sort: optimize average calculation in symMerge and doPivot.

Change code of the form `i + (j-i)/2` to `int(uint(i+j) >> 1)`.

The optimized average calculation uses fewer instructions to calculate
the average without overflowing at the addition.

Analogous to https://golang.org/cl/36332.

name                 old time/op  new time/op  delta
StableString1K-4     49.6µs ± 3%  49.1µs ± 8%    ~     (p=0.659 n=16+19)
StableInt1K-4         160µs ±10%   148µs ± 5%  -7.29%  (p=0.000 n=20+16)
StableInt1K_Slice-4   139µs ± 4%   136µs ± 3%  -2.52%  (p=0.000 n=20+16)
StableInt64K-4       8.84ms ± 6%  8.57ms ± 5%  -3.07%  (p=0.001 n=20+19)
Stable1e2-4           162µs ±19%   147µs ±16%  -8.79%  (p=0.002 n=20+20)
Stable1e4-4          31.0ms ± 5%  30.6ms ± 5%    ~     (p=0.221 n=20+20)
Stable1e6-4           6.37s ± 3%   6.27s ± 2%  -1.67%  (p=0.000 n=19+20)

Change-Id: I1cea0bcb9ace8ef7e48b8fab772e41b4b2170da9
Reviewed-on: https://go-review.googlesource.com/36570
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Russ Cox <rsc@golang.org>

crypto/sha512: Add AVX2 version for AMD64

name          old time/op    new time/op     delta
Hash8Bytes-6     913ns ± 0%      667ns ± 0%  -26.91%  (p=0.000 n=10+10)
Hash1K-6        6.58µs ± 0%     4.23µs ± 0%  -35.69%  (p=0.000 n=10+9)
Hash8K-6        45.9µs ± 0%     28.1µs ± 0%  -38.93%  (p=0.000 n=10+10)

name          old speed      new speed       delta
Hash8Bytes-6  8.76MB/s ± 0%  11.99MB/s ± 0%  +36.87%  (p=0.000 n=10+8)
Hash1K-6       156MB/s ± 0%    242MB/s ± 0%  +55.49%  (p=0.000 n=10+9)
Hash8K-6       178MB/s ± 0%    292MB/s ± 0%  +63.74%  (p=0.000 n=10+10)

Change-Id: Ic9211d68b02935b2195995f264ec57d6bc36f713
Reviewed-on: https://go-review.googlesource.com/36630
Run-TryBot: Ilya Tocar <ilya.tocar@intel.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Russ Cox <rsc@golang.org>

crypto/x509: ignore CN if SAN extension present.

The code previously tested only whether DNS-name SANs were present in a
certificate which is only approximately correct. In fact, /any/ SAN
extension, including one with no DNS names, should cause the CN to be
ignored.

Change-Id: I3d9824918975be6d4817e7cbb48ed1b0c5a2fc8b
Reviewed-on: https://go-review.googlesource.com/36696
Run-TryBot: Adam Langley <agl@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd/{asm,internal/obj/s390x}, math: remove emulated float instructions

The s390x port was based on the ppc64 port and, because of the way the
port was done, inherited some instructions from it. ppc64 supports
3-operand (4-operand for FMADD etc.) floating point instructions
but s390x doesn't (the destination register is always an input) and
so these were emulated.

There is a bug in the emulation of FMADD whereby if the destination
register is also a source for the multiplication it will be
clobbered. This doesn't break any assembly code in the std lib but
could affect future work.

To fix this I have gone through the floating point instructions and
removed all unnecessary 3-/4-operand emulation. The compiler doesn't
need it and assembly writers don't need it, it's just a source of
bugs.

I've also deleted the FNMADD family of emulated instructions. They
aren't used anywhere.

Change-Id: Ic07cedcf141a6a3b43a0c84895460f6cfbf56c04
Reviewed-on: https://go-review.googlesource.com/33350
Run-TryBot: Michael Munday <munday@ca.ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>

math: check overflow in amd64 Exp implementation

Unlike the pure go implementation used by every other architecture,
the amd64 asm implementation of Exp does not fail early if the
argument is known to overflow. Make it fail early.

Cost of the check is < 1ns (on an old Sandy Bridge machine):

name   old time/op  new time/op  delta
Exp-4  18.3ns ± 1%  18.7ns ± 1%  +2.08%  (p=0.000 n=18+20)

Fixes #14932
Fixes #18912

Change-Id: I04b3f9b4ee853822cbdc97feade726fbe2907289
Reviewed-on: https://go-review.googlesource.com/36271
Run-TryBot: Alberto Donizetti <alb.donizetti@gmail.com>
Reviewed-by: Russ Cox <rsc@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

encoding/json: clarify documention for Unmarshal into a pointer.

Fixes #18730.

Change-Id: If3ef28e62f7e449d4c8dc1dfd78f7d6f5a87ed26
Reviewed-on: https://go-review.googlesource.com/36478
Reviewed-by: Russ Cox <rsc@golang.org>

database/sql: ensure driverConns are closed if not returned to pool

Previously if a connection was requested but timed out during the
request and when acquiring the db.Lock the connection request
is fulfilled and the request is unable to be returned to the
connection pool, then then driver connection would not be closed.

No tests were added or modified because I was unable to determine
how to trigger this situation without something invasive.

Change-Id: I9d4dc680e3fdcf63d79d212174a5b8b313f363f1
Reviewed-on: https://go-review.googlesource.com/36641
Reviewed-by: Russ Cox <rsc@golang.org>

runtime/pprof: merge internal/protopprof into pprof package

These are very tightly coupled, and internal/protopprof is small.
There's no point to having a separate package.

Change-Id: I2c8aa49c9e18a7128657bf2b05323860151b5606
Reviewed-on: https://go-review.googlesource.com/36711
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

image: fix the overlap check in Rectangle.Intersect.

This is a re-roll of a previous commit,
a855da29dbd7a80c4d87a421c1f88a8603c020fa, which was rolled back in
14347ee480968c712ea885a4ea62779fd8a0dc44.

It was rolled back because it broke a unit test in image/gif. The
image/gif code was fixed by 9ef65dbe0683634a2e8a557d12267d0309ae1570
"image/gif: fix frame-inside-image bounds checking".

The original commit message:

image: fix the overlap check in Rectangle.Intersect.

The doc comment for Rectangle.Intersect clearly states, "If the two
rectangles do not overlap then the zero rectangle will be returned."
Prior to this fix, calling Intersect on adjacent but non-overlapping
rectangles would return an empty but non-zero rectangle.

The fix essentially changes
if r.Min.X > r.Max.X || r.Min.Y > r.Max.Y { etc }
to
if r.Min.X >= r.Max.X || r.Min.Y >= r.Max.Y { etc }
(note that the > signs have become >= signs), but changing that line to:
if r.Empty() { etc }
seems clearer (and equivalent).

