This isn't everything, but it's a start.
Still producing bit-identical compiler output.

The semantics of the old back ends is preserved,
even when they are probably buggy.
There are some TODOs in gc/gsubr.c to
remove special cases to preserve bugs in 5g and 8g.

Change-Id: I28ae295fbfc94ef9df43e13ab96bd6fc2f194bc4
Reviewed-on: https://go-review.googlesource.com/3802Reviewed-by: Austin Clements <austin@google.com>

benchmark                               old ns/op     new ns/op     delta
BenchmarkFormatFloatDecimal             300           283           -5.67%
BenchmarkFormatFloat                    383           381           -0.52%
BenchmarkFormatFloatExp                 359           357           -0.56%
BenchmarkFormatFloatNegExp              357           358           +0.28%
BenchmarkFormatFloatBig                 468           430           -8.12%
BenchmarkAppendFloatDecimal             104           92.5          -11.06%
BenchmarkAppendFloat                    199           190           -4.52%
BenchmarkAppendFloatExp                 172           167           -2.91%
BenchmarkAppendFloatNegExp              172           169           -1.74%
BenchmarkAppendFloatBig                 280           235           -16.07%
BenchmarkAppendFloat32Integer           104           92.4          -11.15%
BenchmarkAppendFloat32ExactFraction     168           171           +1.79%
BenchmarkAppendFloat32Point             206           199           -3.40%
BenchmarkAppendFloat32Exp               167           167           +0.00%
BenchmarkAppendFloat32NegExp            167           166           -0.60%
BenchmarkAppendFloat64Fixed1            134           129           -3.73%
BenchmarkAppendFloat64Fixed2            144           136           -5.56%
BenchmarkAppendFloat64Fixed3            138           134           -2.90%
BenchmarkAppendFloat64Fixed4            145           138           -4.83%

Change-Id: Ia143840cb34cbd1cebd6b691dd0a45b7264b406c
Reviewed-on: https://go-review.googlesource.com/3920Reviewed-by: Alan Donovan <adonovan@google.com>

Change-Id: Ib3738492a2ec8fc99323e687168b17b7239db6ad
Reviewed-on: https://go-review.googlesource.com/4511Reviewed-by: Alan Donovan <adonovan@google.com>

Change-Id: Ic5f3864bc6d94d60b754e3ccf72b1d40c5c09117
Reviewed-on: https://go-review.googlesource.com/4510Reviewed-by: Rob Pike <r@golang.org>
Reviewed-by: Alan Donovan <adonovan@google.com>

Change-Id: Id98f7333fe6ae1b64e0469c6d01f02360c1f8f55
Reviewed-on: https://go-review.googlesource.com/4481Reviewed-by: Alan Donovan <adonovan@google.com>

This avoids surprises.

Change-Id: Iaae67da2d12e29c4e797ad6313e0895f7ce80cb1
Reviewed-on: https://go-review.googlesource.com/4480Reviewed-by: Rob Pike <r@golang.org>
Reviewed-by: Alan Donovan <adonovan@google.com>

Add local workbufs to the m struct in order to reduce contention.
Add consistency checks for workbuf ownership.
Chain workbufs through call change to avoid swapping them
to and from the m struct.
Adjust the size of the workbuf so that the mutators can
more frequently pass modifications to the GC thus shifting
some work from the STW mark termination phase to the concurrent
mark phase.

Change-Id: I557b53af34ad9972265e0ed9f5996e52d548563d
Reviewed-on: https://go-review.googlesource.com/3972Reviewed-by: Austin Clements <austin@google.com>

Fixes #9791

g.issystem flag setup races with other code wherever we set it.
Even if we set both in parent goroutine and in the system goroutine,
it is still possible that some other goroutine crashes
before the flag is set. We could pass issystem flag to newproc1,
but we start all goroutines with go nowadays.

Instead look at g.startpc to distinguish system goroutines (similar to topofstack).

