go1.3.2

A security bug affects programs that use crypto/tls to implement a TLS server
from Go 1.1 onwards. If the server enables TLS client authentication using
certificates (this is rare) and explicitly sets SessionTicketsDisabled to true
in the tls.Config, then a malicious client can falsely assert ownership of any
client certificate it wishes.

This issue was discovered internally and there is no evidence of exploitation.

Change authored by Adam Langley <agl@golang.org>
https://golang.org/cl/148080043/

.hgtags

@@ -126,6 +126,5 @@ b3d5a20b070a92da2458c5788694d1359b353f4a go1.2rc5
 87dea3f5ebe7510998c84dbeeec89382b7d42f9c go1.2
 0ddbdc3c7ce27e66508fe58ab81ff29324786026 go1.2.1
 9c4fdd8369ca4483fbed1cb8e67f02643ca10f79 go1.2.2
-9c4fdd8369ca4483fbed1cb8e67f02643ca10f79 release
 f8b50ad4cac4d4c4ecf48324b4f512f65e82cc1c go1.3beta1
 9e1652c32289c164126b6171f024afad5665fc9e go1.3beta2

VERSION

@@ -0,0 +1 @@
+go1.3.2

doc/cmd.html

@@ -62,7 +62,7 @@ details.
 </tr>
 
 <tr>
-<td><a href="http://godoc.org/code.google.com/p/go.tools/cmd/cover/">cover</a></td>
+<td><a href="//godoc.org/code.google.com/p/go.tools/cmd/cover/">cover</a></td>
 <td>&nbsp;&nbsp;&nbsp;&nbsp;</td>
 <td>Cover is a program for creating and analyzing the coverage profiles
 generated by <code>"go test -coverprofile"</code>.
@@ -83,13 +83,13 @@ gofmt</a> command with more general options.</td>
 </tr>
 
 <tr>
-<td><a href="http://godoc.org/code.google.com/p/go.tools/cmd/godoc/">godoc</a></td>
+<td><a href="//godoc.org/code.google.com/p/go.tools/cmd/godoc/">godoc</a></td>
 <td>&nbsp;&nbsp;&nbsp;&nbsp;</td>
 <td>Godoc extracts and generates documentation for Go packages.</td>
 </tr>
 
 <tr>
-<td><a href="http://godoc.org/code.google.com/p/go.tools/cmd/vet/">vet</a></td>
+<td><a href="//godoc.org/code.google.com/p/go.tools/cmd/vet/">vet</a></td>
 <td>&nbsp;&nbsp;&nbsp;&nbsp;</td>
 <td>Vet examines Go source code and reports suspicious constructs, such as Printf
 calls whose arguments do not align with the format string.</td>

doc/code.html

@@ -18,7 +18,7 @@ It explains the simplest way to get up and running with your Go installation.
 
 <p>
 A similar explanation is available as a
-<a href="http://www.youtube.com/watch?v=XCsL89YtqCs">screencast</a>.
+<a href="//www.youtube.com/watch?v=XCsL89YtqCs">screencast</a>.
 </p>
 
 
@@ -508,7 +508,7 @@ revision control system such as Git or Mercurial. The <code>go</code> tool uses
 this property to automatically fetch packages from remote repositories.
 For instance, the examples described in this document are also kept in a
 Mercurial repository hosted at Google Code,
-<code><a href="http://code.google.com/p/go.example">code.google.com/p/go.example</a></code>.
+<code><a href="//code.google.com/p/go.example">code.google.com/p/go.example</a></code>.
 If you include the repository URL in the package's import path,
 <code>go get</code> will fetch, build, and install it automatically:
 </p>
@@ -569,8 +569,8 @@ import "code.google.com/p/go.example/newmath"
 <p>
 This convention is the easiest way to make your Go packages available for
 others to use.
-The <a href="http://code.google.com/p/go-wiki/wiki/Projects">Go Wiki</a>
-and <a href="http://godoc.org/">godoc.org</a>
+The <a href="//code.google.com/p/go-wiki/wiki/Projects">Go Wiki</a>
+and <a href="//godoc.org/">godoc.org</a>
 provide lists of external Go projects.
 </p>
 
@@ -584,7 +584,7 @@ For more information on using remote repositories with the <code>go</code> tool,
 
 <p>
 Subscribe to the
-<a href="http://groups.google.com/group/golang-announce">golang-announce</a>
+<a href="//groups.google.com/group/golang-announce">golang-announce</a>
 mailing list to be notified when a new stable version of Go is released.
 </p>
 
@@ -594,7 +594,7 @@ clear, idiomatic Go code.
 </p>
 
 <p>
-Take <a href="http://tour.golang.org/">A Tour of Go</a> to learn the language
+Take <a href="//tour.golang.org/">A Tour of Go</a> to learn the language
 proper.
 </p>
 
@@ -613,10 +613,10 @@ For real-time help, ask the helpful gophers in <code>#go-nuts</code> on the
 
 <p>
 The official mailing list for discussion of the Go language is
-<a href="http://groups.google.com/group/golang-nuts">Go Nuts</a>.
+<a href="//groups.google.com/group/golang-nuts">Go Nuts</a>.
 </p>
 
 <p>
 Report bugs using the
-<a href="http://code.google.com/p/go/issues/list">Go issue tracker</a>.
+<a href="//code.google.com/p/go/issues/list">Go issue tracker</a>.
 </p>

doc/contrib.html

@@ -9,7 +9,7 @@
 
 <p>
 Go is an open source project developed by a team at
-<a href="http://google.com/">Google</a> and many
+<a href="//google.com/">Google</a> and many
 <a href="/CONTRIBUTORS">contributors</a> from the open source community.
 </p>
 
@@ -17,13 +17,13 @@ Go is an open source project developed by a team at
 Go is distributed under a <a href="/LICENSE">BSD-style license</a>.
 </p>
 
-<h3 id="announce"><a href="http://groups.google.com/group/golang-announce">Announcements Mailing List</a></h3>
+<h3 id="announce"><a href="//groups.google.com/group/golang-announce">Announcements Mailing List</a></h3>
 <p>
 A low traffic mailing list for important announcements, such as new releases.
 </p>
 <p>
 We encourage all Go users to subscribe to
-<a href="http://groups.google.com/group/golang-announce">golang-announce</a>.
+<a href="//groups.google.com/group/golang-announce">golang-announce</a>.
 </p>
 
 
@@ -72,9 +72,9 @@ href="https://groups.google.com/group/golang-nuts">golang-nuts</a>.</p>
 <p>A mailing list that receives a message summarizing each checkin to the Go repository.</p>
 
 <h3 id="golang-bugs"><a href="https://groups.google.com/group/golang-bugs">Bugs Mailing List</a></h3>
-<p>A mailing list that receives each update to the Go <a href="http://golang.org/issue">issue tracker</a>.</p>
+<p>A mailing list that receives each update to the Go <a href="//golang.org/issue">issue tracker</a>.</p>
 
-<h3 id="build_status"><a href="http://build.golang.org/">Build Status</a></h3>
+<h3 id="build_status"><a href="//build.golang.org/">Build Status</a></h3>
 <p>View the status of Go builds across the supported operating
 systems and architectures.</p>
 

doc/devel/release.html

@@ -3,7 +3,7 @@
 }-->
 
 <p>This page summarizes the changes between official stable releases of Go.
-The <a href="http://code.google.com/p/go/source/list">Mercurial change log</a>
+The <a href="//code.google.com/p/go/source/list">Mercurial change log</a>
 has the full details.</p>
 
 <p>To update to a specific release, use:</p>
@@ -13,6 +13,25 @@ hg pull
 hg update <i>tag</i>
 </pre>
 
+<h2 id="go1.3">go1.3 (released 2014/06/18)</h2>
+
+<p>
+Go 1.3 is a major release of Go.
+Read the <a href="/doc/go1.3">Go 1.3 Release Notes</a> for more information.
+</p>
+
+<h3 id="go1.3.minor">Minor revisions</h3>
+
+<p>
+go1.3.1 (released 2014/08/13) includes bug fixes to the compiler and the <code>runtime</code>, <code>net</code>, and <code>crypto/rsa</code> packages.
+See the <a href="//code.google.com/p/go/source/list?name=release-branch.go1.3&r=073fc578434bf3e1e22749b559d273c8da728ebb">change history</a> for details.
+</p>
+
+<p>
+go1.3.2 (released 2014/09/25) includes bug fixes to cgo and the crypto/tls packages.
+See the <a href="//code.google.com/p/go/source/list?name=release-branch.go1.3&r=go1.3.2">change history</a> for details.
+</p>
+
 <h2 id="go1.2">go1.2 (released 2013/12/01)</h2>
 
 <p>
@@ -24,12 +43,12 @@ Read the <a href="/doc/go1.2">Go 1.2 Release Notes</a> for more information.
 
 <p>
 go1.2.1 (released 2014/03/02) includes bug fixes to the <code>runtime</code>, <code>net</code>, and <code>database/sql</code> packages.
-See the <a href="https://code.google.com/p/go/source/list?name=release-branch.go1.2&r=7ada9e760ce34e78aee5b476c9621556d0fa5d31">change history</a> for details.
+See the <a href="//code.google.com/p/go/source/list?name=release-branch.go1.2&r=7ada9e760ce34e78aee5b476c9621556d0fa5d31">change history</a> for details.
 </p>
 
 <p>
 go1.2.2 (released 2014/05/05) includes a
-<a href="https://code.google.com/p/go/source/detail?r=bda3619e7a2c&repo=tools">security fix</a>
+<a href="//code.google.com/p/go/source/detail?r=bda3619e7a2c&repo=tools">security fix</a>
 that affects the tour binary included in the binary distributions (thanks to Guillaume T).
 </p>
 
@@ -44,18 +63,18 @@ Read the <a href="/doc/go1.1">Go 1.1 Release Notes</a> for more information.
 
 <p>
 go1.1.1 (released 2013/06/13) includes several compiler and runtime bug fixes.
-See the <a href="https://code.google.com/p/go/source/list?name=release-branch.go1.1&r=43c4a41d24382a56a90e924800c681e435d9e399">change history</a> for details.
+See the <a href="//code.google.com/p/go/source/list?name=release-branch.go1.1&r=43c4a41d24382a56a90e924800c681e435d9e399">change history</a> for details.
 </p>
 
 <p>
 go1.1.2 (released 2013/08/13) includes fixes to the <code>gc</code> compiler
 and <code>cgo</code>, and the <code>bufio</code>, <code>runtime</code>,
 <code>syscall</code>, and <code>time</code> packages.
-See the <a href="https://code.google.com/p/go/source/list?name=release-branch.go1.1&r=a6a9792f94acd4ff686b2bc57383d163608b91cf">change history</a> for details.
+See the <a href="//code.google.com/p/go/source/list?name=release-branch.go1.1&r=a6a9792f94acd4ff686b2bc57383d163608b91cf">change history</a> for details.
 If you use package syscall's <code>Getrlimit</code> and <code>Setrlimit</code>
 functions under Linux on the ARM or 386 architectures, please note change
-<a href="http://golang.org/change/55ac276af5a7">55ac276af5a7</a>
-that fixes <a href="http://golang.org/issue/5949">issue 5949</a>.
+<a href="//golang.org/change/55ac276af5a7">55ac276af5a7</a>
+that fixes <a href="//golang.org/issue/5949">issue 5949</a>.
 </p>
 
 <h2 id="go1">go1 (released 2012/03/28)</h2>
@@ -81,17 +100,17 @@ The go1 release corresponds to
 
 <p>
 go1.0.1 (released 2012/04/25) was issued to
-<a href="https://code.google.com/p/go/source/detail?r=a890477d3dfb">fix</a> an
-<a href="https://code.google.com/p/go/issues/detail?id=3545">escape analysis
-bug</a> that can lead to memory corruption.
+<a href="//golang.org/change/a890477d3dfb">fix</a> an
+<a href="//golang.org/issue/3545">escape analysis bug</a>
+that can lead to memory corruption.
 It also includes several minor code and documentation fixes.
 </p>
 
 <p>
 go1.0.2 (released 2012/06/13) was issued to fix two bugs in the implementation
 of maps using struct or array keys: 
-<a href="http://code.google.com/p/go/issues/detail?id=3695">issue 3695</a> and
-<a href="http://code.google.com/p/go/issues/detail?id=3573">issue 3573</a>.
+<a href="//golang.org/issue/3695">issue 3695</a> and
+<a href="//golang.org/issue/3573">issue 3573</a>.
 It also includes many minor code and documentation fixes.
 </p>
 
@@ -100,7 +119,7 @@ go1.0.3 (released 2012/09/21) includes minor code and documentation fixes.
 </p>
 
 <p>
-See the <a href="http://code.google.com/p/go/source/list?name=release-branch.go1">go1 release branch history</a> for the complete list of changes.
+See the <a href="//code.google.com/p/go/source/list?name=release-branch.go1">go1 release branch history</a> for the complete list of changes.
 </p>
 
 <h2 id="r60">r60 (released 2011/09/07)</h2>
@@ -112,7 +131,7 @@ This section highlights the most significant changes in this release.
 For a more detailed summary, see the
 <a href="weekly.html#2011-08-17">weekly release notes</a>.
 For complete information, see the
-<a href="http://code.google.com/p/go/source/list?r=release-branch.r60">Mercurial change list</a>.
+<a href="//code.google.com/p/go/source/list?r=release-branch.r60">Mercurial change list</a>.
 </p>
 
 <h3 id="r60.lang">Language</h3>
@@ -182,26 +201,26 @@ more accessible.
 
 <p>
 r60.1 includes a 
-<a href="http://code.google.com/p/go/source/detail?r=1824581bf62d">linker
+<a href="//golang.org/change/1824581bf62d">linker
 fix</a>, a pair of
-<a href="http://code.google.com/p/go/source/detail?r=9ef4429c2c64">goplay</a>
-<a href="http://code.google.com/p/go/source/detail?r=d42ed8c3098e">fixes</a>,
+<a href="//golang.org/change/9ef4429c2c64">goplay</a>
+<a href="//golang.org/change/d42ed8c3098e">fixes</a>,
 and a <code>json</code> package
-<a href="http://code.google.com/p/go/source/detail?r=d5e97874fe84">fix</a> and
+<a href="//golang.org/change/d5e97874fe84">fix</a> and
 a new
-<a href="http://code.google.com/p/go/source/detail?r=4f0e6269213f">struct tag
+<a href="//golang.org/change/4f0e6269213f">struct tag
 option</a>.
 </p>
 
 <p>
 r60.2
-<a href="http://code.google.com/p/go/source/detail?r=ff19536042ac">fixes</a>
+<a href="//golang.org/change/ff19536042ac">fixes</a>
 a memory leak involving maps.
 </p>
 
 <p>
 r60.3 fixes a
-<a href="http://code.google.com/p/go/source/detail?r=01fa62f5e4e5">reflect bug</a>.
+<a href="//golang.org/change/01fa62f5e4e5">reflect bug</a>.
 </p>
 
 <h2 id="r59">r59 (released 2011/08/01)</h2>
@@ -213,7 +232,7 @@ This section highlights the most significant changes in this release.
 For a more detailed summary, see the
 <a href="weekly.html#2011-07-07">weekly release notes</a>.
 For complete information, see the
-<a href="http://code.google.com/p/go/source/list?r=release-branch.r59">Mercurial change list</a>.
+<a href="//code.google.com/p/go/source/list?r=release-branch.r59">Mercurial change list</a>.
 </p>
 
 <h3 id="r59.lang">Language</h3>
@@ -223,7 +242,7 @@ This release includes a language change that restricts the use of
 <code>goto</code>.  In essence, a <code>goto</code> statement outside a block
 cannot jump to a label inside that block. Your code may require changes if it
 uses <code>goto</code>.
-See <a href="http://code.google.com/p/go/source/detail?r=dc6d3cf9279d">this
+See <a href="//golang.org/change/dc6d3cf9279d">this
 changeset</a> for how the new rule affected the Go tree.
 </p>
 
@@ -321,13 +340,13 @@ This section highlights the most significant changes in this release.
 For a more detailed summary, see the
 <a href="weekly.html#2011-06-09">weekly release notes</a>.
 For complete information, see the
-<a href="http://code.google.com/p/go/source/list?r=release-branch.r58">Mercurial change list</a>.
+<a href="//code.google.com/p/go/source/list?r=release-branch.r58">Mercurial change list</a>.
 </p>
 
 <h3 id="r58.lang">Language</h3>
 
 <p>
-This release fixes a <a href="http://code.google.com/p/go/source/detail?r=b720749486e1">use of uninitialized memory in programs that misuse <code>goto</code></a>.
+This release fixes a <a href="//golang.org/change/b720749486e1">use of uninitialized memory in programs that misuse <code>goto</code></a>.
 </p>
 
 <h3 id="r58.pkg">Packages</h3>
@@ -386,8 +405,8 @@ the Go tree (and avoid writing Makefiles).
 <h3 id="r58.minor">Minor revisions</h3>
 
 <p>r58.1 adds 
-<a href="http://code.google.com/p/go/source/detail?r=293c25943586">build</a> and
-<a href="http://code.google.com/p/go/source/detail?r=bf17e96b6582">runtime</a>
+<a href="//golang.org/change/293c25943586">build</a> and
+<a href="//golang.org/change/bf17e96b6582">runtime</a>
 changes to make Go run on OS X 10.7 Lion.
 </p>
 
@@ -401,7 +420,7 @@ This section highlights the most significant changes in this release.
 For a more detailed summary, see the
 <a href="weekly.html#2011-04-27">weekly release notes</a>.
 For complete information, see the
-<a href="http://code.google.com/p/go/source/list?r=release-branch.r57">Mercurial change list</a>.
+<a href="//code.google.com/p/go/source/list?r=release-branch.r57">Mercurial change list</a>.
 </p>
 
 <p>The new <a href="/cmd/gofix">gofix</a> tool finds Go programs that use old APIs and rewrites them to use
@@ -412,7 +431,7 @@ future changes to the libraries.
 Gofix can’t
 handle all situations perfectly, so read and test the changes it makes before
 committing them.
-See <a href="http://blog.golang.org/2011/04/introducing-gofix.html">the gofix blog post</a> for more
+See <a href="//blog.golang.org/2011/04/introducing-gofix.html">the gofix blog post</a> for more
 information.</p>
 
 <h3 id="r57.lang">Language</h3>
@@ -458,7 +477,7 @@ For clients, there are new
 <a href="/pkg/http/#Client">Client</a> and <a href="/pkg/http/#Transport">Transport</a>
 abstractions that give more control over HTTP details such as headers sent
 and redirections followed.  These abstractions make it easy to implement
-custom clients that add functionality such as <a href="http://code.google.com/p/goauth2/source/browse/oauth/oauth.go">OAuth2</a>.
+custom clients that add functionality such as <a href="//code.google.com/p/goauth2/source/browse/oauth/oauth.go">OAuth2</a>.
 For servers, <a href="/pkg/http/#ResponseWriter">ResponseWriter</a>
 has dropped its non-essential methods.
 The Hijack and Flush methods are no longer required;
@@ -502,7 +521,7 @@ implements all the possible value methods.
 Instead of a type switch on a Value <code>v</code>, switch on <code>v.Kind()</code>.
 Typeof and NewValue are now called <a href="/pkg/reflect/#Type.TypeOf">TypeOf</a> and <a href="/pkg/reflect/#Value.ValueOf">ValueOf</a>
 To create a writable Value, use <code>New(t).Elem()</code> instead of <code>Zero(t)</code>.
-See <a href="http://code.google.com/p/go/source/detail?r=843855f3c026">the change description</a>
+See <a href="//golang.org/change/843855f3c026">the change description</a>
 for the full details.
 The new API allows a more efficient implementation of Value
 that avoids many of the allocations required by the previous API.
@@ -538,8 +557,8 @@ For other uses, see the <a href="/pkg/runtime/pprof/">runtime/pprof</a> document
 
 <h3 id="r57.minor">Minor revisions</h3>
 
-<p>r57.1 fixes a <a href="http://code.google.com/p/go/source/detail?r=ff2bc62726e7145eb2ecc1e0f076998e4a8f86f0">nil pointer dereference in http.FormFile</a>.</p>
-<p>r57.2 fixes a <a href="http://code.google.com/p/go/source/detail?r=063b0ff67d8277df03c956208abc068076818dae">use of uninitialized memory in programs that misuse <code>goto</code></a>.</p>
+<p>r57.1 fixes a <a href="//golang.org/change/ff2bc62726e7145eb2ecc1e0f076998e4a8f86f0">nil pointer dereference in http.FormFile</a>.</p>
+<p>r57.2 fixes a <a href="//golang.org/change/063b0ff67d8277df03c956208abc068076818dae">use of uninitialized memory in programs that misuse <code>goto</code></a>.</p>
 
 <h2 id="r56">r56 (released 2011/03/16)</h2>
 

doc/devel/weekly.html

@@ -5,7 +5,7 @@
 <p>This page summarizes the changes between tagged weekly snapshots of Go.
 Such snapshots are no longer created. This page remains as a historical reference only.</p>
 
-<p>For recent information, see the <a href="http://code.google.com/p/go/source/list">Mercurial change log</a> and <a href="http://groups.google.com/group/golang-dev/">development mailing list</a>.</p>
+<p>For recent information, see the <a href="//code.google.com/p/go/source/list">Mercurial change log</a> and <a href="//groups.google.com/group/golang-dev/">development mailing list</a>.</p>
 
 <h2 id="2012-03-27">2012-03-27 (<a href="release.html#go1">Go 1</a>)</h2>
 

doc/docs.html

@@ -33,20 +33,20 @@ libraries.
 
 <img class="gopher" src="/doc/gopher/doc.png"/>
 
-<h3 id="go_tour"><a href="http://tour.golang.org/">A Tour of Go</a></h3>
+<h3 id="go_tour"><a href="//tour.golang.org/">A Tour of Go</a></h3>
 <p>
 An interactive introduction to Go in three sections.
 The first section covers basic syntax and data structures; the second discusses
 methods and interfaces; and the third introduces Go's concurrency primitives.
 Each section concludes with a few exercises so you can practice what you've
-learned. You can <a href="http://tour.golang.org/">take the tour online</a> or
-<a href="http://code.google.com/p/go-tour/">install it locally</a>.
+learned. You can <a href="//tour.golang.org/">take the tour online</a> or
+<a href="//code.google.com/p/go-tour/">install it locally</a>.
 </p>
 
 <h3 id="code"><a href="code.html">How to write Go code</a></h3>
 <p>
 Also available as a
-<a href="http://www.youtube.com/watch?v=XCsL89YtqCs">screencast</a>, this doc
+<a href="//www.youtube.com/watch?v=XCsL89YtqCs">screencast</a>, this doc
 explains how to use the <a href="/cmd/go/">go command</a> to fetch, build, and
 install packages, commands, and run tests.
 </p>
@@ -97,10 +97,13 @@ one goroutine can be guaranteed to observe values produced by writes to the
 same variable in a different goroutine.
 </p>
 
+<h3 id="release"><a href="/doc/devel/release.html">Release History</a></h3>
+<p>A summary of the changes between Go releases.</p>
+
 
 <h2 id="articles">Articles</h2>
 
-<h3 id="blog"><a href="http://blog.golang.org/">The Go Blog</a></h3>
+<h3 id="blog"><a href="//blog.golang.org/">The Go Blog</a></h3>
 <p>The official blog of the Go project, featuring news and in-depth articles by
 the Go team and guests.</p>
 
@@ -145,6 +148,7 @@ Guided tours of Go programs.
 <li><a href="/blog/profiling-go-programs">Profiling Go Programs</a></li>
 <li><a href="/doc/articles/race_detector.html">Data Race Detector</a> - a manual for the data race detector.</li>
 <li><a href="/blog/race-detector">Introducing the Go Race Detector</a> - an introduction to the race detector.
+<li><a href="/doc/asm">A Quick Guide to Go's Assembler</a> - an introduction to the assembler used by Go.
 </ul>
 
 <h4 id="articles_more">More</h4>
@@ -165,17 +169,17 @@ interfaces, reflection, and concurrency. Builds a toy web crawler to
 demonstrate these.
 </p>
 
-<h3 id="go_code_that_grows"><a href="http://vimeo.com/53221560">Code that grows with grace</a></h3>
+<h3 id="go_code_that_grows"><a href="//vimeo.com/53221560">Code that grows with grace</a></h3>
 <p>
 One of Go's key design goals is code adaptability; that it should be easy to take a simple design and build upon it in a clean and natural way. In this talk Andrew Gerrand describes a simple "chat roulette" server that matches pairs of incoming TCP connections, and then use Go's concurrency mechanisms, interfaces, and standard library to extend it with a web interface and other features. While the function of the program changes dramatically, Go's flexibility preserves the original design as it grows.
 </p>
 
-<h3 id="go_concurrency_patterns"><a href="http://www.youtube.com/watch?v=f6kdp27TYZs">Go Concurrency Patterns</a></h3>
+<h3 id="go_concurrency_patterns"><a href="//www.youtube.com/watch?v=f6kdp27TYZs">Go Concurrency Patterns</a></h3>
 <p>
 Concurrency is the key to designing high performance network services. Go's concurrency primitives (goroutines and channels) provide a simple and efficient means of expressing concurrent execution. In this talk we see how tricky concurrency problems can be solved gracefully with simple Go code.
 </p>
 
-<h3 id="advanced_go_concurrency_patterns"><a href="http://www.youtube.com/watch?v=QDDwwePbDtw">Advanced Go Concurrency Patterns</a></h3>
+<h3 id="advanced_go_concurrency_patterns"><a href="//www.youtube.com/watch?v=QDDwwePbDtw">Advanced Go Concurrency Patterns</a></h3>
 <p>
 This talk expands on the <i>Go Concurrency Patterns</i> talk to dive deeper into Go's concurrency primitives.
 </p>

doc/effective_go.html

@@ -28,7 +28,7 @@ will be easy for other Go programmers to understand.
 <p>
 This document gives tips for writing clear, idiomatic Go code.
 It augments the <a href="/ref/spec">language specification</a>,
-the <a href="http://tour.golang.org/">Tour of Go</a>,
+the <a href="//tour.golang.org/">Tour of Go</a>,
 and <a href="/doc/code.html">How to Write Go Code</a>,
 all of which you
 should read first.
@@ -43,8 +43,8 @@ only as the core library but also as examples of how to
 use the language.
 Moreover, many of the packages contain working, self-contained
 executable examples you can run directly from the
-<a href="http://golang.org">golang.org</a> web site, such as
-<a href="http://golang.org/pkg/strings/#example_Map">this one</a> (if
+<a href="//golang.org">golang.org</a> web site, such as
+<a href="//golang.org/pkg/strings/#example_Map">this one</a> (if
 necessary, click on the word "Example" to open it up).
 If you have a question about how to approach a problem or how something
 might be implemented, the documentation, code and examples in the
@@ -3214,7 +3214,7 @@ Although the concurrency features of Go can make some problems easy
 to structure as parallel computations, Go is a concurrent language,
 not a parallel one, and not all parallelization problems fit Go's model.
 For a discussion of the distinction, see the talk cited in
-<a href="http://blog.golang.org/2013/01/concurrency-is-not-parallelism.html">this
+<a href="//blog.golang.org/2013/01/concurrency-is-not-parallelism.html">this
 blog post</a>.
 
 <h3 id="leaky_buffer">A leaky buffer</h3>

doc/gccgo_contribute.html

@@ -30,7 +30,7 @@ contribution rules</a>.
 
 <p>
 The master sources for the gccgo frontend may be found at
-<a href="http://code.google.com/p/gofrontend">http://code.google.com/p/gofrontend</a>.
+<a href="//code.google.com/p/gofrontend">http://code.google.com/p/gofrontend</a>.
 The master sources are not buildable by themselves, but only in
 conjunction with GCC (in the future, other compilers may be
 supported).  Changes made to the gccgo frontend are also applied to
@@ -40,7 +40,7 @@ is mirrored to the <code>gcc/go/gofrontend</code> directory in the GCC
 repository, and the <code>gofrontend</code> <code>libgo</code>
 directory is mirrored to the GCC <code>libgo</code> directory.  In
 addition, the <code>test</code> directory
-from <a href="http://code.google.com/p/go">the main Go repository</a>
+from <a href="//code.google.com/p/go">the main Go repository</a>
 is mirrored to the <code>gcc/testsuite/go.test/test</code> directory
 in the GCC repository.
 </p>
@@ -65,7 +65,7 @@ from <code>gcc/go/gofrontend</code> to <code>gcc/go</code>.
 
 <p>
 The run-time library for gccgo is mostly the same as the library
-in <a href="http://code.google.com/p/go">the main Go repository</a>.
+in <a href="//code.google.com/p/go">the main Go repository</a>.
 The library code in the Go repository is periodically merged into
 the <code>libgo/go</code> directory of the <code>gofrontend</code> and
 then the GCC repositories, using the shell

doc/go1.1.html

@@ -304,7 +304,7 @@ The race detector is documented in <a href="/doc/articles/race_detector.html">a
 
 <p>
 Due to the change of the <a href="#int"><code>int</code></a> to 64 bits and
-a new internal <a href="http://golang.org/s/go11func">representation of functions</a>,
+a new internal <a href="//golang.org/s/go11func">representation of functions</a>,
 the arrangement of function arguments on the stack has changed in the gc tool chain.
 Functions written in assembly will need to be revised at least
 to adjust frame pointer offsets.

doc/go1.2.html

@@ -63,7 +63,7 @@ The implementation may inject extra tests into the compiled program to enforce t
 
 <p>
 Further details are in the
-<a href="http://golang.org/s/go12nil">design document</a>.
+<a href="//golang.org/s/go12nil">design document</a>.
 </p>
 
 <p>
@@ -115,7 +115,7 @@ It is possible that future releases of Go may introduce default values for these
 
 <p>
 Further details are in the
-<a href="http://golang.org/s/go12slice">design document</a>.
+<a href="//golang.org/s/go12slice">design document</a>.
 </p>
 
 <p>
@@ -199,7 +199,7 @@ compiler to build any pieces of the linked-to library that are written in C++;
 <p>
 Both binaries are still included with the distribution, but the source code for the
 godoc and vet commands has moved to the
-<a href="http://code.google.com/p/go.tools">go.tools</a> subrepository.
+<a href="//code.google.com/p/go.tools">go.tools</a> subrepository.
 </p>
 
 <p>
@@ -218,7 +218,7 @@ no client Go code depends on the their source and no updating is required.
 </p>
 
 <p>
-The binary distributions available from <a href="http://golang.org">golang.org</a>
+The binary distributions available from <a href="//golang.org">golang.org</a>
 include these binaries, so users of these distributions are unaffected.
 </p>
 
@@ -448,7 +448,7 @@ The new interfaces are called
 and
 <a href="/pkg/encoding/#TextUnmarshaler"><code>TextUnmarshaler</code></a>.
 Full details are in the <a href="/pkg/encoding/">documentation</a> for the package
-and a separate <a href="http://golang.org/s/go12encoding">design document</a>.
+and a separate <a href="//golang.org/s/go12encoding">design document</a>.
 </p>
 
 <h3 id="fmt_indexed_arguments">The fmt package</h3>

doc/go1.3.html

@@ -66,7 +66,7 @@ It runs on the 32-bit Intel architectures (<code>GOARCH=386</code>) and also on
 There is not yet support for Native Client on ARM.
 Note that this is Native Client (NaCl), not Portable Native Client (PNaCl).
 Details about Native Client are <a href="https://developers.google.com/native-client/dev/">here</a>;
-how to set up the Go version is described <a href="http://golang.org/wiki/NativeClient">here</a>.
+how to set up the Go version is described <a href="//golang.org/wiki/NativeClient">here</a>.
 </p>
 
 <h3 id="netbsd">Support for NetBSD</h3>
@@ -118,7 +118,7 @@ than is available, its stack is transferred to a larger single block of memory.
 The overhead of this transfer operation amortizes well and eliminates the old "hot spot"
 problem when a calculation repeatedly steps across a segment boundary.
 Details including performance numbers are in this
-<a href="http://golang.org/s/contigstacks">design document</a>.
+<a href="//golang.org/s/contigstacks">design document</a>.
 </p>
 
 <h3 id="garbage_collector">Changes to the garbage collector</h3>
@@ -141,7 +141,7 @@ Programs that use <a href="/pkg/unsafe/">package unsafe</a> to store pointers
 in integer-typed values are also illegal but more difficult to diagnose during execution.
 Because the pointers are hidden from the runtime, a stack expansion or garbage collection
 may reclaim the memory they point at, creating
-<a href="http://en.wikipedia.org/wiki/Dangling_pointer">dangling pointers</a>.
+<a href="//en.wikipedia.org/wiki/Dangling_pointer">dangling pointers</a>.
 </p>
 
 <p>
@@ -154,7 +154,7 @@ Such code can be identified by <code>go vet</code>.
 