Change-Id: I2e3af1f1686064a573b2e513b39246fe60c03631
Reviewed-on: https://go-review.googlesource.com/36734
Reviewed-by: Rob Pike <r@golang.org>
Run-TryBot: Nigel Tao <nigeltao@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile: use constants directly for fast map access calls

CL 35554 taught order.go to use static variables
for constants that needed to be addressable for runtime routines.
However, there is one class of runtime routines that
do not actually need an addressable value: fast map access routines.
This CL teaches order.go to avoid using static variables
for addressability in those cases.
Instead, it avoids introducing a temp at all,
which the backend would just have to optimize away.

Fixes #19015.

Change-Id: I5ef780c604fac3fb48dabb23a344435e283cb832
Reviewed-on: https://go-review.googlesource.com/36693
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>

image/gif: fix frame-inside-image bounds checking.

The semantics of the Go image.Rectangle type is that the In and
Intersects methods treat empty rectangles specially. There are multiple
valid representations of an empty image.Rectangle. One of them is the
zero image.Rectangle but there are others. They're obviously not all
equal in the == sense, so we shouldn't use != to check GIF's semantics.

This change will allow us to re-roll
a855da29dbd7a80c4d87a421c1f88a8603c020fa "image: fix the overlap check
in Rectangle.Intersect" which was rolled back in
14347ee480968c712ea885a4ea62779fd8a0dc44.

Change-Id: Ie1a0d092510a7bb6170e61adbf334b21361ff9e6
Reviewed-on: https://go-review.googlesource.com/36639
Run-TryBot: Nigel Tao <nigeltao@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rob Pike <r@golang.org>

net/http: improve handling of errors in Dir.Open

The current implementation fails to produce an "IsNotExist" error on some
platforms (unix) for certain situations where it would be expected. This causes
downstream consumers, like FileServer, to emit 500 errors instead of a 404 for
some non-existant paths on certain platforms but not others.

As an example, os.Open("/index.html/foo") on a unix-type system will return
syscall.ENOTDIR, which os.IsNotExist cannot return true for (because the
error code is ambiguous without context). On windows, this same example
would result in os.IsNotExist returning true -- since the returned error is
specific.

This change alters Dir.Open to look up the tree for an "IsPermission" or
"IsNotExist" error to return, or a non-directory, returning os.ErrNotExist in
the last case. For all other error scenarios, the original error is returned.
This ensures that downstream code, like FileServer, receive errors that behave
the same across all platforms.

Fixes #18984

Change-Id: Id7d16591c24cd96afddb6d8ae135ac78da42ed37
Reviewed-on: https://go-review.googlesource.com/36635
Reviewed-by: Yasuhiro MATSUMOTO <mattn.jp@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

spec: refer to "not defined type" rather than "unnamed type" in conversions

We missed this in https://golang.org/cl/36213.
Thanks to Chris Hines for pointing it out.

For #18130.

Change-Id: I6279ab19966c4391c4b4458b21fd2527d3f949dd
Reviewed-on: https://go-review.googlesource.com/36691
Reviewed-by: Ian Lance Taylor <iant@golang.org>

cmd/compile/internal/syntax: removed gcCompat code needed to pass orig. tests

The gcCompat mode was introduced to match the new parser's node position
setup exactly with the positions used by the original parser. Some of the
gcCompat adjustments were required to satisfy syntax error test cases,
and the rest were required to make toolstash cmp pass.

This change removes the former gcCompat adjustments and instead adjusts
the respective test cases as necessary. In some cases this makes the error
lines consistent with the ones reported by gccgo.

Where it has changed, the position associated with a given syntactic construct
is the position (line/col number) of the left-most token belonging to the
construct.

Change-Id: I5b60c00c5999a895c4d6d6e9b383c6405ccf725c
Reviewed-on: https://go-review.googlesource.com/36695
Run-TryBot: Robert Griesemer <gri@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>

crypto/dsa: also use fromHex in TestSignAndVerify.

This change contains a very minor tidy-up to a test.

Change-Id: I3a8c0168bcdcbf90cacbbac2566c8423c92129f8
Reviewed-on: https://go-review.googlesource.com/33726
Reviewed-by: Adam Langley <agl@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Adam Langley <agl@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile: disallow combining nosplit and systemstack

go:systemstack works by tweaking the stack check prologue to check
against a different bound, while go:nosplit removes the stack check
prologue entirely. Hence, they can't be used together. Make the build
fail if they are.

Change-Id: I2d180c4b1d31ff49ec193291ecdd42921d253359
Reviewed-on: https://go-review.googlesource.com/36710
Run-TryBot: Austin Clements <austin@google.com>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

crypto/aes: fix build failure by changing VORL to VOR

Recently, a commit (85ecc51c) changed the instruction from VORL to VOR.

Fixes #19014

Change-Id: I9a7e0b5771842b1abb5afc73dc41d5e7960cf390
Reviewed-on: https://go-review.googlesource.com/36625
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

net/rpc: fix aliasing in TestAcceptExitAfterListenerClose

TestRPC writes to newServer and newServerAddr guarded with a
sync.Once.
TestAcceptExitAfterListenerClose was overwriting those variables,
which caused the second invocation of TestRPC within a single process
to fail.

A second invocation can occur as a result of running the test with
multiple values for the -cpu flag.

fixes #19001.

Change-Id: I291bacf44aefb49c2264ca0290a28248c026f80e
Reviewed-on: https://go-review.googlesource.com/36624
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>

cmd/compile: remove some allocs from CSE

Pick up a few pennies:

* CSE gets run twice for each function,
but the set of Aux values doesn't change.
Avoid populating it twice.

* Don't bother populating auxmap for values
that can't be CSE'd anyway.

name       old alloc/op     new alloc/op     delta
Template       41.0MB ± 0%      40.7MB ± 0%  -0.61%  (p=0.008 n=5+5)
Unicode        32.3MB ± 0%      32.3MB ± 0%  -0.22%  (p=0.008 n=5+5)
GoTypes         122MB ± 0%       121MB ± 0%  -0.55%  (p=0.008 n=5+5)
Compiler        482MB ± 0%       479MB ± 0%  -0.58%  (p=0.008 n=5+5)
SSA             865MB ± 0%       862MB ± 0%  -0.35%  (p=0.008 n=5+5)
Flate          26.5MB ± 0%      26.5MB ± 0%    ~     (p=0.056 n=5+5)
GoParser       32.6MB ± 0%      32.4MB ± 0%  -0.58%  (p=0.008 n=5+5)
Reflect        84.2MB ± 0%      83.8MB ± 0%  -0.57%  (p=0.008 n=5+5)
Tar            27.7MB ± 0%      27.6MB ± 0%  -0.37%  (p=0.008 n=5+5)
XML            44.7MB ± 0%      44.5MB ± 0%  -0.53%  (p=0.008 n=5+5)

name       old allocs/op    new allocs/op    delta
Template         373k ± 0%        373k ± 1%    ~     (p=1.000 n=5+5)
Unicode          326k ± 0%        325k ± 0%    ~     (p=0.548 n=5+5)
GoTypes         1.16M ± 0%       1.16M ± 0%    ~     (p=0.841 n=5+5)
Compiler        4.16M ± 0%       4.15M ± 0%    ~     (p=0.222 n=5+5)
SSA             7.57M ± 0%       7.56M ± 0%  -0.22%  (p=0.008 n=5+5)
Flate            238k ± 1%        239k ± 1%    ~     (p=0.690 n=5+5)
GoParser         304k ± 0%        304k ± 0%    ~     (p=1.000 n=5+5)
Reflect         1.01M ± 0%       1.00M ± 0%  -0.31%  (p=0.016 n=4+5)
Tar              245k ± 0%        245k ± 1%    ~     (p=0.548 n=5+5)
XML              393k ± 0%        391k ± 1%    ~     (p=0.095 n=5+5)