Change-Id: Ia3467968dee27fa07d9fecedd4c2b00928f26645
Reviewed-on: https://go-review.googlesource.com/4113Reviewed-by: Keith Randall <khr@golang.org>

benchmark                                  old allocs     new allocs     delta
BenchmarkSkipValue                         14914          14202          -4.77%

Change-Id: I40e1fe8843cc6a099a2abfcd814ecc2a2d6a5b1f
Reviewed-on: https://go-review.googlesource.com/3744Reviewed-by: Keith Randall <khr@golang.org>

Update #8832

This is probably not the root cause of the issue.
Resolve TODO about setting unusedsince on a wrong span.

Change-Id: I69c87e3d93cb025e3e6fa80a8cffba6ad6ad1395
Reviewed-on: https://go-review.googlesource.com/4390Reviewed-by: Keith Randall <khr@golang.org>

Currently, if there is a VERSION.cache, running make.bash will set
runtime.theVersion to the revision as of the *last* make.bash run
instead of the current make.bash run.

For example,

$ git rev-parse --short HEAD
5c4a86d0
$ ./make.bash
...
$ cat ../VERSION.cache
devel +5c4a86d0 Tue Feb 10 01:46:30 2015 +0000
$ git checkout a1dbb920
$ ./make.bash
...
$ go version
go version devel +5c4a86d0 Tue Feb 10 01:46:30 2015 +0000 linux/amd64
$ ./make.bash
...
$ go version
go version devel +a1dbb920 Tue Feb 10 02:31:27 2015 +0000 linux/amd64

This happens because go tool dist reads the potentially stale
VERSION.cache into goversion during early initialization; then cleans,
which deletes VERSION.cache; then builds the runtime using the stale
revision read in to goversion.  It isn't until make later in the build
process, when make.bash invokes go tool dist again, that VERSION.cache
gets updated with the current revision.

To address this, simply don't bother fetching the version until go
tool dist needs it and don't bother caching the value in memory.  This
is more robust since it interacts with cleaning in the expected ways.
Futhermore, there's no downside to eliminating the in-memory cache;
the file system cache is perfectly reasonable for the whole three
times make.bash consults it.

Change-Id: I8c480100e56bb2db0816e8a088177004d9e87973
Reviewed-on: https://go-review.googlesource.com/4540Reviewed-by: Russ Cox <rsc@golang.org>

If an absolute domain name (i.e. ends in a '.' like "example.com.") is used
with ssl/tls, the certificate will be reported as invalid. In matchHostnames,
the host and patterns are split on '.' and if the lengths of the resulting
slices do not match, the function returns false. When splitting an absolute
domain name on '.', the slice will have an extra empty string at the end. This
empty string should be discarded before comparison, if present.

Fixes #9828

Change-Id: I0e39674b44a6f93b5024497e76cf1b550832a61d
Reviewed-on: https://go-review.googlesource.com/4380Reviewed-by: Adam Langley <agl@golang.org>
TryBot: Adam Langley <agl@golang.org>

Container symbols shouldn't be considered as functions in the functab.
Having them present probably messes up function lookup, as you might get
the descriptor of the container instead of the descriptor of the actual
function on the stack.  It also messed up the findfunctab because these
entries caused off-by-one errors in how functab entries were counted.

Normal code is not affected - it only changes (& hopefully fixes) the
behavior for libraries linked as a unit, like:
  net
  runtime/cgo
  runtime/race

Fixes #9804

Change-Id: I81e036e897571ac96567d59e1f1d7f058ca75e85
Reviewed-on: https://go-review.googlesource.com/4290Reviewed-by: Russ Cox <rsc@golang.org>

Change-Id: I695b89ec2b63233d94c49c4a40a57b50350ec67c
Reviewed-on: https://go-review.googlesource.com/4350
TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

Change-Id: I69f24887601e491d6d722bfeb2952d927df8ad80
Reviewed-on: https://go-review.googlesource.com/4351Reviewed-by: Ian Lance Taylor <iant@golang.org>

Fixes #9814.

Change-Id: I1be49efae0648038f590eeca1262037bf1af3df5
Reviewed-on: https://go-review.googlesource.com/4240Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Russ Cox <rsc@golang.org>

- Frexp, Ldexp are equivalents to the corresponding math functions.
- Set now has the same prec behavior as the other functions
- Copy is a true assignment (replaces old version of Set)
- Cmp now handles infinities
- more tests

Change-Id: I0d33980c08be3095b25d7b3d16bcad1aa7abbd0f
Reviewed-on: https://go-review.googlesource.com/4292Reviewed-by: Alan Donovan <adonovan@google.com>

The sanity checks at the beginning of WriteMsgUDP were too
strict, and did not allow a case sendmsg(2) suppports: sending
to a connected UDP socket.