 <p>
 Iterations over small maps no longer happen in a consistent order.
-Go 1 defines that &ldquo;<a href="http://golang.org/ref/spec#For_statements">The iteration order over maps
+Go 1 defines that &ldquo;<a href="//golang.org/ref/spec#For_statements">The iteration order over maps
 is not specified and is not guaranteed to be the same from one iteration to the next.</a>&rdquo;
 To keep code from depending on map iteration order,
 Go 1.0 started each map iteration at a random index in the map.
@@ -175,7 +175,7 @@ Because only small maps are affected, the problem arises most often in tests.
 <h3 id="liblink">The linker</h3>
 
 <p>
-As part of the general <a href="http://golang.org/s/go13linker">overhaul</a> to
+As part of the general <a href="//golang.org/s/go13linker">overhaul</a> to
 the Go linker, the compilers and linkers have been refactored.
 The linker is still a C program, but now the instruction selection phase that
 was part of the linker has been moved to the compiler through the creation of a new
@@ -295,13 +295,30 @@ For example,
 (The same change was made to the compilers and linkers in <a href="/doc/go1.1#gc_flag">Go 1.1</a>.)
 </p>
 
+<h3 id="godoc">Changes to godoc</h3>
+<p>
+When invoked with the <code>-analysis</code> flag, 
+<a href="//godoc.org/code.google.com/p/go.tools/cmd/godoc">godoc</a>
+now performs sophisticated <a href="/lib/godoc/analysis/help.html">static
+analysis</a> of the code it indexes.  
+The results of analysis are presented in both the source view and the
+package documentation view, and include the call graph of each package
+and the relationships between 
+definitions and references,
+types and their methods,
+interfaces and their implementations,
+send and receive operations on channels,
+functions and their callers, and
+call sites and their callees.
+</p>
+
 <h3 id="misc">Miscellany</h3>
 
 <p>
 The program <code>misc/benchcmp</code> that compares
 performance across benchmarking runs has been rewritten.
 Once a shell and awk script in the main repository, it is now a Go program in the <code>go.tools</code> repo.
-Documentation is <a href="http://godoc.org/code.google.com/p/go.tools/cmd/benchcmp">here</a>.
+Documentation is <a href="//godoc.org/code.google.com/p/go.tools/cmd/benchcmp">here</a>.
 </p>
 
 <p>
@@ -309,7 +326,6 @@ For the few of us that build Go distributions, the tool <code>misc/dist</code> h
 moved and renamed; it now lives in <code>misc/makerelease</code>, still in the main repository.
 </p>
 
-
 <h2 id="performance">Performance</h2>
 
 <p>
@@ -505,6 +521,15 @@ field to specify an end-to-end timeout on requests made using the
 client.
 </li>
 
+<li>
+The <a href="/pkg/net/http/"><code>net/http</code></a> package's
+<a href="/pkg/net/http/#Request.ParseMultipartForm"><code>Request.ParseMultipartForm</code></a>
+method will now return an error if the body's <code>Content-Type</code>
+is not <code>mutipart/form-data</code>.
+Prior to Go 1.3 it would silently fail and return <code>nil</code>.
+Code that relies on the previous behavior should be updated.
+</li>
+
 <li> In the <a href="/pkg/net/"><code>net</code></a> package,
 the <a href="/pkg/net/#Dialer"><code>Dialer</code></a> struct now
 has a <code>KeepAlive</code> option to specify a keep-alive period for the connection.
@@ -539,6 +564,12 @@ Earlier versions of Go handled this situation inconsistently: most instances
 were reported as deadlocks, but some trivial cases exited cleanly instead.
 </li>
 
+<li>
+The runtime/debug package now has a new function
+<a href="/pkg/runtime/debug/#WriteHeapDump"><code>debug.WriteHeapDump</code></a>
+that writes out a description of the heap.
+</li>
+
 <li>
 The <a href="/pkg/strconv/#CanBackquote"><code>CanBackquote</code></a>
 function in the <a href="/pkg/strconv/"><code>strconv</code></a> package

doc/go1.html

@@ -486,7 +486,7 @@ into subdirectories. For instance, <code>utf8</code> and
 <code>utf16</code> now occupy subdirectories of <code>unicode</code>.
 Also, <a href="#subrepo">some packages</a> have moved into
 subrepositories of
-<a href="http://code.google.com/p/go"><code>code.google.com/p/go</code></a>
+<a href="//code.google.com/p/go"><code>code.google.com/p/go</code></a>
 while <a href="#deleted">others</a> have been deleted outright.
 </p>
 
@@ -565,7 +565,7 @@ by hand.
 <p>
 Because they are not standardized, the packages under the <code>exp</code> directory will not be available in the
 standard Go 1 release distributions, although they will be available in source code form
-in <a href="http://code.google.com/p/go/">the repository</a> for
+in <a href="//code.google.com/p/go/">the repository</a> for
 developers who wish to use them.
 </p>
 
@@ -651,7 +651,7 @@ and also the command <code>gotry</code>.
 <em>Updating</em>:
 Code that uses <code>container/vector</code> should be updated to use
 slices directly.  See
-<a href="http://code.google.com/p/go-wiki/wiki/SliceTricks">the Go
+<a href="//code.google.com/p/go-wiki/wiki/SliceTricks">the Go
 Language Community Wiki</a> for some suggestions.
 Code that uses the other packages (there should be almost zero) will need to be rethought.
 </p>
@@ -660,7 +660,7 @@ Code that uses the other packages (there should be almost zero) will need to be
 
 <p>
 Go 1 has moved a number of packages into other repositories, usually sub-repositories of
-<a href="http://code.google.com/p/go/">the main Go repository</a>.
+<a href="//code.google.com/p/go/">the main Go repository</a>.
 This table lists the old and new import paths:
 
 <table class="codetable" frame="border" summary="Sub-repositories">
@@ -1695,7 +1695,7 @@ The compiler will catch code using the old interface.
 The <a href="/pkg/regexp/"><code>regexp</code></a> package has been rewritten.
 It has the same interface but the specification of the regular expressions
 it supports has changed from the old "egrep" form to that of
-<a href="http://code.google.com/p/re2/">RE2</a>.
+<a href="//code.google.com/p/re2/">RE2</a>.
 </p>
 
 <p>
@@ -1912,7 +1912,7 @@ package <a href="/pkg/reflect/"><code>reflect</code></a>.
 <em>Updating</em>:
 Code using these functions must be rewritten to use
 package <a href="/pkg/reflect/"><code>reflect</code></a>.
-The changes to <a href="http://code.google.com/p/go/source/detail?r=2646dc956207">encoding/gob</a> and the <a href="http://code.google.com/p/goprotobuf/source/detail?r=5340ad310031">protocol buffer library</a>
+The changes to <a href="//golang.org/change/2646dc956207">encoding/gob</a> and the <a href="//code.google.com/p/goprotobuf/source/detail?r=5340ad310031">protocol buffer library</a>
 may be helpful as examples.
 </p>
 
@@ -2035,4 +2035,4 @@ They are available for many combinations of architecture and operating system
 Installation details are described on the
 <a href="/doc/install">Getting Started</a> page, while
 the distributions themselves are listed on the
-<a href="http://code.google.com/p/go/downloads/list">downloads page</a>.
+<a href="/dl/">downloads page</a>.

doc/go1compat.html

@@ -145,7 +145,7 @@ developed software based on Go 1.
 
 <p>
 Code in sub-repositories of the main go tree, such as
-<a href="http://code.google.com/p/go.net">code.google.com/p/go.net</a>,
+<a href="//code.google.com/p/go.net">code.google.com/p/go.net</a>,
 may be developed under
 looser compatibility requirements. However, the sub-repositories
 will be tagged as appropriate to identify versions that are compatible

doc/go_faq.html

@@ -57,7 +57,7 @@ software on multicore machines.
 
 <p>
 A much more expansive answer to this question is available in the article,
-<a href="http://talks.golang.org/2012/splash.article">Go at Google:
+<a href="//talks.golang.org/2012/splash.article">Go at Google:
 Language Design in the Service of Software Engineering</a>.
 
 <h3 id="What_is_the_status_of_the_project">
@@ -66,7 +66,7 @@ What is the status of the project?</h3>
 <p>
 Go became a public open source project on November 10, 2009.
 After a couple of years of very active design and development, stability was called for and
-Go 1 was <a href="http://blog.golang.org/2012/03/go-version-1-is-released.html">released</a>
+Go 1 was <a href="//blog.golang.org/2012/03/go-version-1-is-released.html">released</a>
 on March 28, 2012.
 Go 1, which includes a <a href="/ref/spec">language specification</a>,
 <a href="/pkg/">standard libraries</a>,
@@ -163,7 +163,7 @@ language was called for.
 </p>
 
 <p>
-The article <a href="http://talks.golang.org/2012/splash.article">Go at Google</a>
+The article <a href="//talks.golang.org/2012/splash.article">Go at Google</a>
 discusses the background and motivation behind the design of the Go language,
 as well as providing more detail about many of the answers presented in this FAQ.
 </p>
@@ -221,7 +221,7 @@ easier to understand what happens when things combine.
 <p>
 Yes. There are now several Go programs deployed in
 production inside Google.  A public example is the server behind
-<a href="http://golang.org">http://golang.org</a>.
+<a href="//golang.org">golang.org</a>.
 It's just the <a href="/cmd/godoc"><code>godoc</code></a>
 document server running in a production configuration on
 <a href="https://developers.google.com/appengine/">Google App Engine</a>.
@@ -260,7 +260,7 @@ Does Go support Google's protocol buffers?</h3>
 <p>
 A separate open source project provides the necessary compiler plugin and library.
 It is available at
-<a href="http://code.google.com/p/goprotobuf/">http://code.google.com/p/goprotobuf/</a>
+<a href="//code.google.com/p/goprotobuf/">code.google.com/p/goprotobuf/</a>
 </p>
 
 
@@ -270,9 +270,9 @@ Can I translate the Go home page into another language?</h3>
 <p>
 Absolutely. We encourage developers to make Go Language sites in their own languages.
 However, if you choose to add the Google logo or branding to your site
-(it does not appear on <a href="http://golang.org/">golang.org</a>),
+(it does not appear on <a href="//golang.org/">golang.org</a>),
 you will need to abide by the guidelines at
-<a href="http://www.google.com/permissions/guidelines.html">http://www.google.com/permissions/guidelines.html</a>
+<a href="//www.google.com/permissions/guidelines.html">www.google.com/permissions/guidelines.html</a>
 </p>
 
 <h2 id="Design">Design</h2>
@@ -466,7 +466,7 @@ Will you accept my language change?</h3>
 
 <p>
 People often suggest improvements to the language—the
-<a href="http://groups.google.com/group/golang-nuts">mailing list</a>
+<a href="//groups.google.com/group/golang-nuts">mailing list</a>
 contains a rich history of such discussions—but very few of these changes have
 been accepted.
 </p>
@@ -484,7 +484,7 @@ to start talking about what that might be.
 <p>
 Even if your proposal is compatible with the Go 1 spec, it might
 not be in the spirit of Go's design goals.
-The article <i><a href="http://talks.golang.org/2012/splash.article">Go
+The article <i><a href="//talks.golang.org/2012/splash.article">Go
 at Google: Language Design in the Service of Software Engineering</a></i>
 explains Go's origins and the motivation behind its design.
 </p>
@@ -940,9 +940,9 @@ How are libraries documented?</h3>
 There is a program, <code>godoc</code>, written in Go, that extracts
 package documentation from the source code. It can be used on the
 command line or on the web. An instance is running at
-<a href="/pkg/">http://golang.org/pkg/</a>.
+<a href="/pkg/">golang.org/pkg/</a>.
 In fact, <code>godoc</code> implements the full site at
-<a href="/">http://golang.org/</a>.
+<a href="/">golang.org/</a>.
 </p>
 
 <h3 id="Is_there_a_Go_programming_style_guide">
@@ -961,7 +961,7 @@ All the Go code in the repository has been run through <code>gofmt</code>.
 
 <p>
 The document titled
-<a href="http://golang.org/s/comments">Go Code Review Comments</a>
+<a href="//golang.org/s/comments">Go Code Review Comments</a>
 is a collection of very short essays about details of Go idiom that are often
 missed by programmers.
 It is a handy reference for people doing code reviews for Go projects.
@@ -986,14 +986,14 @@ Why does the project use Mercurial and not git?</h3>
 
 <p>
 The Go project, hosted by Google Code at
-<a href="http://code.google.com/p/go">code.google.com/p/go</a>,
+<a href="//code.google.com/p/go">code.google.com/p/go</a>,
 uses Mercurial as its version control system.
 When the project launched,
 Google Code supported only Subversion and Mercurial.
 Mercurial was a better choice because of its plugin mechanism
 that allowed us to create the "codereview" plugin to connect
 the project to the excellent code review tools at
-<a href="http://codereview.appspot.com">codereview.appspot.com</a>.
+<a href="//codereview.appspot.com">codereview.appspot.com</a>.
 </p>
 
 <p>
@@ -1304,7 +1304,7 @@ Do not communicate by sharing memory. Instead, share memory by communicating.
 </p>
 
 <p>
-See the <a href="/doc/codewalk/sharemem/">Share Memory By Communicating</a> code walk and its <a href="http://blog.golang.org/2010/07/share-memory-by-communicating.html">associated article</a> for a detailed discussion of this concept.
+See the <a href="/doc/codewalk/sharemem/">Share Memory By Communicating</a> code walk and its <a href="//blog.golang.org/2010/07/share-memory-by-communicating.html">associated article</a> for a detailed discussion of this concept.
 </p>
 
 <h3 id="Why_no_multi_CPU">
@@ -1321,7 +1321,7 @@ run-time support to utilize more than one OS thread.
 Programs that perform parallel computation should benefit from an increase in
 <code>GOMAXPROCS</code>.
 However, be aware that
-<a href="http://blog.golang.org/2013/01/concurrency-is-not-parallelism.html">concurrency
+<a href="//blog.golang.org/2013/01/concurrency-is-not-parallelism.html">concurrency
 is not parallelism</a>.
 </p>
 
@@ -1358,7 +1358,7 @@ should recognize such cases and optimize its use of OS threads. For now,
 
 <p>
 For more detail on this topic see the talk entitled,
-<a href="http://blog.golang.org/2013/01/concurrency-is-not-parallelism.html">Concurrency
+<a href="//blog.golang.org/2013/01/concurrency-is-not-parallelism.html">Concurrency
 is not Parallelism</a>.
 
 <h2 id="Functions_methods">Functions and Methods</h2>
@@ -1736,7 +1736,7 @@ In any case, Go can often be very competitive.
 There has been significant improvement in the performance of many programs
 as the language and tools have developed.
 See the blog post about
-<a href="http://blog.golang.org/2011/06/profiling-go-programs.html">profiling
+<a href="//blog.golang.org/2011/06/profiling-go-programs.html">profiling
 Go programs</a> for an informative example.
 
 <h2 id="change_from_c">Changes from C</h2>
@@ -1895,7 +1895,7 @@ considerable control over memory layout and allocation, much more than
 is typical in garbage-collected languages. A careful programmer can reduce
 the garbage collection overhead dramatically by using the language well;
 see the article about
-<a href="http://blog.golang.org/2011/06/profiling-go-programs.html">profiling
+<a href="//blog.golang.org/2011/06/profiling-go-programs.html">profiling
 Go programs</a> for a worked example, including a demonstration of Go's
 profiling tools.
 </p>

doc/go_mem.html

@@ -1,6 +1,6 @@
 <!--{
 	"Title": "The Go Memory Model",
-	"Subtitle": "Version of March 6, 2012",
+	"Subtitle": "Version of May 31, 2014",
 	"Path": "/ref/mem"
 }-->
 
@@ -275,17 +275,17 @@ crash, or do something else.)
 </p>
 
 <p class="rule">
-The <i>k</i>th send on a channel with capacity <i>C</i> happens before the <i>k</i>+<i>C</i>th receive from that channel completes.
+The <i>k</i>th receive on a channel with capacity <i>C</i> happens before the <i>k</i>+<i>C</i>th send from that channel completes.
 </p>
 
 <p>
 This rule generalizes the previous rule to buffered channels.
 It allows a counting semaphore to be modeled by a buffered channel:
-the number of items in the channel corresponds to the semaphore count,
-the capacity of the channel corresponds to the semaphore maximum,
+the number of items in the channel corresponds to the number of active uses,
+the capacity of the channel corresponds to the maximum number of simultaneous uses,
 sending an item acquires the semaphore, and receiving an item releases
 the semaphore.
-This is a common idiom for rate-limiting work.
+This is a common idiom for limiting concurrency.
 </p>
 
 <p>

doc/go_spec.html

@@ -23,7 +23,7 @@ TODO
 
 <p>
 This is a reference manual for the Go programming language. For
-more information and other documents, see <a href="/">http://golang.org</a>.
+more information and other documents, see <a href="/">golang.org</a>.
 </p>
 
 <p>

doc/help.html

@@ -20,11 +20,11 @@ Need help with Go? Try these resources.
 <h3 id="wiki"><a href="/wiki">The Go Wiki</a></h3>
 <p>A wiki maintained by the Go community.</p>
 
-<h3 id="mailinglist"><a href="http://groups.google.com/group/golang-nuts">Go Nuts Mailing List</a></h3>
+<h3 id="mailinglist"><a href="//groups.google.com/group/golang-nuts">Go Nuts Mailing List</a></h3>
 <p>
-Search the <a href="http://groups.google.com/group/golang-nuts">golang-nuts</a>
+Search the <a href="//groups.google.com/group/golang-nuts">golang-nuts</a>
 archives and consult the <a href="/doc/go_faq.html">FAQ</a> and
-<a href="http://code.google.com/p/go-wiki/wiki">wiki</a> before posting.
+<a href="//code.google.com/p/go-wiki/wiki">wiki</a> before posting.
 </p>
 
 <h3 id="irc"><a href="irc:irc.freenode.net/go-nuts">Go IRC Channel</a></h3>
@@ -37,7 +37,7 @@ Go IRC channel.</p>
 <h3 id="plus"><a href="https://plus.google.com/101406623878176903605/posts">The Go Programming Language at Google+</a></h3>
 <p>The Go project's Google+ page.</p>
 
-<h3 id="twitter"><a href="http://twitter.com/golang">@golang at Twitter</a></h3>
+<h3 id="twitter"><a href="//twitter.com/golang">@golang at Twitter</a></h3>
 <p>The Go project's official Twitter account.</p>
 <p>Tweeting about your problem with the <code>#golang</code> hashtag usually
 generates some helpful responses.</p>

doc/install-source.html

@@ -81,7 +81,7 @@ The full set of supported combinations is listed in the discussion of
 
 <p>
 The Go tool chain is written in C. To build it, you need a C compiler installed. 
-Please refer to the <a href="http://code.google.com/p/go-wiki/wiki/InstallFromSource#Install_C_tools">InstallFromSource</a>
+Please refer to the <a href="//golang.org/wiki/InstallFromSource#Install_C_tools">InstallFromSource</a>
 page on the Go community Wiki for operating system specific instructions.
 </p>
 
@@ -283,16 +283,16 @@ You can access the latter commands with
 The usual community resources such as
 <code>#go-nuts</code> on the <a href="http://freenode.net/">Freenode</a> IRC server
 and the
-<a href="http://groups.google.com/group/golang-nuts">Go Nuts</a>
+<a href="//groups.google.com/group/golang-nuts">Go Nuts</a>
 mailing list have active developers that can help you with problems
 with your installation or your development work.
 For those who wish to keep up to date,
-there is another mailing list, <a href="http://groups.google.com/group/golang-checkins">golang-checkins</a>,
+there is another mailing list, <a href="//groups.google.com/group/golang-checkins">golang-checkins</a>,
 that receives a message summarizing each checkin to the Go repository.
 </p>
 
 <p>
-Bugs can be reported using the <a href="http://code.google.com/p/go/issues/list">Go issue tracker</a>.
+Bugs can be reported using the <a href="//code.google.com/p/go/issues/list">Go issue tracker</a>.
 </p>
 
 
@@ -306,7 +306,7 @@ The Go project maintains a stable tag in its Mercurial repository:
 <p>
 The <code>release</code> tag refers to the current stable release of Go.
 Most Go users should use this version. New releases are announced on the
-<a href="http://groups.google.com/group/golang-announce">golang-announce</a>
+<a href="//groups.google.com/group/golang-announce">golang-announce</a>
 mailing list.
 </p>
 
@@ -457,7 +457,7 @@ installs all commands there.
 </p>
 
 <li><code>$GO386</code> (for <code>386</code> only, default is auto-detected
-if built natively, <code>387</code> if not)
+if built on either <code>386</code> or <code>amd64</code>, <code>387</code> otherwise)
 <p>
 This controls the code generated by 8g to use either the 387 floating-point unit
 (set to <code>387</code>) or SSE2 instructions (set to <code>sse2</code>) for
@@ -482,8 +482,8 @@ should target. If you are compiling on the target system, its value will be auto
 <p>
 If in doubt, leave this variable unset, and adjust it if required
 when you first run the Go executable.
-The <a href="http://code.google.com/p/go-wiki/wiki/GoArm">GoARM</a> page
-on the <a href="http://code.google.com/p/go-wiki/w/list">Go community wiki</a>
+The <a href="//golang.org/wiki/GoArm">GoARM</a> page
+on the <a href="//golang.org/wiki">Go community wiki</a>
 contains further details regarding Go's ARM support.
 </p>
 

doc/install.html

@@ -6,7 +6,7 @@
 <h2 id="download">Download the Go distribution</h2>
 
 <p>
-<a href="https://code.google.com/p/go/wiki/Downloads?tm=2" id="start" class="download" target="_blank">
+<a href="/dl/" id="start" class="download" target="_blank">
 <span class="big">Download Go</span>
 <span class="desc">Click here to visit the downloads page</span>
 </a>
@@ -39,9 +39,9 @@ proceeding. If your OS or architecture is not on the list, it's possible that
 
 <table class="codetable" frame="border" summary="requirements">
 <tr>
-<th align="middle">Operating system</th>
-<th align="middle">Architectures</th>
-<th align="middle">Notes</th>
+<th align="center">Operating system</th>
+<th align="center">Architectures</th>
+<th align="center">Notes</th>
 </tr>
 <tr><td colspan="3"><hr></td></tr>
 <tr><td>FreeBSD 8 or later</td> <td>amd64, 386, arm</td> <td>Debian GNU/kFreeBSD not supported; FreeBSD/ARM needs FreeBSD 10 or later</td></tr>
@@ -270,12 +270,12 @@ For real-time help, ask the helpful gophers in <code>#go-nuts</code> on the
 
 <p>
 The official mailing list for discussion of the Go language is
-<a href="http://groups.google.com/group/golang-nuts">Go Nuts</a>.
+<a href="//groups.google.com/group/golang-nuts">Go Nuts</a>.
 </p>
 
 <p>
 Report bugs using the
-<a href="http://code.google.com/p/go/issues/list">Go issue tracker</a>.
+<a href="//golang.org/issue">Go issue tracker</a>.
 </p>
 
 

doc/root.html

@@ -26,7 +26,7 @@ Hello, 世界
 <div class="buttons">
 <a class="run" href="#" title="Run this code [shift-enter]">Run</a>
 <a class="share" href="#" title="Share this code">Share</a>
-<a class="tour" href="http://tour.golang.org/" title="Learn Go from your browser">Tour</a>
+<a class="tour" href="//tour.golang.org/" title="Learn Go from your browser">Tour</a>
 </div>
 <div class="toys">
 <select>
@@ -78,7 +78,7 @@ Linux, Mac OS X, Windows, and more.
 
 <div id="blog">
 <div class="rootHeading">Featured articles</div>
-<div class="read"><a href="http://blog.golang.org/">Read more</a></div>
+<div class="read"><a href="//blog.golang.org/">Read more</a></div>
 </div>
 
 </div>
@@ -124,7 +124,7 @@ window.initFuncs.push(function() {
 			"outputEl":      "#learn .output",
 			"runEl":         "#learn .run",
 			"shareEl":       "#learn .share",
-			"shareRedirect": "http://play.golang.org/p/",
+			"shareRedirect": "//play.golang.org/p/",
 			"toysEl":        "#learn .toys select"
 		});
 	} else {
@@ -133,7 +133,7 @@ window.initFuncs.push(function() {
 
 	// Load blog feed.
 	$('<script/>').attr('text', 'text/javascript')
-		.attr('src', 'http://blog.golang.org/.json?jsonp=feedLoaded')
+		.attr('src', '//blog.golang.org/.json?jsonp=feedLoaded')
 		.appendTo('body');
 
 	// Set the video at random.

doc/tos.html

@@ -5,7 +5,7 @@
 <p>
 The Go website (the "Website") is hosted by Google.
 By using and/or visiting the Website, you consent to be bound by Google's general
-<a href="http://www.google.com/intl/en/policies/terms/">Terms of Service</a>
+<a href="//www.google.com/intl/en/policies/terms/">Terms of Service</a>
 and Google's general
-<a href="http://www.google.com/intl/en/privacy/privacy-policy.html">Privacy Policy</a>.
+<a href="//www.google.com/intl/en/privacy/privacy-policy.html">Privacy Policy</a>.
 </p>

include/u.h

@@ -40,6 +40,7 @@ extern "C" {
 #endif
 #define _BSD_SOURCE 1
 #define _NETBSD_SOURCE 1	/* NetBSD */
+#define _DEFAULT_SOURCE 1	/* glibc > 2.19 */
 #define _SVID_SOURCE 1
 #if !defined(__APPLE__) && !defined(__OpenBSD__)
 #	define _XOPEN_SOURCE 1000

misc/cgo/test/cgo_test.go

@@ -53,5 +53,6 @@ func Test5986(t *testing.T)                { test5986(t) }
 func Test7665(t *testing.T)                { test7665(t) }
 func TestNaming(t *testing.T)              { testNaming(t) }
 func Test7560(t *testing.T)                { test7560(t) }
+func Test7978(t *testing.T)                { test7978(t) }
 
 func BenchmarkCgoCall(b *testing.B) { benchCgoCall(b) }

misc/cgo/test/issue7978.go

@@ -0,0 +1,103 @@
+// Copyright 2014 The Go Authors.  All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// Issue 7978.  Stack tracing didn't work during cgo code after calling a Go
+// callback.  Make sure GC works and the stack trace is correct.
+
+package cgotest
+
+/*
+#include <stdint.h>
+
+void issue7978cb(void);
+
+// use ugly atomic variable sync since that doesn't require calling back into
+// Go code or OS dependencies
+static void issue7978c(uint32_t *sync) {
+	while(__sync_fetch_and_add(sync, 0) != 0)
+		;
+	__sync_fetch_and_add(sync, 1);
+	while(__sync_fetch_and_add(sync, 0) != 2)
+		;
+	issue7978cb();
+	__sync_fetch_and_add(sync, 1);
+	while(__sync_fetch_and_add(sync, 0) != 6)
+		;
+}
+*/
+import "C"
+
+import (
+	"os"
+	"runtime"
+	"strings"
+	"sync/atomic"
+	"testing"
+)
+
+var issue7978sync uint32
+
+func issue7978check(t *testing.T, wantFunc string, badFunc string, depth int) {
+	runtime.GC()
+	buf := make([]byte, 65536)
+	trace := string(buf[:runtime.Stack(buf, true)])
+	for _, goroutine := range strings.Split(trace, "\n\n") {
+		if strings.Contains(goroutine, "test.issue7978go") {
+			trace := strings.Split(goroutine, "\n")
+			// look for the expected function in the stack
+			for i := 0; i < depth; i++ {
+				if badFunc != "" && strings.Contains(trace[1+2*i], badFunc) {
+					t.Errorf("bad stack: found %s in the stack:\n%s", badFunc, goroutine)
+					return
+				}
+				if strings.Contains(trace[1+2*i], wantFunc) {
+					return
+				}
+			}
+			t.Errorf("bad stack: didn't find %s in the stack:\n%s", wantFunc, goroutine)
+			return
+		}
+	}
+	t.Errorf("bad stack: goroutine not found. Full stack dump:\n%s", trace)
+}
+
+func issue7978wait(store uint32, wait uint32) {
+	if store != 0 {
+		atomic.StoreUint32(&issue7978sync, store)
+	}
+	for atomic.LoadUint32(&issue7978sync) != wait {
+		runtime.Gosched()
+	}
+}
+
+//export issue7978cb
+func issue7978cb() {
+	issue7978wait(3, 4)
+}
+
+func issue7978go() {
+	C.issue7978c((*C.uint32_t)(&issue7978sync))
+	issue7978wait(7, 8)
+}
+
+func test7978(t *testing.T) {
+	if os.Getenv("GOTRACEBACK") != "2" {
+		t.Fatal("GOTRACEBACK must be 2")
+	}
+	issue7978sync = 0
+	go issue7978go()
+	// test in c code, before callback
+	issue7978wait(0, 1)
+	issue7978check(t, "runtime.cgocall(", "", 1)
+	// test in go code, during callback
+	issue7978wait(2, 3)
+	issue7978check(t, "test.issue7978cb(", "test.issue7978go", 4)
+	// test in c code, after callback
+	issue7978wait(4, 5)
+	issue7978check(t, "runtime.cgocall(", "runtime.cgocallback", 1)
+	// test in go code, after return from cgo
+	issue7978wait(6, 7)
+	issue7978check(t, "test.issue7978go(", "", 4)
+	atomic.StoreUint32(&issue7978sync, 8)
+}

misc/cgo/test/issue8148.go

@@ -0,0 +1,31 @@
+// Copyright 2014 The Go Authors.  All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// Issue 8148.  A typedef of an unnamed struct didn't work when used
+// with an exported Go function.  No runtime test; just make sure it
+// compiles.
+
+package cgotest
+
+/*
+typedef struct { int i; } T;
+
+int issue8148Callback(T*);
+
+static int get() {
+	T t;
+	t.i = 42;
+	return issue8148Callback(&t);
+}
+*/
+import "C"
+
+//export issue8148Callback
+func issue8148Callback(t *C.T) C.int {
+	return t.i
+}
+
+func Issue8148() int {
+	return int(C.get())
+}

misc/cgo/test/issue8331.h

@@ -2,8 +2,6 @@
 // Use of this source code is governed by a BSD-style
 // license that can be found in the LICENSE file.
 
-// Placeholder to keep build building.
-
-package main
-
-func main() {}
+typedef struct {
+	int i;
+} issue8331;

misc/cgo/test/issue8331a.go

@@ -0,0 +1,15 @@
+// Copyright 2014 The Go Authors.  All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// Issue 8331.  A typedef of an unnamed struct is the same struct when
+// #include'd twice.  No runtime test; just make sure it compiles.
+
+package cgotest
+
+// #include "issue8331.h"
+import "C"
+
+func issue8331a() C.issue8331 {
+	return issue8331Var
+}

misc/cgo/test/issue8331b.go

@@ -0,0 +1,13 @@
+// Copyright 2014 The Go Authors.  All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// Issue 8331.  A typedef of an unnamed struct is the same struct when
+// #include'd twice.  No runtime test; just make sure it compiles.
+
+package cgotest
+
+// #include "issue8331.h"
+import "C"
+
+var issue8331Var C.issue8331

misc/cgo/test/issue8441.go

@@ -0,0 +1,27 @@
+// Copyright 2014 The Go Authors.  All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// Issue 8368 and 8441.  Recursive struct definitions didn't work.
+// No runtime test; just make sure it compiles.
+
+package cgotest
+
+/*
+typedef struct one one;
+typedef struct two two;
+struct one {
+	two *x;
+};
+struct two {
+	one *x;
+};
+*/
+import "C"
+
+func issue8368(one *C.struct_one, two *C.struct_two) {
+}
+
+func issue8441(one *C.one, two *C.two) {
+	issue8441(two.x, one.x)
+}

misc/nacl/README

@@ -54,12 +54,10 @@ The script's name has a special format, go_$GOOS_$GOARCH_exec, so cmd/go can fin
 
 In short, if the support scripts are in place, the cmd/go tool can be used as per normal.
 
-# Build the Go toolchain.
+# Build and test Go for NaCl
+
+NaCl does not permit direct file system access. Instead, package syscall provides a simulated file system served by in-memory data. The script nacltest.bash is the NaCl equivalent of all.bash. It builds NaCl with an in-memory file system containing files needed for tests, and then it runs the tests.
 