Change-Id: I78f1ffe129bd8fd590b7511717dd2bf9f5ecbd6d
Reviewed-on: https://go-review.googlesource.com/36690
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Daniel Martí <mvdan@mvdan.cc>
Reviewed-by: Cherry Zhang <cherryyz@google.com>

crypto/aes: improve performance for aes on ppc64le

Add asm implementation for AES in order to make use of VMX cryptographic
acceleration instructions for POWER8. There is a speed boost of over 10
times using those instructions:

Fixes #18076

                        old ns/op  new ns/op  delta
BenchmarkEncrypt-20     337        30.3       -91.00%
BenchmarkDecrypt-20     347        30.5a      -91.21%
BenchmarkExpand-20      1180       130        -88.98%

                        old MB/s   new MB/s   speedup
BenchmarkEncrypt-20     47.38      527.68     11.13x
BenchmarkDecrypt-20     46.05      524.45     11.38x

Change-Id: Ifa4d1b508f4803cc72dcaad97acc8495d651b019
Reviewed-on: https://go-review.googlesource.com/33587
Run-TryBot: Lynn Boger <laboger@linux.vnet.ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Lynn Boger <laboger@linux.vnet.ibm.com>

crypto/tls: don't hold lock when closing underlying net.Conn.

There's no need to hold the handshake lock across this call and it can
lead to deadlocks if the net.Conn calls back into the tls.Conn.

Fixes #18426.

Change-Id: Ib1b2813cce385949d970f8ad2e52cfbd1390e624
Reviewed-on: https://go-review.googlesource.com/36561
Run-TryBot: Adam Langley <agl@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd/compile: rules change to use ANDN more effectively on ppc64x

Currently there are cases where an XOR with -1 followed by an AND
is generanted when it could be done with just an ANDN instruction.

Changes to PPC64.rules and required files allows this change
in generated code.  Examples of this occur in sha3 among others.

Fixes: #18918
Change-Id: I647cb9b4a4aaeebb27db85f8bf75487d78f720c9
Reviewed-on: https://go-review.googlesource.com/36218
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Chase <drchase@google.com>
Reviewed-by: Carlos Eduardo Seo <cseo@linux.vnet.ibm.com>

runtime: enable/disable SIGPROF if needed when profiling

This ensures that SIGPROF is handled correctly when using
runtime/pprof in a c-archive or c-shared library.

Separate profiler handling into pre-process changes and per-thread
changes. Simplify the Windows code slightly accordingly.

Fixes #18220.

Change-Id: I5060f7084c91ef0bbe797848978bdc527c312777
Reviewed-on: https://go-review.googlesource.com/34018
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
Run-TryBot: Austin Clements <austin@google.com>

crypto/x509: sort the list of fields used by CreateCertificateRequest.

Change-Id: I67589cb9e728e6c7df5ef6e981189193154338d3
Reviewed-on: https://go-review.googlesource.com/36559
Reviewed-by: Ian Lance Taylor <iant@golang.org>

crypto/x509: CreateCertificateRequest reads ExtraExtensions, not Extensions.

Fixes #18899.

Change-Id: I6a4bf0aad9cf1dbe6691ba4e4c478fcb33c44528
Reviewed-on: https://go-review.googlesource.com/36558
Reviewed-by: Ian Lance Taylor <iant@golang.org>

crypto/x509: sort the list of fields used by CreateCertificate.

Change-Id: I20f4419ca377ee9428075e42db0bad46a75d983f
Reviewed-on: https://go-review.googlesource.com/36557
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

crypto/x509: document AuthorityKeyId and don't mutate it.

The AuthorityKeyId value from the template was used by
CreateCertificate, but that wasn't documented. Also, CreateCertificate
would stash a value in the template if it needed to override it, which
was wrong: template should be read-only.

Fixes #18962.

Change-Id: Ida15c54c341e5bbf553756e8aa65021d8085f453
Reviewed-on: https://go-review.googlesource.com/36556
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

plugin: remove unnecessary import "C" from example

It seems that it is not needed to import the pseudo package "C"
for the plugin to be built correctly.
Removing it to avoid confusion.

Change-Id: I62838a953ad2889881bfbfd1a36141661565f033
Reviewed-on: https://go-review.googlesource.com/36638
Reviewed-by: David Crawshaw <crawshaw@golang.org>

syscall: remove "use" function and calls from generated code.

Update syscall code generators to set build tags.

Regenerate zsyscall files, which makes the following changes:
- remove calls to "use"
- update build tags, adding missing ones in some cases
- "stat" renamed to "st" in some cases
- "libc_Utimes" renamed "libc_utimes" in one case

I'll mirror this change to x/sys/unix once committed.

Change-Id: Ic07e0ae1433dd133eb57e8dd2a3b86a62aab4eda
Reviewed-on: https://go-review.googlesource.com/36616
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd/asm, cmd/internal/obj/ppc64: Add ISA 2.05, 2.06 and 2.07 instructions.

This change adds instructions from ISA 2.05, 2.06 and 2.07 that are frequently
used in assembly optimizations for ppc64.

It also fixes two problems:

  * the implementation of RLDICR[CC]/RLDICL[CC] did not consider all possible
  cases for the bit mask.
  * removed two non-existing instructions that were added by mistake in the VMX
  implementation (VORL/VANDL).

Change-Id: Iaef4e5c6a5240c2156c6c0f28ad3bcd8780e9830
Reviewed-on: https://go-review.googlesource.com/36230
Run-TryBot: Lynn Boger <laboger@linux.vnet.ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Lynn Boger <laboger@linux.vnet.ibm.com>

text/template: fix method lookup on addressable nil pointer

Fixes #18816.

Change-Id: I4f8f1cac2680dbde492c56d3a5a038577605e7c1
Reviewed-on: https://go-review.googlesource.com/36542
Run-TryBot: Russ Cox <rsc@golang.org>
Reviewed-by: Rob Pike <r@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

time: optimize Now on darwin, windows

Fetch both monotonic and wall time together when possible.
Avoids skew and is cheaper.

Also shave a few ns off in conversion in package time.