This fixes the sanity checks. Either the socket is unconnected,
and a destination addresses is required (what all existing callers
must have been doing), or the socket is connected and an explicit
destination address must not be used.

Fixes #9807

Change-Id: I08d4ec3c2bf830335c402acfc0680c841cfcec71
Reviewed-on: https://go-review.googlesource.com/3951Reviewed-by: Mikio Hara <mikioh.mikioh@gmail.com>

Change-Id: I5cedd9e53f4e020aea74d498d0db88d79a95260c
Reviewed-on: https://go-review.googlesource.com/2718Reviewed-by: Minux Ma <minux@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

I think the test was meant to test requests to 'localhost:80' instead
of 'localhost:80:80'. It passes even with 'localhost:80:80' because
net.SplitHostPort fails inside useProxy. Please comment if you want to
leave old 'localhost:80' is the list too to check old code path.

Change-Id: Ic4cd21901563449e3d4e2f4c8caf723f4ca15bac
u
Reviewed-on: https://go-review.googlesource.com/4293Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Change-Id: Iad1764707d173a09467fd36e8c49a58147f12219
Reviewed-on: https://go-review.googlesource.com/4320Reviewed-by: Minux Ma <minux@golang.org>

This CL introduces new methods for 'context' type, so we can
manipulate its values in an architecture independent way.

Use new methods to replace both 386 and amd64 versions of
dosigprof with single piece of code.

There is more similar code to be converted in the following CLs.

Also remove os_windows_386.go and os_windows_amd64.go. These
contain unused functions.

Change-Id: I28f76aeb97f6e4249843d30d3d0c33fb233d3f7f
Reviewed-on: https://go-review.googlesource.com/2790Reviewed-by: Minux Ma <minux@golang.org>

I am an idiot but the failure to implement this means we can decide
exactly what its design should be for 1.5

Change-Id: Ie2b025fcd899d306ddeddd09d1d0e8f9a99ab7a8
Reviewed-on: https://go-review.googlesource.com/4291Reviewed-by: Minux Ma <minux@golang.org>

Fixes #5395

Change-Id: I4322bc8a974d04d9bae6b48c71c5d32d9252973c
Reviewed-on: https://go-review.googlesource.com/3024Reviewed-by: Alex Brainman <alex.brainman@gmail.com>

Fixes #9732
Fixes #9819

Rather than detecting vfp support via catching SIGILL signals,
parse the contents of /proc/cpuinfo.

As the GOARM values for NaCl and freebsd are hard coded, this parsing
logic only needs to support linux/arm.

This change also fixes the nacl/arm build which is broken because the
first stage of nacltest.bash is executed with GOARM=5, embedding that
into 5g.

The second stage of nacltest.bash correctly detects GOARM=7, but this is
ignored as we pass --no-clean at that point, and thus do not replace
the compiler.

Lastyly, include a fix to error message in nacltest.bash

Change-Id: I13f306ff07a99b44b493fade72ac00d0d5097e1c
Reviewed-on: https://go-review.googlesource.com/3981Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

CL 2118 makes the assumption that all references to runtime.tlsg
should be accompanied by a declaration of runtime.tlsg if its type
should be a normal variable, instead of a placeholder for TLS
relocation.

Because if runtime.tlsg is not declared by the runtime package,
the type of runtime.tlsg will be zero, so fix the check in liblink
to look for 0 instead of STLSBSS (the type will be initialized by
cmd/ld, but cmd/ld doesn't run during assembly).

Change-Id: I691ac5c3faea902f8b9a0b963e781b22e7b269a7
Reviewed-on: https://go-review.googlesource.com/4030Reviewed-by: David Crawshaw <crawshaw@golang.org>

Change-Id: I208c8ac44d1a8882d8fdeb18347dc20941e20374
Reviewed-on: https://go-review.googlesource.com/4250Reviewed-by: Alan Donovan <adonovan@google.com>

"go test -v" buffers output if more than one package is
being tested to avoid mixing the outputs from multiple
tests running in parallel. It currently enables streaming
if there's only a single package under test.