 	% cd go/src
-	% env GOOS=nacl GOARCH=amd64p32 ./make.bash
+	% env GOARCH=amd64p32 ./nacltest.bash
 
-# Test the Go toolchain.
-
-	% cd go/src
-	% env GOOS=nacl GOARCH=amd64p32 ./run.bash

misc/nacl/testzip.proto

@@ -13,9 +13,6 @@ go	src=..
 			gofmt
 				testdata
 					+
-			link
-				testdata
-					+
 		pkg
 			archive
 				tar

src/cmd/6g/ggen.c

@@ -47,7 +47,7 @@ defframe(Prog *ptxt)
 
 		if(lo != hi && n->xoffset + n->type->width >= lo - 2*widthreg) {
 			// merge with range we already have
-			lo = rnd(n->xoffset, widthreg);
+			lo = n->xoffset;
 			continue;
 		}
 		// zero old range
@@ -943,7 +943,7 @@ cgen_hmul(Node *nl, Node *nr, Node *res)
 	if(t->width == 1) {
 		// byte multiply behaves differently.
 		nodreg(&ax, t, D_AH);
-		nodreg(&dx, t, D_DL);
+		nodreg(&dx, t, D_DX);
 		gmove(&ax, &dx);
 	}
 	nodreg(&dx, t, D_DX);

src/cmd/6g/peep.c

@@ -838,6 +838,11 @@ copyu(Prog *p, Adr *v, Adr *s)
 static int
 copyas(Adr *a, Adr *v)
 {
+	if(D_AL <= a->type && a->type <= D_R15B)
+		fatal("use of byte register");
+	if(D_AL <= v->type && v->type <= D_R15B)
+		fatal("use of byte register");
+
 	if(a->type != v->type)
 		return 0;
 	if(regtyp(v))

src/cmd/8g/ggen.c

@@ -991,7 +991,7 @@ cgen_hmul(Node *nl, Node *nr, Node *res)
 	if(t->width == 1) {
 		// byte multiply behaves differently.
 		nodreg(&ax, t, D_AH);
-		nodreg(&dx, t, D_DL);
+		nodreg(&dx, t, D_DX);
 		gmove(&ax, &dx);
 	}
 	nodreg(&dx, t, D_DX);

src/cmd/8g/peep.c

@@ -634,6 +634,11 @@ copyu(Prog *p, Adr *v, Adr *s)
 static int
 copyas(Adr *a, Adr *v)
 {
+	if(D_AL <= a->type && a->type <= D_BL)
+		fatal("use of byte register");
+	if(D_AL <= v->type && v->type <= D_BL)
+		fatal("use of byte register");
+
 	if(a->type != v->type)
 		return 0;
 	if(regtyp(v))

src/cmd/cgo/gcc.go

@@ -552,8 +552,8 @@ func (p *Package) loadDWARF(f *File, names []*Name) {
 				n.Const = fmt.Sprintf("%#x", enumVal[i])
 			}
 		}
+		conv.FinishType(pos)
 	}
-
 }
 
 // mangleName does name mangling to translate names
@@ -926,6 +926,12 @@ type typeConv struct {
 	m       map[dwarf.Type]*Type
 	typedef map[string]ast.Expr
 
+	// Map from types to incomplete pointers to those types.
+	ptrs map[dwarf.Type][]*Type
+
+	// Fields to be processed by godefsField after completing pointers.
+	todoFlds [][]*ast.Field
+
 	// Predeclared types.
 	bool                                   ast.Expr
 	byte                                   ast.Expr // denotes padding
@@ -950,6 +956,7 @@ func (c *typeConv) Init(ptrSize, intSize int64) {
 	c.ptrSize = ptrSize
 	c.intSize = intSize
 	c.m = make(map[dwarf.Type]*Type)
+	c.ptrs = make(map[dwarf.Type][]*Type)
 	c.bool = c.Ident("bool")
 	c.byte = c.Ident("byte")
 	c.int8 = c.Ident("int8")
@@ -1029,6 +1036,32 @@ func (tr *TypeRepr) Set(repr string, fargs ...interface{}) {
 	tr.FormatArgs = fargs
 }
 
+// FinishType completes any outstanding type mapping work.
+// In particular, it resolves incomplete pointer types and also runs
+// godefsFields on any new struct types.
+func (c *typeConv) FinishType(pos token.Pos) {
+	// Completing one pointer type might produce more to complete.
+	// Keep looping until they're all done.
+	for len(c.ptrs) > 0 {
+		for dtype := range c.ptrs {
+			// Note Type might invalidate c.ptrs[dtype].
+			t := c.Type(dtype, pos)
+			for _, ptr := range c.ptrs[dtype] {
+				ptr.Go.(*ast.StarExpr).X = t.Go
+				ptr.C.Set("%s*", t.C)
+			}
+			delete(c.ptrs, dtype)
+		}
+	}
+
+	// Now that pointer types are completed, we can invoke godefsFields
+	// to rewrite struct definitions.
+	for _, fld := range c.todoFlds {
+		godefsFields(fld)
+	}
+	c.todoFlds = nil
+}
+
 // Type returns a *Type with the same memory layout as
 // dtype when used as the type of a variable or a struct field.
 func (c *typeConv) Type(dtype dwarf.Type, pos token.Pos) *Type {
@@ -1068,13 +1101,12 @@ func (c *typeConv) Type(dtype dwarf.Type, pos token.Pos) *Type {
 			t.Go = c.Opaque(t.Size)
 			break
 		}
-		gt := &ast.ArrayType{
-			Len: c.intExpr(dt.Count),
-		}
-		t.Go = gt // publish before recursive call
 		sub := c.Type(dt.Type, pos)
 		t.Align = sub.Align
-		gt.Elt = sub.Go
+		t.Go = &ast.ArrayType{
+			Len: c.intExpr(dt.Count),
+			Elt: sub.Go,
+		}
 		t.C.Set("__typeof__(%s[%d])", sub.C, dt.Count)
 
 	case *dwarf.BoolType:
@@ -1184,11 +1216,10 @@ func (c *typeConv) Type(dtype dwarf.Type, pos token.Pos) *Type {
 			break
 		}
 
-		gt := &ast.StarExpr{}
-		t.Go = gt // publish before recursive call
-		sub := c.Type(dt.Type, pos)
-		gt.X = sub.Go
-		t.C.Set("%s*", sub.C)
+		// Placeholder initialization; completed in FinishType.
+		t.Go = &ast.StarExpr{}
+		t.C.Set("<incomplete>*")
+		c.ptrs[dt.Type] = append(c.ptrs[dt.Type], t)
 
 	case *dwarf.QualType:
 		// Ignore qualifier.
@@ -1265,11 +1296,12 @@ func (c *typeConv) Type(dtype dwarf.Type, pos token.Pos) *Type {
 		}
 		name := c.Ident("_Ctype_" + dt.Name)
 		goIdent[name.Name] = name
-		t.Go = name // publish before recursive call
 		sub := c.Type(dt.Type, pos)
+		t.Go = name
 		t.Size = sub.Size
 		t.Align = sub.Align
-		if _, ok := typedef[name.Name]; !ok {
+		oldType := typedef[name.Name]
+		if oldType == nil {
 			tt := *t
 			tt.Go = sub.Go
 			typedef[name.Name] = &tt
@@ -1281,6 +1313,20 @@ func (c *typeConv) Type(dtype dwarf.Type, pos token.Pos) *Type {
 		// In -godefs and -cdefs mode, do this for all typedefs.
 		if isStructUnionClass(sub.Go) || *godefs || *cdefs {
 			t.Go = sub.Go
+
+			if isStructUnionClass(sub.Go) {
+				// Use the typedef name for C code.
+				typedef[sub.Go.(*ast.Ident).Name].C = t.C
+			}
+
+			// If we've seen this typedef before, and it
+			// was an anonymous struct/union/class before
+			// too, use the old definition.
+			// TODO: it would be safer to only do this if
+			// we verify that the types are the same.
+			if oldType != nil && isStructUnionClass(oldType.Go) {
+				t.Go = oldType.Go
+			}
 		}
 
 	case *dwarf.UcharType:
@@ -1589,7 +1635,7 @@ func (c *typeConv) Struct(dt *dwarf.StructType, pos token.Pos) (expr *ast.Struct
 	csyntax = buf.String()
 
 	if *godefs || *cdefs {
-		godefsFields(fld)
+		c.todoFlds = append(c.todoFlds, fld)
 	}
 	expr = &ast.StructType{Fields: &ast.FieldList{List: fld}}
 	return

src/cmd/gc/array.c

@@ -108,20 +108,6 @@ arrayadd(Array *array, void *element)
 	arrayset(array, array->length - 1, element);
 }
 
-int32
-arrayindexof(Array *array, void *element)
-{
-	void *p;
-	int32 i;
-
-	for(i = 0; i < array->length; i++) {
-		p = arrayget(array, i);
-		if(memcmp(p, &element, array->size) == 0)
-			return i;
-	}
-	return -1;
-}
-
 void
 arraysort(Array *array, int (*cmp)(const void*, const void*))
 {

src/cmd/gc/bv.c

@@ -9,6 +9,8 @@
 enum {
 	WORDSIZE = sizeof(uint32),
 	WORDBITS = 32,
+	WORDMASK = WORDBITS - 1,
+	WORDSHIFT = 5,
 };
 
 static uintptr
@@ -94,13 +96,35 @@ bvconcat(Bvec *src1, Bvec *src2)
 int
 bvget(Bvec *bv, int32 i)
 {
-	uint32 mask, word;
-
 	if(i < 0 || i >= bv->n)
 		fatal("bvget: index %d is out of bounds with length %d\n", i, bv->n);
-	mask = 1U << (i % WORDBITS);
-	word = bv->b[i / WORDBITS] & mask;
-	return word ? 1 : 0;
+	return (bv->b[i>>WORDSHIFT] >> (i&WORDMASK)) & 1;
+}
+
+// bvnext returns the smallest index >= i for which bvget(bv, i) == 1.
+// If there is no such index, bvnext returns -1.
+int
+bvnext(Bvec *bv, int32 i)
+{
+	uint32 w;
+
+	// Jump i ahead to next word with bits.
+	if((bv->b[i>>WORDSHIFT]>>(i&WORDMASK)) == 0) {
+		i &= ~WORDMASK;
+		i += WORDBITS;
+		while(i < bv->n && bv->b[i>>WORDSHIFT] == 0)
+			i += WORDBITS;
+	}
+	if(i >= bv->n)
+		return -1;
+	
+	// Find 1 bit.
+	w = bv->b[i>>WORDSHIFT]>>(i&WORDMASK);
+	while((w&1) == 0) {
+		w>>=1;
+		i++;
+	}
+	return i;
 }
 
 int
@@ -109,7 +133,7 @@ bvisempty(Bvec *bv)
 	int32 i;
 
 	for(i = 0; i < bv->n; i += WORDBITS)
-		if(bv->b[i / WORDBITS] != 0)
+		if(bv->b[i>>WORDSHIFT] != 0)
 			return 0;
 	return 1;
 }

src/cmd/gc/esc.c

@@ -204,6 +204,13 @@ struct EscState {
 	// flow to.
 	Node	theSink;
 	
+	// If an analyzed function is recorded to return
+	// pieces obtained via indirection from a parameter,
+	// and later there is a call f(x) to that function,
+	// we create a link funcParam <- x to record that fact.
+	// The funcParam node is handled specially in escflood.
+	Node	funcParam;	
+	
 	NodeList*	dsts;		// all dst nodes
 	int	loopdepth;	// for detecting nested loop scopes
 	int	pdepth;		// for debug printing in recursions.
@@ -269,7 +276,13 @@ analyze(NodeList *all, int recursive)
 	e->theSink.sym = lookup(".sink");
 	e->theSink.escloopdepth = -1;
 	e->recursive = recursive;
-
+	
+	e->funcParam.op = ONAME;
+	e->funcParam.orig = &e->funcParam;
+	e->funcParam.class = PAUTO;
+	e->funcParam.sym = lookup(".param");
+	e->funcParam.escloopdepth = 10000000;
+	
 	for(l=all; l; l=l->next)
 		if(l->n->op == ODCLFUNC)
 			l->n->esc = EscFuncPlanned;
@@ -429,6 +442,18 @@ esc(EscState *e, Node *n, Node *up)
 	if(n->op == OFOR || n->op == ORANGE)
 		e->loopdepth++;
 
+	// type switch variables have no ODCL.
+	// process type switch as declaration.
+	// must happen before processing of switch body,
+	// so before recursion.
+	if(n->op == OSWITCH && n->ntest && n->ntest->op == OTYPESW) {
+		for(ll=n->list; ll; ll=ll->next) {  // cases
+			// ll->n->nname is the variable per case
+			if(ll->n->nname)
+				ll->n->nname->escloopdepth = e->loopdepth;
+		}
+	}
+
 	esc(e, n->left, n);
 	esc(e, n->right, n);
 	esc(e, n->ntest, n);
@@ -645,13 +670,24 @@ esc(EscState *e, Node *n, Node *up)
 		// current loop depth is an upper bound on actual loop depth
 		// of addressed value.
 		n->escloopdepth = e->loopdepth;
-		// for &x, use loop depth of x.
+		// for &x, use loop depth of x if known.
+		// it should always be known, but if not, be conservative
+		// and keep the current loop depth.
 		if(n->left->op == ONAME) {
 			switch(n->left->class) {
 			case PAUTO:
+				if(n->left->escloopdepth != 0)
+					n->escloopdepth = n->left->escloopdepth;
+				break;
 			case PPARAM:
 			case PPARAMOUT:
-				n->escloopdepth = n->left->escloopdepth;
+				// PPARAM is loop depth 1 always.
+				// PPARAMOUT is loop depth 0 for writes
+				// but considered loop depth 1 for address-of,
+				// so that writing the address of one result
+				// to another (or the same) result makes the
+				// first result move to the heap.
+				n->escloopdepth = 1;
 				break;
 			}
 		}
@@ -822,12 +858,17 @@ escassignfromtag(EscState *e, Strlit *note, NodeList *dsts, Node *src)
 		escassign(e, &e->theSink, src);
 		return em;
 	}
-	
+
 	if(em == EscNone)
 		return em;
+	
+	// If content inside parameter (reached via indirection)
+	// escapes back to results, mark as such.
+	if(em & EscContentEscapes)
+		escassign(e, &e->funcParam, src);
 
 	em0 = em;
-	for(em >>= EscBits; em && dsts; em >>= 1, dsts=dsts->next)
+	for(em >>= EscReturnBits; em && dsts; em >>= 1, dsts=dsts->next)
 		if(em & 1)
 			escassign(e, dsts->n, src);
 
@@ -1090,19 +1131,30 @@ escwalk(EscState *e, int level, Node *dst, Node *src)
 
 	// Input parameter flowing to output parameter?
 	if(dst->op == ONAME && dst->class == PPARAMOUT && dst->vargen <= 20) {
-		if(src->op == ONAME && src->class == PPARAM && level == 0 && src->curfn == dst->curfn) {
-			if(src->esc != EscScope && src->esc != EscHeap) {
+		if(src->op == ONAME && src->class == PPARAM && src->curfn == dst->curfn && src->esc != EscScope && src->esc != EscHeap) {
+			if(level == 0) {
 				if(debug['m'])
 					warnl(src->lineno, "leaking param: %hN to result %S", src, dst->sym);
 				if((src->esc&EscMask) != EscReturn)
 					src->esc = EscReturn;
-				src->esc |= 1<<((dst->vargen-1) + EscBits);
+				src->esc |= 1<<((dst->vargen-1) + EscReturnBits);
+				goto recurse;
+			} else if(level > 0) {
+				if(debug['m'])
+					warnl(src->lineno, "%N leaking param %hN content to result %S", src->curfn->nname, src, dst->sym);
+				if((src->esc&EscMask) != EscReturn)
+					src->esc = EscReturn;
+				src->esc |= EscContentEscapes;
 				goto recurse;
 			}
 		}
 	}
 
-	leaks = (level <= 0) && (dst->escloopdepth < src->escloopdepth);
+	// The second clause is for values pointed at by an object passed to a call
+	// that returns something reached via indirect from the object.
+	// We don't know which result it is or how many indirects, so we treat it as leaking.
+	leaks = level <= 0 && dst->escloopdepth < src->escloopdepth ||
+		level < 0 && dst == &e->funcParam && haspointers(src->type);
 
 	switch(src->op) {
 	case ONAME:

src/cmd/gc/go.h

@@ -236,8 +236,10 @@ enum
 	EscNone,
 	EscReturn,
 	EscNever,
-	EscBits = 4,
+	EscBits = 3,
 	EscMask = (1<<EscBits) - 1,
+	EscContentEscapes = 1<<EscBits, // value obtained by indirect of parameter escapes to some returned result
+	EscReturnBits = EscBits+1,
 };
 
 struct	Node
@@ -1015,7 +1017,6 @@ int32	arraylength(Array *array);
 void*	arrayget(Array *array, int32 index);
 void	arrayset(Array *array, int32 index, void *element);
 void	arrayadd(Array *array, void *element);
-int32	arrayindexof(Array* array, void *element);
 void	arraysort(Array* array, int (*cmp)(const void*, const void*));
 
 /*
@@ -1041,6 +1042,7 @@ int	bvcmp(Bvec *bv1, Bvec *bv2);
 void	bvcopy(Bvec *dst, Bvec *src);
 Bvec*	bvconcat(Bvec *src1, Bvec *src2);
 int	bvget(Bvec *bv, int32 i);
+int32	bvnext(Bvec *bv, int32 i);
 int	bvisempty(Bvec *bv);
 void	bvnot(Bvec *bv);
 void	bvor(Bvec *dst, Bvec *src1, Bvec *src2);

src/cmd/gc/inl.c

@@ -802,6 +802,7 @@ inlvar(Node *var)
 	n->class = PAUTO;
 	n->used = 1;
 	n->curfn = curfn;   // the calling function, not the called one
+	n->addrtaken = var->addrtaken;
 
 	// esc pass wont run if we're inlining into a iface wrapper
 	// luckily, we can steal the results from the target func

src/cmd/gc/plive.c

@@ -283,13 +283,30 @@ getvariables(Node *fn)
 			// For arguments and results, the bitmap covers all variables,
 			// so we must include all the variables, even the ones without
 			// pointers.
+			//
+			// The Node.opt field is available for use by optimization passes.
+			// We use it to hold the index of the node in the variables array, plus 1
+			// (so that 0 means the Node is not in the variables array).
+			// Each pass should clear opt when done, but you never know,
+			// so clear them all ourselves too.
+			// The Node.curfn field is supposed to be set to the current function
+			// already, but for some compiler-introduced names it seems not to be,
+			// so fix that here.
+			// Later, when we want to find the index of a node in the variables list,
+			// we will check that n->curfn == curfn and n->opt > 0. Then n->opt - 1
+			// is the index in the variables list.
+			ll->n->opt = nil;
+			ll->n->curfn = curfn;
 			switch(ll->n->class) {
 			case PAUTO:
-				if(haspointers(ll->n->type))
+				if(haspointers(ll->n->type)) {
+					ll->n->opt = (void*)(uintptr)(arraylength(result)+1);
 					arrayadd(result, &ll->n);
+				}
 				break;
 			case PPARAM:
 			case PPARAMOUT:
+				ll->n->opt = (void*)(uintptr)(arraylength(result)+1);
 				arrayadd(result, &ll->n);
 				break;
 			}
@@ -718,14 +735,16 @@ progeffects(Prog *prog, Array *vars, Bvec *uevar, Bvec *varkill, Bvec *avarinit)
 	}
 	if(info.flags & (LeftRead | LeftWrite | LeftAddr)) {
 		from = &prog->from;
-		if (from->node != nil && from->sym != nil) {
+		if (from->node != nil && from->sym != nil && from->node->curfn == curfn) {
 			switch(from->node->class & ~PHEAP) {
 			case PAUTO:
 			case PPARAM:
 			case PPARAMOUT:
-				pos = arrayindexof(vars, from->node);
+				pos = (int)(uintptr)from->node->opt - 1; // index in vars
 				if(pos == -1)
 					goto Next;
+				if(pos >= arraylength(vars) || *(Node**)arrayget(vars, pos) != from->node)
+					fatal("bad bookkeeping in liveness %N %d", from->node, pos);
 				if(from->node->addrtaken) {
 					bvset(avarinit, pos);
 				} else {
@@ -741,14 +760,16 @@ progeffects(Prog *prog, Array *vars, Bvec *uevar, Bvec *varkill, Bvec *avarinit)
 Next:
 	if(info.flags & (RightRead | RightWrite | RightAddr)) {
 		to = &prog->to;
-		if (to->node != nil && to->sym != nil) {
+		if (to->node != nil && to->sym != nil && to->node->curfn == curfn) {
 			switch(to->node->class & ~PHEAP) {
 			case PAUTO:
 			case PPARAM:
 			case PPARAMOUT:
-				pos = arrayindexof(vars, to->node);
+				pos = (int)(uintptr)to->node->opt - 1; // index in vars
 				if(pos == -1)
 					goto Next1;
+				if(pos >= arraylength(vars) || *(Node**)arrayget(vars, pos) != to->node)
+					fatal("bad bookkeeping in liveness %N %d", to->node, pos);
 				if(to->node->addrtaken) {
 					if(prog->as != AVARKILL)
 						bvset(avarinit, pos);
@@ -1020,6 +1041,9 @@ checkptxt(Node *fn, Prog *firstp)
 {
 	Prog *p;
 
+	if(debuglive == 0)
+		return;
+
 	for(p = firstp; p != P; p = p->link) {
 		if(0)
 			print("analyzing '%P'\n", p);
@@ -1172,21 +1196,17 @@ twobitlivepointermap(Liveness *lv, Bvec *liveout, Array *vars, Bvec *args, Bvec
 	vlong xoffset;
 	int32 i;
 
-	for(i = 0; i < arraylength(vars); i++) {
+	for(i = 0; (i = bvnext(liveout, i)) >= 0; i++) {
 		node = *(Node**)arrayget(vars, i);
 		switch(node->class) {
 		case PAUTO:
-			if(bvget(liveout, i)) {
-				xoffset = node->xoffset + stkptrsize;
-				twobitwalktype1(node->type, &xoffset, locals);
-			}
+			xoffset = node->xoffset + stkptrsize;
+			twobitwalktype1(node->type, &xoffset, locals);
 			break;
 		case PPARAM:
 		case PPARAMOUT:
-			if(bvget(liveout, i)) {
-				xoffset = node->xoffset;
-				twobitwalktype1(node->type, &xoffset, args);
-			}
+			xoffset = node->xoffset;
+			twobitwalktype1(node->type, &xoffset, args);
 			break;
 		}
 	}
@@ -1937,6 +1957,7 @@ liveness(Node *fn, Prog *firstp, Sym *argssym, Sym *livesym)
 	Array *cfg, *vars;
 	Liveness *lv;
 	int debugdelta;
+	NodeList *l;
 
 	// Change name to dump debugging information only for a specific function.
 	debugdelta = 0;
@@ -1977,6 +1998,9 @@ liveness(Node *fn, Prog *firstp, Sym *argssym, Sym *livesym)
 	twobitwritesymbol(lv->argslivepointers, argssym);
 
 	// Free everything.
+	for(l=fn->dcl; l != nil; l = l->next)
+		if(l->n != N)
+			l->n->opt = nil;
 	freeliveness(lv);
 	arrayfree(vars);
 	freecfg(cfg);

src/cmd/gc/subr.c

@@ -2493,6 +2493,7 @@ genwrapper(Type *rcvr, Type *method, Sym *newnam, int iface)
 	Type *tpad, *methodrcvr;
 	int isddd;
 	Val v;
+	static int linehistdone = 0;
 
 	if(0 && debug['r'])
 		print("genwrapper rcvrtype=%T method=%T newnam=%S\n",
@@ -2500,7 +2501,11 @@ genwrapper(Type *rcvr, Type *method, Sym *newnam, int iface)
 
 	lexlineno++;
 	lineno = lexlineno;
-	linehist("<autogenerated>", 0, 0);
+	if (linehistdone == 0) {
+		// All the wrappers can share the same linehist entry.
+		linehist("<autogenerated>", 0, 0);
+		linehistdone = 1;
+	}
 
 	dclcontext = PEXTERN;
 	markdcl();

src/cmd/ld/dwarf.c

@@ -1624,13 +1624,13 @@ writelines(void)
 			}
 			putpclcdelta(s->value + pcline.pc - pc, pcline.value - line);
 
-			pc = epc;
+			pc = s->value + pcline.pc;
+			line = pcline.value;
 			if(pcfile.nextpc < pcline.nextpc)
 				epc = pcfile.nextpc;
 			else
 				epc = pcline.nextpc;
 			epc += s->value;
-			line = pcline.value;
 		}
 
 		da = 0;

src/cmd/link/auto.go

@@ -1,117 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-// Automatic symbol generation.
-
-// TODO(rsc): Handle go.typelink, go.track symbols.
-// TODO(rsc): Do not handle $f64. and $f32. symbols. Instead, generate those
-// from the compiler and assemblers as dupok data, and then remove autoData below.
-package main
-
-import (
-	"debug/goobj"
-	"strconv"
-	"strings"
-)
-
-// linkerDefined lists the symbols supplied by other parts of the linker
-// (runtime.go and layout.go).
-var linkerDefined = map[string]bool{
-	"bss":        true,
-	"data":       true,
-	"ebss":       true,
-	"edata":      true,
-	"efunctab":   true,
-	"end":        true,
-	"enoptrbss":  true,
-	"enoptrdata": true,
-	"erodata":    true,
-	"etext":      true,
-	"etypelink":  true,
-	"functab":    true,
-	"gcbss":      true,
-	"gcdata":     true,
-	"noptrbss":   true,
-	"noptrdata":  true,
-	"pclntab":    true,
-	"rodata":     true,
-	"text":       true,
-	"typelink":   true,
-}
-
-// isAuto reports whether sym is an automatically-generated data or constant symbol.
-func (p *Prog) isAuto(sym goobj.SymID) bool {
-	return strings.HasPrefix(sym.Name, "go.weak.") ||
-		strings.HasPrefix(sym.Name, "$f64.") ||
-		strings.HasPrefix(sym.Name, "$f32.") ||
-		linkerDefined[sym.Name]
-}
-
-// autoData defines the automatically generated data symbols needed by p.
-func (p *Prog) autoData() {
-	for sym := range p.Missing {
-		switch {
-		// Floating-point constants that need to be loaded from memory are
-		// written as $f64.{16 hex digits} or $f32.{8 hex digits}; the hex digits
-		// give the IEEE bit pattern of the constant. As far as the layout into
-		// memory is concerned, we interpret these as uint64 or uint32 constants.
-		case strings.HasPrefix(sym.Name, "$f64."), strings.HasPrefix(sym.Name, "$f32."):
-			size := 64
-			if sym.Name[2:4] == "32" {
-				size = 32
-			}
-			delete(p.Missing, sym)
-			fbits, err := strconv.ParseUint(sym.Name[len("$f64."):], 16, size)
-			if err != nil {
-				p.errorf("unexpected floating point symbol %s", sym)
-				continue
-			}
-			data := make([]byte, size/8)
-			if size == 64 {
-				p.byteorder.PutUint64(data, fbits)
-			} else {
-				p.byteorder.PutUint32(data, uint32(fbits))
-			}
-			p.addSym(&Sym{
-				Sym: &goobj.Sym{
-					SymID: sym,
-					Kind:  goobj.SRODATA,
-					Size:  size / 8,
-				},
-				Bytes: data,
-			})
-		}
-	}
-}
-
-// autoConst defines the automatically generated constant symbols needed by p.
-func (p *Prog) autoConst() {
-	for sym := range p.Missing {
-		switch {
-		case strings.HasPrefix(sym.Name, "go.weak."):
-			// weak symbol resolves to actual symbol if present, or else nil.
-			delete(p.Missing, sym)
-			targ := sym
-			targ.Name = sym.Name[len("go.weak."):]
-			var addr Addr
-			if s := p.Syms[targ]; s != nil {
-				addr = s.Addr
-			}
-			p.defineConst(sym.Name, addr)
-		}
-	}
-}
-
-// defineConst defines a new symbol with the given name and constant address.
-func (p *Prog) defineConst(name string, addr Addr) {
-	sym := goobj.SymID{Name: name}
-	p.addSym(&Sym{
-		Sym: &goobj.Sym{
-			SymID: sym,
-			Kind:  goobj.SCONST,
-		},
-		Package: nil,
-		Addr:    addr,
-	})
-}

src/cmd/link/auto_test.go

@@ -1,72 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-// Test for auto-generated symbols.
-
-// There is no test for $f64. and $f32. symbols, because those are
-// not possible to write in the assembler syntax. Instead of changing
-// the assembler to allow that, we plan to change the compilers
-// not to generate such symbols (plain dupok data is sufficient).
-
-package main
-
-import (
-	"bytes"
-	"debug/goobj"
-	"testing"
-)
-
-// Each test case is an object file, generated from a corresponding .s file.
-// The image of the autotab symbol should be a sequence of pairs of
-// identical 8-byte sequences.
-var autoTests = []string{
-	"testdata/autosection.6",
-	"testdata/autoweak.6",
-}
-
-func TestAuto(t *testing.T) {
-	for _, obj := range autoTests {
-		p := Prog{GOOS: "darwin", GOARCH: "amd64", StartSym: "start"}
-		p.omitRuntime = true
-		p.Error = func(s string) { t.Error(s) }
-		var buf bytes.Buffer
-		p.link(&buf, obj)
-		if p.NumError > 0 {
-			continue // already reported
-		}
-
-		const name = "autotab"
-		sym := p.Syms[goobj.SymID{Name: name}]
-		if sym == nil {
-			t.Errorf("%s is missing %s symbol", obj, name)
-			return
-		}
-		if sym.Size == 0 {
-			return
-		}
-
-		seg := sym.Section.Segment
-		off := sym.Addr - seg.VirtAddr
-		data := seg.Data[off : off+Addr(sym.Size)]
-		if len(data)%16 != 0 {
-			t.Errorf("%s: %s.Size = %d, want multiple of 16", obj, name, len(data))
-			return
-		}
-	Data:
-		for i := 0; i < len(data); i += 16 {
-			have := p.byteorder.Uint64(data[i : i+8])
-			want := p.byteorder.Uint64(data[i+8 : i+16])
-			if have != want {
-				// Look for relocation so we can explain what went wrong.
-				for _, r := range sym.Reloc {
-					if r.Offset == i {
-						t.Errorf("%s: %s+%#x: %s: have %#x want %#x", obj, name, i, r.Sym, have, want)
-						continue Data
-					}
-				}
-				t.Errorf("%s: %s+%#x: have %#x want %#x", obj, name, i, have, want)
-			}
-		}
-	}
-}

src/cmd/link/dead.go

@@ -1,74 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-// Removal of dead code and data.
-
-package main
-
-import "debug/goobj"
-
-// dead removes unreachable code and data from the program.
-// It is basically a mark-sweep garbage collection: traverse all the
-// symbols reachable from the entry (startSymID) and then delete
-// the rest.
-func (p *Prog) dead() {
-	p.Dead = make(map[goobj.SymID]bool)
-	reachable := make(map[goobj.SymID]bool)
-	p.walkDead(p.startSym, reachable)
-
-	for sym := range p.Syms {
-		if !reachable[sym] {
-			delete(p.Syms, sym)
-			p.Dead[sym] = true
-		}
-	}
-
-	for sym := range p.Missing {
-		if !reachable[sym] {
-			delete(p.Missing, sym)
-			p.Dead[sym] = true
-		}
-	}
-
-	p.SymOrder = removeDead(p.SymOrder, reachable)
-
-	for _, pkg := range p.Packages {
-		pkg.Syms = removeDead(pkg.Syms, reachable)
-	}
-}
-
-// walkDead traverses the symbols reachable from sym, adding them to reachable.
-// The caller has verified that reachable[sym] = false.
-func (p *Prog) walkDead(sym goobj.SymID, reachable map[goobj.SymID]bool) {
-	reachable[sym] = true
-	s := p.Syms[sym]
-	if s == nil {
-		return
-	}
-	for i := range s.Reloc {
-		r := &s.Reloc[i]
-		if !reachable[r.Sym] {
-			p.walkDead(r.Sym, reachable)
-		}
-	}
-	if s.Func != nil {
-		for _, fdata := range s.Func.FuncData {
-			if fdata.Sym.Name != "" && !reachable[fdata.Sym] {
-				p.walkDead(fdata.Sym, reachable)
-			}
-		}
-	}
-}
-
-// removeDead removes unreachable (dead) symbols from syms,
-// returning a shortened slice using the same underlying array.
-func removeDead(syms []*Sym, reachable map[goobj.SymID]bool) []*Sym {
-	keep := syms[:0]
-	for _, sym := range syms {
-		if reachable[sym.SymID] {
-			keep = append(keep, sym)
-		}
-	}
-	return keep
-}

src/cmd/link/dead_test.go

@@ -1,97 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-package main
-
-import (
-	"debug/goobj"
-	"reflect"
-	"strings"
-	"testing"
-)
-
-// Each test case is an object file, generated from a corresponding .s file.
-// The symbols in the object file with a dead_ prefix are the ones that
-// should be removed from the program.
-var deadTests = []string{
-	"testdata/dead.6",
-}
-
-func TestDead(t *testing.T) {
-	for _, obj := range deadTests {
-		p := Prog{GOOS: "darwin", GOARCH: "amd64", StartSym: "start"}
-		p.omitRuntime = true
-		p.Error = func(s string) { t.Error(s) }
-		p.init()
-		p.scan(obj)
-		if p.NumError > 0 {
-			continue // already reported
-		}
-		origSyms := copyMap(p.Syms)
-		origMissing := copyMap(p.Missing)
-		origSymOrder := copySlice(p.SymOrder)
-		origPkgSyms := copySlice(p.Packages["main"].Syms)
-		p.dead()
-		checkDeadMap(t, obj, "p.Syms", origSyms, p.Syms)
-		checkDeadMap(t, obj, "p.Missing", origMissing, p.Missing)
-		checkDeadSlice(t, obj, "p.SymOrder", origSymOrder, p.SymOrder)
-		checkDeadSlice(t, obj, `p.Packages["main"].Syms`, origPkgSyms, p.Packages["main"].Syms)
-	}
-}
-
-func copyMap(m interface{}) interface{} {
-	v := reflect.ValueOf(m)
-	out := reflect.MakeMap(v.Type())
-	for _, key := range v.MapKeys() {
-		out.SetMapIndex(key, v.MapIndex(key))
-	}
-	return out.Interface()
-}
-
-func checkDeadMap(t *testing.T, obj, name string, old, new interface{}) {
-	vold := reflect.ValueOf(old)
-	vnew := reflect.ValueOf(new)
-	for _, vid := range vold.MapKeys() {
-		id := vid.Interface().(goobj.SymID)
-		if strings.HasPrefix(id.Name, "dead_") {
-			if vnew.MapIndex(vid).IsValid() {
-				t.Errorf("%s: %s contains unnecessary symbol %s", obj, name, id)
-			}
-		} else {
-			if !vnew.MapIndex(vid).IsValid() {
-				t.Errorf("%s: %s is missing symbol %s", obj, name, id)
-			}
-		}
-	}
-	for _, vid := range vnew.MapKeys() {
-		id := vid.Interface().(goobj.SymID)
-		if !vold.MapIndex(vid).IsValid() {
-			t.Errorf("%s: %s contains unexpected symbol %s", obj, name, id)
-		}
-	}
-}
-
-func copySlice(x []*Sym) (out []*Sym) {
-	return append(out, x...)
-}
-
-func checkDeadSlice(t *testing.T, obj, name string, old, new []*Sym) {
-	for i, s := range old {
-		if strings.HasPrefix(s.Name, "dead_") {
-			continue
-		}
-		if len(new) == 0 {
-			t.Errorf("%s: %s is missing symbol %s\nhave%v\nwant%v", obj, name, s, new, old[i:])
-			return
-		}
-		if new[0].SymID != s.SymID {
-			t.Errorf("%s: %s is incorrect: have %s, want %s\nhave%v\nwant%v", obj, name, new[0].SymID, s.SymID, new, old[i:])
-			return
-		}
-		new = new[1:]
-	}
-	if len(new) > 0 {
-		t.Errorf("%s: %s has unexpected symbols: %v", obj, name, new)
-	}
-}

src/cmd/link/debug.go

@@ -1,11 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-// Generation of debug data structures (in the executable but not mapped at run time).
-// See also runtime.go.
-
-package main
-
-func (p *Prog) debug() {
-}

src/cmd/link/hex_test.go

@@ -1,74 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-package main
-
-import (
-	"encoding/hex"
-	"fmt"
-	"io/ioutil"
-	"regexp"
-	"strconv"
-	"strings"
-	"testing"
-)
-
-// mustParseHexdumpFile returns a block of data generated by
-// parsing the hex dump in the named file.
-// If the file cannot be read or does not contain a valid hex dump,
-// mustParseHexdumpFile calls t.Fatal.
-func mustParseHexdumpFile(t *testing.T, file string) []byte {
-	hex, err := ioutil.ReadFile(file)
-	if err != nil {
-		t.Fatal(err)
-	}
-	data, err := parseHexdump(string(hex))
-	if err != nil {
-		t.Fatal(err)
-	}
-	return data
-}
-
-// parseHexdump parses the hex dump in text, which should be the
-// output of "hexdump -C" or Plan 9's "xd -b",
-// and returns the original data used to produce the dump.
-// It is meant to enable storing golden binary files as text, so that
-// changes to the golden files can be seen during code reviews.
-func parseHexdump(text string) ([]byte, error) {
-	var out []byte
-	for _, line := range strings.Split(text, "\n") {
-		if i := strings.Index(line, "|"); i >= 0 { // remove text dump
-			line = line[:i]
-		}
-		f := strings.Fields(line)
-		if len(f) > 1+16 {
-			return nil, fmt.Errorf("parsing hex dump: too many fields on line %q", line)
-		}
-		if len(f) == 0 || len(f) == 1 && f[0] == "*" { // all zeros block omitted
-			continue
-		}
-		addr64, err := strconv.ParseUint(f[0], 16, 0)
-		if err != nil {
-			return nil, fmt.Errorf("parsing hex dump: invalid address %q", f[0])
-		}
-		addr := int(addr64)
-		if len(out) < addr {
-			out = append(out, make([]byte, addr-len(out))...)
-		}
-		for _, x := range f[1:] {
-			val, err := strconv.ParseUint(x, 16, 8)
-			if err != nil {
-				return nil, fmt.Errorf("parsing hexdump: invalid hex byte %q", x)
-			}
-			out = append(out, byte(val))
-		}
-	}
-	return out, nil
-}
-
-func hexdump(data []byte) string {
-	text := hex.Dump(data) + fmt.Sprintf("%08x\n", len(data))
-	text = regexp.MustCompile(`\n([0-9a-f]+(\s+00){16}.*\n)+`).ReplaceAllString(text, "\n*\n")
-	return text
-}