Compared to current implementation (after monotonic changes):

name   old time/op  new time/op  delta
Now    19.6ns ± 1%   9.7ns ± 1%  -50.63%  (p=0.000 n=41+49) darwin/amd64
Now    23.5ns ± 4%  10.6ns ± 5%  -54.61%  (p=0.000 n=30+28) windows/amd64
Now    54.5ns ± 5%  29.8ns ± 9%  -45.40%  (p=0.000 n=27+29) windows/386

More importantly, compared to Go 1.8:

name   old time/op  new time/op  delta
Now     9.5ns ± 1%   9.7ns ± 1%   +1.94%  (p=0.000 n=41+49) darwin/amd64
Now    12.9ns ± 5%  10.6ns ± 5%  -17.73%  (p=0.000 n=30+28) windows/amd64
Now    15.3ns ± 5%  29.8ns ± 9%  +94.36%  (p=0.000 n=30+29) windows/386

This brings time.Now back in line with Go 1.8 on darwin/amd64 and windows/amd64.

It's not obvious why windows/386 is still noticeably worse than Go 1.8,
but it's better than before this CL. The windows/386 speed is not too
important; the changes just keep the two architectures similar.

Change-Id: If69b94970c8a1a57910a371ee91e0d4e82e46c5d
Reviewed-on: https://go-review.googlesource.com/36428
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

database/sql: replace the expr of timeunit * N with N * timeunit in test

Change-Id: I97981b30a9629916f896cb989cc2a42a8bdbef47
Reviewed-on: https://go-review.googlesource.com/36672
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

database/sql: fix nits in test

Change-Id: I451b33d8da8d97917f2b257e6a25392c6e6582db
Reviewed-on: https://go-review.googlesource.com/36671
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd/compile/internal/parser: improved syntax error for incorrect if/for/switch header

Starting the error message with "expecting" rather than "missing"
causes the syntax error mechanism to add additional helpful info
(it recognizes "expecting" but not "missing").

Fixes #17328.

Change-Id: I8482ca5e5a6a6b22e0ed0d831b7328e264156334
Reviewed-on: https://go-review.googlesource.com/36637
Run-TryBot: Robert Griesemer <gri@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

time: format negative monotonic times correctly in Time.String

Fixes #18993

Change-Id: Ia1fa20b6d82384b07e9ba5512b909439e0bec2a5
Reviewed-on: https://go-review.googlesource.com/36611
Run-TryBot: Caleb Spare <cespare@gmail.com>
Reviewed-by: Russ Cox <rsc@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile/internal/syntax: differentiate between ';' and '\n' in syntax errors

Towards better syntax error messages: With this change, the parser knows whether
a semicolon was an actual ';' in the source, or whether it was an automatically
inserted semicolon as result of a '\n' or EOF. Using this information in error
messages makes them more understandable.

For #17328.

Change-Id: I8cd9accee8681b62569d0ecef922d38682b401eb
Reviewed-on: https://go-review.googlesource.com/36636
Reviewed-by: Matthew Dempsky <mdempsky@google.com>

database/sql: do not exhaust connection pool on conn request timeout

Previously if a context was canceled while it was waiting for a
connection request, that connection request would leak.

To prevent this remove the pending connection request if the
context is canceled and ensure no connection has been sent on the channel.
This requires a change to how the connection requests are represented in the DB.

Fixes #18995

Change-Id: I9a274b48b8f4f7ca46cdee166faa38f56d030852
Reviewed-on: https://go-review.googlesource.com/36563
Reviewed-by: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

database/sql: fix typo

Change-Id: I09fdcebb939417f18af09ed57f24460724cab64f
Reviewed-on: https://go-review.googlesource.com/36632
Reviewed-by: Ian Lance Taylor <iant@golang.org>

net: merge FreeBSD and DragonFly sendfile support

The two files were identical except for comments.

Change-Id: Ifc300026c8e4584afa50a7b669099eaff146ea5d
Reviewed-on: https://go-review.googlesource.com/36631
Run-TryBot: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/gofmt: fix diff on Plan 9

On Plan 9, GNU diff is called ape/diff.

Fixes #18999.

Change-Id: I7cf6c23c97bcc47172bbf838fd9dd72aefa4c18b
Reviewed-on: https://go-review.googlesource.com/36650
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: David du Colombier <0intro@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

net/http: don't modify Request in StripPrefix

As of https://golang.org/cl/21530, rules are updated to state
that Handlers shouldn't modify the provided Request. This change
updates StripPrefix to follow that rule.

Resolves #18952.

Change-Id: I29bbb580722e871131fa75a97e6e038ec64fdfcd
Reviewed-on: https://go-review.googlesource.com/36483
Reviewed-by: Matt Layher <mdlayher@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Matt Layher <mdlayher@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile: handle DOT STRUCTLIT for zero-valued struct in SSA

CL 35261 makes SSA handle zero-valued STRUCTLIT, but DOT operation
was not handled.

Fixes #18994.

Change-Id: Ic7976036acca1523b0b14afac4d170797e8aee20
Reviewed-on: https://go-review.googlesource.com/36565
Run-TryBot: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile: include linknames in export data

This lets the compiler inline functions that contain a linknamed symbol.
Previously, the net/http tests would fail to build with -l=4 because
the compiler inlined functions that call net.byteIndex (which is
linknamed to strings.IndexByte).

This changes only the compiler-specific export data, so we don't need to
bump the export format version number.

The following benchmark results show how the size of package export data
is impacted by this change. These benchmarks were created by compiling
the go1 benchmark and running `go tool pack x` to extract the export
data from the resulting .a files.

