings" +) + +// ToSymbolFunc returns a function that may be used to convert a +// package path into a string suitable for use as a symbol. +// cmd is the gccgo/GoLLVM compiler in use, and tmpdir is a temporary +// directory to pass to ioutil.TempFile. +// For example, this returns a function that converts "net/http" +// into a string like "net..z2fhttp". The actual string varies for +// different gccgo/GoLLVM versions, which is why this returns a function +// that does the conversion appropriate for the compiler in use. +func ToSymbolFunc(cmd, tmpdir string) (func(string) string, error) { + // To determine the scheme used by cmd, we compile a small + // file and examine the assembly code. Older versions of gccgo + // use a simple mangling scheme where there can be collisions + // between packages whose paths are different but mangle to + // the same string. More recent versions use a new mangler + // that avoids these collisions. + const filepat = "*_gccgo_manglechck.go" + f, err := ioutil.TempFile(tmpdir, filepat) + if err != nil { + return nil, err + } + gofilename := f.Name() + f.Close() + defer os.Remove(gofilename) + + if err := ioutil.WriteFile(gofilename, []byte(mangleCheckCode), 0644); err != nil { + return nil, err + } + + command := exec.Command(cmd, "-S", "-o", "-", gofilename) + buf, err := command.Output() + if err != nil { + return nil, err + } + + // New mangling: expect go.l..u00e4ufer.Run + // Old mangling: expect go.l__ufer.Run + if bytes.Contains(buf, []byte("go.l..u00e4ufer.Run")) { + return toSymbolV2, nil + } else if bytes.Contains(buf, []byte("go.l__ufer.Run")) { + return toSymbolV1, nil + } else { + return nil, errors.New(cmd + ": unrecognized mangling scheme") + } +} + +// mangleCheckCode is the package we compile to determine the mangling scheme. +const mangleCheckCode = ` +package läufer +func Run(x int) int { + return 1 +} +` + +// toSymbolV1 converts a package path using the original mangling scheme. +func toSymbolV1(ppath string) string { + clean := func(r rune) rune { + switch { + case 'A' <= r && r <= 'Z', 'a' <= r && r <= 'z', + '0' <= r && r <= '9': + return r + } + return '_' + } + return strings.Map(clean, ppath) +} + +// toSymbolV2 converts a package path using the newer mangling scheme. +func toSymbolV2(ppath string) string { + // This has to build at boostrap time, so it has to build + // with Go 1.4, so we don't use strings.Builder. + bsl := make([]byte, 0, len(ppath)) + changed := false + for _, c := range ppath { + if ('A' <= c && c <= 'Z') || ('a' <= c && c <= 'z') || ('0' <= c && c <= '9') || c == '_' { + bsl = append(bsl, byte(c)) + continue + } + var enc string + switch { + case c == '.': + enc = ".x2e" + case c < 0x80: + enc = fmt.Sprintf("..z%02x", c) + case c < 0x10000: + enc = fmt.Sprintf("..u%04x", c) + default: + enc = fmt.Sprintf("..U%08x", c) + } + bsl = append(bsl, enc...) + changed = true + } + if !changed { + return ppath + } + return string(bsl) +} diff --git a/src/cmd/internal/pkgpath/pkgpath_test.go b/src/cmd/internal/pkgpath/pkgpath_test.go new file mode 100644 index 0000000000..7355f81bae --- /dev/null +++ b/src/cmd/internal/pkgpath/pkgpath_test.go @@ -0,0 +1,121 @@ +// Copyright 2020 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 pkgpath + +import ( + "os" + "testing" +) + +const testEnvName = "GO_PKGPATH_TEST_COMPILER" + +// This init function supports TestToSymbolFunc. For simplicity, +// we use the test binary itself as a sample gccgo driver. +// We set an environment variable to specify how it should behave. +func init() { + switch os.Getenv(testEnvName) { + case "": + return + case "v1": + os.Stdout.WriteString(`.string "go.l__ufer.Run"`) + os.Exit(0) + case "v2": + os.Stdout.WriteString(`.string "go.l..u00e4ufer.Run"`) + os.Exit(0) + case "error": + os.Stdout.WriteString(`unknown string`) + os.Exit(0) + } +} + +func TestToSymbolFunc(t *testing.T) { + const input = "pä世🜃" + tests := []struct { + env string + fail bool + mangled string + }{ + { + env: "v1", + mangled: "p___", + }, + { + env: "v2", + mangled: "p..u00e4..u4e16..U0001f703", + }, + { + env: "error", + fail: true, + }, + } + + cmd := os.Args[0] + tmpdir := t.TempDir() + + defer os.Unsetenv(testEnvName) + + for _, test := range tests { + t.Run(test.env, func(t *testing.T) { + os.Setenv(testEnvName, test.env) + + fn, err := ToSymbolFunc(cmd, tmpdir) + if err != nil { + if !test.fail { + t.Errorf("ToSymbolFunc(%q, %q): unexpected error %v", cmd, tmpdir, err) + } + } else if test.fail { + t.Errorf("ToSymbolFunc(%q, %q) succeeded but expected to fail", cmd, tmpdir) + } else if got, want := fn(input), test.mangled; got != want { + t.Errorf("ToSymbolFunc(%q, %q)(%q) = %q, want %q", cmd, tmpdir, input, got, want) + } + }) + } +} + +var symbolTests = []struct { + input, v1, v2 string +}{ + { + "", + "", + "", + }, + { + "bytes", + "bytes", + "bytes", + }, + { + "net/http", + "net_http", + "net..z2fhttp", + }, + { + "golang.org/x/net/http", + "golang_org_x_net_http", + "golang.x2eorg..z2fx..z2fnet..z2fhttp", + }, + { + "pä世.🜃", + "p____", + "p..u00e4..u4e16.x2e..U0001f703", + }, +} + +func TestV1(t *testing.T) { + for _, test := range symbolTests { + if got, want := toSymbolV1(test.input), test.v1; got != want { + t.Errorf("toSymbolV1(%q) = %q, want %q", test.input, got, want) + } + } +} + +func TestV2(t *testing.T) { + for _, test := range symbolTests { + if got, want := toSymbolV2(test.input), test.v2; got != want { + t.Errorf("toSymbolV2(%q) = %q, want %q", test.input, got, want) + } + } +} -- cgit v1.3 From c47eac7db00e03776c3975025184e1938fbced75 Mon Sep 17 00:00:00 2001 From: Ian Lance Taylor Date: Fri, 20 Nov 2020 12:54:18 -0800 Subject: cmd/cgo, cmd/internal/pkgpath: support gofrontend mangler v3 The gofrontend mangling scheme used by gccgo and GoLLVM has changed again. Support the new version. This is a port of the relevant parts of https://golang.org/cl/271726. For #41862 Change-Id: I9c961c8e17ec960a83a23e1d49ea900962b63393 Reviewed-on: https://go-review.googlesource.com/c/go/+/272127 Trust: Ian Lance Taylor Run-TryBot: Ian Lance Taylor Reviewed-by: Than McIntosh TryBot-Result: Go Bot --- src/cmd/cgo/out.go | 23 ++++++----- src/cmd/internal/pkgpath/pkgpath.go | 68 ++++++++++++++++++++++++++++++-- src/cmd/internal/pkgpath/pkgpath_test.go | 22 ++++++++++- 3 files changed, 98 insertions(+), 15 deletions(-) (limited to 'src/cmd/internal/pkgpath') diff --git a/src/cmd/cgo/out.go b/src/cmd/cgo/out.go index bb963799f6..11c53facf8 100644 --- a/src/cmd/cgo/out.go +++ b/src/cmd/cgo/out.go @@ -186,7 +186,7 @@ func (p *Package) writeDefs() { panic(fmt.Errorf("invalid var kind %q", n.Kind)) } if *gccgo { - fmt.Fprintf(fc, `extern void *%s __asm__("%s.