src/cmd/link/layout.go

@@ -1,180 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-// Executable image layout - address assignment.
-
-package main
-
-import (
-	"debug/goobj"
-)
-
-// A layoutSection describes a single section to add to the
-// final executable. Go binaries only have a fixed set of possible
-// sections, and the symbol kind determines the section.
-type layoutSection struct {
-	Segment string
-	Section string
-	Kind    goobj.SymKind
-	Index   int
-}
-
-// layout defines the layout of the generated Go executable.
-// The order of entries here is the order in the executable.
-// Entries with the same Segment name must be contiguous.
-var layout = []layoutSection{
-	{Segment: "text", Section: "text", Kind: goobj.STEXT},
-	{Segment: "rodata", Section: "rodata", Kind: goobj.SRODATA},
-	{Segment: "rodata", Section: "functab", Kind: goobj.SPCLNTAB},
-	{Segment: "rodata", Section: "typelink", Kind: goobj.STYPELINK},
-	{Segment: "data", Section: "noptrdata", Kind: goobj.SNOPTRDATA},
-	{Segment: "data", Section: "data", Kind: goobj.SDATA},
-	{Segment: "data", Section: "bss", Kind: goobj.SBSS},
-	{Segment: "data", Section: "noptrbss", Kind: goobj.SNOPTRBSS},
-
-	// Later:
-	//	{"rodata", "type", goobj.STYPE},
-	//	{"rodata", "string", goobj.SSTRING},
-	//	{"rodata", "gostring", goobj.SGOSTRING},
-	//	{"rodata", "gofunc", goobj.SGOFUNC},
-}
-
-// layoutByKind maps from SymKind to an entry in layout.
-var layoutByKind []*layoutSection
-
-func init() {
-	// Build index from symbol type to layout entry.
-	max := 0
-	for _, sect := range layout {
-		if max <= int(sect.Kind) {
-			max = int(sect.Kind) + 1
-		}
-	}
-	layoutByKind = make([]*layoutSection, max)
-	for i := range layout {
-		sect := &layout[i]
-		layoutByKind[sect.Kind] = sect
-		sect.Index = i
-	}
-}
-
-// layout arranges symbols into sections and sections into segments,
-// and then it assigns addresses to segments, sections, and symbols.
-func (p *Prog) layout() {
-	sections := make([]*Section, len(layout))
-
-	// Assign symbols to sections using index, creating sections as needed.
-	// Could keep sections separated by type during input instead.
-	for _, sym := range p.SymOrder {
-		kind := sym.Kind
-		if kind < 0 || int(kind) >= len(layoutByKind) || layoutByKind[kind] == nil {
-			p.errorf("%s: unexpected symbol kind %v", sym.SymID, kind)
-			continue
-		}
-		lsect := layoutByKind[kind]
-		sect := sections[lsect.Index]
-		if sect == nil {
-			sect = &Section{
-				Name:  lsect.Section,
-				Align: 1,
-			}
-			sections[lsect.Index] = sect
-		}
-		if sym.Data.Size > 0 || len(sym.Bytes) > 0 {
-			sect.InFile = true
-		}
-		sym.Section = sect
-		sect.Syms = append(sect.Syms, sym)
-
-		// TODO(rsc): Incorporate alignment information.
-		// First that information needs to be added to the object files.
-		//
-		// if sect.Align < Addr(sym.Align) {
-		//	sect.Align = Addr(sym.Align)
-		// }
-	}
-
-	// Assign sections to segments, creating segments as needed.
-	var seg *Segment
-	for i, sect := range sections {
-		if sect == nil {
-			continue
-		}
-		segName := layout[i].Segment
-
-		// Special case: Mach-O does not support "rodata" segment,
-		// so store read-only data in text segment.
-		if p.GOOS == "darwin" && segName == "rodata" {
-			segName = "text"
-		}
-
-		if seg == nil || seg.Name != segName {
-			seg = &Segment{
-				Name: segName,
-			}
-			p.Segments = append(p.Segments, seg)
-		}
-		sect.Segment = seg
-		seg.Sections = append(seg.Sections, sect)
-	}
-
-	// Assign addresses.
-
-	// TODO(rsc): This choice needs to be informed by both
-	// the formatter and the target architecture.
-	// And maybe eventually a command line flag (sigh).
-	const segAlign = 4096
-
-	// TODO(rsc): Use a larger amount on most systems, which will let the
-	// compiler eliminate more nil checks.
-	if p.UnmappedSize == 0 {
-		p.UnmappedSize = segAlign
-	}
-
-	// TODO(rsc): addr := Addr(0) when generating a shared library or PIE.
-	addr := p.UnmappedSize
-
-	// Account for initial file header.
-	hdrVirt, hdrFile := p.formatter.headerSize(p)
-	addr += hdrVirt
-
-	// Assign addresses to segments, sections, symbols.
-	// Assign sizes to segments, sections.
-	startVirt := addr
-	startFile := hdrFile
-	for _, seg := range p.Segments {
-		addr = round(addr, segAlign)
-		seg.VirtAddr = addr
-		seg.FileOffset = startFile + seg.VirtAddr - startVirt
-		for _, sect := range seg.Sections {
-			addr = round(addr, sect.Align)
-			sect.VirtAddr = addr
-			for _, sym := range sect.Syms {
-				// TODO(rsc): Respect alignment once we have that information.
-				sym.Addr = addr
-				addr += Addr(sym.Size)
-			}
-			sect.Size = addr - sect.VirtAddr
-			if sect.InFile {
-				seg.FileSize = addr - seg.VirtAddr
-			}
-		}
-		seg.VirtSize = addr - seg.VirtAddr
-	}
-
-	// Define symbols for section names.
-	var progEnd Addr
-	for i, sect := range sections {
-		name := layout[i].Section
-		var start, end Addr
-		if sect != nil {
-			start = sect.VirtAddr
-			end = sect.VirtAddr + sect.Size
-		}
-		p.defineConst(name, start)
-		p.defineConst("e"+name, end)
-		progEnd = end
-	}
-	p.defineConst("end", progEnd)
-}

src/cmd/link/layout_test.go

@@ -1,45 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-package main
-
-import (
-	"bytes"
-	"strings"
-	"testing"
-)
-
-func TestLayout(t *testing.T) {
-	p := Prog{GOOS: "darwin", GOARCH: "amd64", StartSym: "text_start"}
-	p.omitRuntime = true
-	p.Error = func(s string) { t.Error(s) }
-	var buf bytes.Buffer
-	const obj = "testdata/layout.6"
-	p.link(&buf, obj)
-	if p.NumError > 0 {
-		return // already reported
-	}
-	if len(p.Dead) > 0 {
-		t.Errorf("%s: unexpected dead symbols %v", obj, p.Dead)
-		return
-	}
-
-	for _, sym := range p.SymOrder {
-		if p.isAuto(sym.SymID) {
-			continue
-		}
-		if sym.Section == nil {
-			t.Errorf("%s: symbol %s is missing section", obj, sym)
-			continue
-		}
-		i := strings.Index(sym.Name, "_")
-		if i < 0 {
-			t.Errorf("%s: unexpected symbol %s", obj, sym)
-			continue
-		}
-		if sym.Section.Name != sym.Name[:i] {
-			t.Errorf("%s: symbol %s in section %s, want %s", obj, sym, sym.Section.Name, sym.Name[:i])
-		}
-	}
-}

src/cmd/link/link_test.go

@@ -1,35 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-package main
-
-import (
-	"bytes"
-	"debug/goobj"
-	"io/ioutil"
-	"testing"
-)
-
-func TestLinkHello(t *testing.T) {
-	p := &Prog{
-		GOOS:     "darwin",
-		GOARCH:   "amd64",
-		Error:    func(s string) { t.Error(s) },
-		StartSym: "_rt0_go",
-	}
-	var buf bytes.Buffer
-	p.link(&buf, "testdata/hello.6")
-	if p.NumError > 0 {
-		return
-	}
-	if p.Syms[goobj.SymID{"_rt0_go", 0}] == nil || p.Syms[goobj.SymID{"hello", 1}] == nil {
-		t.Errorf("Syms = %v, want at least [_rt0_go hello<1>]", p.Syms)
-	}
-
-	// uncomment to leave file behind for execution:
-	if false {
-		ioutil.WriteFile("a.out", buf.Bytes(), 0777)
-	}
-	checkGolden(t, buf.Bytes(), "testdata/link.hello.darwin.amd64")
-}

src/cmd/link/load.go

@@ -1,116 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-// Loading of code and data fragments from package files into final image.
-
-package main
-
-import "os"
-
-// load allocates segment images, populates them with data
-// read from package files, and applies relocations to the data.
-func (p *Prog) load() {
-	// TODO(rsc): mmap the output file and store the data directly.
-	// That will make writing the output file more efficient.
-	for _, seg := range p.Segments {
-		seg.Data = make([]byte, seg.FileSize)
-	}
-	for _, pkg := range p.Packages {
-		p.loadPackage(pkg)
-	}
-}
-
-// loadPackage loads and relocates data for all the
-// symbols needed in the given package.
-func (p *Prog) loadPackage(pkg *Package) {
-	if pkg.File == "" {
-		// This "package" contains internally generated symbols only.
-		// All such symbols have a sym.Bytes field holding the actual data
-		// (if any), plus relocations.
-		for _, sym := range pkg.Syms {
-			if sym.Bytes == nil {
-				continue
-			}
-			seg := sym.Section.Segment
-			off := sym.Addr - seg.VirtAddr
-			data := seg.Data[off : off+Addr(sym.Size)]
-			copy(data, sym.Bytes)
-			p.relocateSym(sym, data)
-		}
-		return
-	}
-
-	// Package stored in file.
-	f, err := os.Open(pkg.File)
-	if err != nil {
-		p.errorf("%v", err)
-		return
-	}
-	defer f.Close()
-
-	// TODO(rsc): Mmap file into memory.
-
-	for _, sym := range pkg.Syms {
-		if sym.Data.Size == 0 {
-			continue
-		}
-		// TODO(rsc): If not using mmap, at least coalesce nearby reads.
-		if sym.Section == nil {
-			p.errorf("internal error: missing section for %s", sym.Name)
-		}
-		seg := sym.Section.Segment
-		off := sym.Addr - seg.VirtAddr
-		if off >= Addr(len(seg.Data)) || off+Addr(sym.Data.Size) > Addr(len(seg.Data)) {
-			p.errorf("internal error: allocated space for %s too small: %d bytes for %d+%d (%d)", sym, len(seg.Data), off, sym.Data.Size, sym.Size)
-		}
-		data := seg.Data[off : off+Addr(sym.Data.Size)]
-		_, err := f.ReadAt(data, sym.Data.Offset)
-		if err != nil {
-			p.errorf("reading %v: %v", sym.SymID, err)
-		}
-		p.relocateSym(sym, data)
-	}
-}
-
-// TODO(rsc): Define full enumeration for relocation types.
-const (
-	R_ADDR    = 1
-	R_SIZE    = 2
-	R_CALL    = 3
-	R_CALLARM = 4
-	R_CALLIND = 5
-	R_CONST   = 6
-	R_PCREL   = 7
-)
-
-// relocateSym applies relocations to sym's data.
-func (p *Prog) relocateSym(sym *Sym, data []byte) {
-	for i := range sym.Reloc {
-		r := &sym.Reloc[i]
-		targ := p.Syms[r.Sym]
-		if targ == nil {
-			p.errorf("%v: reference to undefined symbol %v", sym, r.Sym)
-			continue
-		}
-		val := targ.Addr + Addr(r.Add)
-		switch r.Type {
-		default:
-			p.errorf("%v: unknown relocation type %d", sym, r.Type)
-		case R_ADDR, R_CALLIND:
-			// ok
-		case R_PCREL, R_CALL:
-			val -= sym.Addr + Addr(r.Offset+r.Size)
-		}
-		frag := data[r.Offset : r.Offset+r.Size]
-		switch r.Size {
-		default:
-			p.errorf("%v: unknown relocation size %d", sym, r.Size)
-		case 4:
-			// TODO(rsc): Check for overflow?
-			p.byteorder.PutUint32(frag, uint32(val))
-		case 8:
-			p.byteorder.PutUint64(frag, uint64(val))
-		}
-	}
-}

src/cmd/link/macho.go

@@ -1,380 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-// Mach-O (Darwin) object file writing.
-
-package main
-
-import (
-	"debug/macho"
-	"encoding/binary"
-	"io"
-	"strings"
-)
-
-// machoFormat is the implementation of formatter.
-type machoFormat struct{}
-
-// machoHeader and friends are data structures
-// corresponding to the Mach-O file header
-// to be written to disk.
-
-const (
-	macho64Bit     = 1 << 24
-	machoSubCPU386 = 3
-)
-
-// machoArch describes a Mach-O target architecture.
-type machoArch struct {
-	CPU    uint32
-	SubCPU uint32
-}
-
-// machoHeader is the Mach-O file header.
-type machoHeader struct {
-	machoArch
-	FileType uint32
-	Loads    []*machoLoad
-	Segments []*machoSegment
-	p        *Prog // for reporting errors
-}
-
-// machoLoad is a Mach-O load command.
-type machoLoad struct {
-	Type uint32
-	Data []uint32
-}
-
-// machoSegment is a Mach-O segment.
-type machoSegment struct {
-	Name       string
-	VirtAddr   Addr
-	VirtSize   Addr
-	FileOffset Addr
-	FileSize   Addr
-	Prot1      uint32
-	Prot2      uint32
-	Flags      uint32
-	Sections   []*machoSection
-}
-
-// machoSection is a Mach-O section, inside a segment.
-type machoSection struct {
-	Name    string
-	Segment string
-	Addr    Addr
-	Size    Addr
-	Offset  uint32
-	Align   uint32
-	Reloc   uint32
-	Nreloc  uint32
-	Flags   uint32
-	Res1    uint32
-	Res2    uint32
-}
-
-// layout positions the segments and sections in p
-// to make room for the Mach-O file header.
-// That is, it edits their VirtAddr fields to adjust for the presence
-// of the Mach-O header at the beginning of the address space.
-func (machoFormat) headerSize(p *Prog) (virt, file Addr) {
-	var h machoHeader
-	h.init(p)
-	size := Addr(h.size())
-	size = round(size, 4096)
-	p.HeaderSize = size
-	return size, size
-}
-
-// write writes p to w as a Mach-O executable.
-// layout(p) must have already been called,
-// and the number, sizes, and addresses of the segments
-// and sections must not have been modified since the call.
-func (machoFormat) write(w io.Writer, p *Prog) {
-	var h machoHeader
-	h.init(p)
-	off := Addr(0)
-	enc := h.encode()
-	w.Write(enc)
-	off += Addr(len(enc))
-	for _, seg := range p.Segments {
-		if seg.FileOffset < off {
-			h.p.errorf("mach-o error: invalid file offset")
-		}
-		w.Write(make([]byte, int(seg.FileOffset-off)))
-		if seg.FileSize != Addr(len(seg.Data)) {
-			h.p.errorf("mach-o error: invalid file size")
-		}
-		w.Write(seg.Data)
-		off = seg.FileOffset + Addr(len(seg.Data))
-	}
-}
-
-// Conversion of Prog to macho data structures.
-
-// machoArches maps from GOARCH to machoArch.
-var machoArches = map[string]machoArch{
-	"amd64": {
-		CPU:    uint32(macho.CpuAmd64),
-		SubCPU: uint32(machoSubCPU386),
-	},
-}
-
-// init initializes the header h to describe p.
-func (h *machoHeader) init(p *Prog) {
-	h.p = p
-	h.Segments = nil
-	h.Loads = nil
-	var ok bool
-	h.machoArch, ok = machoArches[p.GOARCH]
-	if !ok {
-		p.errorf("mach-o: unknown target GOARCH %q", p.GOARCH)
-		return
-	}
-	h.FileType = uint32(macho.TypeExec)
-
-	mseg := h.addSegment(p, "__PAGEZERO", nil)
-	mseg.VirtSize = p.UnmappedSize
-
-	for _, seg := range p.Segments {
-		h.addSegment(p, "__"+strings.ToUpper(seg.Name), seg)
-	}
-
-	var data []uint32
-	switch h.CPU {
-	default:
-		p.errorf("mach-o: unknown cpu %#x for GOARCH %q", h.CPU, p.GOARCH)
-	case uint32(macho.CpuAmd64):
-		data = make([]uint32, 2+42)
-		data[0] = 4                  // thread type
-		data[1] = 42                 // word count
-		data[2+32] = uint32(p.Entry) // RIP register, in two parts
-		data[2+32+1] = uint32(p.Entry >> 32)
-	}
-
-	h.Loads = append(h.Loads, &machoLoad{
-		Type: uint32(macho.LoadCmdUnixThread),
-		Data: data,
-	})
-}
-
-// addSegment adds to h a Mach-O segment like seg with the given name.
-func (h *machoHeader) addSegment(p *Prog, name string, seg *Segment) *machoSegment {
-	mseg := &machoSegment{
-		Name: name,
-	}
-	h.Segments = append(h.Segments, mseg)
-	if seg == nil {
-		return mseg
-	}
-
-	mseg.VirtAddr = seg.VirtAddr
-	mseg.VirtSize = seg.VirtSize
-	mseg.FileOffset = round(seg.FileOffset, 4096)
-	mseg.FileSize = seg.FileSize
-
-	if name == "__TEXT" {
-		// Initially RWX, then just RX
-		mseg.Prot1 = 7
-		mseg.Prot2 = 5
-
-		// Text segment maps Mach-O header, needed by dynamic linker.
-		mseg.VirtAddr -= p.HeaderSize
-		mseg.VirtSize += p.HeaderSize
-		mseg.FileOffset -= p.HeaderSize
-		mseg.FileSize += p.HeaderSize
-	} else {
-		// RW
-		mseg.Prot1 = 3
-		mseg.Prot2 = 3
-	}
-
-	for _, sect := range seg.Sections {
-		h.addSection(mseg, seg, sect)
-	}
-	return mseg
-}
-
-// addSection adds to mseg a Mach-O section like sect, inside seg, with the given name.
-func (h *machoHeader) addSection(mseg *machoSegment, seg *Segment, sect *Section) {
-	msect := &machoSection{
-		Name:    "__" + sect.Name,
-		Segment: mseg.Name,
-		// Reloc: sect.RelocOffset,
-		// NumReloc: sect.RelocLen / 8,
-		Addr: sect.VirtAddr,
-		Size: sect.Size,
-	}
-	mseg.Sections = append(mseg.Sections, msect)
-
-	for 1<<msect.Align < sect.Align {
-		msect.Align++
-	}
-
-	if off := sect.VirtAddr - seg.VirtAddr; off < seg.FileSize {
-		// Data in file.
-		if sect.Size > seg.FileSize-off {
-			h.p.errorf("mach-o error: section crosses file boundary")
-		}
-		msect.Offset = uint32(seg.FileOffset + off)
-	} else {
-		// Zero filled.
-		msect.Flags |= 1
-	}
-
-	if sect.Name == "text" {
-		msect.Flags |= 0x400 // contains executable instructions
-	}
-}
-
-// A machoWriter helps write Mach-O headers.
-// It is basically a buffer with some helper routines for writing integers.
-type machoWriter struct {
-	dst   []byte
-	tmp   [8]byte
-	order binary.ByteOrder
-	is64  bool
-	p     *Prog
-}
-
-// if64 returns x if w is writing a 64-bit object file; otherwise it returns y.
-func (w *machoWriter) if64(x, y interface{}) interface{} {
-	if w.is64 {
-		return x
-	}
-	return y
-}
-
-// encode encodes each of the given arguments into the writer.
-// It encodes uint32, []uint32, uint64, and []uint64 by writing each value
-// in turn in the correct byte order for the output file.
-// It encodes an Addr as a uint64 if writing a 64-bit output file, or else as a uint32.
-// It encodes []byte and string by writing the raw bytes (no length prefix).
-// It skips nil values in the args list.
-func (w *machoWriter) encode(args ...interface{}) {
-	for _, arg := range args {
-		switch arg := arg.(type) {
-		default:
-			w.p.errorf("mach-o error: cannot encode %T", arg)
-		case nil:
-			// skip
-		case []byte:
-			w.dst = append(w.dst, arg...)
-		case string:
-			w.dst = append(w.dst, arg...)
-		case uint32:
-			w.order.PutUint32(w.tmp[:], arg)
-			w.dst = append(w.dst, w.tmp[:4]...)
-		case []uint32:
-			for _, x := range arg {
-				w.order.PutUint32(w.tmp[:], x)
-				w.dst = append(w.dst, w.tmp[:4]...)
-			}
-		case uint64:
-			w.order.PutUint64(w.tmp[:], arg)
-			w.dst = append(w.dst, w.tmp[:8]...)
-		case Addr:
-			if w.is64 {
-				w.order.PutUint64(w.tmp[:], uint64(arg))
-				w.dst = append(w.dst, w.tmp[:8]...)
-			} else {
-				if Addr(uint32(arg)) != arg {
-					w.p.errorf("mach-o error: truncating address %#x to uint32", arg)
-				}
-				w.order.PutUint32(w.tmp[:], uint32(arg))
-				w.dst = append(w.dst, w.tmp[:4]...)
-			}
-		}
-	}
-}
-
-// segmentSize returns the size of the encoding of seg in bytes.
-func (w *machoWriter) segmentSize(seg *machoSegment) int {
-	if w.is64 {
-		return 18*4 + 20*4*len(seg.Sections)
-	}
-	return 14*4 + 22*4*len(seg.Sections)
-}
-
-// zeroPad returns the string s truncated or padded with NULs to n bytes.
-func zeroPad(s string, n int) string {
-	if len(s) >= n {
-		return s[:n]
-	}
-	return s + strings.Repeat("\x00", n-len(s))
-}
-
-// size returns the encoded size of the header.
-func (h *machoHeader) size() int {
-	// Could write separate code, but encoding is cheap; encode and throw it away.
-	return len(h.encode())
-}
-
-// encode returns the Mach-O encoding of the header.
-func (h *machoHeader) encode() []byte {
-	w := &machoWriter{p: h.p}
-	w.is64 = h.CPU&macho64Bit != 0
-	w.order = w.p.byteorder
-
-	loadSize := 0
-	for _, seg := range h.Segments {
-		loadSize += w.segmentSize(seg)
-	}
-	for _, l := range h.Loads {
-		loadSize += 4 * (2 + len(l.Data))
-	}
-
-	w.encode(
-		w.if64(macho.Magic64, macho.Magic32),
-		uint32(h.CPU),
-		uint32(h.SubCPU),
-		uint32(h.FileType),
-		uint32(len(h.Loads)+len(h.Segments)),
-		uint32(loadSize),
-		uint32(1),
-		w.if64(uint32(0), nil),
-	)
-
-	for _, seg := range h.Segments {
-		w.encode(
-			w.if64(uint32(macho.LoadCmdSegment64), uint32(macho.LoadCmdSegment)),
-			uint32(w.segmentSize(seg)),
-			zeroPad(seg.Name, 16),
-			seg.VirtAddr,
-			seg.VirtSize,
-			seg.FileOffset,
-			seg.FileSize,
-			seg.Prot1,
-			seg.Prot2,
-			uint32(len(seg.Sections)),
-			seg.Flags,
-		)
-		for _, sect := range seg.Sections {
-			w.encode(
-				zeroPad(sect.Name, 16),
-				zeroPad(seg.Name, 16),
-				sect.Addr,
-				sect.Size,
-				sect.Offset,
-				sect.Align,
-				sect.Reloc,
-				sect.Nreloc,
-				sect.Flags,
-				sect.Res1,
-				sect.Res2,
-				w.if64(uint32(0), nil),
-			)
-		}
-	}
-
-	for _, load := range h.Loads {
-		w.encode(
-			load.Type,
-			uint32(4*(2+len(load.Data))),
-			load.Data,
-		)
-	}
-
-	return w.dst
-}

src/cmd/link/macho_test.go

@@ -1,407 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-package main
-
-import (
-	"bytes"
-	"debug/macho"
-	"encoding/binary"
-	"fmt"
-	"io/ioutil"
-	"strings"
-	"testing"
-)
-
-// Test macho writing by checking that each generated prog can be written
-// and then read back using debug/macho to get the same prog.
-// Also check against golden testdata file.
-var machoWriteTests = []struct {
-	name   string
-	golden bool
-	prog   *Prog
-}{
-	// amd64 exit 9
-	{
-		name:   "exit9",
-		golden: true,
-		prog: &Prog{
-			GOARCH:       "amd64",
-			GOOS:         "darwin",
-			UnmappedSize: 0x1000,
-			Entry:        0x1000,
-			Segments: []*Segment{
-				{
-					Name:       "text",
-					VirtAddr:   0x1000,
-					VirtSize:   13,
-					FileOffset: 0,
-					FileSize:   13,
-					Data: []byte{
-						0xb8, 0x01, 0x00, 0x00, 0x02, // MOVL $0x2000001, AX
-						0xbf, 0x09, 0x00, 0x00, 0x00, // MOVL $9, DI
-						0x0f, 0x05, // SYSCALL
-						0xf4, // HLT
-					},
-					Sections: []*Section{
-						{
-							Name:     "text",
-							VirtAddr: 0x1000,
-							Size:     13,
-							Align:    64,
-						},
-					},
-				},
-			},
-		},
-	},
-
-	// amd64 write hello world & exit 9
-	{
-		name:   "hello",
-		golden: true,
-		prog: &Prog{
-			GOARCH:       "amd64",
-			GOOS:         "darwin",
-			UnmappedSize: 0x1000,
-			Entry:        0x1000,
-			Segments: []*Segment{
-				{
-					Name:       "text",
-					VirtAddr:   0x1000,
-					VirtSize:   35,
-					FileOffset: 0,
-					FileSize:   35,
-					Data: []byte{
-						0xb8, 0x04, 0x00, 0x00, 0x02, // MOVL $0x2000001, AX
-						0xbf, 0x01, 0x00, 0x00, 0x00, // MOVL $1, DI
-						0xbe, 0x00, 0x30, 0x00, 0x00, // MOVL $0x3000, SI
-						0xba, 0x0c, 0x00, 0x00, 0x00, // MOVL $12, DX
-						0x0f, 0x05, // SYSCALL
-						0xb8, 0x01, 0x00, 0x00, 0x02, // MOVL $0x2000001, AX
-						0xbf, 0x09, 0x00, 0x00, 0x00, // MOVL $9, DI
-						0x0f, 0x05, // SYSCALL
-						0xf4, // HLT
-					},
-					Sections: []*Section{
-						{
-							Name:     "text",
-							VirtAddr: 0x1000,
-							Size:     35,
-							Align:    64,
-						},
-					},
-				},
-				{
-					Name:       "data",
-					VirtAddr:   0x2000,
-					VirtSize:   12,
-					FileOffset: 0x1000,
-					FileSize:   12,
-					Data:       []byte("hello world\n"),
-					Sections: []*Section{
-						{
-							Name:     "data",
-							VirtAddr: 0x2000,
-							Size:     12,
-							Align:    64,
-						},
-					},
-				},
-			},
-		},
-	},
-
-	// amd64 write hello world from rodata & exit 0
-	{
-		name:   "helloro",
-		golden: true,
-		prog: &Prog{
-			GOARCH:       "amd64",
-			GOOS:         "darwin",
-			UnmappedSize: 0x1000,
-			Entry:        0x1000,
-			Segments: []*Segment{
-				{
-					Name:       "text",
-					VirtAddr:   0x1000,
-					VirtSize:   0x100c,
-					FileOffset: 0,
-					FileSize:   0x100c,
-					Data: concat(
-						[]byte{
-							0xb8, 0x04, 0x00, 0x00, 0x02, // MOVL $0x2000001, AX
-							0xbf, 0x01, 0x00, 0x00, 0x00, // MOVL $1, DI
-							0xbe, 0x00, 0x30, 0x00, 0x00, // MOVL $0x3000, SI
-							0xba, 0x0c, 0x00, 0x00, 0x00, // MOVL $12, DX
-							0x0f, 0x05, // SYSCALL
-							0xb8, 0x01, 0x00, 0x00, 0x02, // MOVL $0x2000001, AX
-							0xbf, 0x00, 0x00, 0x00, 0x00, // MOVL $0, DI
-							0x0f, 0x05, // SYSCALL
-							0xf4, // HLT
-						},
-						make([]byte, 0x1000-35),
-						[]byte("hello world\n"),
-					),
-					Sections: []*Section{
-						{
-							Name:     "text",
-							VirtAddr: 0x1000,
-							Size:     35,
-							Align:    64,
-						},
-						{
-							Name:     "rodata",
-							VirtAddr: 0x2000,
-							Size:     12,
-							Align:    64,
-						},
-					},
-				},
-			},
-		},
-	},
-}
-
-func concat(xs ...[]byte) []byte {
-	var out []byte
-	for _, x := range xs {
-		out = append(out, x...)
-	}
-	return out
-}
-
-func TestMachoWrite(t *testing.T) {
-	for _, tt := range machoWriteTests {
-		name := tt.prog.GOARCH + "." + tt.name
-		prog := cloneProg(tt.prog)
-		prog.init()
-		var f machoFormat
-		vsize, fsize := f.headerSize(prog)
-		shiftProg(prog, vsize, fsize)
-		var buf bytes.Buffer
-		f.write(&buf, prog)
-		if false { // enable to debug
-			ioutil.WriteFile("a.out", buf.Bytes(), 0777)
-		}
-		read, err := machoRead(machoArches[tt.prog.GOARCH], buf.Bytes())
-		if err != nil {
-			t.Errorf("%s: reading mach-o output:\n\t%v", name, err)
-			continue
-		}
-		diffs := diffProg(read, prog)
-		if diffs != nil {
-			t.Errorf("%s: mismatched prog:\n\t%s", name, strings.Join(diffs, "\n\t"))
-			continue
-		}
-		if !tt.golden {
-			continue
-		}
-		checkGolden(t, buf.Bytes(), "testdata/macho."+name)
-	}
-}
-
-// machoRead reads the mach-o file in data and returns a corresponding prog.
-func machoRead(arch machoArch, data []byte) (*Prog, error) {
-	f, err := macho.NewFile(bytes.NewReader(data))
-	if err != nil {
-		return nil, err
-	}
-
-	var errors []string
-	errorf := func(format string, args ...interface{}) {
-		errors = append(errors, fmt.Sprintf(format, args...))
-	}
-
-	magic := uint32(0xFEEDFACE)
-	if arch.CPU&macho64Bit != 0 {
-		magic |= 1
-	}
-	if f.Magic != magic {
-		errorf("header: Magic = %#x, want %#x", f.Magic, magic)
-	}
-	if f.Cpu != macho.CpuAmd64 {
-		errorf("header: CPU = %#x, want %#x", f.Cpu, macho.CpuAmd64)
-	}
-	if f.SubCpu != 3 {
-		errorf("header: SubCPU = %#x, want %#x", f.SubCpu, 3)
-	}
-	if f.Type != 2 {
-		errorf("header: FileType = %d, want %d", f.Type, 2)
-	}
-	if f.Flags != 1 {
-		errorf("header: Flags = %d, want %d", f.Flags, 1)
-	}
-
-	msects := f.Sections
-	var limit uint64
-	prog := new(Prog)
-	for _, load := range f.Loads {
-		switch load := load.(type) {
-		default:
-			errorf("unexpected macho load %T %x", load, load.Raw())
-
-		case macho.LoadBytes:
-			if len(load) < 8 || len(load)%4 != 0 {
-				errorf("unexpected load length %d", len(load))
-				continue
-			}
-			cmd := f.ByteOrder.Uint32(load)
-			switch macho.LoadCmd(cmd) {
-			default:
-				errorf("unexpected macho load cmd %s", macho.LoadCmd(cmd))
-			case macho.LoadCmdUnixThread:
-				data := make([]uint32, len(load[8:])/4)
-				binary.Read(bytes.NewReader(load[8:]), f.ByteOrder, data)
-				if len(data) != 44 {
-					errorf("macho thread len(data) = %d, want 42", len(data))
-					continue
-				}
-				if data[0] != 4 {
-					errorf("macho thread type = %d, want 4", data[0])
-				}
-				if data[1] != uint32(len(data))-2 {
-					errorf("macho thread desc len = %d, want %d", data[1], uint32(len(data))-2)
-					continue
-				}
-				for i, val := range data[2:] {
-					switch i {
-					default:
-						if val != 0 {
-							errorf("macho thread data[%d] = %#x, want 0", i, val)
-						}
-					case 32:
-						prog.Entry = Addr(val)
-					case 33:
-						prog.Entry |= Addr(val) << 32
-					}
-				}
-			}
-
-		case *macho.Segment:
-			if load.Addr < limit {
-				errorf("segments out of order: %q at %#x after %#x", load.Name, load.Addr, limit)
-			}
-			limit = load.Addr + load.Memsz
-			if load.Name == "__PAGEZERO" || load.Addr == 0 && load.Filesz == 0 {
-				if load.Name != "__PAGEZERO" {
-					errorf("segment with Addr=0, Filesz=0 is named %q, want %q", load.Name, "__PAGEZERO")
-				} else if load.Addr != 0 || load.Filesz != 0 {
-					errorf("segment %q has Addr=%#x, Filesz=%d, want Addr=%#x, Filesz=%d", load.Name, load.Addr, load.Filesz, 0, 0)
-				}
-				prog.UnmappedSize = Addr(load.Memsz)
-				continue
-			}
-
-			if !strings.HasPrefix(load.Name, "__") {
-				errorf("segment name %q does not begin with %q", load.Name, "__")
-			}
-			if strings.ToUpper(load.Name) != load.Name {
-				errorf("segment name %q is not all upper case", load.Name)
-			}
-
-			seg := &Segment{
-				Name:       strings.ToLower(strings.TrimPrefix(load.Name, "__")),
-				VirtAddr:   Addr(load.Addr),
-				VirtSize:   Addr(load.Memsz),
-				FileOffset: Addr(load.Offset),
-				FileSize:   Addr(load.Filesz),
-			}
-			prog.Segments = append(prog.Segments, seg)
-
-			data, err := load.Data()
-			if err != nil {
-				errorf("loading data from %q: %v", load.Name, err)
-			}
-			seg.Data = data
-
-			var maxprot, prot uint32
-			if load.Name == "__TEXT" {
-				maxprot, prot = 7, 5
-			} else {
-				maxprot, prot = 3, 3
-			}
-			if load.Maxprot != maxprot || load.Prot != prot {
-				errorf("segment %q protection is %d, %d, want %d, %d",
-					load.Name, load.Maxprot, load.Prot, maxprot, prot)
-			}
-
-			for len(msects) > 0 && msects[0].Addr < load.Addr+load.Memsz {
-				msect := msects[0]
-				msects = msects[1:]
-
-				if msect.Offset > 0 && prog.HeaderSize == 0 {
-					prog.HeaderSize = Addr(msect.Offset)
-					if seg.FileOffset != 0 {
-						errorf("initial segment %q does not map header", load.Name)
-					}
-					seg.VirtAddr += prog.HeaderSize
-					seg.VirtSize -= prog.HeaderSize
-					seg.FileOffset += prog.HeaderSize
-					seg.FileSize -= prog.HeaderSize
-					seg.Data = seg.Data[prog.HeaderSize:]
-				}
-
-				if msect.Addr < load.Addr {
-					errorf("section %q at address %#x is missing segment", msect.Name, msect.Addr)
-					continue
-				}
-
-				if !strings.HasPrefix(msect.Name, "__") {
-					errorf("section name %q does not begin with %q", msect.Name, "__")
-				}
-				if strings.ToLower(msect.Name) != msect.Name {
-					errorf("section name %q is not all lower case", msect.Name)
-				}
-				if msect.Seg != load.Name {
-					errorf("section %q is lists segment name %q, want %q",
-						msect.Name, msect.Seg, load.Name)
-				}
-				if uint64(msect.Offset) != uint64(load.Offset)+msect.Addr-load.Addr {
-					errorf("section %q file offset is %#x, want %#x",
-						msect.Name, msect.Offset, load.Offset+msect.Addr-load.Addr)
-				}
-				if msect.Reloff != 0 || msect.Nreloc != 0 {
-					errorf("section %q has reloff %d,%d, want %d,%d",
-						msect.Name, msect.Reloff, msect.Nreloc, 0, 0)
-				}
-				flags := uint32(0)
-				if msect.Name == "__text" {
-					flags = 0x400
-				}
-				if msect.Offset == 0 {
-					flags = 1
-				}
-				if msect.Flags != flags {
-					errorf("section %q flags = %#x, want %#x", msect.Name, msect.Flags, flags)
-				}
-				sect := &Section{
-					Name:     strings.ToLower(strings.TrimPrefix(msect.Name, "__")),
-					VirtAddr: Addr(msect.Addr),
-					Size:     Addr(msect.Size),
-					Align:    1 << msect.Align,
-				}
-				seg.Sections = append(seg.Sections, sect)
-			}
-		}
-	}
-
-	for _, msect := range msects {
-		errorf("section %q has no segment", msect.Name)
-	}
-
-	limit = 0
-	for _, msect := range f.Sections {
-		if msect.Addr < limit {
-			errorf("sections out of order: %q at %#x after %#x", msect.Name, msect.Addr, limit)
-		}
-		limit = msect.Addr + msect.Size
-	}
-
-	err = nil
-	if errors != nil {
-		err = fmt.Errorf("%s", strings.Join(errors, "\n\t"))
-	}
-	return prog, err
-}