name                                          old bytes   new bytes   delta
bufio                                        3.48k ± 0%  3.58k ± 0%  +2.90%
bytes                                        5.05k ± 0%  5.16k ± 0%  +2.16%
compress/bzip2                               2.61k ± 0%  2.68k ± 0%  +2.68%
compress/flate                               5.07k ± 0%  5.14k ± 0%  +1.40%
compress/gzip                                8.26k ± 0%  8.40k ± 0%  +1.70%
container/list                               1.69k ± 0%  1.76k ± 0%  +4.07%
context                                      3.93k ± 0%  4.01k ± 0%  +1.86%
crypto                                       1.03k ± 0%  1.03k ± 0%  +0.39%
crypto/aes                                     475 ± 0%    475 ± 0%  +0.00%
crypto/cipher                                1.18k ± 0%  1.18k ± 0%  +0.00%
crypto/des                                     502 ± 0%    502 ± 0%  +0.00%
crypto/dsa                                   5.71k ± 0%  5.77k ± 0%  +1.16%
crypto/ecdsa                                 6.67k ± 0%  6.75k ± 0%  +1.08%
crypto/elliptic                              6.28k ± 0%  6.35k ± 0%  +1.07%
crypto/hmac                                    464 ± 0%    464 ± 0%  +0.00%
crypto/internal/cipherhw                       313 ± 0%    313 ± 0%  +0.00%
crypto/md5                                     691 ± 0%    695 ± 0%  +0.58%
crypto/rand                                  5.37k ± 0%  5.43k ± 0%  +1.23%
crypto/rc4                                     512 ± 0%    512 ± 0%  +0.00%
crypto/rsa                                   7.05k ± 0%  7.12k ± 0%  +1.05%
crypto/sha1                                    756 ± 0%    760 ± 0%  +0.53%
crypto/sha256                                  523 ± 0%    523 ± 0%  +0.00%
crypto/sha512                                  662 ± 0%    662 ± 0%  +0.00%
crypto/subtle                                  835 ± 0%    873 ± 0%  +4.55%
crypto/tls                                   28.1k ± 0%  28.5k ± 0%  +1.30%
crypto/x509                                  17.7k ± 0%  17.9k ± 0%  +1.04%
crypto/x509/pkix                             9.75k ± 0%  9.90k ± 0%  +1.50%
encoding                                       473 ± 0%    473 ± 0%  +0.00%
encoding/asn1                                1.41k ± 0%  1.42k ± 0%  +1.00%
encoding/base64                              1.67k ± 0%  1.69k ± 0%  +0.90%
encoding/binary                              2.65k ± 0%  2.76k ± 0%  +4.07%
encoding/gob                                 13.3k ± 0%  13.5k ± 0%  +1.65%
encoding/hex                                   854 ± 0%    857 ± 0%  +0.35%
encoding/json                                11.9k ± 0%  12.1k ± 0%  +1.71%
encoding/pem                                   484 ± 0%    484 ± 0%  +0.00%
errors                                         360 ± 0%    361 ± 0%  +0.28%
flag                                         7.32k ± 0%  7.42k ± 0%  +1.48%
fmt                                          1.42k ± 0%  1.42k ± 0%  +0.00%
go/ast                                       15.7k ± 0%  15.8k ± 0%  +1.07%
go/parser                                    7.48k ± 0%  7.59k ± 0%  +1.55%
go/scanner                                   3.88k ± 0%  3.94k ± 0%  +1.39%
go/token                                     3.51k ± 0%  3.53k ± 0%  +0.60%
hash                                           507 ± 0%    507 ± 0%  +0.00%
hash/crc32                                     685 ± 0%    685 ± 0%  +0.00%
internal/nettrace                              474 ± 0%    474 ± 0%  +0.00%
internal/pprof/profile                       8.29k ± 0%  8.36k ± 0%  +0.89%
internal/race                                  511 ± 0%    511 ± 0%  +0.00%
internal/singleflight                          966 ± 0%    969 ± 0%  +0.31%
internal/syscall/unix                          427 ± 0%    427 ± 0%  +0.00%
io                                           3.48k ± 0%  3.52k ± 0%  +1.15%
io/ioutil                                    5.30k ± 0%  5.38k ± 0%  +1.53%
log                                          4.46k ± 0%  4.53k ± 0%  +1.59%
math                                         3.72k ± 0%  3.75k ± 0%  +0.75%
math/big                                     8.91k ± 0%  9.01k ± 0%  +1.15%
math/rand                                    1.29k ± 0%  1.30k ± 0%  +0.46%
mime                                         2.59k ± 0%  2.63k ± 0%  +1.55%
mime/multipart                               3.61k ± 0%  3.68k ± 0%  +1.80%
mime/quotedprintable                         2.20k ± 0%  2.25k ± 0%  +2.50%
net                                          21.1k ± 0%  21.3k ± 0%  +1.10%
net/http                                     56.6k ± 0%  57.3k ± 0%  +1.28%
net/http/httptest                            33.6k ± 0%  34.1k ± 0%  +1.38%
net/http/httptrace                           14.4k ± 0%  14.5k ± 0%  +1.29%
net/http/internal                            2.70k ± 0%  2.77k ± 0%  +2.59%
net/textproto                                4.51k ± 0%  4.60k ± 0%  +1.82%
net/url                                      1.71k ± 0%  1.73k ± 0%  +1.41%
os                                           11.3k ± 0%  11.4k ± 0%  +1.36%
path                                           587 ± 0%    589 ± 0%  +0.34%
path/filepath                                4.46k ± 0%  4.55k ± 0%  +1.88%
reflect                                      6.39k ± 0%  6.43k ± 0%  +0.72%
regexp                                       5.82k ± 0%  5.88k ± 0%  +1.12%
regexp/syntax                                3.22k ± 0%  3.24k ± 0%  +0.62%
runtime                                      12.9k ± 0%  13.2k ± 0%  +1.94%
runtime/cgo                                    229 ± 0%    229 ± 0%  +0.00%
runtime/debug                                3.66k ± 0%  3.72k ± 0%  +1.86%
runtime/internal/atomic                        905 ± 0%    905 ± 0%  +0.00%
runtime/internal/sys                         2.00k ± 0%  2.05k ± 0%  +2.55%
runtime/pprof                                4.16k ± 0%  4.23k ± 0%  +1.66%
runtime/pprof/internal/protopprof            11.5k ± 0%  11.7k ± 0%  +1.27%
runtime/trace                                  354 ± 0%    354 ± 0%  +0.00%
sort                                         1.63k ± 0%  1.68k ± 0%  +2.94%
strconv                                      1.84k ± 0%  1.85k ± 0%  +0.54%
strings                                      3.87k ± 0%  3.97k ± 0%  +2.48%
sync                                         1.51k ± 0%  1.52k ± 0%  +0.33%
sync/atomic                                  1.58k ± 0%  1.60k ± 0%  +1.27%
syscall                                      53.2k ± 0%  53.3k ± 0%  +0.20%
testing                                      8.14k ± 0%  8.26k ± 0%  +1.49%
testing/internal/testdeps                      597 ± 0%    598 ± 0%  +0.17%
text/tabwriter                               3.09k ± 0%  3.14k ± 0%  +1.85%
text/template                                15.4k ± 0%  15.7k ± 0%  +1.89%
text/template/parse                          8.90k ± 0%  9.12k ± 0%  +2.46%
time                                         5.75k ± 0%  5.86k ± 0%  +1.86%
unicode                                      4.62k ± 0%  4.62k ± 0%  +0.07%
unicode/utf16                                  693 ± 0%    706 ± 0%  +1.88%
unicode/utf8                                 1.05k ± 0%  1.07k ± 0%  +1.14%
vendor/golang_org/x/crypto/chacha20poly1305  1.25k ± 0%  1.26k ± 0%  +0.64%
vendor/golang_org/x/crypto/curve25519          392 ± 0%    392 ± 0%  +0.00%
vendor/golang_org/x/crypto/poly1305            426 ± 0%    426 ± 0%  +0.00%
vendor/golang_org/x/net/http2/hpack          4.19k ± 0%  4.26k ± 0%  +1.69%
vendor/golang_org/x/net/idna                   355 ± 0%    355 ± 0%  +0.00%
vendor/golang_org/x/net/lex/httplex            609 ± 0%    615 ± 0%  +0.99%
vendor/golang_org/x/text/transform           1.31k ± 0%  1.31k ± 0%  +0.08%
vendor/golang_org/x/text/unicode/norm        5.78k ± 0%  5.90k ± 0%  +2.06%
vendor/golang_org/x/text/width               1.24k ± 0%  1.24k ± 0%  +0.16%
[Geo mean]                                    2.49k       2.52k       +1.10%

Fixes #18167.

Change-Id: Ia5b7e70adc9652c7ee9954ca2efc1c59fa79be2b
Reviewed-on: https://go-review.googlesource.com/33911
Run-TryBot: David Lazar <lazard@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
Reviewed-by: David Chase <drchase@google.com>

net/http/pprof: return error when requested profile duration exceeds WriteTimeout

Updates Profile and Trace handlers to reject requests for durations >=
WriteTimeout.