It is ok to stream output from multiple tests if we know
that they're not going to be running in parallel.

To see the difference: go test -v -p=1 runtime fmt -short

Change-Id: Idc24575c899eac30d553e0bf52b86f90e189392d
Reviewed-on: https://go-review.googlesource.com/4153Reviewed-by: Ian Lance Taylor <iant@golang.org>

- better and more consistent documentation
- more functions implemented
- more tests

Change-Id: If4c591e7af4ec5434fbb411a48dd0f8add993720
Reviewed-on: https://go-review.googlesource.com/4140Reviewed-by: Alan Donovan <adonovan@google.com>

This is an alternative to http://golang.org/cl/4150,
and is motivated by a review comment on that CL.

testKillProcess() tries to build and run the Go equivalent
for "sleep 1". This doesn't work for testing cross compilers
since the Go compiler is not available on the targets. This
change embeds the "sleep 1" functionality within the "os.test"
binary itself.

Change-Id: I6bad513deaa6c9e2704e70319098eb4983f1bb23
Reviewed-on: https://go-review.googlesource.com/4190Reviewed-by: Ian Lance Taylor <iant@golang.org>

When we use dialMulti we also allocate dialSingle closure for no reason.

Change-Id: I074282a9d6e2c2a1063ab311a1b95e10fe65219f
Reviewed-on: https://go-review.googlesource.com/4119Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Change-Id: Ib46a42fc873066b1cc00368fe43648f08dce48bd
Reviewed-on: https://go-review.googlesource.com/4200Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

There are no behavior changes in this CL, only specifying the status
quo. A follow-up CL, https://go-review.googlesource.com/#/c/2660/, will
change the EncodeToken behavior.

Change-Id: I6ecbcfb05ae681de71fa1099d054df2826ed4acb
Reviewed-on: https://go-review.googlesource.com/4167Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

The Addr might be a stack variable with uninitialized fields.

Fixes #9777.

Change-Id: I799786e3d8b2e17e069725bc66a076cf9ca11f93
Signed-off-by: Shenghou Ma <minux@golang.org>
Reviewed-on: https://go-review.googlesource.com/3932Reviewed-by: Dave Cheney <dave@cheney.net>
Reviewed-by: Keith Randall <khr@golang.org>

There was a small typo in the comment before the Stable function.

Change-Id: Ia6fa5272aa7869124a637d2eeda81c4f35ef46c8
Reviewed-on: https://go-review.googlesource.com/4201Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Change-Id: I66d923abbef13ba35c76c6f6da5b14c6d534127c
Reviewed-on: https://go-review.googlesource.com/4165Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Reviewed-by: David du Colombier <0intro@gmail.com>

If we cannot load timezone information for a reason other than the
zoneinfo file not existing, return it since that will be much more
useful in debugging failures than "unknown time zone XYZ".

Fixes #9723.

Change-Id: I3aa5774859cec28e584d16bcc1fef0705d95288c
Reviewed-on: https://go-review.googlesource.com/3984Reviewed-by: Ian Lance Taylor <iant@golang.org>

Reverts https://golang.org/cl/119530044 (OS X 10.10 Yosemite beta
14A299l workaround), since it was fixed in the final Yosemite release.

I verified that the C program http://swtch.com/~rsc/readdirbug.c
passes on Yosemite.

Adds a new test to the os package too, to verify that reading a
regular file as a directory fails.

Fixes #9789 (ReadDir: no error if dirname is a file)

Change-Id: I75286cef88fbb2ebccf045b479e33c810749dcbc
Reviewed-on: https://go-review.googlesource.com/4164Reviewed-by: Dave Cheney <dave@cheney.net>

This reverts commit 11d1c05f.
See #9296 for details.

Change-Id: I89a36351cb007836662f28a611af5616818b95fe
Reviewed-on: https://go-review.googlesource.com/1536Reviewed-by: Minux Ma <minux@golang.org>
Reviewed-by: Russ Cox <rsc@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

Change-Id: If697ab554e6cb5545d99c6b103ed8bc54f69ed48
Reviewed-on: https://go-review.googlesource.com/4161Reviewed-by: Andrew Gerrand <adg@golang.org>