src/cmd/link/pclntab.go

@@ -1,479 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-// Generation of runtime function information (pclntab).
-
-package main
-
-import (
-	"debug/goobj"
-	"encoding/binary"
-	"os"
-	"sort"
-)
-
-var zerofunc goobj.Func
-
-// pclntab collects the runtime function data for each function that will
-// be listed in the binary and builds a single table describing all functions.
-// This table is used at run time for stack traces and to look up PC-specific
-// information during garbage collection. The symbol created is named
-// "pclntab" for historical reasons; the scope of the table has grown to
-// include more than just PC/line number correspondences.
-// The table format is documented at http://golang.org/s/go12symtab.
-func (p *Prog) pclntab() {
-	// Count number of functions going into the binary,
-	// so that we can size the initial index correctly.
-	nfunc := 0
-	for _, sym := range p.SymOrder {
-		if sym.Kind != goobj.STEXT {
-			continue
-		}
-		nfunc++
-	}
-
-	// Table header.
-	buf := new(SymBuffer)
-	buf.Init(p)
-	buf.SetSize(8 + p.ptrsize)
-	off := 0
-	off = buf.Uint32(off, 0xfffffffb)
-	off = buf.Uint8(off, 0)
-	off = buf.Uint8(off, 0)
-	off = buf.Uint8(off, uint8(p.pcquantum))
-	off = buf.Uint8(off, uint8(p.ptrsize))
-	off = buf.Uint(off, uint64(nfunc), p.ptrsize)
-	indexOff := off
-	off += (nfunc*2 + 1) * p.ptrsize // function index, to be filled in
-	off += 4                         // file table start offset, to be filled in
-	buf.SetSize(off)
-
-	// One-file cache for reading PCData tables from package files.
-	// TODO(rsc): Better I/O strategy.
-	var (
-		file  *os.File
-		fname string
-	)
-
-	// Files gives the file numbering for source file names recorded
-	// in the binary.
-	files := make(map[string]int)
-
-	// Build the table, build the index, and build the file name numbering.
-	// The loop here must visit functions in the same order that they will
-	// be stored in the binary, or else binary search over the index will fail.
-	// The runtime checks that the index is sorted properly at program start time.
-	var lastSym *Sym
-	for _, sym := range p.SymOrder {
-		if sym.Kind != goobj.STEXT {
-			continue
-		}
-		lastSym = sym
-
-		// Treat no recorded function information same as all zeros.
-		f := sym.Func
-		if f == nil {
-			f = &zerofunc
-		}
-
-		// Open package file if needed, for reading PC data.
-		if fname != sym.Package.File {
-			if file != nil {
-				file.Close()
-			}
-			var err error
-			file, err = os.Open(sym.Package.File)
-			if err != nil {
-				p.errorf("%v: %v", sym, err)
-				return
-			}
-			fname = sym.Package.File
-		}
-
-		// off is the offset of the table entry where we're going to write
-		// the encoded form of Func.
-		// indexOff is the current position in the table index;
-		// we add an entry in the index pointing at off.
-		off = (buf.Size() + p.ptrsize - 1) &^ (p.ptrsize - 1)
-		indexOff = buf.Addr(indexOff, sym.SymID, 0)
-		indexOff = buf.Uint(indexOff, uint64(off), p.ptrsize)
-
-		// The Func encoding starts with a header giving offsets
-		// to data blobs, and then the data blobs themselves.
-		// end gives the current write position for the data blobs.
-		end := off + p.ptrsize + 3*4 + 5*4 + len(f.PCData)*4 + len(f.FuncData)*p.ptrsize
-		if len(f.FuncData) > 0 {
-			end += -end & (p.ptrsize - 1)
-		}
-		buf.SetSize(end)
-
-		// entry uintptr
-		// name int32
-		// args int32
-		// frame int32
-		//
-		// The frame recorded in the object file is
-		// the frame size used in an assembly listing, which does
-		// not include the caller PC on the stack.
-		// The frame size we want to list here is the delta from
-		// this function's SP to its caller's SP, which does include
-		// the caller PC. Add p.ptrsize to f.Frame to adjust.
-		// TODO(rsc): Record the same frame size in the object file.
-		off = buf.Addr(off, sym.SymID, 0)
-		off = buf.Uint32(off, uint32(addString(buf, sym.Name)))
-		off = buf.Uint32(off, uint32(f.Args))
-		off = buf.Uint32(off, uint32(f.Frame+p.ptrsize))
-
-		// pcdata
-		off = buf.Uint32(off, uint32(addPCTable(p, buf, file, f.PCSP)))
-		off = buf.Uint32(off, uint32(addPCFileTable(p, buf, file, f.PCFile, sym, files)))
-		off = buf.Uint32(off, uint32(addPCTable(p, buf, file, f.PCLine)))
-		off = buf.Uint32(off, uint32(len(f.PCData)))
-		off = buf.Uint32(off, uint32(len(f.FuncData)))
-		for _, pcdata := range f.PCData {
-			off = buf.Uint32(off, uint32(addPCTable(p, buf, file, pcdata)))
-		}
-
-		// funcdata
-		if len(f.FuncData) > 0 {
-			off += -off & (p.ptrsize - 1) // must be pointer-aligned
-			for _, funcdata := range f.FuncData {
-				if funcdata.Sym.Name == "" {
-					off = buf.Uint(off, uint64(funcdata.Offset), p.ptrsize)
-				} else {
-					off = buf.Addr(off, funcdata.Sym, funcdata.Offset)
-				}
-			}
-		}
-
-		if off != end {
-			p.errorf("internal error: invalid math in pclntab: off=%#x end=%#x", off, end)
-			break
-		}
-	}
-	if file != nil {
-		file.Close()
-	}
-
-	// Final entry of index is end PC of last function.
-	indexOff = buf.Addr(indexOff, lastSym.SymID, int64(lastSym.Size))
-
-	// Start file table.
-	// Function index is immediately followed by offset to file table.
-	off = (buf.Size() + p.ptrsize - 1) &^ (p.ptrsize - 1)
-	buf.Uint32(indexOff, uint32(off))
-
-	// File table is an array of uint32s.
-	// The first entry gives 1+n, the size of the array.
-	// The following n entries hold offsets to string data.
-	// File number n uses the string pointed at by entry n.
-	// File number 0 is invalid.
-	buf.SetSize(off + (1+len(files))*4)
-	buf.Uint32(off, uint32(1+len(files)))
-	var filestr []string
-	for file := range files {
-		filestr = append(filestr, file)
-	}
-	sort.Strings(filestr)
-	for _, file := range filestr {
-		id := files[file]
-		buf.Uint32(off+4*id, uint32(addString(buf, file)))
-	}
-
-	pclntab := &Sym{
-		Sym: &goobj.Sym{
-			SymID: goobj.SymID{Name: "pclntab"},
-			Kind:  goobj.SPCLNTAB,
-			Size:  buf.Size(),
-			Reloc: buf.Reloc(),
-		},
-		Bytes: buf.Bytes(),
-	}
-	p.addSym(pclntab)
-}
-
-// addString appends the string s to the buffer b.
-// It returns the offset of the beginning of the string in the buffer.
-func addString(b *SymBuffer, s string) int {
-	off := b.Size()
-	b.SetSize(off + len(s) + 1)
-	copy(b.data[off:], s)
-	return off
-}
-
-// addPCTable appends the PC-data table stored in the file f at the location loc
-// to the symbol buffer b. It returns the offset of the beginning of the table
-// in the buffer.
-func addPCTable(p *Prog, b *SymBuffer, f *os.File, loc goobj.Data) int {
-	if loc.Size == 0 {
-		return 0
-	}
-	off := b.Size()
-	b.SetSize(off + int(loc.Size))
-	_, err := f.ReadAt(b.data[off:off+int(loc.Size)], loc.Offset)
-	if err != nil {
-		p.errorf("%v", err)
-	}
-	return off
-}
-
-// addPCFileTable is like addPCTable, but it renumbers the file names referred to by the table
-// to use the global numbering maintained in the files map. It adds new files to the
-// map as necessary.
-func addPCFileTable(p *Prog, b *SymBuffer, f *os.File, loc goobj.Data, sym *Sym, files map[string]int) int {
-	if loc.Size == 0 {
-		return 0
-	}
-	off := b.Size()
-
-	src := make([]byte, loc.Size)
-	_, err := f.ReadAt(src, loc.Offset)
-	if err != nil {
-		p.errorf("%v", err)
-		return 0
-	}
-
-	filenum := make([]int, len(sym.Func.File))
-	for i, name := range sym.Func.File {
-		num := files[name]
-		if num == 0 {
-			num = len(files) + 1
-			files[name] = num
-		}
-		filenum[i] = num
-	}
-
-	var dst []byte
-	newval := int32(-1)
-	var it PCIter
-	for it.Init(p, src); !it.Done; it.Next() {
-		// value delta
-		oldval := it.Value
-		val := oldval
-		if oldval != -1 {
-			if oldval < 0 || int(oldval) >= len(filenum) {
-				p.errorf("%s: corrupt pc-file table", sym)
-				break
-			}
-			val = int32(filenum[oldval])
-		}
-		dv := val - newval
-		newval = val
-		uv := uint32(dv<<1) ^ uint32(dv>>31)
-		dst = appendVarint(dst, uv)
-
-		// pc delta
-		dst = appendVarint(dst, it.NextPC-it.PC)
-	}
-	if it.Corrupt {
-		p.errorf("%s: corrupt pc-file table", sym)
-	}
-
-	// terminating value delta
-	dst = appendVarint(dst, 0)
-
-	b.SetSize(off + len(dst))
-	copy(b.data[off:], dst)
-	return off
-}
-
-// A SymBuffer is a buffer for preparing the data image of a
-// linker-generated symbol.
-type SymBuffer struct {
-	data    []byte
-	reloc   []goobj.Reloc
-	order   binary.ByteOrder
-	ptrsize int
-}
-
-// Init initializes the buffer for writing.
-func (b *SymBuffer) Init(p *Prog) {
-	b.data = nil
-	b.reloc = nil
-	b.order = p.byteorder
-	b.ptrsize = p.ptrsize
-}
-
-// Bytes returns the buffer data.
-func (b *SymBuffer) Bytes() []byte {
-	return b.data
-}
-
-// SetSize sets the buffer's data size to n bytes.
-func (b *SymBuffer) SetSize(n int) {
-	for cap(b.data) < n {
-		b.data = append(b.data[:cap(b.data)], 0)
-	}
-	b.data = b.data[:n]
-}
-
-// Size returns the buffer's data size.
-func (b *SymBuffer) Size() int {
-	return len(b.data)
-}
-
-// Reloc returns the buffered relocations.
-func (b *SymBuffer) Reloc() []goobj.Reloc {
-	return b.reloc
-}
-
-// Uint8 sets the uint8 at offset off to v.
-// It returns the offset just beyond v.
-func (b *SymBuffer) Uint8(off int, v uint8) int {
-	b.data[off] = v
-	return off + 1
-}
-
-// Uint16 sets the uint16 at offset off to v.
-// It returns the offset just beyond v.
-func (b *SymBuffer) Uint16(off int, v uint16) int {
-	b.order.PutUint16(b.data[off:], v)
-	return off + 2
-}
-
-// Uint32 sets the uint32 at offset off to v.
-// It returns the offset just beyond v.
-func (b *SymBuffer) Uint32(off int, v uint32) int {
-	b.order.PutUint32(b.data[off:], v)
-	return off + 4
-}
-
-// Uint64 sets the uint64 at offset off to v.
-// It returns the offset just beyond v.
-func (b *SymBuffer) Uint64(off int, v uint64) int {
-	b.order.PutUint64(b.data[off:], v)
-	return off + 8
-}
-
-// Uint sets the size-byte unsigned integer at offset off to v.
-// It returns the offset just beyond v.
-func (b *SymBuffer) Uint(off int, v uint64, size int) int {
-	switch size {
-	case 1:
-		return b.Uint8(off, uint8(v))
-	case 2:
-		return b.Uint16(off, uint16(v))
-	case 4:
-		return b.Uint32(off, uint32(v))
-	case 8:
-		return b.Uint64(off, v)
-	}
-	panic("invalid use of SymBuffer.SetUint")
-}
-
-// Addr sets the pointer-sized address at offset off to refer
-// to symoff bytes past the start of sym. It returns the offset
-// just beyond the address.
-func (b *SymBuffer) Addr(off int, sym goobj.SymID, symoff int64) int {
-	b.reloc = append(b.reloc, goobj.Reloc{
-		Offset: off,
-		Size:   b.ptrsize,
-		Sym:    sym,
-		Add:    int(symoff),
-		Type:   R_ADDR,
-	})
-	return off + b.ptrsize
-}
-
-// A PCIter implements iteration over PC-data tables.
-//
-//	var it PCIter
-//	for it.Init(p, data); !it.Done; it.Next() {
-//		it.Value holds from it.PC up to (but not including) it.NextPC
-//	}
-//	if it.Corrupt {
-//		data was malformed
-//	}
-//
-type PCIter struct {
-	PC        uint32
-	NextPC    uint32
-	Value     int32
-	Done      bool
-	Corrupt   bool
-	p         []byte
-	start     bool
-	pcquantum uint32
-}
-
-// Init initializes the iteration.
-// On return, if it.Done is true, the iteration is over.
-// Otherwise it.Value applies in the pc range [it.PC, it.NextPC).
-func (it *PCIter) Init(p *Prog, buf []byte) {
-	it.p = buf
-	it.PC = 0
-	it.NextPC = 0
-	it.Value = -1
-	it.start = true
-	it.pcquantum = uint32(p.pcquantum)
-	it.Done = false
-	it.Next()
-}
-
-// Next steps forward one entry in the table.
-// On return, if it.Done is true, the iteration is over.
-// Otherwise it.Value applies in the pc range [it.PC, it.NextPC).
-func (it *PCIter) Next() {
-	it.PC = it.NextPC
-	if it.Done {
-		return
-	}
-	if len(it.p) == 0 {
-		it.Done = true
-		return
-	}
-
-	// value delta
-	uv, p, ok := decodeVarint(it.p)
-	if !ok {
-		it.Done = true
-		it.Corrupt = true
-		return
-	}
-	it.p = p
-	if uv == 0 && !it.start {
-		it.Done = true
-		return
-	}
-	it.start = false
-	sv := int32(uv)>>1 ^ int32(uv)<<31>>31
-	it.Value += sv
-
-	// pc delta
-	uv, it.p, ok = decodeVarint(it.p)
-	if !ok {
-		it.Done = true
-		it.Corrupt = true
-		return
-	}
-	it.NextPC = it.PC + uv*it.pcquantum
-}
-
-// decodeVarint decodes an unsigned varint from p,
-// reporting the value, the remainder of the data, and
-// whether the decoding was successful.
-func decodeVarint(p []byte) (v uint32, rest []byte, ok bool) {
-	for shift := uint(0); ; shift += 7 {
-		if len(p) == 0 {
-			return
-		}
-		c := uint32(p[0])
-		p = p[1:]
-		v |= (c & 0x7F) << shift
-		if c&0x80 == 0 {
-			break
-		}
-	}
-	return v, p, true
-}
-
-// appendVarint appends an unsigned varint encoding of v to p
-// and returns the resulting slice.
-func appendVarint(p []byte, v uint32) []byte {
-	for ; v >= 0x80; v >>= 7 {
-		p = append(p, byte(v)|0x80)
-	}
-	p = append(p, byte(v))
-	return p
-}

src/cmd/link/pclntab_test.go

@@ -1,334 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-package main
-
-import (
-	"bytes"
-	"debug/goobj"
-	"fmt"
-	"math/rand"
-	"sort"
-	"strings"
-	"testing"
-)
-
-// Test of pcln table encoding.
-// testdata/genpcln.go generates an assembly file with
-// pseudorandom values for the data that pclntab stores.
-// This test recomputes the same pseudorandom stream
-// and checks that the final linked binary uses those values
-// as well.
-func TestPclntab(t *testing.T) {
-	p := &Prog{
-		GOOS:        "darwin",
-		GOARCH:      "amd64",
-		Error:       func(s string) { t.Error(s) },
-		StartSym:    "start",
-		omitRuntime: true,
-	}
-	var buf bytes.Buffer
-	p.link(&buf, "testdata/pclntab.6")
-	if p.NumError > 0 {
-		return
-	}
-
-	// The algorithm for computing values here must match
-	// the one in testdata/genpcln.go.
-	for f := 0; f < 3; f++ {
-		file := "input"
-		line := 1
-		rnd := rand.New(rand.NewSource(int64(f)))
-		args := rnd.Intn(100) * 8
-		frame := 32 + rnd.Intn(32)/8*8
-		size := 200 + rnd.Intn(100)*8
-
-		name := fmt.Sprintf("func%d", f)
-		r, off, fargs, fframe, ok := findFunc(t, p, name)
-		if !ok {
-			continue // error already printed
-		}
-		if fargs != args {
-			t.Errorf("%s: args=%d, want %d", name, fargs, args)
-		}
-		if fframe != frame+8 {
-			t.Errorf("%s: frame=%d, want %d", name, fframe, frame+8)
-		}
-
-		// Check FUNCDATA 1.
-		fdata, ok := loadFuncdata(t, r, name, off, 1)
-		if ok {
-			fsym := p.Syms[goobj.SymID{Name: fmt.Sprintf("funcdata%d", f)}]
-			if fsym == nil {
-				t.Errorf("funcdata%d is missing in binary", f)
-			} else if fdata != fsym.Addr {
-				t.Errorf("%s: funcdata 1 = %#x, want %#x", name, fdata, fsym.Addr)
-			}
-		}
-
-		// Walk code checking pcdata values.
-		spadj := 0
-		pcdata1 := -1
-		pcdata2 := -1
-
-		checkPCSP(t, r, name, off, 0, 0)
-		checkPCData(t, r, name, off, 0, 0, -1)
-		checkPCData(t, r, name, off, 0, 1, -1)
-		checkPCData(t, r, name, off, 0, 2, -1)
-
-		firstpc := 4
-		for i := 0; i < size; i++ {
-			pc := firstpc + i // skip SP adjustment to allocate frame
-			if i >= 0x100 && t.Failed() {
-				break
-			}
-			// Possible SP adjustment.
-			checkPCSP(t, r, name, off, pc, frame+spadj)
-			if rnd.Intn(100) == 0 {
-				checkPCFileLine(t, r, name, off, pc, file, line)
-				checkPCData(t, r, name, off, pc, 1, pcdata1)
-				checkPCData(t, r, name, off, pc, 2, pcdata2)
-				i += 1
-				pc = firstpc + i
-				checkPCFileLine(t, r, name, off, pc-1, file, line)
-				checkPCData(t, r, name, off, pc-1, 1, pcdata1)
-				checkPCData(t, r, name, off, pc-1, 2, pcdata2)
-				checkPCSP(t, r, name, off, pc-1, frame+spadj)
-
-				if spadj <= -32 || spadj < 32 && rnd.Intn(2) == 0 {
-					spadj += 8
-				} else {
-					spadj -= 8
-				}
-				checkPCSP(t, r, name, off, pc, frame+spadj)
-			}
-
-			// Possible PCFile change.
-			if rnd.Intn(100) == 0 {
-				file = fmt.Sprintf("file%d.s", rnd.Intn(10))
-				line = rnd.Intn(100) + 1
-			}
-
-			// Possible PCLine change.
-			if rnd.Intn(10) == 0 {
-				line = rnd.Intn(1000) + 1
-			}
-
-			// Possible PCData $1 change.
-			if rnd.Intn(100) == 0 {
-				pcdata1 = rnd.Intn(1000)
-			}
-
-			// Possible PCData $2 change.
-			if rnd.Intn(100) == 0 {
-				pcdata2 = rnd.Intn(1000)
-			}
-
-			if i == 0 {
-				checkPCFileLine(t, r, name, off, 0, file, line)
-				checkPCFileLine(t, r, name, off, pc-1, file, line)
-			}
-			checkPCFileLine(t, r, name, off, pc, file, line)
-			checkPCData(t, r, name, off, pc, 1, pcdata1)
-			checkPCData(t, r, name, off, pc, 2, pcdata2)
-		}
-	}
-}
-
-// findFunc finds the function information in the pclntab of p
-// for the function with the given name.
-// It returns a symbol reader for pclntab, the offset of the function information
-// within that symbol, and the args and frame values read out of the information.
-func findFunc(t *testing.T, p *Prog, name string) (r *SymReader, off, args, frame int, ok bool) {
-	tabsym := p.Syms[goobj.SymID{Name: "pclntab"}]
-	if tabsym == nil {
-		t.Errorf("pclntab is missing in binary")
-		return
-	}
-
-	r = new(SymReader)
-	r.Init(p, tabsym)
-
-	// pclntab must with 8-byte header
-	if r.Uint32(0) != 0xfffffffb || r.Uint8(4) != 0 || r.Uint8(5) != 0 || r.Uint8(6) != uint8(p.pcquantum) || r.Uint8(7) != uint8(p.ptrsize) {
-		t.Errorf("pclntab has incorrect header %.8x", r.data[:8])
-		return
-	}
-
-	sym := p.Syms[goobj.SymID{Name: name}]
-	if sym == nil {
-		t.Errorf("%s is missing in the binary", name)
-		return
-	}
-
-	// index is nfunc addr0 off0 addr1 off1 ... addr_nfunc (sentinel)
-	nfunc := int(r.Addr(8))
-	i := sort.Search(nfunc, func(i int) bool {
-		return r.Addr(8+p.ptrsize*(1+2*i)) >= sym.Addr
-	})
-	if entry := r.Addr(8 + p.ptrsize*(1+2*i)); entry != sym.Addr {
-		indexTab := make([]Addr, 2*nfunc+1)
-		for j := range indexTab {
-			indexTab[j] = r.Addr(8 + p.ptrsize*(1+j))
-		}
-		t.Errorf("pclntab is missing entry for %s (%#x): %#x", name, sym.Addr, indexTab)
-		return
-	}
-
-	off = int(r.Addr(8 + p.ptrsize*(1+2*i+1)))
-
-	// func description at off is
-	//	entry addr
-	//	nameoff uint32
-	//	args uint32
-	//	frame uint32
-	//	pcspoff uint32
-	//	pcfileoff uint32
-	//	pclineoff uint32
-	//	npcdata uint32
-	//	nfuncdata uint32
-	//	pcdata npcdata*uint32
-	//	funcdata nfuncdata*addr
-	//
-	if entry := r.Addr(off); entry != sym.Addr {
-		t.Errorf("pclntab inconsistent: entry for %s addr=%#x has entry=%#x", name, sym.Addr, entry)
-		return
-	}
-	nameoff := int(r.Uint32(off + p.ptrsize))
-	args = int(r.Uint32(off + p.ptrsize + 1*4))
-	frame = int(r.Uint32(off + p.ptrsize + 2*4))
-
-	fname := r.String(nameoff)
-	if fname != name {
-		t.Errorf("pclntab inconsistent: entry for %s addr=%#x has name %q", name, sym.Addr, fname)
-	}
-
-	ok = true // off, args, frame are usable
-	return
-}
-
-// loadFuncdata returns the funcdata #fnum value
-// loaded from the function information for name.
-func loadFuncdata(t *testing.T, r *SymReader, name string, off int, fnum int) (Addr, bool) {
-	npcdata := int(r.Uint32(off + r.p.ptrsize + 6*4))
-	nfuncdata := int(r.Uint32(off + r.p.ptrsize + 7*4))
-	if fnum >= nfuncdata {
-		t.Errorf("pclntab(%s): no funcdata %d (only < %d)", name, fnum, nfuncdata)
-		return 0, false
-	}
-	fdataoff := off + r.p.ptrsize + (8+npcdata)*4 + fnum*r.p.ptrsize
-	fdataoff += fdataoff & 4
-	return r.Addr(fdataoff), true
-}
-
-// checkPCSP checks that the PCSP table in the function information at off
-// lists spadj as the sp delta for pc.
-func checkPCSP(t *testing.T, r *SymReader, name string, off, pc, spadj int) {
-	pcoff := r.Uint32(off + r.p.ptrsize + 3*4)
-	pcval, ok := readPCData(t, r, name, "PCSP", pcoff, pc)
-	if !ok {
-		return
-	}
-	if pcval != spadj {
-		t.Errorf("pclntab(%s): at pc=+%#x, pcsp=%d, want %d", name, pc, pcval, spadj)
-	}
-}
-
-// checkPCSP checks that the PCFile and PCLine tables in the function information at off
-// list file, line as the file name and line number for pc.
-func checkPCFileLine(t *testing.T, r *SymReader, name string, off, pc int, file string, line int) {
-	pcfileoff := r.Uint32(off + r.p.ptrsize + 4*4)
-	pclineoff := r.Uint32(off + r.p.ptrsize + 5*4)
-	pcfilenum, ok1 := readPCData(t, r, name, "PCFile", pcfileoff, pc)
-	pcline, ok2 := readPCData(t, r, name, "PCLine", pclineoff, pc)
-	if !ok1 || !ok2 {
-		return
-	}
-	nfunc := int(r.Addr(8))
-	filetaboff := r.Uint32(8 + r.p.ptrsize*2*(nfunc+1))
-	nfile := int(r.Uint32(int(filetaboff)))
-	if pcfilenum <= 0 || pcfilenum >= nfile {
-		t.Errorf("pclntab(%s): at pc=+%#x, filenum=%d (invalid; nfile=%d)", name, pc, pcfilenum, nfile)
-	}
-	pcfile := r.String(int(r.Uint32(int(filetaboff) + pcfilenum*4)))
-	if !strings.HasSuffix(pcfile, file) {
-		t.Errorf("pclntab(%s): at pc=+%#x, file=%q, want %q", name, pc, pcfile, file)
-	}
-	if pcline != line {
-		t.Errorf("pclntab(%s): at pc=+%#x, line=%d, want %d", name, pc, pcline, line)
-	}
-}
-
-// checkPCData checks that the PCData#pnum table in the function information at off
-// list val as the value for pc.
-func checkPCData(t *testing.T, r *SymReader, name string, off, pc, pnum, val int) {
-	pcoff := r.Uint32(off + r.p.ptrsize + (8+pnum)*4)
-	pcval, ok := readPCData(t, r, name, fmt.Sprintf("PCData#%d", pnum), pcoff, pc)
-	if !ok {
-		return
-	}
-	if pcval != val {
-		t.Errorf("pclntab(%s): at pc=+%#x, pcdata#%d=%d, want %d", name, pc, pnum, pcval, val)
-	}
-}
-
-// readPCData reads the PCData table offset off
-// to obtain and return the value associated with pc.
-func readPCData(t *testing.T, r *SymReader, name, pcdataname string, pcoff uint32, pc int) (int, bool) {
-	var it PCIter
-	for it.Init(r.p, r.data[pcoff:]); !it.Done; it.Next() {
-		if it.PC <= uint32(pc) && uint32(pc) < it.NextPC {
-			return int(it.Value), true
-		}
-	}
-	if it.Corrupt {
-		t.Errorf("pclntab(%s): %s: corrupt pcdata table", name, pcdataname)
-	}
-	return 0, false
-}
-
-// A SymReader provides typed access to the data for a symbol.
-type SymReader struct {
-	p    *Prog
-	data []byte
-}
-
-func (r *SymReader) Init(p *Prog, sym *Sym) {
-	seg := sym.Section.Segment
-	off := sym.Addr - seg.VirtAddr
-	data := seg.Data[off : off+Addr(sym.Size)]
-	r.p = p
-	r.data = data
-}
-
-func (r *SymReader) Uint8(off int) uint8 {
-	return r.data[off]
-}
-
-func (r *SymReader) Uint16(off int) uint16 {
-	return r.p.byteorder.Uint16(r.data[off:])
-}
-
-func (r *SymReader) Uint32(off int) uint32 {
-	return r.p.byteorder.Uint32(r.data[off:])
-}
-
-func (r *SymReader) Uint64(off int) uint64 {
-	return r.p.byteorder.Uint64(r.data[off:])
-}
-
-func (r *SymReader) Addr(off int) Addr {
-	if r.p.ptrsize == 4 {
-		return Addr(r.Uint32(off))
-	}
-	return Addr(r.Uint64(off))
-}
-
-func (r *SymReader) String(off int) string {
-	end := off
-	for r.data[end] != '\x00' {
-		end++
-	}
-	return string(r.data[off:end])
-}

src/cmd/link/prog.go

@@ -1,220 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-package main
-
-import (
-	"debug/goobj"
-	"encoding/binary"
-	"fmt"
-	"go/build"
-	"io"
-	"os"
-	"runtime"
-)
-
-// A Prog holds state for constructing an executable (program) image.
-//
-// The usual sequence of operations on a Prog is:
-//
-//	p.init()
-//	p.scan(file)
-//	p.dead()
-//	p.runtime()
-//	p.layout()
-//	p.load()
-//	p.debug()
-//	p.write(w)
-//
-// p.init is in this file. The rest of the methods are in files
-// named for the method. The convenience method p.link runs
-// this sequence.
-//
-type Prog struct {
-	// Context
-	GOOS     string       // target operating system
-	GOARCH   string       // target architecture
-	Format   string       // desired file format ("elf", "macho", ...)
-	Error    func(string) // called to report an error (if set)
-	NumError int          // number of errors printed
-	StartSym string
-
-	// Derived context
-	arch
-	formatter   formatter
-	startSym    goobj.SymID
-	pkgdir      string
-	omitRuntime bool // do not load runtime package
-
-	// Input
-	Packages   map[string]*Package  // loaded packages, by import path
-	Syms       map[goobj.SymID]*Sym // defined symbols, by symbol ID
-	Missing    map[goobj.SymID]bool // missing symbols
-	Dead       map[goobj.SymID]bool // symbols removed as dead
-	SymOrder   []*Sym               // order syms were scanned
-	MaxVersion int                  // max SymID.Version, for generating fresh symbol IDs
-
-	// Output
-	UnmappedSize Addr       // size of unmapped region at address 0
-	HeaderSize   Addr       // size of object file header
-	Entry        Addr       // virtual address where execution begins
-	Segments     []*Segment // loaded memory segments
-}
-
-// An arch describes architecture-dependent settings.
-type arch struct {
-	byteorder binary.ByteOrder
-	ptrsize   int
-	pcquantum int
-}
-
-// A formatter takes care of the details of generating a particular
-// kind of executable file.
-type formatter interface {
-	// headerSize returns the footprint of the header for p
-	// in both virtual address space and file bytes.
-	// The footprint does not include any bytes stored at the
-	// end of the file.
-	headerSize(p *Prog) (virt, file Addr)
-
-	// write writes the executable file for p to w.
-	write(w io.Writer, p *Prog)
-}
-
-// An Addr represents a virtual memory address, a file address, or a size.
-// It must be a uint64, not a uintptr, so that a 32-bit linker can still generate a 64-bit binary.
-// It must be unsigned in order to link programs placed at very large start addresses.
-// Math involving Addrs must be checked carefully not to require negative numbers.
-type Addr uint64
-
-// A Package is a Go package loaded from a file.
-type Package struct {
-	*goobj.Package        // table of contents
-	File           string // file name for reopening
-	Syms           []*Sym // symbols defined by this package
-}
-
-// A Sym is a symbol defined in a loaded package.
-type Sym struct {
-	*goobj.Sym          // symbol metadata from package file
-	Package    *Package // package defining symbol
-	Section    *Section // section where symbol is placed in output program
-	Addr       Addr     // virtual address of symbol in output program
-	Bytes      []byte   // symbol data, for internally defined symbols
-}
-
-// A Segment is a loaded memory segment.
-// A Prog is expected to have segments named "text" and optionally "data",
-// in that order, before any other segments.
-type Segment struct {
-	Name       string     // name of segment: "text", "data", ...
-	VirtAddr   Addr       // virtual memory address of segment base
-	VirtSize   Addr       // size of segment in memory
-	FileOffset Addr       // file offset of segment base
-	FileSize   Addr       // size of segment in file; can be less than VirtSize
-	Sections   []*Section // sections inside segment
-	Data       []byte     // raw data of segment image
-}
-
-// A Section is part of a loaded memory segment.
-type Section struct {
-	Name     string   // name of section: "text", "rodata", "noptrbss", and so on
-	VirtAddr Addr     // virtual memory address of section base
-	Size     Addr     // size of section in memory
-	Align    Addr     // required alignment
-	InFile   bool     // section has image data in file (like data, unlike bss)
-	Syms     []*Sym   // symbols stored in section
-	Segment  *Segment // segment containing section
-}
-
-func (p *Prog) errorf(format string, args ...interface{}) {
-	if p.Error != nil {
-		p.Error(fmt.Sprintf(format, args...))
-	} else {
-		fmt.Fprintf(os.Stderr, format+"\n", args...)
-	}
-	p.NumError++
-}
-
-// link is the one-stop convenience method for running a link.
-// It writes to w the object file generated from using mainFile as the main package.
-func (p *Prog) link(w io.Writer, mainFile string) {
-	p.init()
-	p.scan(mainFile)
-	if p.NumError > 0 {
-		return
-	}
-	p.dead()
-	p.runtime()
-	p.autoData()
-	p.layout()
-	p.autoConst()
-	if p.NumError > 0 {
-		return
-	}
-	p.load()
-	if p.NumError > 0 {
-		return
-	}
-	p.debug()
-	if p.NumError > 0 {
-		return
-	}
-	p.write(w)
-}
-
-// init initializes p for use by the other methods.
-func (p *Prog) init() {
-	// Set default context if not overridden.
-	if p.GOOS == "" {
-		p.GOOS = build.Default.GOOS
-	}
-	if p.GOARCH == "" {
-		p.GOARCH = build.Default.GOARCH
-	}
-	if p.Format == "" {
-		p.Format = goosFormat[p.GOOS]
-		if p.Format == "" {
-			p.errorf("no default file format for GOOS %q", p.GOOS)
-			return
-		}
-	}
-	if p.StartSym == "" {
-		p.StartSym = fmt.Sprintf("_rt0_%s_%s", p.GOARCH, p.GOOS)
-	}
-
-	// Derive internal context.
-	p.formatter = formatters[p.Format]
-	if p.formatter == nil {
-		p.errorf("unknown output file format %q", p.Format)
-		return
-	}
-	p.startSym = goobj.SymID{Name: p.StartSym}
-	arch, ok := arches[p.GOARCH]
-	if !ok {
-		p.errorf("unknown GOOS %q", p.GOOS)
-		return
-	}
-	p.arch = arch
-
-	p.pkgdir = fmt.Sprintf("%s/pkg/%s_%s", runtime.GOROOT(), p.GOOS, p.GOARCH)
-}
-
-// goosFormat records the default format for each known GOOS value.
-var goosFormat = map[string]string{
-	"darwin": "darwin",
-}
-
-// formatters records the format implementation for each known format value.
-var formatters = map[string]formatter{
-	"darwin": machoFormat{},
-}
-
-var arches = map[string]arch{
-	"amd64": {
-		byteorder: binary.LittleEndian,
-		ptrsize:   8,
-		pcquantum: 1,
-	},
-}