Modifies go tool pprof to print the body of the http response when
status != 200.

Fixes #18755

Change-Id: I6faed21685693caf39f315f003039538114937b0
Reviewed-on: https://go-review.googlesource.com/35564
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

crypto/tls: fix link to more info about channel bindings

Link in the description of TLSUnique field of ConnectionState struct
leads to an article that is no longer available, so this commit
replaces it with link to a copy of the very same article on another
site.

Fixes #18842.

Change-Id: I8f8d298c4774dc0fbbad5042db0684bb3220aee8
Reviewed-on: https://go-review.googlesource.com/36052
Reviewed-by: Filippo Valsorda <hi@filippo.io>
Reviewed-by: Adam Langley <agl@golang.org>

doc: remove the confusing use of CL

CL (change list) pops out of nowhere and confuses the
reader. Use "change" instead to be consistent with the
rest of the document.

Fixes #18989.

Change-Id: I525a63a195dc6bb992c8ad0f10c2f2e1b2b952df
Reviewed-on: https://go-review.googlesource.com/36564
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd/compile: remove unnecessary type conversions on s390x

Some rules insert MOVDreg ops to ensure that type changes are kept.
If there is no type change (or the input is constant) then the MOVDreg
can be omitted, allowing further optimization.

Reduces the size of the .text section in the asm tool by ~33KB.

Change-Id: I386883bb35b843c7b99a269cd6840dca77cf4371
Reviewed-on: https://go-review.googlesource.com/36547
Run-TryBot: Michael Munday <munday@ca.ibm.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>

go/constant: use new math/big.IsInt and isUint predicates

Slightly cleaner and more readable code.

Change-Id: I35263dbf338861b0a1bd62d59417b6a2c6a4e670
Reviewed-on: https://go-review.googlesource.com/36562
Run-TryBot: Robert Griesemer <gri@golang.org>
Reviewed-by: Alan Donovan <adonovan@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/gofmt: use actual filename in gofmt -d output

By using actual filename, diff output of "gofmt -d" can be used with
other commands like "diffstat" and "patch".

Example:
  $ gofmt -d path/to/file.go | diffstat
  $ gofmt -d path/to/file.go > gofmt.patch
  $ patch -u -p0 < gofmt.patch

Fixes #18932

Change-Id: I21ce15eb77870d72f2c14bfd5e7c21e2c77dc9ab
Reviewed-on: https://go-review.googlesource.com/36374
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

bytes, strings: optimize Split*

The relevant benchmark results on linux/amd64:

bytes:

SplitSingleByteSeparator-4   25.7ms ± 5%   9.1ms ± 4%  -64.40%  (p=0.000 n=10+10)
SplitMultiByteSeparator-4    13.8ms ±20%   4.3ms ± 8%  -69.23%  (p=0.000 n=10+10)
SplitNSingleByteSeparator-4  1.88µs ± 9%  0.88µs ± 4%  -53.25%  (p=0.000 n=10+10)
SplitNMultiByteSeparator-4   4.83µs ±10%  1.32µs ± 9%  -72.65%  (p=0.000 n=10+10)

strings:

name                         old time/op  new time/op  delta
SplitSingleByteSeparator-4   21.4ms ± 8%   8.5ms ± 5%  -60.19%  (p=0.000 n=10+10)
SplitMultiByteSeparator-4    13.2ms ± 9%   3.9ms ± 4%  -70.29%  (p=0.000 n=10+10)
SplitNSingleByteSeparator-4  1.54µs ± 5%  0.75µs ± 7%  -51.21%  (p=0.000 n=10+10)
SplitNMultiByteSeparator-4   3.57µs ± 8%  1.01µs ±11%  -71.76%  (p=0.000 n=10+10)

Fixes #18973

Change-Id: Ie4bc010c6cc389001e72eab530497c81e5b26f34
Reviewed-on: https://go-review.googlesource.com/36510
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

database/sql: record the context error in Rows if canceled

Previously it was intended that Rows.Scan would return
an error and Rows.Err would return nil. This was problematic
because drivers could not differentiate between a normal
Rows.Close or a context cancel close.

The alternative is to require drivers to return a Scan to return
an error if the driver is closed while there are still rows to be read.
This is currently not how several drivers currently work and may be
difficult to detect when there are additional rows.

At the same time guard the the Rows.lasterr and prevent a close
while a Rows operation is active.

For the drivers that do not have Context methods, do not check for
context cancelation after the operation, but before for any operation
that may modify the database state.

Fixes #18961

Change-Id: I49a25318ecd9f97a35d5b50540ecd850c01cfa5e
Reviewed-on: https://go-review.googlesource.com/36485
Reviewed-by: Russ Cox <rsc@golang.org>
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

crypto/tls: document that only tickets are supported.

This change clarifies that only ticket-based resumption is supported by
crypto/tls. It's not clear where to document this for a server,
although perhaps it's obvious there because there's nowhere to plug in
the storage that would be needed by SessionID-based resumption.

Fixes #18607

Change-Id: Iaaed53e8d8f2f45c2f24c0683052df4be6340922
Reviewed-on: https://go-review.googlesource.com/36560
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

bytes: use Index in Count

Similar to https://go-review.googlesource.com/28586,
but for package bytes instead of strings.
This provides simpler code and some performance gain.
Also update strings.Count to use the same code.

On AMD64 with heavily optimized Index I see:

name             old time/op    new time/op     delta
Count/10-6         47.3ns ± 0%     36.8ns ± 0%    -22.35%  (p=0.000 n=10+10)
Count/32-6          286ns ± 0%       38ns ± 0%    -86.71%  (p=0.000 n=10+10)
Count/4K-6         50.1µs ± 0%      4.4µs ± 0%    -91.18%  (p=0.000 n=10+10)
Count/4M-6         48.1ms ± 1%      4.5ms ± 0%    -90.56%  (p=0.000 n=10+9)
Count/64M-6         784ms ± 0%       73ms ± 0%    -90.73%  (p=0.000 n=10+10)
CountEasy/10-6     28.4ns ± 0%     31.0ns ± 0%     +9.23%  (p=0.000 n=10+10)
CountEasy/32-6     30.6ns ± 0%     37.0ns ± 0%    +20.92%  (p=0.000 n=10+10)
CountEasy/4K-6      186ns ± 0%      198ns ± 0%     +6.45%  (p=0.000 n=9+10)
CountEasy/4M-6      233µs ± 2%      234µs ± 2%       ~     (p=0.912 n=10+10)
CountEasy/64M-6    6.70ms ± 0%     6.68ms ± 1%       ~     (p=0.762 n=8+10)