%s");`, n.Mangle, gccgoSymbolPrefix, n.Mangle) + fmt.Fprintf(fc, `extern void *%s __asm__("%s.%s");`, n.Mangle, gccgoSymbolPrefix, gccgoToSymbol(n.Mangle)) fmt.Fprintf(&gccgoInit, "\t%s = &%s;\n", n.Mangle, n.C) fmt.Fprintf(fc, "\n") } @@ -1148,7 +1148,7 @@ func (p *Package) writeGccgoExports(fgo2, fm, fgcc, fgcch io.Writer) { // will not be able to link against it from the C // code. goName := "Cgoexp_" + exp.ExpName - fmt.Fprintf(fgcc, `extern %s %s %s __asm__("%s.%s");`, cRet, goName, cParams, gccgoSymbolPrefix, goName) + fmt.Fprintf(fgcc, `extern %s %s %s __asm__("%s.%s");`, cRet, goName, cParams, gccgoSymbolPrefix, gccgoToSymbol(goName)) fmt.Fprint(fgcc, "\n") fmt.Fprint(fgcc, "\nCGO_NO_SANITIZE_THREAD\n") @@ -1182,7 +1182,7 @@ func (p *Package) writeGccgoExports(fgo2, fm, fgcc, fgcch io.Writer) { fmt.Fprint(fgcc, "}\n") // Dummy declaration for _cgo_main.c - fmt.Fprintf(fm, `char %s[1] __asm__("%s.%s");`, goName, gccgoSymbolPrefix, goName) + fmt.Fprintf(fm, `char %s[1] __asm__("%s.%s");`, goName, gccgoSymbolPrefix, gccgoToSymbol(goName)) fmt.Fprint(fm, "\n") // For gccgo we use a wrapper function in Go, in order @@ -1266,9 +1266,8 @@ func (p *Package) writeExportHeader(fgcch io.Writer) { fmt.Fprintf(fgcch, "%s\n", p.gccExportHeaderProlog()) } -// gccgoPkgpathToSymbol converts a package path to a mangled packagepath -// symbol. -func gccgoPkgpathToSymbol(ppath string) string { +// gccgoToSymbol converts a name to a mangled symbol for gccgo. +func gccgoToSymbol(ppath string) string { if gccgoMangler == nil { var err error cmd := os.Getenv("GCCGO") @@ -1293,12 +1292,12 @@ func (p *Package) gccgoSymbolPrefix() string { } if *gccgopkgpath != "" { - return gccgoPkgpathToSymbol(*gccgopkgpath) + return gccgoToSymbol(*gccgopkgpath) } if *gccgoprefix == "" && p.PackageName == "main" { return "main" } - prefix := gccgoPkgpathToSymbol(*gccgoprefix) + prefix := gccgoToSymbol(*gccgoprefix) if prefix == "" { prefix = "go" } @@ -1687,8 +1686,12 @@ void _cgoPREFIX_Cfunc__Cmalloc(void *v) { ` func (p *Package) cPrologGccgo() string { - return strings.Replace(strings.Replace(cPrologGccgo, "PREFIX", cPrefix, -1), - "GCCGOSYMBOLPREF", p.gccgoSymbolPrefix(), -1) + r := strings.NewReplacer( + "PREFIX", cPrefix, + "GCCGOSYMBOLPREF", p.gccgoSymbolPrefix(), + "_cgoCheckPointer", gccgoToSymbol("_cgoCheckPointer"), + "_cgoCheckResult", gccgoToSymbol("_cgoCheckResult")) + return r.Replace(cPrologGccgo) } const cPrologGccgo = ` diff --git a/src/cmd/internal/pkgpath/pkgpath.go b/src/cmd/internal/pkgpath/pkgpath.go index 0b24468be6..40a040a81a 100644 --- a/src/cmd/internal/pkgpath/pkgpath.go +++ b/src/cmd/internal/pkgpath/pkgpath.go @@ -50,9 +50,12 @@ func ToSymbolFunc(cmd, tmpdir string) (func(string) string, error) { return nil, err } - // New mangling: expect go.l..u00e4ufer.Run - // Old mangling: expect go.l__ufer.Run - if bytes.Contains(buf, []byte("go.l..u00e4ufer.Run")) { + // Original mangling: go.l__ufer.Run + // Mangling v2: go.l..u00e4ufer.Run + // Mangling v3: go_0l_u00e4ufer.Run + if bytes.Contains(buf, []byte("go_0l_u00e4ufer.Run")) { + return toSymbolV3, nil + } else if bytes.Contains(buf, []byte("go.l..u00e4ufer.Run")) { return toSymbolV2, nil } else if bytes.Contains(buf, []byte("go.l__ufer.Run")) { return toSymbolV1, nil @@ -82,7 +85,7 @@ func toSymbolV1(ppath string) string { return strings.Map(clean, ppath) } -// toSymbolV2 converts a package path using the newer mangling scheme. +// toSymbolV2 converts a package path using the second mangling scheme. func toSymbolV2(ppath string) string { // This has to build at boostrap time, so it has to build // with Go 1.4, so we don't use strings.Builder. @@ -112,3 +115,60 @@ func toSymbolV2(ppath string) string { } return string(bsl) } + +// v3UnderscoreCodes maps from a character that supports an underscore +// encoding to the underscore encoding character. +var v3UnderscoreCodes = map[byte]byte{ + '_': '_', + '.': '0', + '/': '1', + '*': '2', + ',': '3', + '{': '4', + '}': '5', + '[': '6', + ']': '7', + '(': '8', + ')': '9', + '"': 'a', + ' ': 'b', + ';': 'c', +} + +// toSymbolV3 converts a package path using the third mangling scheme. +func toSymbolV3(ppath string) string { + // This has to build at boostrap time, so it has to build + // with Go 1.4, so we don't use strings.Builder. + bsl := make([]byte, 0, len(ppath)) + changed := false + for _, c := range ppath { + if ('A' <= c && c <= 'Z') || ('a' <= c && c <= 'z') || ('0' <= c && c <= '9') { + bsl = append(bsl, byte(c)) + continue + } + + if c < 0x80 { + if u, ok := v3UnderscoreCodes[byte(c)]; ok { + bsl = append(bsl, '_', u) + changed = true + continue + } + } + + var enc string + switch { + case c < 0x80: + enc = fmt.Sprintf("_x%02x", c) + case c < 0x10000: + enc = fmt.Sprintf("_u%04x", c) + default: + enc = fmt.Sprintf("_U%08x", c) + } + bsl = append(bsl, enc...) + changed = true + } + if !changed { + return ppath + } + return string(bsl) +} diff --git a/src/cmd/internal/pkgpath/pkgpath_test.go b/src/cmd/internal/pkgpath/pkgpath_test.go index 7355f81bae..232e803a60 100644 --- a/src/cmd/internal/pkgpath/pkgpath_test.go +++ b/src/cmd/internal/pkgpath/pkgpath_test.go @@ -24,6 +24,9 @@ func init() { case "v2": os.Stdout.WriteString(`.string "go.l..u00e4ufer.Run"`) os.Exit(0) + case "v3": + os.Stdout.WriteString(`.string "go_0l_u00e4ufer.Run"`) + os.Exit(0) case "error": os.Stdout.WriteString(`unknown string`) os.Exit(0) @@ -45,6 +48,10 @@ func TestToSymbolFunc(t *testing.T) { env: "v2", mangled: "p..u00e4..u4e16..U0001f703", }, + { + env: "v3", + mangled: "p_u00e4_u4e16_U0001f703", + }, { env: "error", fail: true, @@ -75,32 +82,37 @@ func TestToSymbolFunc(t *testing.T) { } var symbolTests = []struct { - input, v1, v2 string + input, v1, v2, v3 string }{ { "", "", "", + "", }, { "bytes", "bytes", "bytes", + "bytes", }, { "net/http", "net_http", "net..z2fhttp", + "net_1http", }, { "golang.org/x/net/http", "golang_org_x_net_http", "golang.x2eorg..z2fx..z2fnet..z2fhttp", + "golang_0org_1x_1net_1http", }, { "pä世.🜃", "p____", "p..u00e4..u4e16.x2e..U0001f703", + "p_u00e4_u4e16_0_U0001f703", }, } @@ -119,3 +131,11 @@ func TestV2(t *testing.T) { } } } + +func TestV3(t *testing.T) { + for _, test := range symbolTests { + if got, want := toSymbolV3(test.input), test.v3; got != want { + t.Errorf("toSymbolV3(%q) = %q, want %q", test.input, got, want) + } + } +} -- cgit v1.3