src/cmd/link/prog_test.go

@@ -1,163 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-package main
-
-import (
-	"bytes"
-	"fmt"
-	"io/ioutil"
-	"testing"
-)
-
-// shiftProg adjusts the addresses in p.
-// It adds vdelta to all virtual addresses and fdelta to all file offsets.
-func shiftProg(p *Prog, vdelta Addr, fdelta Addr) {
-	p.Entry += vdelta
-	for _, seg := range p.Segments {
-		seg.FileOffset += fdelta
-		seg.VirtAddr += vdelta
-		for _, sect := range seg.Sections {
-			sect.VirtAddr += vdelta
-			for _, sym := range sect.Syms {
-				sym.Addr += vdelta
-			}
-		}
-	}
-}
-
-// diffProg returns a list of differences between p and q,
-// assuming p is being checked and q is the correct answer.
-func diffProg(p, q *Prog) []string {
-	var errors []string
-	if p.UnmappedSize != q.UnmappedSize {
-		errors = append(errors, fmt.Sprintf("p.UnmappedSize = %#x, want %#x", p.UnmappedSize, q.UnmappedSize))
-	}
-	if p.HeaderSize != q.HeaderSize {
-		errors = append(errors, fmt.Sprintf("p.HeaderSize = %#x, want %#x", p.HeaderSize, q.HeaderSize))
-	}
-	if p.Entry != q.Entry {
-		errors = append(errors, fmt.Sprintf("p.Entry = %#x, want %#x", p.Entry, q.Entry))
-	}
-	for i := 0; i < len(p.Segments) || i < len(q.Segments); i++ {
-		if i >= len(p.Segments) {
-			errors = append(errors, fmt.Sprintf("p missing segment %q", q.Segments[i].Name))
-			continue
-		}
-		if i >= len(q.Segments) {
-			errors = append(errors, fmt.Sprintf("p has extra segment %q", p.Segments[i].Name))
-			continue
-		}
-		pseg := p.Segments[i]
-		qseg := q.Segments[i]
-		if pseg.Name != qseg.Name {
-			errors = append(errors, fmt.Sprintf("segment %d Name = %q, want %q", i, pseg.Name, qseg.Name))
-			continue // probably out of sync
-		}
-		if pseg.VirtAddr != qseg.VirtAddr {
-			errors = append(errors, fmt.Sprintf("segment %q VirtAddr = %#x, want %#x", pseg.Name, pseg.VirtAddr, qseg.VirtAddr))
-		}
-		if pseg.VirtSize != qseg.VirtSize {
-			errors = append(errors, fmt.Sprintf("segment %q VirtSize = %#x, want %#x", pseg.Name, pseg.VirtSize, qseg.VirtSize))
-		}
-		if pseg.FileOffset != qseg.FileOffset {
-			errors = append(errors, fmt.Sprintf("segment %q FileOffset = %#x, want %#x", pseg.Name, pseg.FileOffset, qseg.FileOffset))
-		}
-		if pseg.FileSize != qseg.FileSize {
-			errors = append(errors, fmt.Sprintf("segment %q FileSize = %#x, want %#x", pseg.Name, pseg.FileSize, qseg.FileSize))
-		}
-		if len(pseg.Data) != len(qseg.Data) {
-			errors = append(errors, fmt.Sprintf("segment %q len(Data) = %d, want %d", pseg.Name, len(pseg.Data), len(qseg.Data)))
-		} else if !bytes.Equal(pseg.Data, qseg.Data) {
-			errors = append(errors, fmt.Sprintf("segment %q Data mismatch:\n\thave %x\n\twant %x", pseg.Name, pseg.Data, qseg.Data))
-		}
-
-		for j := 0; j < len(pseg.Sections) || j < len(qseg.Sections); j++ {
-			if j >= len(pseg.Sections) {
-				errors = append(errors, fmt.Sprintf("segment %q missing section %q", pseg.Name, qseg.Sections[i].Name))
-				continue
-			}
-			if j >= len(qseg.Sections) {
-				errors = append(errors, fmt.Sprintf("segment %q has extra section %q", pseg.Name, pseg.Sections[i].Name))
-				continue
-			}
-			psect := pseg.Sections[j]
-			qsect := qseg.Sections[j]
-			if psect.Name != qsect.Name {
-				errors = append(errors, fmt.Sprintf("segment %q, section %d Name = %q, want %q", pseg.Name, j, psect.Name, qsect.Name))
-				continue // probably out of sync
-			}
-
-			if psect.VirtAddr != qsect.VirtAddr {
-				errors = append(errors, fmt.Sprintf("segment %q section %q VirtAddr = %#x, want %#x", pseg.Name, psect.Name, psect.VirtAddr, qsect.VirtAddr))
-			}
-			if psect.Size != qsect.Size {
-				errors = append(errors, fmt.Sprintf("segment %q section %q Size = %#x, want %#x", pseg.Name, psect.Name, psect.Size, qsect.Size))
-			}
-			if psect.Align != qsect.Align {
-				errors = append(errors, fmt.Sprintf("segment %q section %q Align = %#x, want %#x", pseg.Name, psect.Name, psect.Align, qsect.Align))
-			}
-		}
-	}
-
-	return errors
-}
-
-// cloneProg returns a deep copy of p.
-func cloneProg(p *Prog) *Prog {
-	q := new(Prog)
-	*q = *p
-	q.Segments = make([]*Segment, len(p.Segments))
-	for i, seg := range p.Segments {
-		q.Segments[i] = cloneSegment(seg)
-	}
-	return q
-}
-
-// cloneSegment returns a deep copy of seg.
-func cloneSegment(seg *Segment) *Segment {
-	t := new(Segment)
-	*t = *seg
-	t.Sections = make([]*Section, len(seg.Sections))
-	for i, sect := range seg.Sections {
-		t.Sections[i] = cloneSection(sect)
-	}
-	t.Data = make([]byte, len(seg.Data))
-	copy(t.Data, seg.Data)
-	return t
-}
-
-// cloneSection returns a deep copy of section.
-func cloneSection(sect *Section) *Section {
-	// At the moment, there's nothing we need to make a deep copy of.
-	t := new(Section)
-	*t = *sect
-	return t
-}
-
-const saveMismatch = true
-
-// checkGolden checks that data matches the named file.
-// If not, it reports the error to the test.
-func checkGolden(t *testing.T, data []byte, name string) {
-	golden := mustParseHexdumpFile(t, name)
-	if !bytes.Equal(data, golden) {
-		if saveMismatch {
-			ioutil.WriteFile(name+".raw", data, 0666)
-			ioutil.WriteFile(name+".hex", []byte(hexdump(data)), 0666)
-		}
-		// TODO(rsc): A better diff would be nice, as needed.
-		i := 0
-		for i < len(data) && i < len(golden) && data[i] == golden[i] {
-			i++
-		}
-		if i >= len(data) {
-			t.Errorf("%s: output file shorter than expected: have %d bytes, want %d", name, len(data), len(golden))
-		} else if i >= len(golden) {
-			t.Errorf("%s: output file larger than expected: have %d bytes, want %d", name, len(data), len(golden))
-		} else {
-			t.Errorf("%s: output file differs at byte %d: have %#02x, want %#02x", name, i, data[i], golden[i])
-		}
-	}
-}

src/cmd/link/runtime.go

@@ -1,28 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-// Generation of runtime-accessible data structures.
-// See also debug.go.
-
-package main
-
-import "debug/goobj"
-
-func (p *Prog) runtime() {
-	p.pclntab()
-
-	// TODO: Implement garbage collection data.
-	p.addSym(&Sym{
-		Sym: &goobj.Sym{
-			SymID: goobj.SymID{Name: "gcdata"},
-			Kind:  goobj.SRODATA,
-		},
-	})
-	p.addSym(&Sym{
-		Sym: &goobj.Sym{
-			SymID: goobj.SymID{Name: "gcbss"},
-			Kind:  goobj.SRODATA,
-		},
-	})
-}

src/cmd/link/scan.go

@@ -1,187 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-// Initial scan of packages making up a program.
-
-// TODO(rsc): Rename goobj.SymID.Version to StaticID to avoid confusion with the ELF meaning of version.
-// TODO(rsc): Fix file format so that SBSS/SNOPTRBSS with data is listed as SDATA/SNOPTRDATA.
-// TODO(rsc): Parallelize scan to overlap file i/o where possible.
-
-package main
-
-import (
-	"debug/goobj"
-	"os"
-	"sort"
-	"strings"
-)
-
-// scan scans all packages making up the program, starting with package main defined in mainfile.
-func (p *Prog) scan(mainfile string) {
-	p.initScan()
-	p.scanFile("main", mainfile)
-	if len(p.Missing) > 0 && !p.omitRuntime {
-		p.scanImport("runtime")
-	}
-
-	var missing []string
-	for sym := range p.Missing {
-		if !p.isAuto(sym) {
-			missing = append(missing, sym.String())
-		}
-	}
-
-	if missing != nil {
-		sort.Strings(missing)
-		for _, sym := range missing {
-			p.errorf("undefined: %s", sym)
-		}
-	}
-
-	// TODO(rsc): Walk import graph to diagnose cycles.
-}
-
-// initScan initializes the Prog fields needed by scan.
-func (p *Prog) initScan() {
-	p.Packages = make(map[string]*Package)
-	p.Syms = make(map[goobj.SymID]*Sym)
-	p.Missing = make(map[goobj.SymID]bool)
-	p.Missing[p.startSym] = true
-}
-
-// scanFile reads file to learn about the package with the given import path.
-func (p *Prog) scanFile(pkgpath string, file string) {
-	pkg := &Package{
-		File: file,
-	}
-	p.Packages[pkgpath] = pkg
-
-	f, err := os.Open(file)
-	if err != nil {
-		p.errorf("%v", err)
-		return
-	}
-	gp, err := goobj.Parse(f, pkgpath)
-	f.Close()
-	if err != nil {
-		p.errorf("reading %s: %v", file, err)
-		return
-	}
-
-	// TODO(rsc): Change debug/goobj to record package name as gp.Name.
-	// TODO(rsc): If pkgpath == "main", check that gp.Name == "main".
-
-	pkg.Package = gp
-
-	for _, gs := range gp.Syms {
-		// TODO(rsc): Fix file format instead of this workaround.
-		if gs.Data.Size > 0 {
-			switch gs.Kind {
-			case goobj.SBSS:
-				gs.Kind = goobj.SDATA
-			case goobj.SNOPTRBSS:
-				gs.Kind = goobj.SNOPTRDATA
-			}
-		}
-
-		if gs.Version != 0 {
-			gs.Version += p.MaxVersion
-		}
-		for i := range gs.Reloc {
-			r := &gs.Reloc[i]
-			if r.Sym.Version != 0 {
-				r.Sym.Version += p.MaxVersion
-			}
-			if p.Syms[r.Sym] == nil {
-				p.Missing[r.Sym] = true
-			}
-		}
-		if gs.Func != nil {
-			for i := range gs.Func.FuncData {
-				fdata := &gs.Func.FuncData[i]
-				if fdata.Sym.Name != "" {
-					if fdata.Sym.Version != 0 {
-						fdata.Sym.Version += p.MaxVersion
-					}
-					if p.Syms[fdata.Sym] == nil {
-						p.Missing[fdata.Sym] = true
-					}
-				}
-			}
-		}
-		if old := p.Syms[gs.SymID]; old != nil {
-			// Duplicate definition of symbol. Is it okay?
-			// TODO(rsc): Write test for this code.
-			switch {
-			// If both symbols are BSS (no data), take max of sizes
-			// but otherwise ignore second symbol.
-			case old.Data.Size == 0 && gs.Data.Size == 0:
-				if old.Size < gs.Size {
-					old.Size = gs.Size
-				}
-				continue
-
-			// If one is in BSS and one is not, use the one that is not.
-			case old.Data.Size > 0 && gs.Data.Size == 0:
-				continue
-			case gs.Data.Size > 0 && old.Data.Size == 0:
-				break // install gs as new symbol below
-
-			// If either is marked as DupOK, we can keep either one.
-			// Keep the one that we saw first.
-			case old.DupOK || gs.DupOK:
-				continue
-
-			// Otherwise, there's an actual conflict:
-			default:
-				p.errorf("symbol %s defined in both %s and %s %v %v", gs.SymID, old.Package.File, file, old.Data, gs.Data)
-				continue
-			}
-		}
-		s := &Sym{
-			Sym:     gs,
-			Package: pkg,
-		}
-		p.addSym(s)
-		delete(p.Missing, gs.SymID)
-
-		if s.Data.Size > int64(s.Size) {
-			p.errorf("%s: initialized data larger than symbol (%d > %d)", s, s.Data.Size, s.Size)
-		}
-	}
-	p.MaxVersion += pkg.MaxVersion
-
-	for i, pkgpath := range pkg.Imports {
-		// TODO(rsc): Fix file format to drop .a from recorded import path.
-		pkgpath = strings.TrimSuffix(pkgpath, ".a")
-		pkg.Imports[i] = pkgpath
-
-		p.scanImport(pkgpath)
-	}
-}
-
-func (p *Prog) addSym(s *Sym) {
-	pkg := s.Package
-	if pkg == nil {
-		pkg = p.Packages[""]
-		if pkg == nil {
-			pkg = &Package{}
-			p.Packages[""] = pkg
-		}
-		s.Package = pkg
-	}
-	pkg.Syms = append(pkg.Syms, s)
-	p.Syms[s.SymID] = s
-	p.SymOrder = append(p.SymOrder, s)
-}
-
-// scanImport finds the object file for the given import path and then scans it.
-func (p *Prog) scanImport(pkgpath string) {
-	if p.Packages[pkgpath] != nil {
-		return // already loaded
-	}
-
-	// TODO(rsc): Implement correct search to find file.
-	p.scanFile(pkgpath, p.pkgdir+"/"+pkgpath+".a")
-}

src/cmd/link/testdata/Makefile

@@ -1,15 +0,0 @@
-ALL=\
-	autosection.6\
-	autoweak.6\
-	dead.6\
-	hello.6\
-	layout.6\
-	pclntab.6\
-
-all: $(ALL)
-
-%.6: %.s
-	go tool 6a $*.s
-
-pclntab.s: genpcln.go
-	go run genpcln.go >pclntab.s

src/cmd/link/testdata/autosection.s

@@ -1,60 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-// Test of section-named symbols.
-
-#include "../../ld/textflag.h"
-
-TEXT start(SB),7,$0
-	MOVQ $autotab(SB),AX
-	MOVQ $autoptr(SB),AX
-	RET
-
-GLOBL zero(SB), $8
-
-GLOBL zeronoptr(SB), NOPTR, $16
-
-// text
-DATA autotab+0x00(SB)/8, $text(SB)
-DATA autotab+0x08(SB)/8, $start(SB)
-DATA autotab+0x10(SB)/8, $etext(SB)
-DATA autotab+0x18(SB)/8, $start+16(SB)
-
-// data
-DATA autotab+0x20(SB)/8, $data(SB)
-DATA autotab+0x28(SB)/8, $autotab(SB)
-DATA autotab+0x30(SB)/8, $edata(SB)
-DATA autotab+0x38(SB)/8, $nonzero+4(SB)
-
-// bss
-DATA autotab+0x40(SB)/8, $bss(SB)
-DATA autotab+0x48(SB)/8, $zero(SB)
-DATA autotab+0x50(SB)/8, $ebss(SB)
-DATA autotab+0x58(SB)/8, $zero+8(SB)
-
-// noptrdata
-DATA autotab+0x60(SB)/8, $noptrdata(SB)
-DATA autotab+0x68(SB)/8, $nonzeronoptr(SB)
-DATA autotab+0x70(SB)/8, $enoptrdata(SB)
-DATA autotab+0x78(SB)/8, $nonzeronoptr+8(SB)
-
-// noptrbss
-DATA autotab+0x80(SB)/8, $noptrbss(SB)
-DATA autotab+0x88(SB)/8, $zeronoptr(SB)
-DATA autotab+0x90(SB)/8, $enoptrbss(SB)
-DATA autotab+0x98(SB)/8, $zeronoptr+16(SB)
-
-// end
-DATA autotab+0xa0(SB)/8, $end(SB)
-DATA autotab+0xa8(SB)/8, $zeronoptr+16(SB)
-
-GLOBL autotab(SB), $0xb0
-
-DATA nonzero(SB)/4, $1
-GLOBL nonzero(SB), $4
-
-DATA nonzeronoptr(SB)/8, $2
-GLOBL nonzeronoptr(SB), NOPTR, $8
-
-GLOBL autoptr(SB), $0

src/cmd/link/testdata/autoweak.s

@@ -1,30 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-// Test of go.weak symbols.
-
-TEXT start(SB),7,$0
-	MOVQ $autotab(SB),AX
-	MOVQ $autoptr(SB),AX
-	RET
-
-// go.weak.sym should resolve to sym, because sym is in the binary.
-DATA autotab+0(SB)/8, $go·weak·sym(SB)
-DATA autotab+8(SB)/8, $sym(SB)
-
-// go.weak.missingsym should resolve to 0, because missingsym is not in the binary.
-DATA autotab+16(SB)/8, $go·weak·missingsym(SB)
-DATA autotab+24(SB)/8, $0
-
-// go.weak.deadsym should resolve to 0, because deadsym is discarded during dead code removal
-DATA autotab+32(SB)/8, $go·weak·deadsym(SB)
-DATA autotab+40(SB)/8, $0
-
-GLOBL autotab(SB), $48
-
-GLOBL sym(SB), $1
-
-GLOBL deadsym(SB), $1
-
-GLOBL autoptr(SB), $0

src/cmd/link/testdata/dead.s

@@ -1,49 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-// Test of dead code removal.
-// Symbols with names beginning with dead_ should be discarded.
-// Others should be kept.
-
-TEXT start(SB),7,$0	// start symbol
-	MOVQ $data1<>(SB), AX
-	CALL text1(SB)
-	MOVQ $text2(SB), BX
-	RET
-
-TEXT text1(SB),7,$0
-	FUNCDATA $1, funcdata+4(SB)
-	RET
-
-TEXT text2(SB),7,$0
-	MOVQ $edata(SB),BX
-	RET
-
-DATA data1<>+0(SB)/8, $data2(SB)
-DATA data1<>+8(SB)/8, $data3(SB)
-GLOBL data1<>(SB), $16
-GLOBL data2(SB), $1
-GLOBL data3(SB), $1
-GLOBL funcdata(SB), $8
-
-TEXT dead_start(SB),7,$0
-	MOVQ $dead_data1(SB), AX
-	CALL dead_text1(SB)
-	MOVQ $dead_text2(SB), BX
-	RET
-
-TEXT dead_text1(SB),7,$0
-	FUNCDATA $1, dead_funcdata+4(SB)
-	RET
-
-TEXT dead_text2(SB),7,$0
-	RET
-
-DATA dead_data1+0(SB)/8, $dead_data2(SB)
-DATA dead_data1+8(SB)/8, $dead_data3(SB)
-GLOBL dead_data1(SB), $16
-GLOBL dead_data2(SB), $1
-GLOBL dead_data3(SB), $1
-GLOBL dead_funcdata(SB), $8
-

src/cmd/link/testdata/genpcln.go

@@ -1,112 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-// This program generates a .s file using a pseudorandom
-// value stream for the runtime function data.
-// The pclntab test checks that the linked copy
-// still has the same pseudorandom value stream.
-
-package main
-
-import (
-	"fmt"
-	"math/rand"
-)
-
-func main() {
-	fmt.Printf("// generated by genpcln.go; do not edit\n\n")
-	for f := 0; f < 3; f++ {
-		r := rand.New(rand.NewSource(int64(f)))
-		file := "input"
-		line := 1
-		args := r.Intn(100) * 8
-		frame := 32 + r.Intn(32)/8*8
-		fmt.Printf("#line %d %q\n", line, file)
-		fmt.Printf("TEXT func%d(SB),7,$%d-%d\n", f, frame, args)
-		fmt.Printf("\tFUNCDATA $1, funcdata%d(SB)\n", f)
-		fmt.Printf("#line %d %q\n", line, file)
-		size := 200 + r.Intn(100)*8
-		spadj := 0
-		flushed := 0
-		firstpc := 4
-		flush := func(i int) {
-			for i-flushed >= 10 {
-				fmt.Printf("#line %d %q\n", line, file)
-				fmt.Printf("/*%#04x*/\tMOVQ $0x123456789, AX\n", firstpc+flushed)
-				flushed += 10
-			}
-			for i-flushed >= 5 {
-				fmt.Printf("#line %d %q\n", line, file)
-				fmt.Printf("/*%#04x*/\tMOVL $0x1234567, AX\n", firstpc+flushed)
-				flushed += 5
-			}
-			for i-flushed > 0 {
-				fmt.Printf("#line %d %q\n", line, file)
-				fmt.Printf("/*%#04x*/\tBYTE $0\n", firstpc+flushed)
-				flushed++
-			}
-		}
-		for i := 0; i < size; i++ {
-			// Possible SP adjustment.
-			if r.Intn(100) == 0 {
-				flush(i)
-				fmt.Printf("#line %d %q\n", line, file)
-				if spadj <= -32 || spadj < 32 && r.Intn(2) == 0 {
-					spadj += 8
-					fmt.Printf("/*%#04x*/\tPUSHQ AX\n", firstpc+i)
-				} else {
-					spadj -= 8
-					fmt.Printf("/*%#04x*/\tPOPQ AX\n", firstpc+i)
-				}
-				i += 1
-				flushed = i
-			}
-
-			// Possible PCFile change.
-			if r.Intn(100) == 0 {
-				flush(i)
-				file = fmt.Sprintf("file%d.s", r.Intn(10))
-				line = r.Intn(100) + 1
-			}
-
-			// Possible PCLine change.
-			if r.Intn(10) == 0 {
-				flush(i)
-				line = r.Intn(1000) + 1
-			}
-
-			// Possible PCData $1 change.
-			if r.Intn(100) == 0 {
-				flush(i)
-				fmt.Printf("/*%6s*/\tPCDATA $1, $%d\n", "", r.Intn(1000))
-			}
-
-			// Possible PCData $2 change.
-			if r.Intn(100) == 0 {
-				flush(i)
-				fmt.Printf("/*%6s*/\tPCDATA $2, $%d\n", "", r.Intn(1000))
-			}
-		}
-		flush(size)
-		for spadj < 0 {
-			fmt.Printf("\tPUSHQ AX\n")
-			spadj += 8
-		}
-		for spadj > 0 {
-			fmt.Printf("\tPOPQ AX\n")
-			spadj -= 8
-		}
-		fmt.Printf("\tRET\n")
-
-		fmt.Printf("\n")
-		fmt.Printf("GLOBL funcdata%d(SB), $16\n", f)
-	}
-
-	fmt.Printf("\nTEXT start(SB),7,$0\n")
-	for f := 0; f < 3; f++ {
-		fmt.Printf("\tCALL func%d(SB)\n", f)
-	}
-	fmt.Printf("\tMOVQ $pclntab(SB), AX\n")
-	fmt.Printf("\n\tRET\n")
-}

src/cmd/link/testdata/hello.s

@@ -1,15 +0,0 @@
-TEXT _rt0_go(SB),7,$0
-	MOVL $1, DI
-	MOVL $hello<>(SB), SI
-	MOVL $12, DX
-	MOVL $0x2000004, AX
-	SYSCALL
-	MOVL $0, DI
-	MOVL $0x2000001, AX
-	SYSCALL
-	RET
-
-DATA hello<>+0(SB)/4, $"hell"
-DATA hello<>+4(SB)/4, $"o wo"
-DATA hello<>+8(SB)/4, $"rld\n"
-GLOBL hello<>(SB), $12

src/cmd/link/testdata/layout.s

@@ -1,29 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-// Test of section assignment in layout.go.
-// Each symbol should end up in the section named by the symbol name prefix (up to the underscore).
-
-#include "../../ld/textflag.h"
-
-TEXT text_start(SB),7,$0
-	MOVQ $rodata_sym(SB), AX
-	MOVQ $noptrdata_sym(SB), AX
-	MOVQ $data_sym(SB), AX
-	MOVQ $bss_sym(SB), AX
-	MOVQ $noptrbss_sym(SB), AX
-	RET
-
-DATA rodata_sym(SB)/4, $1
-GLOBL rodata_sym(SB), RODATA, $4
-
-DATA noptrdata_sym(SB)/4, $1
-GLOBL noptrdata_sym(SB), NOPTR, $4
-
-DATA data_sym(SB)/4, $1
-GLOBL data_sym(SB), $4
-
-GLOBL bss_sym(SB), $4
-
-GLOBL noptrbss_sym(SB), NOPTR, $4

src/cmd/link/testdata/link.hello.darwin.amd64

@@ -1,57 +0,0 @@
-00000000  cf fa ed fe 07 00 00 01  03 00 00 00 02 00 00 00  |................|
-00000010  04 00 00 00 d0 02 00 00  01 00 00 00 00 00 00 00  |................|
-00000020  19 00 00 00 48 00 00 00  5f 5f 50 41 47 45 5a 45  |....H...__PAGEZE|
-00000030  52 4f 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |RO..............|
-00000040  00 10 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-*
-00000060  00 00 00 00 00 00 00 00  19 00 00 00 38 01 00 00  |............8...|
-00000070  5f 5f 54 45 58 54 00 00  00 00 00 00 00 00 00 00  |__TEXT..........|
-00000080  00 10 00 00 00 00 00 00  d6 10 00 00 00 00 00 00  |................|
-00000090  00 00 00 00 00 00 00 00  d6 10 00 00 00 00 00 00  |................|
-000000a0  07 00 00 00 05 00 00 00  03 00 00 00 00 00 00 00  |................|
-000000b0  5f 5f 74 65 78 74 00 00  00 00 00 00 00 00 00 00  |__text..........|
-000000c0  5f 5f 54 45 58 54 00 00  00 00 00 00 00 00 00 00  |__TEXT..........|
-000000d0  00 20 00 00 00 00 00 00  20 00 00 00 00 00 00 00  |. ...... .......|
-000000e0  00 10 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-000000f0  00 04 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-00000100  5f 5f 72 6f 64 61 74 61  00 00 00 00 00 00 00 00  |__rodata........|
-00000110  5f 5f 54 45 58 54 00 00  00 00 00 00 00 00 00 00  |__TEXT..........|
-00000120  20 20 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |  ..............|
-00000130  20 10 00 00 00 00 00 00  00 00 00 00 00 00 00 00  | ...............|
-*
-00000150  5f 5f 66 75 6e 63 74 61  62 00 00 00 00 00 00 00  |__functab.......|
-00000160  5f 5f 54 45 58 54 00 00  00 00 00 00 00 00 00 00  |__TEXT..........|
-00000170  20 20 00 00 00 00 00 00  b6 00 00 00 00 00 00 00  |  ..............|
-00000180  20 10 00 00 00 00 00 00  00 00 00 00 00 00 00 00  | ...............|
-*
-000001a0  19 00 00 00 98 00 00 00  5f 5f 44 41 54 41 00 00  |........__DATA..|
-000001b0  00 00 00 00 00 00 00 00  00 30 00 00 00 00 00 00  |.........0......|
-000001c0  0c 00 00 00 00 00 00 00  00 20 00 00 00 00 00 00  |......... ......|
-000001d0  0c 00 00 00 00 00 00 00  03 00 00 00 03 00 00 00  |................|
-000001e0  01 00 00 00 00 00 00 00  5f 5f 64 61 74 61 00 00  |........__data..|
-000001f0  00 00 00 00 00 00 00 00  5f 5f 44 41 54 41 00 00  |........__DATA..|
-00000200  00 00 00 00 00 00 00 00  00 30 00 00 00 00 00 00  |.........0......|
-00000210  0c 00 00 00 00 00 00 00  00 20 00 00 00 00 00 00  |......... ......|
-*
-00000230  00 00 00 00 00 00 00 00  05 00 00 00 b8 00 00 00  |................|
-00000240  04 00 00 00 2a 00 00 00  00 00 00 00 00 00 00 00  |....*...........|
-*
-000002c0  00 00 00 00 00 00 00 00  00 20 00 00 00 00 00 00  |......... ......|
-*
-00001000  bf 01 00 00 00 be 00 30  00 00 ba 0c 00 00 00 b8  |.......0........|
-00001010  04 00 00 02 0f 05 31 ff  b8 01 00 00 02 0f 05 c3  |......1.........|
-00001020  fb ff ff ff 00 00 01 08  01 00 00 00 00 00 00 00  |................|
-00001030  00 20 00 00 00 00 00 00  30 00 00 00 00 00 00 00  |. ......0.......|
-00001040  20 20 00 00 00 00 00 00  80 00 00 00 00 00 00 00  |  ..............|
-00001050  00 20 00 00 00 00 00 00  58 00 00 00 00 00 00 80  |. ......X.......|
-00001060  08 00 00 00 60 00 00 00  63 00 00 00 66 00 00 00  |....`...c...f...|
-00001070  00 00 00 00 00 00 00 00  5f 72 74 30 5f 67 6f 00  |........_rt0_go.|
-00001080  02 20 00 04 20 00 06 05  02 05 02 05 02 05 02 02  |. .. ...........|
-00001090  02 02 02 05 02 02 02 01  00 00 00 00 00 00 00 00  |................|
-000010a0  02 00 00 00 88 00 00 00  2f 55 73 65 72 73 2f 72  |......../Users/r|
-000010b0  73 63 2f 67 2f 67 6f 2f  73 72 63 2f 63 6d 64 2f  |sc/g/go/src/cmd/|
-000010c0  6c 69 6e 6b 2f 74 65 73  74 64 61 74 61 2f 68 65  |link/testdata/he|
-000010d0  6c 6c 6f 2e 73 00 00 00  00 00 00 00 00 00 00 00  |llo.s...........|
-*
-00002000  68 65 6c 6c 6f 20 77 6f  72 6c 64 0a              |hello world.|
-0000200c

src/cmd/link/testdata/macho.amd64.exit9

@@ -1,24 +0,0 @@
-00000000  cf fa ed fe 07 00 00 01  03 00 00 00 02 00 00 00  |................|
-00000010  03 00 00 00 98 01 00 00  01 00 00 00 00 00 00 00  |................|
-00000020  19 00 00 00 48 00 00 00  5f 5f 50 41 47 45 5a 45  |....H...__PAGEZE|
-00000030  52 4f 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |RO..............|
-00000040  00 10 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-00000050  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-00000060  00 00 00 00 00 00 00 00  19 00 00 00 98 00 00 00  |................|
-00000070  5f 5f 54 45 58 54 00 00  00 00 00 00 00 00 00 00  |__TEXT..........|
-00000080  00 10 00 00 00 00 00 00  0d 10 00 00 00 00 00 00  |................|
-00000090  00 00 00 00 00 00 00 00  0d 10 00 00 00 00 00 00  |................|
-000000a0  07 00 00 00 05 00 00 00  01 00 00 00 00 00 00 00  |................|
-000000b0  5f 5f 74 65 78 74 00 00  00 00 00 00 00 00 00 00  |__text..........|
-000000c0  5f 5f 54 45 58 54 00 00  00 00 00 00 00 00 00 00  |__TEXT..........|
-000000d0  00 20 00 00 00 00 00 00  0d 00 00 00 00 00 00 00  |. ..............|
-000000e0  00 10 00 00 06 00 00 00  00 00 00 00 00 00 00 00  |................|
-000000f0  00 04 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-00000100  05 00 00 00 b8 00 00 00  04 00 00 00 2a 00 00 00  |............*...|
-00000110  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-*
-00000190  00 20 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |. ..............|
-000001a0  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-*
-00001000  b8 01 00 00 02 bf 09 00  00 00 0f 05 f4           |.............|
-0000100d

src/cmd/link/testdata/macho.amd64.hello

@@ -1,39 +0,0 @@
-00000000  cf fa ed fe 07 00 00 01  03 00 00 00 02 00 00 00  |................|
-00000010  04 00 00 00 30 02 00 00  01 00 00 00 00 00 00 00  |....0...........|
-00000020  19 00 00 00 48 00 00 00  5f 5f 50 41 47 45 5a 45  |....H...__PAGEZE|
-00000030  52 4f 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |RO..............|
-00000040  00 10 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-00000050  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-00000060  00 00 00 00 00 00 00 00  19 00 00 00 98 00 00 00  |................|
-00000070  5f 5f 54 45 58 54 00 00  00 00 00 00 00 00 00 00  |__TEXT..........|
-00000080  00 10 00 00 00 00 00 00  23 10 00 00 00 00 00 00  |........#.......|
-00000090  00 00 00 00 00 00 00 00  23 10 00 00 00 00 00 00  |........#.......|
-000000a0  07 00 00 00 05 00 00 00  01 00 00 00 00 00 00 00  |................|
-000000b0  5f 5f 74 65 78 74 00 00  00 00 00 00 00 00 00 00  |__text..........|
-000000c0  5f 5f 54 45 58 54 00 00  00 00 00 00 00 00 00 00  |__TEXT..........|
-000000d0  00 20 00 00 00 00 00 00  23 00 00 00 00 00 00 00  |. ......#.......|
-000000e0  00 10 00 00 06 00 00 00  00 00 00 00 00 00 00 00  |................|
-000000f0  00 04 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-00000100  19 00 00 00 98 00 00 00  5f 5f 44 41 54 41 00 00  |........__DATA..|
-00000110  00 00 00 00 00 00 00 00  00 30 00 00 00 00 00 00  |.........0......|
-00000120  0c 00 00 00 00 00 00 00  00 20 00 00 00 00 00 00  |......... ......|
-00000130  0c 00 00 00 00 00 00 00  03 00 00 00 03 00 00 00  |................|
-00000140  01 00 00 00 00 00 00 00  5f 5f 64 61 74 61 00 00  |........__data..|
-00000150  00 00 00 00 00 00 00 00  5f 5f 44 41 54 41 00 00  |........__DATA..|
-00000160  00 00 00 00 00 00 00 00  00 30 00 00 00 00 00 00  |.........0......|
-00000170  0c 00 00 00 00 00 00 00  00 20 00 00 06 00 00 00  |......... ......|
-00000180  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-00000190  00 00 00 00 00 00 00 00  05 00 00 00 b8 00 00 00  |................|
-000001a0  04 00 00 00 2a 00 00 00  00 00 00 00 00 00 00 00  |....*...........|
-000001b0  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-*
-00000220  00 00 00 00 00 00 00 00  00 20 00 00 00 00 00 00  |......... ......|
-00000230  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-*
-00001000  b8 04 00 00 02 bf 01 00  00 00 be 00 30 00 00 ba  |............0...|
-00001010  0c 00 00 00 0f 05 b8 01  00 00 02 bf 09 00 00 00  |................|
-00001020  0f 05 f4 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-00001030  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-*
-00002000  68 65 6c 6c 6f 20 77 6f  72 6c 64 0a              |hello world.|
-0000200c

src/cmd/link/testdata/macho.amd64.helloro

@@ -1,34 +0,0 @@
-00000000  cf fa ed fe 07 00 00 01  03 00 00 00 02 00 00 00  |................|
-00000010  03 00 00 00 e8 01 00 00  01 00 00 00 00 00 00 00  |................|
-00000020  19 00 00 00 48 00 00 00  5f 5f 50 41 47 45 5a 45  |....H...__PAGEZE|
-00000030  52 4f 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |RO..............|
-00000040  00 10 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-00000050  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-00000060  00 00 00 00 00 00 00 00  19 00 00 00 e8 00 00 00  |................|
-00000070  5f 5f 54 45 58 54 00 00  00 00 00 00 00 00 00 00  |__TEXT..........|
-00000080  00 10 00 00 00 00 00 00  0c 20 00 00 00 00 00 00  |......... ......|
-00000090  00 00 00 00 00 00 00 00  0c 20 00 00 00 00 00 00  |......... ......|
-000000a0  07 00 00 00 05 00 00 00  02 00 00 00 00 00 00 00  |................|
-000000b0  5f 5f 74 65 78 74 00 00  00 00 00 00 00 00 00 00  |__text..........|
-000000c0  5f 5f 54 45 58 54 00 00  00 00 00 00 00 00 00 00  |__TEXT..........|
-000000d0  00 20 00 00 00 00 00 00  23 00 00 00 00 00 00 00  |. ......#.......|
-000000e0  00 10 00 00 06 00 00 00  00 00 00 00 00 00 00 00  |................|
-000000f0  00 04 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-00000100  5f 5f 72 6f 64 61 74 61  00 00 00 00 00 00 00 00  |__rodata........|
-00000110  5f 5f 54 45 58 54 00 00  00 00 00 00 00 00 00 00  |__TEXT..........|
-00000120  00 30 00 00 00 00 00 00  0c 00 00 00 00 00 00 00  |.0..............|
-00000130  00 20 00 00 06 00 00 00  00 00 00 00 00 00 00 00  |. ..............|
-00000140  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-00000150  05 00 00 00 b8 00 00 00  04 00 00 00 2a 00 00 00  |............*...|
-00000160  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-*
-000001e0  00 20 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |. ..............|
-000001f0  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-*
-00001000  b8 04 00 00 02 bf 01 00  00 00 be 00 30 00 00 ba  |............0...|
-00001010  0c 00 00 00 0f 05 b8 01  00 00 02 bf 00 00 00 00  |................|
-00001020  0f 05 f4 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-00001030  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
-*
-00002000  68 65 6c 6c 6f 20 77 6f  72 6c 64 0a              |hello world.|
-0000200c

src/cmd/link/util.go

@@ -1,11 +0,0 @@
-// Copyright 2014 The Go Authors.  All rights reserved.
-// Use of this source code is governed by a BSD-style
-// license that can be found in the LICENSE file.
-
-package main
-
-// round returns size rounded up to the next multiple of align;
-// align must be a power of two.
-func round(size, align Addr) Addr {
-	return (size + align - 1) &^ (align - 1)
-}

src/cmd/nm/debug_goobj.go

@@ -1,3 +1,8 @@
+// DO NOT EDIT. Generated by code.google.com/p/rsc/cmd/bundle
+// bundle -p main -x goobj_ debug/goobj
+
+/* read.go */
+
 // Copyright 2013 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.
@@ -7,7 +12,8 @@
 // TODO(rsc): Decide where this package should live. (golang.org/issue/6932)
 // TODO(rsc): Decide the appropriate integer types for various fields.
 // TODO(rsc): Write tests. (File format still up in the air a little.)
-package goobj
+
+package main
 
 import (
 	"bufio"
@@ -20,7 +26,7 @@ import (
 )
 