name             old speed      new speed       delta
Count/10-6        211MB/s ± 0%    272MB/s ± 0%    +28.77%  (p=0.000 n=10+9)
Count/32-6        112MB/s ± 0%    842MB/s ± 0%   +652.84%  (p=0.000 n=10+10)
Count/4K-6       81.8MB/s ± 0%  927.6MB/s ± 0%  +1033.63%  (p=0.000 n=10+9)
Count/4M-6       87.2MB/s ± 1%  924.0MB/s ± 0%   +959.25%  (p=0.000 n=10+9)
Count/64M-6      85.6MB/s ± 0%  922.9MB/s ± 0%   +978.31%  (p=0.000 n=10+10)
CountEasy/10-6    352MB/s ± 0%    322MB/s ± 0%     -8.41%  (p=0.000 n=10+10)
CountEasy/32-6   1.05GB/s ± 0%   0.87GB/s ± 0%    -17.35%  (p=0.000 n=9+10)
CountEasy/4K-6   22.0GB/s ± 0%   20.6GB/s ± 0%     -6.33%  (p=0.000 n=10+10)
CountEasy/4M-6   18.0GB/s ± 2%   18.0GB/s ± 2%       ~     (p=0.912 n=10+10)
CountEasy/64M-6  10.0GB/s ± 0%   10.0GB/s ± 1%       ~     (p=0.762 n=8+10)

On 386, without asm version of Index:

Count/10-6         57.0ns ± 0%     56.9ns ± 0%   -0.11%  (p=0.006 n=10+9)
Count/32-6          340ns ± 0%      274ns ± 0%  -19.48%  (p=0.000 n=10+9)
Count/4K-6         49.5µs ± 0%     37.1µs ± 0%  -24.96%  (p=0.000 n=10+10)
Count/4M-6         51.1ms ± 0%     38.2ms ± 0%  -25.21%  (p=0.000 n=10+10)
Count/64M-6         818ms ± 0%      613ms ± 0%  -25.07%  (p=0.000 n=8+10)
CountEasy/10-6     60.0ns ± 0%     70.4ns ± 0%  +17.34%  (p=0.000 n=10+10)
CountEasy/32-6     81.1ns ± 0%     94.0ns ± 0%  +15.97%  (p=0.000 n=9+10)
CountEasy/4K-6     4.37µs ± 0%     4.39µs ± 0%   +0.30%  (p=0.000 n=10+9)
CountEasy/4M-6     4.43ms ± 0%     4.43ms ± 0%     ~     (p=0.579 n=10+10)
CountEasy/64M-6    70.9ms ± 0%     70.9ms ± 0%     ~     (p=0.912 n=10+10)

name             old speed      new speed       delta
Count/10-6        176MB/s ± 0%    176MB/s ± 0%   +0.10%  (p=0.000 n=10+9)
Count/32-6       93.9MB/s ± 0%  116.5MB/s ± 0%  +24.06%  (p=0.000 n=10+9)
Count/4K-6       82.7MB/s ± 0%  110.3MB/s ± 0%  +33.26%  (p=0.000 n=10+10)
Count/4M-6       82.1MB/s ± 0%  109.7MB/s ± 0%  +33.70%  (p=0.000 n=10+10)
Count/64M-6      82.0MB/s ± 0%  109.5MB/s ± 0%  +33.46%  (p=0.000 n=8+10)
CountEasy/10-6    167MB/s ± 0%    142MB/s ± 0%  -14.75%  (p=0.000 n=9+10)
CountEasy/32-6    395MB/s ± 0%    340MB/s ± 0%  -13.77%  (p=0.000 n=10+10)
CountEasy/4K-6    936MB/s ± 0%    934MB/s ± 0%   -0.29%  (p=0.000 n=10+9)
CountEasy/4M-6    947MB/s ± 0%    946MB/s ± 0%     ~     (p=0.591 n=10+10)
CountEasy/64M-6   947MB/s ± 0%    947MB/s ± 0%     ~     (p=0.867 n=10+10)

Change-Id: Ia76b247372b6f5b5d23a9f10253a86536a5153b3
Reviewed-on: https://go-review.googlesource.com/36489
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

hash/crc32: use sub-benchmarks

Change-Id: Iae68a097a6897f1616f94fdc3548837ef200e66f
Reviewed-on: https://go-review.googlesource.com/36541
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Joe Tsai <thebrokentoaster@gmail.com>

time: bound file reads and validate LoadLocation argument

Fixes #18985

Change-Id: I956117f47d1d2b453b4786c7b78c1c944defeca0
Reviewed-on: https://go-review.googlesource.com/36551
Reviewed-by: Ian Lance Taylor <iant@golang.org>

cmd/gofmt: clear pattern match map at the correct time

We need to clear the pattern match map after the recursive rewrite
applications, otherwise there might be lingering entries that cause
match to fail.

Fixes #18987.

Change-Id: I7913951c455c98932bda790861db6a860ebad032
Reviewed-on: https://go-review.googlesource.com/36546
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Robert Griesemer <gri@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

runtime: use atomic ops for fwdSig, make sigtable immutable

The fwdSig array is accessed by the signal handler, which may run in
parallel with other threads manipulating it via the os/signal package.
Use atomic accesses to ensure that there are no problems.

Move the _SigHandling flag out of the sigtable array. This makes sigtable
immutable and safe to read from the signal handler.

Change-Id: Icfa407518c4ebe1da38580920ced764898dfc9ad
Reviewed-on: https://go-review.googlesource.com/36321
Run-TryBot: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

runtime: update android time_now call

This was broken in https://golang.org/cl/36255

Change-Id: Ib23323a745a650ac51b0ead161076f97efe6d7b7
Reviewed-on: https://go-review.googlesource.com/36543
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd/go: clarify that tag lists are space-separated

Apparently the current documentation is confusing users that
quickly skim the flags list at the top. Make very clear that
build tags are space-separated.

Updates #18800

Change-Id: I473552c5a2b70ca03d8bbbd2c76805f7f82b49a2
Reviewed-on: https://go-review.googlesource.com/35951
Reviewed-by: Daniel Martí <mvdan@mvdan.cc>
Reviewed-by: Minux Ma <minux@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

os: make Stdin.Stat() return ModeCharDevice if Stdin is console

CL 20845 changed Stdin.Stat() so it returns ModeNamedPipe.
But introduced TestStatStdin does not test what Stdin.Stat()
returns when Stdin is console.

This CL adjusts both TestStatStdin and Stdin.Stat
implementations to handle console. Return ModeCharDevice
from Stdin.Stat() when Stdin is console on windows,
just like it does on unix.

Fixes #14853.

Change-Id: I54d73caee2aea45a99618d11600d8e82fe20d0c0
Reviewed-on: https://go-review.googlesource.com/34090
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

encoding/json: add Valid for checking validity of input bytes

Fixes #18086

Change-Id: Idc501dd37893e04a01c6ed9920147d24c0c1fa18
Reviewed-on: https://go-review.googlesource.com/34202
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

math/big: add IsInt64/IsUint64 predicates

Change-Id: Ia5ed3919cb492009ac8f66d175b47a69f83ee4f1
Reviewed-on: https://go-review.googlesource.com/36487
Reviewed-by: Alan Donovan <adonovan@google.com>

cmd/internal/obj: remove ATYPE

In cmd/compile, we can directly construct obj.Auto to represent local
variables and attach them to the function's obj.LSym.