 // A SymKind describes the kind of memory represented by a symbol.
-type SymKind int
+type goobj_SymKind int
 
 // This list is taken from include/link.h.
 
@@ -28,108 +34,108 @@ type SymKind int
 // TODO(rsc): Give idiomatic Go names.
 // TODO(rsc): Reduce the number of symbol types in the object files.
 const (
-	_ SymKind = iota
+	_ goobj_SymKind = iota
 
 	// readonly, executable
-	STEXT
-	SELFRXSECT
+	goobj_STEXT
+	goobj_SELFRXSECT
 
 	// readonly, non-executable
-	STYPE
-	SSTRING
-	SGOSTRING
-	SGOFUNC
-	SRODATA
-	SFUNCTAB
-	STYPELINK
-	SSYMTAB // TODO: move to unmapped section
-	SPCLNTAB
-	SELFROSECT
+	goobj_STYPE
+	goobj_SSTRING
+	goobj_SGOSTRING
+	goobj_SGOFUNC
+	goobj_SRODATA
+	goobj_SFUNCTAB
+	goobj_STYPELINK
+	goobj_SSYMTAB // TODO: move to unmapped section
+	goobj_SPCLNTAB
+	goobj_SELFROSECT
 
 	// writable, non-executable
-	SMACHOPLT
-	SELFSECT
-	SMACHO // Mach-O __nl_symbol_ptr
-	SMACHOGOT
-	SNOPTRDATA
-	SINITARR
-	SDATA
-	SWINDOWS
-	SBSS
-	SNOPTRBSS
-	STLSBSS
+	goobj_SMACHOPLT
+	goobj_SELFSECT
+	goobj_SMACHO // Mach-O __nl_symbol_ptr
+	goobj_SMACHOGOT
+	goobj_SNOPTRDATA
+	goobj_SINITARR
+	goobj_SDATA
+	goobj_SWINDOWS
+	goobj_SBSS
+	goobj_SNOPTRBSS
+	goobj_STLSBSS
 
 	// not mapped
-	SXREF
-	SMACHOSYMSTR
-	SMACHOSYMTAB
-	SMACHOINDIRECTPLT
-	SMACHOINDIRECTGOT
-	SFILE
-	SFILEPATH
-	SCONST
-	SDYNIMPORT
-	SHOSTOBJ
+	goobj_SXREF
+	goobj_SMACHOSYMSTR
+	goobj_SMACHOSYMTAB
+	goobj_SMACHOINDIRECTPLT
+	goobj_SMACHOINDIRECTGOT
+	goobj_SFILE
+	goobj_SFILEPATH
+	goobj_SCONST
+	goobj_SDYNIMPORT
+	goobj_SHOSTOBJ
 )
 
-var symKindStrings = []string{
-	SBSS:              "SBSS",
-	SCONST:            "SCONST",
-	SDATA:             "SDATA",
-	SDYNIMPORT:        "SDYNIMPORT",
-	SELFROSECT:        "SELFROSECT",
-	SELFRXSECT:        "SELFRXSECT",
-	SELFSECT:          "SELFSECT",
-	SFILE:             "SFILE",
-	SFILEPATH:         "SFILEPATH",
-	SFUNCTAB:          "SFUNCTAB",
-	SGOFUNC:           "SGOFUNC",
-	SGOSTRING:         "SGOSTRING",
-	SHOSTOBJ:          "SHOSTOBJ",
-	SINITARR:          "SINITARR",
-	SMACHO:            "SMACHO",
-	SMACHOGOT:         "SMACHOGOT",
-	SMACHOINDIRECTGOT: "SMACHOINDIRECTGOT",
-	SMACHOINDIRECTPLT: "SMACHOINDIRECTPLT",
-	SMACHOPLT:         "SMACHOPLT",
-	SMACHOSYMSTR:      "SMACHOSYMSTR",
-	SMACHOSYMTAB:      "SMACHOSYMTAB",
-	SNOPTRBSS:         "SNOPTRBSS",
-	SNOPTRDATA:        "SNOPTRDATA",
-	SPCLNTAB:          "SPCLNTAB",
-	SRODATA:           "SRODATA",
-	SSTRING:           "SSTRING",
-	SSYMTAB:           "SSYMTAB",
-	STEXT:             "STEXT",
-	STLSBSS:           "STLSBSS",
-	STYPE:             "STYPE",
-	STYPELINK:         "STYPELINK",
-	SWINDOWS:          "SWINDOWS",
-	SXREF:             "SXREF",
+var goobj_symKindStrings = []string{
+	goobj_SBSS:              "SBSS",
+	goobj_SCONST:            "SCONST",
+	goobj_SDATA:             "SDATA",
+	goobj_SDYNIMPORT:        "SDYNIMPORT",
+	goobj_SELFROSECT:        "SELFROSECT",
+	goobj_SELFRXSECT:        "SELFRXSECT",
+	goobj_SELFSECT:          "SELFSECT",
+	goobj_SFILE:             "SFILE",
+	goobj_SFILEPATH:         "SFILEPATH",
+	goobj_SFUNCTAB:          "SFUNCTAB",
+	goobj_SGOFUNC:           "SGOFUNC",
+	goobj_SGOSTRING:         "SGOSTRING",
+	goobj_SHOSTOBJ:          "SHOSTOBJ",
+	goobj_SINITARR:          "SINITARR",
+	goobj_SMACHO:            "SMACHO",
+	goobj_SMACHOGOT:         "SMACHOGOT",
+	goobj_SMACHOINDIRECTGOT: "SMACHOINDIRECTGOT",
+	goobj_SMACHOINDIRECTPLT: "SMACHOINDIRECTPLT",
+	goobj_SMACHOPLT:         "SMACHOPLT",
+	goobj_SMACHOSYMSTR:      "SMACHOSYMSTR",
+	goobj_SMACHOSYMTAB:      "SMACHOSYMTAB",
+	goobj_SNOPTRBSS:         "SNOPTRBSS",
+	goobj_SNOPTRDATA:        "SNOPTRDATA",
+	goobj_SPCLNTAB:          "SPCLNTAB",
+	goobj_SRODATA:           "SRODATA",
+	goobj_SSTRING:           "SSTRING",
+	goobj_SSYMTAB:           "SSYMTAB",
+	goobj_STEXT:             "STEXT",
+	goobj_STLSBSS:           "STLSBSS",
+	goobj_STYPE:             "STYPE",
+	goobj_STYPELINK:         "STYPELINK",
+	goobj_SWINDOWS:          "SWINDOWS",
+	goobj_SXREF:             "SXREF",
 }
 
-func (k SymKind) String() string {
-	if k < 0 || int(k) >= len(symKindStrings) {
+func (k goobj_SymKind) String() string {
+	if k < 0 || int(k) >= len(goobj_symKindStrings) {
 		return fmt.Sprintf("SymKind(%d)", k)
 	}
-	return symKindStrings[k]
+	return goobj_symKindStrings[k]
 }
 
 // A Sym is a named symbol in an object file.
-type Sym struct {
-	SymID         // symbol identifier (name and version)
-	Kind  SymKind // kind of symbol
-	DupOK bool    // are duplicate definitions okay?
-	Size  int     // size of corresponding data
-	Type  SymID   // symbol for Go type information
-	Data  Data    // memory image of symbol
-	Reloc []Reloc // relocations to apply to Data
-	Func  *Func   // additional data for functions
+type goobj_Sym struct {
+	goobj_SymID               // symbol identifier (name and version)
+	Kind        goobj_SymKind // kind of symbol
+	DupOK       bool          // are duplicate definitions okay?
+	Size        int           // size of corresponding data
+	Type        goobj_SymID   // symbol for Go type information
+	Data        goobj_Data    // memory image of symbol
+	Reloc       []goobj_Reloc // relocations to apply to Data
+	Func        *goobj_Func   // additional data for functions
 }
 
 // A SymID - the combination of Name and Version - uniquely identifies
 // a symbol within a package.
-type SymID struct {
+type goobj_SymID struct {
 	// Name is the name of a symbol.
 	Name string
 
@@ -141,7 +147,7 @@ type SymID struct {
 	Version int
 }
 
-func (s SymID) String() string {
+func (s goobj_SymID) String() string {
 	if s.Version == 0 {
 		return s.Name
 	}
@@ -151,20 +157,20 @@ func (s SymID) String() string {
 // A Data is a reference to data stored in an object file.
 // It records the offset and size of the data, so that a client can
 // read the data only if necessary.
-type Data struct {
+type goobj_Data struct {
 	Offset int64
 	Size   int64
 }
 
 // A Reloc describes a relocation applied to a memory image to refer
 // to an address within a particular symbol.
-type Reloc struct {
+type goobj_Reloc struct {
 	// The bytes at [Offset, Offset+Size) within the memory image
 	// should be updated to refer to the address Add bytes after the start
 	// of the symbol Sym.
 	Offset int
 	Size   int
-	Sym    SymID
+	Sym    goobj_SymID
 	Add    int
 
 	// The Type records the form of address expected in the bytes
@@ -175,7 +181,7 @@ type Reloc struct {
 
 // A Var describes a variable in a function stack frame: a declared
 // local variable, an input argument, or an output result.
-type Var struct {
+type goobj_Var struct {
 	// The combination of Name, Kind, and Offset uniquely
 	// identifies a variable in a function stack frame.
 	// Using fewer of these - in particular, using only Name - does not.
@@ -183,57 +189,57 @@ type Var struct {
 	Kind   int    // TODO(rsc): Define meaning.
 	Offset int    // Frame offset. TODO(rsc): Define meaning.
 
-	Type SymID // Go type for variable.
+	Type goobj_SymID // Go type for variable.
 }
 
 // Func contains additional per-symbol information specific to functions.
-type Func struct {
-	Args     int        // size in bytes of argument frame: inputs and outputs
-	Frame    int        // size in bytes of local variable frame
-	Leaf     bool       // function omits save of link register (ARM)
-	NoSplit  bool       // function omits stack split prologue
-	Var      []Var      // detail about local variables
-	PCSP     Data       // PC → SP offset map
-	PCFile   Data       // PC → file number map (index into File)
-	PCLine   Data       // PC → line number map
-	PCData   []Data     // PC → runtime support data map
-	FuncData []FuncData // non-PC-specific runtime support data
-	File     []string   // paths indexed by PCFile
+type goobj_Func struct {
+	Args     int              // size in bytes of argument frame: inputs and outputs
+	Frame    int              // size in bytes of local variable frame
+	Leaf     bool             // function omits save of link register (ARM)
+	NoSplit  bool             // function omits stack split prologue
+	Var      []goobj_Var      // detail about local variables
+	PCSP     goobj_Data       // PC → SP offset map
+	PCFile   goobj_Data       // PC → file number map (index into File)
+	PCLine   goobj_Data       // PC → line number map
+	PCData   []goobj_Data     // PC → runtime support data map
+	FuncData []goobj_FuncData // non-PC-specific runtime support data
+	File     []string         // paths indexed by PCFile
 }
 
 // TODO: Add PCData []byte and PCDataIter (similar to liblink).
 
 // A FuncData is a single function-specific data value.
-type FuncData struct {
-	Sym    SymID // symbol holding data
-	Offset int64 // offset into symbol for funcdata pointer
+type goobj_FuncData struct {
+	Sym    goobj_SymID // symbol holding data
+	Offset int64       // offset into symbol for funcdata pointer
 }
 
 // A Package is a parsed Go object file or archive defining a Go package.
-type Package struct {
-	ImportPath string   // import path denoting this package
-	Imports    []string // packages imported by this package
-	Syms       []*Sym   // symbols defined by this package
-	MaxVersion int      // maximum Version in any SymID in Syms
+type goobj_Package struct {
+	ImportPath string       // import path denoting this package
+	Imports    []string     // packages imported by this package
+	Syms       []*goobj_Sym // symbols defined by this package
+	MaxVersion int          // maximum Version in any SymID in Syms
 }
 
 var (
-	archiveHeader = []byte("!<arch>\n")
-	archiveMagic  = []byte("`\n")
-	goobjHeader   = []byte("go objec") // truncated to size of archiveHeader
+	goobj_archiveHeader = []byte("!<arch>\n")
+	goobj_archiveMagic  = []byte("`\n")
+	goobj_goobjHeader   = []byte("go objec") // truncated to size of archiveHeader
 
-	errCorruptArchive   = errors.New("corrupt archive")
-	errTruncatedArchive = errors.New("truncated archive")
-	errNotArchive       = errors.New("unrecognized archive format")
+	goobj_errCorruptArchive   = errors.New("corrupt archive")
+	goobj_errTruncatedArchive = errors.New("truncated archive")
+	goobj_errNotArchive       = errors.New("unrecognized archive format")
 
-	errCorruptObject   = errors.New("corrupt object file")
-	errTruncatedObject = errors.New("truncated object file")
-	errNotObject       = errors.New("unrecognized object file format")
+	goobj_errCorruptObject   = errors.New("corrupt object file")
+	goobj_errTruncatedObject = errors.New("truncated object file")
+	goobj_errNotObject       = errors.New("unrecognized object file format")
 )
 
 // An objReader is an object file reader.
-type objReader struct {
-	p         *Package
+type goobj_objReader struct {
+	p         *goobj_Package
 	b         *bufio.Reader
 	f         io.ReadSeeker
 	err       error
@@ -251,7 +257,7 @@ type objReader struct {
 //
 // See ../../../cmd/ld/lib.c:/^pathtoprefix and
 // ../../../cmd/gc/subr.c:/^pathtoprefix.
-func importPathToPrefix(s string) string {
+func goobj_importPathToPrefix(s string) string {
 	// find index of last slash, if any, or else -1.
 	// used for determining whether an index is after the last slash.
 	slash := strings.LastIndex(s, "/")
@@ -284,21 +290,21 @@ func importPathToPrefix(s string) string {
 }
 
 // init initializes r to read package p from f.
-func (r *objReader) init(f io.ReadSeeker, p *Package) {
+func (r *goobj_objReader) init(f io.ReadSeeker, p *goobj_Package) {
 	r.f = f
 	r.p = p
 	r.offset, _ = f.Seek(0, 1)
 	r.limit, _ = f.Seek(0, 2)
 	f.Seek(r.offset, 0)
 	r.b = bufio.NewReader(f)
-	r.pkgprefix = importPathToPrefix(p.ImportPath) + "."
+	r.pkgprefix = goobj_importPathToPrefix(p.ImportPath) + "."
 }
 
 // error records that an error occurred.
 // It returns only the first error, so that an error
 // caused by an earlier error does not discard information
 // about the earlier error.
-func (r *objReader) error(err error) error {
+func (r *goobj_objReader) error(err error) error {
 	if r.err == nil {
 		if err == io.EOF {
 			err = io.ErrUnexpectedEOF
@@ -314,7 +320,7 @@ func (r *objReader) error(err error) error {
 // A sequence of 0 bytes will eventually terminate any
 // parsing state in the object file. In particular, it ends the
 // reading of a varint.
-func (r *objReader) readByte() byte {
+func (r *goobj_objReader) readByte() byte {
 	if r.err != nil {
 		return 0
 	}
@@ -339,7 +345,7 @@ func (r *objReader) readByte() byte {
 // If an error occurs, read returns the error but also
 // records it, so it is safe for callers to ignore the result
 // as long as delaying the report is not a problem.
-func (r *objReader) readFull(b []byte) error {
+func (r *goobj_objReader) readFull(b []byte) error {
 	if r.err != nil {
 		return r.err
 	}
@@ -355,12 +361,12 @@ func (r *objReader) readFull(b []byte) error {
 }
 
 // readInt reads a zigzag varint from the input file.
-func (r *objReader) readInt() int {
+func (r *goobj_objReader) readInt() int {
 	var u uint64
 
 	for shift := uint(0); ; shift += 7 {
 		if shift >= 64 {
-			r.error(errCorruptObject)
+			r.error(goobj_errCorruptObject)
 			return 0
 		}
 		c := r.readByte()
@@ -372,14 +378,14 @@ func (r *objReader) readInt() int {
 
 	v := int64(u>>1) ^ (int64(u) << 63 >> 63)
 	if int64(int(v)) != v {
-		r.error(errCorruptObject) // TODO
+		r.error(goobj_errCorruptObject) // TODO
 		return 0
 	}
 	return int(v)
 }
 
 // readString reads a length-delimited string from the input file.
-func (r *objReader) readString() string {
+func (r *goobj_objReader) readString() string {
 	n := r.readInt()
 	buf := make([]byte, n)
 	r.readFull(buf)
@@ -387,7 +393,7 @@ func (r *objReader) readString() string {
 }
 
 // readSymID reads a SymID from the input file.
-func (r *objReader) readSymID() SymID {
+func (r *goobj_objReader) readSymID() goobj_SymID {
 	name, vers := r.readString(), r.readInt()
 
 	// In a symbol name in an object file, "". denotes the
@@ -403,19 +409,19 @@ func (r *objReader) readSymID() SymID {
 		vers = r.p.MaxVersion
 	}
 
-	return SymID{name, vers}
+	return goobj_SymID{name, vers}
 }
 
 // readData reads a data reference from the input file.
-func (r *objReader) readData() Data {
+func (r *goobj_objReader) readData() goobj_Data {
 	n := r.readInt()
-	d := Data{Offset: r.offset, Size: int64(n)}
+	d := goobj_Data{Offset: r.offset, Size: int64(n)}
 	r.skip(int64(n))
 	return d
 }
 
 // skip skips n bytes in the input.
-func (r *objReader) skip(n int64) {
+func (r *goobj_objReader) skip(n int64) {
 	if n < 0 {
 		r.error(fmt.Errorf("debug/goobj: internal error: misuse of skip"))
 	}
@@ -444,14 +450,14 @@ func (r *objReader) skip(n int64) {
 
 // Parse parses an object file or archive from r,
 // assuming that its import path is pkgpath.
-func Parse(r io.ReadSeeker, pkgpath string) (*Package, error) {
+func goobj_Parse(r io.ReadSeeker, pkgpath string) (*goobj_Package, error) {
 	if pkgpath == "" {
 		pkgpath = `""`
 	}
-	p := new(Package)
+	p := new(goobj_Package)
 	p.ImportPath = pkgpath
 
-	var rd objReader
+	var rd goobj_objReader
 	rd.init(r, p)
 	err := rd.readFull(rd.tmp[:8])
 	if err != nil {
@@ -463,14 +469,14 @@ func Parse(r io.ReadSeeker, pkgpath string) (*Package, error) {
 
 	switch {
 	default:
-		return nil, errNotObject
+		return nil, goobj_errNotObject
 
-	case bytes.Equal(rd.tmp[:8], archiveHeader):
+	case bytes.Equal(rd.tmp[:8], goobj_archiveHeader):
 		if err := rd.parseArchive(); err != nil {
 			return nil, err
 		}
-	case bytes.Equal(rd.tmp[:8], goobjHeader):
-		if err := rd.parseObject(goobjHeader); err != nil {
+	case bytes.Equal(rd.tmp[:8], goobj_goobjHeader):
+		if err := rd.parseObject(goobj_goobjHeader); err != nil {
 			return nil, err
 		}
 	}
@@ -480,7 +486,7 @@ func Parse(r io.ReadSeeker, pkgpath string) (*Package, error) {
 
 // trimSpace removes trailing spaces from b and returns the corresponding string.
 // This effectively parses the form used in archive headers.
-func trimSpace(b []byte) string {
+func goobj_trimSpace(b []byte) string {
 	return string(bytes.TrimRight(b, " "))
 }
 
@@ -488,7 +494,7 @@ func trimSpace(b []byte) string {
 // TODO(rsc): Need to skip non-Go object files.
 // TODO(rsc): Maybe record table of contents in r.p so that
 // linker can avoid having code to parse archives too.
-func (r *objReader) parseArchive() error {
+func (r *goobj_objReader) parseArchive() error {
 	for r.offset < r.limit {
 		if err := r.readFull(r.tmp[:60]); err != nil {
 			return err
@@ -512,20 +518,20 @@ func (r *objReader) parseArchive() error {
 		// The file data that follows is padded to an even number of bytes:
 		// if size is odd, an extra padding byte is inserted betw the next header.
 		if len(data) < 60 {
-			return errTruncatedArchive
+			return goobj_errTruncatedArchive
 		}
-		if !bytes.Equal(data[58:60], archiveMagic) {
-			return errCorruptArchive
+		if !bytes.Equal(data[58:60], goobj_archiveMagic) {
+			return goobj_errCorruptArchive
 		}
-		name := trimSpace(data[0:16])
-		size, err := strconv.ParseInt(trimSpace(data[48:58]), 10, 64)
+		name := goobj_trimSpace(data[0:16])
+		size, err := strconv.ParseInt(goobj_trimSpace(data[48:58]), 10, 64)
 		if err != nil {
-			return errCorruptArchive
+			return goobj_errCorruptArchive
 		}
 		data = data[60:]
 		fsize := size + size&1
 		if fsize < 0 || fsize < size {
-			return errCorruptArchive
+			return goobj_errCorruptArchive
 		}
 		switch name {
 		case "__.SYMDEF", "__.GOSYMDEF", "__.PKGDEF":
@@ -553,7 +559,7 @@ func (r *objReader) parseArchive() error {
 // and then the part we want to parse begins.
 // The format of that part is defined in a comment at the top
 // of src/liblink/objfile.c.
-func (r *objReader) parseObject(prefix []byte) error {
+func (r *goobj_objReader) parseObject(prefix []byte) error {
 	// TODO(rsc): Maybe use prefix and the initial input to
 	// record the header line from the file, which would
 	// give the architecture and other version information.
@@ -563,7 +569,7 @@ func (r *objReader) parseObject(prefix []byte) error {
 	for {
 		c1, c2, c3 = c2, c3, r.readByte()
 		if c3 == 0 { // NUL or EOF, either is bad
-			return errCorruptObject
+			return goobj_errCorruptObject
 		}
 		if c1 == '\n' && c2 == '!' && c3 == '\n' {
 			break
@@ -572,12 +578,12 @@ func (r *objReader) parseObject(prefix []byte) error {
 
 	r.readFull(r.tmp[:8])
 	if !bytes.Equal(r.tmp[:8], []byte("\x00\x00go13ld")) {
-		return r.error(errCorruptObject)
+		return r.error(goobj_errCorruptObject)
 	}
 
 	b := r.readByte()
 	if b != 1 {
-		return r.error(errCorruptObject)
+		return r.error(goobj_errCorruptObject)
 	}
 
 	// Direct package dependencies.
@@ -593,20 +599,20 @@ func (r *objReader) parseObject(prefix []byte) error {
 	for {
 		if b := r.readByte(); b != 0xfe {
 			if b != 0xff {
-				return r.error(errCorruptObject)
+				return r.error(goobj_errCorruptObject)
 			}
 			break
 		}
 
 		typ := r.readInt()
-		s := &Sym{SymID: r.readSymID()}
+		s := &goobj_Sym{goobj_SymID: r.readSymID()}
 		r.p.Syms = append(r.p.Syms, s)
-		s.Kind = SymKind(typ)
+		s.Kind = goobj_SymKind(typ)
 		s.DupOK = r.readInt() != 0
 		s.Size = r.readInt()
 		s.Type = r.readSymID()
 		s.Data = r.readData()
-		s.Reloc = make([]Reloc, r.readInt())
+		s.Reloc = make([]goobj_Reloc, r.readInt())
 		for i := range s.Reloc {
 			rel := &s.Reloc[i]
 			rel.Offset = r.readInt()
@@ -618,14 +624,14 @@ func (r *objReader) parseObject(prefix []byte) error {
 			r.readSymID() // Xsym - ignored
 		}
 
-		if s.Kind == STEXT {
-			f := new(Func)
+		if s.Kind == goobj_STEXT {
+			f := new(goobj_Func)
 			s.Func = f
 			f.Args = r.readInt()
 			f.Frame = r.readInt()
 			f.Leaf = r.readInt() != 0
 			f.NoSplit = r.readInt() != 0
-			f.Var = make([]Var, r.readInt())
+			f.Var = make([]goobj_Var, r.readInt())
 			for i := range f.Var {
 				v := &f.Var[i]
 				v.Name = r.readSymID().Name
@@ -637,11 +643,11 @@ func (r *objReader) parseObject(prefix []byte) error {
 			f.PCSP = r.readData()
 			f.PCFile = r.readData()
 			f.PCLine = r.readData()
-			f.PCData = make([]Data, r.readInt())
+			f.PCData = make([]goobj_Data, r.readInt())
 			for i := range f.PCData {
 				f.PCData[i] = r.readData()
 			}
-			f.FuncData = make([]FuncData, r.readInt())
+			f.FuncData = make([]goobj_FuncData, r.readInt())
 			for i := range f.FuncData {
 				f.FuncData[i].Sym = r.readSymID()
 			}
@@ -657,7 +663,7 @@ func (r *objReader) parseObject(prefix []byte) error {
 
 	r.readFull(r.tmp[:7])
 	if !bytes.Equal(r.tmp[:7], []byte("\xffgo13ld")) {
-		return r.error(errCorruptObject)
+		return r.error(goobj_errCorruptObject)
 	}
 
 	return nil

src/cmd/nm/goobj.go

@@ -7,12 +7,11 @@
 package main
 
 import (
-	"debug/goobj"
 	"fmt"
 	"os"
 )
 
-func goobjName(id goobj.SymID) string {
+func goobjName(id goobj_SymID) string {
 	if id.Version == 0 {
 		return id.Name
 	}
@@ -20,28 +19,28 @@ func goobjName(id goobj.SymID) string {
 }
 
 func goobjSymbols(f *os.File) []Sym {
-	pkg, err := goobj.Parse(f, `""`)
+	pkg, err := goobj_Parse(f, `""`)
 	if err != nil {
 		errorf("parsing %s: %v", f.Name(), err)
 		return nil
 	}
 
-	seen := make(map[goobj.SymID]bool)
+	seen := make(map[goobj_SymID]bool)
 
 	var syms []Sym
 	for _, s := range pkg.Syms {
-		seen[s.SymID] = true
-		sym := Sym{Addr: uint64(s.Data.Offset), Name: goobjName(s.SymID), Size: int64(s.Size), Type: s.Type.Name, Code: '?'}
+		seen[s.goobj_SymID] = true
+		sym := Sym{Addr: uint64(s.Data.Offset), Name: goobjName(s.goobj_SymID), Size: int64(s.Size), Type: s.Type.Name, Code: '?'}
 		switch s.Kind {
-		case goobj.STEXT, goobj.SELFRXSECT:
+		case goobj_STEXT, goobj_SELFRXSECT:
 			sym.Code = 'T'
-		case goobj.STYPE, goobj.SSTRING, goobj.SGOSTRING, goobj.SGOFUNC, goobj.SRODATA, goobj.SFUNCTAB, goobj.STYPELINK, goobj.SSYMTAB, goobj.SPCLNTAB, goobj.SELFROSECT:
+		case goobj_STYPE, goobj_SSTRING, goobj_SGOSTRING, goobj_SGOFUNC, goobj_SRODATA, goobj_SFUNCTAB, goobj_STYPELINK, goobj_SSYMTAB, goobj_SPCLNTAB, goobj_SELFROSECT:
 			sym.Code = 'R'
-		case goobj.SMACHOPLT, goobj.SELFSECT, goobj.SMACHO, goobj.SMACHOGOT, goobj.SNOPTRDATA, goobj.SINITARR, goobj.SDATA, goobj.SWINDOWS:
+		case goobj_SMACHOPLT, goobj_SELFSECT, goobj_SMACHO, goobj_SMACHOGOT, goobj_SNOPTRDATA, goobj_SINITARR, goobj_SDATA, goobj_SWINDOWS:
 			sym.Code = 'D'
-		case goobj.SBSS, goobj.SNOPTRBSS, goobj.STLSBSS:
+		case goobj_SBSS, goobj_SNOPTRBSS, goobj_STLSBSS:
 			sym.Code = 'B'
-		case goobj.SXREF, goobj.SMACHOSYMSTR, goobj.SMACHOSYMTAB, goobj.SMACHOINDIRECTPLT, goobj.SMACHOINDIRECTGOT, goobj.SFILE, goobj.SFILEPATH, goobj.SCONST, goobj.SDYNIMPORT, goobj.SHOSTOBJ:
+		case goobj_SXREF, goobj_SMACHOSYMSTR, goobj_SMACHOSYMTAB, goobj_SMACHOINDIRECTPLT, goobj_SMACHOINDIRECTGOT, goobj_SFILE, goobj_SFILEPATH, goobj_SCONST, goobj_SDYNIMPORT, goobj_SHOSTOBJ:
 			sym.Code = 'X' // should not see
 		}
 		if s.Version != 0 {

src/nacltest.bash

@@ -7,7 +7,8 @@
 # Builds a test file system and embeds it into package syscall
 # in every generated binary.
 #
-# Assumes that sel_ldr binaries are in $PATH; see ../misc/nacl/README.
+# Assumes that sel_ldr binaries and go_nacl_$GOARCH_exec scripts are in $PATH;
+# see ../misc/nacl/README.
 
 set -e
 ulimit -c 0
@@ -32,6 +33,11 @@ amd64p32)
 	exit 1
 esac
 
+if ! which go_nacl_${naclGOARCH}_exec >/dev/null; then
+	echo "cannot find go_nacl_${naclGOARCH}_exec, see ../misc/nacl/README." 1>&2
+	exit 1
+fi
+
 # Run host build to get toolchain for running zip generator.
 unset GOOS GOARCH
 if [ ! -f make.bash ]; then

src/pkg/compress/gzip/gunzip.go

@@ -94,7 +94,11 @@ func NewReader(r io.Reader) (*Reader, error) {
 // This permits reusing a Reader rather than allocating a new one.
 func (z *Reader) Reset(r io.Reader) error {
 	z.r = makeReader(r)
-	z.digest.Reset()
+	if z.digest == nil {
+		z.digest = crc32.NewIEEE()
+	} else {
+		z.digest.Reset()
+	}
 	z.size = 0
 	z.err = nil
 	return z.readHeader(true)

src/pkg/compress/gzip/gunzip_test.go

@@ -353,3 +353,17 @@ func TestIssue6550(t *testing.T) {
 		// ok
 	}
 }
+
+func TestInitialReset(t *testing.T) {
+	var r Reader
+	if err := r.Reset(bytes.NewReader(gunzipTests[1].gzip)); err != nil {
+		t.Error(err)
+	}
+	var buf bytes.Buffer
+	if _, err := io.Copy(&buf, &r); err != nil {
+		t.Error(err)
+	}
+	if s := buf.String(); s != gunzipTests[1].raw {
+		t.Errorf("got %q want %q", s, gunzipTests[1].raw)
+	}
+}

src/pkg/compress/gzip/gzip.go

@@ -245,7 +245,8 @@ func (z *Writer) Flush() error {
 	return z.err
 }
 
-// Close closes the Writer. It does not close the underlying io.Writer.
+// Close closes the Writer, flushing any unwritten data to the underlying
+// io.Writer, but does not close the underlying io.Writer.
 func (z *Writer) Close() error {
 	if z.err != nil {
 		return z.err

src/pkg/compress/zlib/writer.go

@@ -174,7 +174,8 @@ func (z *Writer) Flush() error {
 	return z.err
 }
 
-// Calling Close does not close the wrapped io.Writer originally passed to NewWriter.
+// Close closes the Writer, flushing any unwritten data to the underlying
+// io.Writer, but does not close the underlying io.Writer.
 func (z *Writer) Close() error {
 	if !z.wroteHeader {
 		z.err = z.writeHeader()

src/pkg/crypto/rsa/pkcs1v15.go

@@ -53,11 +53,14 @@ func DecryptPKCS1v15(rand io.Reader, priv *PrivateKey, ciphertext []byte) (out [
 	if err := checkPub(&priv.PublicKey); err != nil {
 		return nil, err
 	}
-	valid, out, err := decryptPKCS1v15(rand, priv, ciphertext)
-	if err == nil && valid == 0 {
-		err = ErrDecryption
+	valid, out, index, err := decryptPKCS1v15(rand, priv, ciphertext)
+	if err != nil {
+		return
 	}
-
+	if valid == 0 {
+		return nil, ErrDecryption
+	}
+	out = out[index:]
 	return
 }
 
@@ -80,21 +83,32 @@ func DecryptPKCS1v15SessionKey(rand io.Reader, priv *PrivateKey, ciphertext []by
 	}
 	k := (priv.N.BitLen() + 7) / 8
 	if k-(len(key)+3+8) < 0 {
-		err = ErrDecryption
-		return
+		return ErrDecryption
 	}
 
-	valid, msg, err := decryptPKCS1v15(rand, priv, ciphertext)
+	valid, em, index, err := decryptPKCS1v15(rand, priv, ciphertext)
 	if err != nil {
 		return
 	}
 
-	valid &= subtle.ConstantTimeEq(int32(len(msg)), int32(len(key)))
-	subtle.ConstantTimeCopy(valid, key, msg)
+	if len(em) != k {
+		// This should be impossible because decryptPKCS1v15 always
+		// returns the full slice.
+		return ErrDecryption
+	}
+
+	valid &= subtle.ConstantTimeEq(int32(len(em)-index), int32(len(key)))
+	subtle.ConstantTimeCopy(valid, key, em[len(em)-len(key):])
 	return
 }
 
-func decryptPKCS1v15(rand io.Reader, priv *PrivateKey, ciphertext []byte) (valid int, msg []byte, err error) {
+// decryptPKCS1v15 decrypts ciphertext using priv and blinds the operation if
+// rand is not nil. It returns one or zero in valid that indicates whether the
+// plaintext was correctly structured. In either case, the plaintext is
+// returned in em so that it may be read independently of whether it was valid
+// in order to maintain constant memory access patterns. If the plaintext was
+// valid then index contains the index of the original message in em.
+func decryptPKCS1v15(rand io.Reader, priv *PrivateKey, ciphertext []byte) (valid int, em []byte, index int, err error) {
 	k := (priv.N.BitLen() + 7) / 8
 	if k < 11 {
 		err = ErrDecryption
@@ -107,7 +121,7 @@ func decryptPKCS1v15(rand io.Reader, priv *PrivateKey, ciphertext []byte) (valid
 		return
 	}
 
-	em := leftPad(m.Bytes(), k)
+	em = leftPad(m.Bytes(), k)
 	firstByteIsZero := subtle.ConstantTimeByteEq(em[0], 0)
 	secondByteIsTwo := subtle.ConstantTimeByteEq(em[1], 2)
 
@@ -115,8 +129,7 @@ func decryptPKCS1v15(rand io.Reader, priv *PrivateKey, ciphertext []byte) (valid
 	// octets, followed by a 0, followed by the message.
 	//   lookingForIndex: 1 iff we are still looking for the zero.
 	//   index: the offset of the first zero byte.
-	var lookingForIndex, index int
-	lookingForIndex = 1
+	lookingForIndex := 1
 
 	for i := 2; i < len(em); i++ {
 		equals0 := subtle.ConstantTimeByteEq(em[i], 0)
@@ -129,8 +142,8 @@ func decryptPKCS1v15(rand io.Reader, priv *PrivateKey, ciphertext []byte) (valid
 	validPS := subtle.ConstantTimeLessOrEq(2+8, index)
 
 	valid = firstByteIsZero & secondByteIsTwo & (^lookingForIndex & 1) & validPS
-	msg = em[index+1:]
-	return
+	index = subtle.ConstantTimeSelect(valid, index+1, 0)
+	return valid, em, index, nil
 }
 
 // nonZeroRandomBytes fills the given slice with non-zero random octets.

src/pkg/crypto/rsa/pkcs1v15_test.go

@@ -227,6 +227,26 @@ func TestUnpaddedSignature(t *testing.T) {
 	}
 }
 
+func TestShortSessionKey(t *testing.T) {
+	// This tests that attempting to decrypt a session key where the
+	// ciphertext is too small doesn't run outside the array bounds.
+	ciphertext, err := EncryptPKCS1v15(rand.Reader, &rsaPrivateKey.PublicKey, []byte{1})
+	if err != nil {
+		t.Fatalf("Failed to encrypt short message: %s", err)
+	}
+
+	var key [32]byte
+	if err := DecryptPKCS1v15SessionKey(nil, rsaPrivateKey, ciphertext, key[:]); err != nil {
+		t.Fatalf("Failed to decrypt short message: %s", err)
+	}
+
+	for _, v := range key {
+		if v != 0 {
+			t.Fatal("key was modified when ciphertext was invalid")
+		}
+	}
+}
+
 // In order to generate new test vectors you'll need the PEM form of this key:
 // -----BEGIN RSA PRIVATE KEY-----
 // MIIBOgIBAAJBALKZD0nEffqM1ACuak0bijtqE2QrI/KLADv7l3kK3ppMyCuLKoF0

src/pkg/crypto/subtle/constant_time.go

@@ -49,9 +49,14 @@ func ConstantTimeEq(x, y int32) int {
 	return int(z & 1)
 }
 
-// ConstantTimeCopy copies the contents of y into x iff v == 1. If v == 0, x is left unchanged.
-// Its behavior is undefined if v takes any other value.
+// ConstantTimeCopy copies the contents of y into x (a slice of equal length)
+// if v == 1. If v == 0, x is left unchanged. Its behavior is undefined if v
+// takes any other value.
 func ConstantTimeCopy(v int, x, y []byte) {
+	if len(x) != len(y) {
+		panic("subtle: slices have different lengths")
+	}
+
 	xmask := byte(v - 1)
 	ymask := byte(^(v - 1))
 	for i := 0; i < len(x); i++ {

src/pkg/crypto/tls/handshake_server.go

@@ -214,6 +214,10 @@ Curves:
 func (hs *serverHandshakeState) checkForResumption() bool {
 	c := hs.c
 
+	if c.config.SessionTicketsDisabled {
+		return false
+	}
+
 	var ok bool
 	if hs.sessionState, ok = c.decryptTicket(hs.clientHello.sessionTicket); !ok {
 		return false

src/pkg/crypto/tls/handshake_server_test.go

@@ -557,6 +557,32 @@ func TestResumption(t *testing.T) {
 	runServerTestTLS12(t, test)
 }
 
+func TestResumptionDisabled(t *testing.T) {
+	sessionFilePath := tempFile("")
+	defer os.Remove(sessionFilePath)
+
+	config := *testConfig
+
+	test := &serverTest{
+		name:    "IssueTicketPreDisable",
+		command: []string{"openssl", "s_client", "-cipher", "RC4-SHA", "-sess_out", sessionFilePath},
+		config:  &config,
+	}
+	runServerTestTLS12(t, test)
+
+	config.SessionTicketsDisabled = true
+
+	test = &serverTest{
+		name:    "ResumeDisabled",
+		command: []string{"openssl", "s_client", "-cipher", "RC4-SHA", "-sess_in", sessionFilePath},
+		config:  &config,
+	}
+	runServerTestTLS12(t, test)
+
+	// One needs to manually confirm that the handshake in the golden data
+	// file for ResumeDisabled does not include a resumption handshake.
+}
+
 // cert.pem and key.pem were generated with generate_cert.go
 // Thus, they have no ExtKeyUsage fields and trigger an error
 // when verification is turned on.

src/pkg/crypto/tls/testdata/Server-TLSv12-IssueTicketPreDisable

@@ -0,0 +1,87 @@
+>>> Flow 1 (client to server)
+00000000  16 03 01 00 60 01 00 00  5c 03 03 54 23 54 02 17  |....`...\..T#T..|
+00000010  f3 53 13 3d 48 88 c3 19  b9 d1 3d 33 7f f5 99 56  |.S.=H.....=3...V|
+00000020  04 71 1b d9 d5 64 8a 0d  4a 54 00 00 00 04 00 05  |.q...d..JT......|
+00000030  00 ff 01 00 00 2f 00 23  00 00 00 0d 00 22 00 20  |...../.#.....". |
+00000040  06 01 06 02 06 03 05 01  05 02 05 03 04 01 04 02  |................|
+00000050  04 03 03 01 03 02 03 03  02 01 02 02 02 03 01 01  |................|
+00000060  00 0f 00 01 01                                    |.....|
+>>> Flow 2 (server to client)
+00000000  16 03 03 00 35 02 00 00  31 03 03 00 00 00 00 00  |....5...1.......|
+00000010  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
+00000020  00 00 00 00 00 00 00 00  00 00 00 00 00 05 00 00  |................|
+00000030  09 00 23 00 00 ff 01 00  01 00 16 03 03 02 be 0b  |..#.............|
+00000040  00 02 ba 00 02 b7 00 02  b4 30 82 02 b0 30 82 02  |.........0...0..|
+00000050  19 a0 03 02 01 02 02 09  00 85 b0 bb a4 8a 7f b8  |................|
+00000060  ca 30 0d 06 09 2a 86 48  86 f7 0d 01 01 05 05 00  |.0...*.H........|
+00000070  30 45 31 0b 30 09 06 03  55 04 06 13 02 41 55 31  |0E1.0...U....AU1|
+00000080  13 30 11 06 03 55 04 08  13 0a 53 6f 6d 65 2d 53  |.0...U....Some-S|
+00000090  74 61 74 65 31 21 30 1f  06 03 55 04 0a 13 18 49  |tate1!0...U....I|
+000000a0  6e 74 65 72 6e 65 74 20  57 69 64 67 69 74 73 20  |nternet Widgits |
+000000b0  50 74 79 20 4c 74 64 30  1e 17 0d 31 30 30 34 32  |Pty Ltd0...10042|
+000000c0  34 30 39 30 39 33 38 5a  17 0d 31 31 30 34 32 34  |4090938Z..110424|
+000000d0  30 39 30 39 33 38 5a 30  45 31 0b 30 09 06 03 55  |090938Z0E1.0...U|
+000000e0  04 06 13 02 41 55 31 13  30 11 06 03 55 04 08 13  |....AU1.0...U...|
+000000f0  0a 53 6f 6d 65 2d 53 74  61 74 65 31 21 30 1f 06  |.Some-State1!0..|
+00000100  03 55 04 0a 13 18 49 6e  74 65 72 6e 65 74 20 57  |.U....Internet W|
+00000110  69 64 67 69 74 73 20 50  74 79 20 4c 74 64 30 81  |idgits Pty Ltd0.|
+00000120  9f 30 0d 06 09 2a 86 48  86 f7 0d 01 01 01 05 00  |.0...*.H........|
+00000130  03 81 8d 00 30 81 89 02  81 81 00 bb 79 d6 f5 17  |....0.......y...|
+00000140  b5 e5 bf 46 10 d0 dc 69  be e6 2b 07 43 5a d0 03  |...F...i..+.CZ..|
+00000150  2d 8a 7a 43 85 b7 14 52  e7 a5 65 4c 2c 78 b8 23  |-.zC...R..eL,x.#|
+00000160  8c b5 b4 82 e5 de 1f 95  3b 7e 62 a5 2c a5 33 d6  |........;~b.,.3.|
+00000170  fe 12 5c 7a 56 fc f5 06  bf fa 58 7b 26 3f b5 cd  |..\zV.....X{&?..|
+00000180  04 d3 d0 c9 21 96 4a c7  f4 54 9f 5a bf ef 42 71  |....!.J..T.Z..Bq|
+00000190  00 fe 18 99 07 7f 7e 88  7d 7d f1 04 39 c4 a2 2e  |......~.}}..9...|
+000001a0  db 51 c9 7c e3 c0 4c 3b  32 66 01 cf af b1 1d b8  |.Q.|..L;2f......|
+000001b0  71 9a 1d db db 89 6b ae  da 2d 79 02 03 01 00 01  |q.....k..-y.....|
+000001c0  a3 81 a7 30 81 a4 30 1d  06 03 55 1d 0e 04 16 04  |...0..0...U.....|
+000001d0  14 b1 ad e2 85 5a cf cb  28 db 69 ce 23 69 de d3  |.....Z..(.i.#i..|
+000001e0  26 8e 18 88 39 30 75 06  03 55 1d 23 04 6e 30 6c  |&...90u..U.#.n0l|
+000001f0  80 14 b1 ad e2 85 5a cf  cb 28 db 69 ce 23 69 de  |......Z..(.i.#i.|
+00000200  d3 26 8e 18 88 39 a1 49  a4 47 30 45 31 0b 30 09  |.&...9.I.G0E1.0.|
+00000210  06 03 55 04 06 13 02 41  55 31 13 30 11 06 03 55  |..U....AU1.0...U|
+00000220  04 08 13 0a 53 6f 6d 65  2d 53 74 61 74 65 31 21  |....Some-State1!|
+00000230  30 1f 06 03 55 04 0a 13  18 49 6e 74 65 72 6e 65  |0...U....Interne|
+00000240  74 20 57 69 64 67 69 74  73 20 50 74 79 20 4c 74  |t Widgits Pty Lt|
+00000250  64 82 09 00 85 b0 bb a4  8a 7f b8 ca 30 0c 06 03  |d...........0...|
+00000260  55 1d 13 04 05 30 03 01  01 ff 30 0d 06 09 2a 86  |U....0....0...*.|
+00000270  48 86 f7 0d 01 01 05 05  00 03 81 81 00 08 6c 45  |H.............lE|
+00000280  24 c7 6b b1 59 ab 0c 52  cc f2 b0 14 d7 87 9d 7a  |$.k.Y..R.......z|
+00000290  64 75 b5 5a 95 66 e4 c5  2b 8e ae 12 66 1f eb 4f  |du.Z.f..+...f..O|
+000002a0  38 b3 6e 60 d3 92 fd f7  41 08 b5 25 13 b1 18 7a  |8.n`....A..%...z|
+000002b0  24 fb 30 1d ba ed 98 b9  17 ec e7 d7 31 59 db 95  |$.0.........1Y..|
+000002c0  d3 1d 78 ea 50 56 5c d5  82 5a 2d 5a 5f 33 c4 b6  |..x.PV\..Z-Z_3..|
+000002d0  d8 c9 75 90 96 8c 0f 52  98 b5 cd 98 1f 89 20 5f  |..u....R...... _|
+000002e0  f2 a0 1c a3 1b 96 94 dd  a9 fd 57 e9 70 e8 26 6d  |..........W.p.&m|
+000002f0  71 99 9b 26 6e 38 50 29  6c 90 a7 bd d9 16 03 03  |q..&n8P)l.......|
+00000300  00 04 0e 00 00 00                                 |......|
+>>> Flow 3 (client to server)
+00000000  16 03 03 00 86 10 00 00  82 00 80 27 e9 a4 f7 e7  |...........'....|
+00000010  df 25 de 84 8c 1f d6 e6  c3 11 28 55 9a c1 91 37  |.%........(U...7|
+00000020  84 f5 ba f8 80 0d ca 50  cb 1e 72 f7 97 6f c2 b2  |.......P..r..o..|
+00000030  04 4d 13 7c e0 6e a0 1f  91 e1 38 1b a2 c0 55 16  |.M.|.n....8...U.|
+00000040  7f 29 fc ed 1c 1a cf 72  14 c3 00 c1 dd 36 36 af  |.).....r.....66.|
+00000050  a6 e4 a8 be ba ec 13 d0  1e d0 1d fd e1 5b 27 fd  |.............['.|
+00000060  9a da 2e 12 c8 b0 b9 c2  b9 76 ec 7f 3c 98 b6 63  |.........v..<..c|
+00000070  bc da f0 07 7a 3d e7 61  f4 2f 12 80 3b f9 3b cc  |....z=.a./..;.;.|
+00000080  05 c8 2f 7e 28 b2 73 bf  97 61 29 14 03 03 00 01  |../~(.s..a).....|
+00000090  01 16 03 03 00 24 17 59  a9 45 53 46 33 96 50 dd  |.....$.Y.ESF3.P.|
+000000a0  3e 23 aa 91 38 f8 56 4a  2f 1a f2 b1 44 9b ce 17  |>#..8.VJ/...D...|
+000000b0  6b 8a 89 76 bc 67 b8 8b  ba 90                    |k..v.g....|
+>>> Flow 4 (server to client)
+00000000  16 03 03 00 72 04 00 00  6e 00 00 00 00 00 68 00  |....r...n.....h.|
+00000010  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 65  |...............e|
+00000020  ea 4b d1 ef ba 2d db 0c  ba 9a d4 20 76 57 c8 ec  |.K...-..... vW..|
+00000030  dc 2d 77 fb fb 3b 93 5f  53 e0 14 4f 90 fb d6 55  |.-w..;._S..O...U|
+00000040  57 8c 8d 0d 25 ea 5d 0d  f2 91 e5 12 22 12 ec 7b  |W...%.]....."..{|
+00000050  5f b6 6e fd 07 59 23 24  fc b1 97 ca ea 56 a5 c2  |_.n..Y#$.....V..|
+00000060  a0 e4 9e 99 64 f2 64 d0  75 7a 46 63 e3 dc 21 ed  |....d.d.uzFc..!.|
+00000070  78 56 e9 e1 ab 66 80 14  03 03 00 01 01 16 03 03  |xV...f..........|
+00000080  00 24 fc 14 68 07 17 1f  df b7 84 cb fd c1 e0 e4  |.$..h...........|
+00000090  f2 1a ea 34 b5 00 7f 70  be c8 1c 0a d6 55 e3 57  |...4...p.....U.W|
+000000a0  50 4e 6d 7d 8a 5d 17 03  03 00 21 24 27 50 40 c1  |PNm}.]....!$'P@.|
+000000b0  c5 bd c7 9f 95 d9 ba 2e  7b 0e db ea a7 31 81 05  |........{....1..|
+000000c0  75 43 b1 63 cf b8 55 92  ef 76 98 a9 15 03 03 00  |uC.c..U..v......|
+000000d0  16 d7 ea 3c 79 e7 a6 2f  61 39 ec 4e 95 86 48 5e  |...<y../a9.N..H^|
+000000e0  75 a0 9e 41 42 89 67                              |u..AB.g|

src/pkg/crypto/tls/testdata/Server-TLSv12-ResumeDisabled

@@ -0,0 +1,87 @@
+>>> Flow 1 (client to server)
+00000000  16 03 01 00 e8 01 00 00  e4 03 03 54 23 54 02 a5  |...........T#T..|
+00000010  10 11 0f 6d e5 2d 2f e8  bb 52 b1 38 3f 65 01 43  |...m.-/..R.8?e.C|
+00000020  36 cc 48 f6 09 22 a1 85  20 28 3c 20 35 8b fe 7a  |6.H..".. (< 5..z|
+00000030  41 3b 59 3a 5d b9 b3 21  f0 62 e9 0d 7b af f5 5d  |A;Y:]..!.b..{..]|
+00000040  fa 65 1a 40 c8 ca cd 74  8c ef d2 fb 00 04 00 05  |.e.@...t........|
+00000050  00 ff 01 00 00 97 00 23  00 68 00 00 00 00 00 00  |.......#.h......|
+00000060  00 00 00 00 00 00 00 00  00 00 65 ea 4b d1 ef ba  |..........e.K...|
+00000070  2d db 0c ba 9a d4 20 76  57 c8 ec dc 2d 77 fb fb  |-..... vW...-w..|
+00000080  3b 93 5f 53 e0 14 4f 90  fb d6 55 57 8c 8d 0d 25  |;._S..O...UW...%|
+00000090  ea 5d 0d f2 91 e5 12 22  12 ec 7b 5f b6 6e fd 07  |.]....."..{_.n..|
+000000a0  59 23 24 fc b1 97 ca ea  56 a5 c2 a0 e4 9e 99 64  |Y#$.....V......d|
+000000b0  f2 64 d0 75 7a 46 63 e3  dc 21 ed 78 56 e9 e1 ab  |.d.uzFc..!.xV...|
+000000c0  66 80 00 0d 00 22 00 20  06 01 06 02 06 03 05 01  |f....". ........|
+000000d0  05 02 05 03 04 01 04 02  04 03 03 01 03 02 03 03  |................|
+000000e0  02 01 02 02 02 03 01 01  00 0f 00 01 01           |.............|
+>>> Flow 2 (server to client)
+00000000  16 03 03 00 31 02 00 00  2d 03 03 00 00 00 00 00  |....1...-.......|
+00000010  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
+00000020  00 00 00 00 00 00 00 00  00 00 00 00 00 05 00 00  |................|
+00000030  05 ff 01 00 01 00 16 03  03 02 be 0b 00 02 ba 00  |................|
+00000040  02 b7 00 02 b4 30 82 02  b0 30 82 02 19 a0 03 02  |.....0...0......|
+00000050  01 02 02 09 00 85 b0 bb  a4 8a 7f b8 ca 30 0d 06  |.............0..|
+00000060  09 2a 86 48 86 f7 0d 01  01 05 05 00 30 45 31 0b  |.*.H........0E1.|
+00000070  30 09 06 03 55 04 06 13  02 41 55 31 13 30 11 06  |0...U....AU1.0..|
+00000080  03 55 04 08 13 0a 53 6f  6d 65 2d 53 74 61 74 65  |.U....Some-State|
+00000090  31 21 30 1f 06 03 55 04  0a 13 18 49 6e 74 65 72  |1!0...U....Inter|
+000000a0  6e 65 74 20 57 69 64 67  69 74 73 20 50 74 79 20  |net Widgits Pty |
+000000b0  4c 74 64 30 1e 17 0d 31  30 30 34 32 34 30 39 30  |Ltd0...100424090|
+000000c0  39 33 38 5a 17 0d 31 31  30 34 32 34 30 39 30 39  |938Z..1104240909|
+000000d0  33 38 5a 30 45 31 0b 30  09 06 03 55 04 06 13 02  |38Z0E1.0...U....|
+000000e0  41 55 31 13 30 11 06 03  55 04 08 13 0a 53 6f 6d  |AU1.0...U....Som|
+000000f0  65 2d 53 74 61 74 65 31  21 30 1f 06 03 55 04 0a  |e-State1!0...U..|
+00000100  13 18 49 6e 74 65 72 6e  65 74 20 57 69 64 67 69  |..Internet Widgi|
+00000110  74 73 20 50 74 79 20 4c  74 64 30 81 9f 30 0d 06  |ts Pty Ltd0..0..|
+00000120  09 2a 86 48 86 f7 0d 01  01 01 05 00 03 81 8d 00  |.*.H............|
+00000130  30 81 89 02 81 81 00 bb  79 d6 f5 17 b5 e5 bf 46  |0.......y......F|
+00000140  10 d0 dc 69 be e6 2b 07  43 5a d0 03 2d 8a 7a 43  |...i..+.CZ..-.zC|
+00000150  85 b7 14 52 e7 a5 65 4c  2c 78 b8 23 8c b5 b4 82  |...R..eL,x.#....|
+00000160  e5 de 1f 95 3b 7e 62 a5  2c a5 33 d6 fe 12 5c 7a  |....;~b.,.3...\z|
+00000170  56 fc f5 06 bf fa 58 7b  26 3f b5 cd 04 d3 d0 c9  |V.....X{&?......|
+00000180  21 96 4a c7 f4 54 9f 5a  bf ef 42 71 00 fe 18 99  |!.J..T.Z..Bq....|
+00000190  07 7f 7e 88 7d 7d f1 04  39 c4 a2 2e db 51 c9 7c  |..~.}}..9....Q.||
+000001a0  e3 c0 4c 3b 32 66 01 cf  af b1 1d b8 71 9a 1d db  |..L;2f......q...|
+000001b0  db 89 6b ae da 2d 79 02  03 01 00 01 a3 81 a7 30  |..k..-y........0|
+000001c0  81 a4 30 1d 06 03 55 1d  0e 04 16 04 14 b1 ad e2  |..0...U.........|
+000001d0  85 5a cf cb 28 db 69 ce  23 69 de d3 26 8e 18 88  |.Z..(.i.#i..&...|
+000001e0  39 30 75 06 03 55 1d 23  04 6e 30 6c 80 14 b1 ad  |90u..U.#.n0l....|
+000001f0  e2 85 5a cf cb 28 db 69  ce 23 69 de d3 26 8e 18  |..Z..(.i.#i..&..|
+00000200  88 39 a1 49 a4 47 30 45  31 0b 30 09 06 03 55 04  |.9.I.G0E1.0...U.|
+00000210  06 13 02 41 55 31 13 30  11 06 03 55 04 08 13 0a  |...AU1.0...U....|
+00000220  53 6f 6d 65 2d 53 74 61  74 65 31 21 30 1f 06 03  |Some-State1!0...|
+00000230  55 04 0a 13 18 49 6e 74  65 72 6e 65 74 20 57 69  |U....Internet Wi|
+00000240  64 67 69 74 73 20 50 74  79 20 4c 74 64 82 09 00  |dgits Pty Ltd...|
+00000250  85 b0 bb a4 8a 7f b8 ca  30 0c 06 03 55 1d 13 04  |........0...U...|
+00000260  05 30 03 01 01 ff 30 0d  06 09 2a 86 48 86 f7 0d  |.0....0...*.H...|
+00000270  01 01 05 05 00 03 81 81  00 08 6c 45 24 c7 6b b1  |..........lE$.k.|
+00000280  59 ab 0c 52 cc f2 b0 14  d7 87 9d 7a 64 75 b5 5a  |Y..R.......zdu.Z|
+00000290  95 66 e4 c5 2b 8e ae 12  66 1f eb 4f 38 b3 6e 60  |.f..+...f..O8.n`|
+000002a0  d3 92 fd f7 41 08 b5 25  13 b1 18 7a 24 fb 30 1d  |....A..%...z$.0.|
+000002b0  ba ed 98 b9 17 ec e7 d7  31 59 db 95 d3 1d 78 ea  |........1Y....x.|
+000002c0  50 56 5c d5 82 5a 2d 5a  5f 33 c4 b6 d8 c9 75 90  |PV\..Z-Z_3....u.|
+000002d0  96 8c 0f 52 98 b5 cd 98  1f 89 20 5f f2 a0 1c a3  |...R...... _....|
+000002e0  1b 96 94 dd a9 fd 57 e9  70 e8 26 6d 71 99 9b 26  |......W.p.&mq..&|
+000002f0  6e 38 50 29 6c 90 a7 bd  d9 16 03 03 00 04 0e 00  |n8P)l...........|
+00000300  00 00                                             |..|
+>>> Flow 3 (client to server)
+00000000  16 03 03 00 86 10 00 00  82 00 80 ae 02 dd 1f 1a  |................|
+00000010  86 83 f5 2f 82 46 4b 29  58 aa a1 b3 56 8b 4e 40  |.../.FK)X...V.N@|
+00000020  ef 23 65 67 ad 48 e5 e1  fd ae dd bf 68 fd bd a6  |.#eg.H......h...|
+00000030  13 a0 7e 05 ab f7 20 e1  6a 4e d1 37 93 08 1d c9  |..~... .jN.7....|
+00000040  37 e0 b5 34 28 bf 20 45  45 da 0f 7e 51 a7 c6 ae  |7..4(. EE..~Q...|
+00000050  61 6c 07 1b 73 ef da 6e  25 c4 ed be e3 3f da ae  |al..s..n%....?..|
+00000060  cd 3c 17 9c 2e ee fb 47  9d b3 a1 b2 c3 5d e0 83  |.<.....G.....]..|
+00000070  74 20 37 2d 72 d6 d0 4d  58 0e 26 1c 50 22 95 08  |t 7-r..MX.&.P"..|
+00000080  7d e0 5f 86 99 9e 2c 2e  a7 a0 7f 14 03 03 00 01  |}._...,.........|
+00000090  01 16 03 03 00 24 a2 ab  41 25 a5 cf 04 18 1d 98  |.....$..A%......|
+000000a0  88 6c 59 21 86 33 54 f4  35 b4 21 6e a5 29 d5 6e  |.lY!.3T.5.!n.).n|
+000000b0  3d 08 72 b0 af 46 b5 8f  6b 86                    |=.r..F..k.|
+>>> Flow 4 (server to client)
+00000000  14 03 03 00 01 01 16 03  03 00 24 59 20 4d c2 17  |..........$Y M..|
+00000010  8b 3c 9b 33 d9 f9 ef fb  80 18 1f 67 a7 58 12 89  |.<.3.......g.X..|
+00000020  4e 73 0f 2d 7b e6 c4 a6  79 73 01 da 22 e8 54 17  |Ns.-{...ys..".T.|
+00000030  03 03 00 21 36 ca 64 0f  4a 12 a5 50 3d 97 bb 39  |...!6.d.J..P=..9|
+00000040  02 fc ed d1 82 6a 9a 2e  21 79 f6 e1 b3 cc 32 db  |.....j..!y....2.|
+00000050  0f 5d b3 fb a5 15 03 03  00 16 51 f4 be 57 7a df  |.]........Q..Wz.|
+00000060  f1 f2 bd b5 51 5e 45 80  be 0b 9a 0c d1 19 3c 79  |....Q^E.......<y|

src/pkg/crypto/tls/ticket.go

@@ -153,7 +153,8 @@ func (c *Conn) encryptTicket(state *sessionState) ([]byte, error) {
 }
 
 func (c *Conn) decryptTicket(encrypted []byte) (*sessionState, bool) {
-	if len(encrypted) < aes.BlockSize+sha256.Size {
+	if c.config.SessionTicketsDisabled ||
+		len(encrypted) < aes.BlockSize+sha256.Size {
 		return nil, false
 	}
 

src/pkg/crypto/tls/tls.go

@@ -28,7 +28,7 @@ func Server(conn net.Conn, config *Config) *Conn {
 
 // Client returns a new TLS client side connection
 // using conn as the underlying transport.
-// The config cannot be nil: users must set either ServerHostname or
+// The config cannot be nil: users must set either ServerName or
 // InsecureSkipVerify in the config.
 func Client(conn net.Conn, config *Config) *Conn {
 	return &Conn{conn: conn, config: config, isClient: true}

src/pkg/debug/elf/file.go

@@ -522,13 +522,17 @@ func (f *File) applyRelocations(dst []byte, rels []byte) error {
 	if f.Class == ELFCLASS64 && f.Machine == EM_X86_64 {
 		return f.applyRelocationsAMD64(dst, rels)
 	}
+	if f.Class == ELFCLASS32 && f.Machine == EM_386 {
+		return f.applyRelocations386(dst, rels)
+	}
 
 	return errors.New("not implemented")
 }
 
 func (f *File) applyRelocationsAMD64(dst []byte, rels []byte) error {
-	if len(rels)%Sym64Size != 0 {
-		return errors.New("length of relocation section is not a multiple of Sym64Size")
+	// 24 is the size of Rela64.
+	if len(rels)%24 != 0 {
+		return errors.New("length of relocation section is not a multiple of 24")
 	}
 
 	symbols, _, err := f.getSymbols(SHT_SYMTAB)
@@ -570,6 +574,43 @@ func (f *File) applyRelocationsAMD64(dst []byte, rels []byte) error {
 	return nil
 }
 
+func (f *File) applyRelocations386(dst []byte, rels []byte) error {
+	// 8 is the size of Rel32.
+	if len(rels)%8 != 0 {
+		return errors.New("length of relocation section is not a multiple of 8")
+	}
+
+	symbols, _, err := f.getSymbols(SHT_SYMTAB)
+	if err != nil {
+		return err
+	}
+
+	b := bytes.NewReader(rels)
+	var rel Rel32
+
+	for b.Len() > 0 {
+		binary.Read(b, f.ByteOrder, &rel)
+		symNo := rel.Info >> 8
+		t := R_386(rel.Info & 0xff)
+
+		if symNo == 0 || symNo > uint32(len(symbols)) {
+			continue
+		}
+		sym := &symbols[symNo-1]
+
+		if t == R_386_32 {
+			if rel.Off+4 >= uint32(len(dst)) {
+				continue
+			}
+			val := f.ByteOrder.Uint32(dst[rel.Off : rel.Off+4])
+			val += uint32(sym.Value)
+			f.ByteOrder.PutUint32(dst[rel.Off:rel.Off+4], val)
+		}
+	}
+
+	return nil
+}
+
 func (f *File) DWARF() (*dwarf.Data, error) {
 	// There are many other DWARF sections, but these
 	// are the required ones, and the debug/dwarf package
@@ -603,6 +644,19 @@ func (f *File) DWARF() (*dwarf.Data, error) {
 		}
 	}
 
+	// When using clang we need to process relocations even for 386.
+	rel := f.Section(".rel.debug_info")
+	if rel != nil && rel.Type == SHT_REL && f.Machine == EM_386 {
+		data, err := rel.Data()
+		if err != nil {
+			return nil, err
+		}
+		err = f.applyRelocations(dat[1], data)
+		if err != nil {
+			return nil, err
+		}
+	}
+
 	abbrev, info, str := dat[0], dat[1], dat[2]
 	d, err := dwarf.New(abbrev, nil, nil, info, nil, nil, nil, str)
 	if err != nil {