In preparation for being able to emit more precise DWARF info based on
other compiler available information (e.g., lexical scoping).

Change-Id: I9c4225ec59306bec42552838493022e0e9d70228
Reviewed-on: https://go-review.googlesource.com/36420
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Heschi Kreinick <heschi@google.com>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>

runtime/pprof: document that profile names should not contain spaces.

Change-Id: I967d897e812bee63b32bc2a7dcf453861b89b7e3
Reviewed-on: https://go-review.googlesource.com/36533
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd/compile: do not use statictmp for zeroing

Also fixes #18687.

Change-Id: I7c6d47c71e632adf4c16937a29074621f771844c
Reviewed-on: https://go-review.googlesource.com/35261
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>

cmd/compile/internal/ssa: use *obj.LSym in ExternSymbol

Change-Id: I713120f90fd1d2df6698c40622ccac6eae907919
Reviewed-on: https://go-review.googlesource.com/36423
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>

cmd/internal/dwarf: use []*Var instead of linked lists

Passes toolstash -cmp.

Change-Id: I202b29495ca1aaf3c52879fa99fdc0a4b86703af
Reviewed-on: https://go-review.googlesource.com/36419
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

runtime/pprof: clarify CPU profile's captured during the lifetime of the prog

Fixes #18504.

Change-Id: I3716fc58fc98472eea15ce3617aee3890670c276
Reviewed-on: https://go-review.googlesource.com/36430
Reviewed-by: Russ Cox <rsc@golang.org>

time: delete incorrect docs about day-of-month checks.

Documentation was introduced by CL https://golang.org/cl/14123
but that behavior was changed later by CL https://golang.org/cl/17710.
This CL deletes the stale paragraph.

Fixes #18980

Change-Id: Ib434f1eac6fc814fde1be112a8f52afe6e3e0fcc
Reviewed-on: https://go-review.googlesource.com/36532
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd/go, go/build: better defenses against GOPATH=GOROOT

Fixes #18863.

Change-Id: I0723563cd23728b0d43ebcc25979bf8d21e2a72c
Reviewed-on: https://go-review.googlesource.com/36427
Run-TryBot: Russ Cox <rsc@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

runtime: fix confusion between _MaxMem and _MaxArena32

Currently both _MaxMem and _MaxArena32 represent the maximum arena
size on 32-bit hosts (except on MIPS32 where _MaxMem is confusingly
smaller than _MaxArena32).

Clean up sysAlloc so that it always uses _MaxMem, which is the maximum
arena size on both 32- and 64-bit architectures and is the arena size
we allocate auxiliary structures for. This lets us simplify and unify
some code paths and eliminate _MaxArena32.

Fixes #18651. mheap.sysAlloc currently assumes that if the arena is
small, we must be on a 32-bit machine and can therefore grow the arena
to _MaxArena32. This breaks down on darwin/arm64, where _MaxMem is
only 2 GB. As a result, on darwin/arm64, we only reserve spans and
bitmap space for a 2 GB heap, and if the application tries to allocate
beyond that, sysAlloc takes the 32-bit path, tries to grow the arena
beyond 2 GB, and panics when it tries to grow the spans array
allocation past its reserved size. This has probably been a problem
for several releases now, but was only noticed recently because
mapSpans didn't check the bounds on the span reservation until
recently. Most likely it corrupted the bitmap before. By using _MaxMem
consistently, we avoid thinking that we can grow the arena larger than
we have auxiliary structures for.

Change-Id: Ifef28cb746a3ead4b31c1d7348495c2242fef520
Reviewed-on: https://go-review.googlesource.com/35253
Reviewed-by: David Crawshaw <crawshaw@golang.org>
Reviewed-by: Elias Naur <elias.naur@gmail.com>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

runtime: simplify and cleanup mallocinit

mallocinit has evolved organically. Make a pass to clean it up in
various ways:

1. Merge the computation of spansSize and bitmapSize. These were
   computed on every loop iteration of two different loops, but always
   have the same value, which can be derived directly from _MaxMem.
   This also avoids over-reserving these on MIPS, were _MaxArena32 is
   larger than _MaxMem.

2. Remove the ulimit -v logic. It's been disabled for many releases
   and the dead code paths to support it are even more wrong now than
   they were when it was first disabled, since now we *must* reserve
   spans and bitmaps for the full address space.

3. Make it clear that we're using a simple linear allocation to lay
   out the spans, bitmap, and arena spaces. Previously there were a
   lot of redundant pointer computations. Now we just bump p1 up as we
   reserve the spaces.

In preparation for #18651.

Updates #5049 (respect ulimit).

Change-Id: Icbe66570d3a7a17bea227dc54fb3c4978b52a3af
Reviewed-on: https://go-review.googlesource.com/35252
Reviewed-by: Russ Cox <rsc@golang.org>

runtime: make _MaxMem an untyped constant

Currently _MaxMem is a uintptr, which is going to complicate some
further changes. Make it untyped so we'll be able to do untyped math
on it before truncating it to a uintptr.

The runtime assembly is identical before and after this change on
{linux,windows}/{amd64,386}.

Updates #18651.

Change-Id: I0f64511faa9e0aa25179a556ab9f185ebf8c9cf8
Reviewed-on: https://go-review.googlesource.com/35251
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rick Hudson <rlh@golang.org>
Reviewed-by: David Crawshaw <crawshaw@golang.org>

cmd/compile: cmd/internal/obj: cull dead code

This code is dead as a result of

* removing the Follow pass
* moving rotation detection from walk to ssa

Change-Id: I14599c85bedb4e3148347b547e724187920182c4
Reviewed-on: https://go-review.googlesource.com/36484
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile: do not use "oaslit" for global

The compiler did not emit write barrier for assigning global with
struct literal, like global = T{} where T contains pointer.

The relevant code path is:
walkexpr OAS var_ OSTRUCTLIT
    oaslit
        anylit OSTRUCTLIT
            walkexpr OAS var_ nil
            return without adding write barrier
    return true
break (without adding write barrier)

This CL makes oaslit not apply to globals. See also CL
https://go-review.googlesource.com/c/36355/ for an alternative
fix.

The downside of this is that it generates static data for zeroing
struct now. Also this only covers global. If there is any lurking
bug with implicit zeroing other than globals, this doesn't fix.

Fixes #18956.

Change-Id: Ibcd27e4fae3aa38390ffa94a32a9dd7a802e4b37
Reviewed-on: https://go-review.googlesource.com/36410
Reviewed-by: Russ Cox <rsc@golang.org>
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

crypto/x509: check for new tls-ca-bundle.pem last

We added CentOS 7's /etc/pki/ca-trust/extracted/pem/tls-ca-bundle.pem
to the list in response to #17549 - not being able to find any certs otherwise.

Now we have #18813, where CentOS 6 apparently has both that file
and /etc/pki/tls/certs/ca-bundle.crt, and the latter is complete while
the former is not.

Moving the new CentOS 7 file to the bottom of the list should fix both
problems: the CentOS 7 system that didn't have any of the other files
in the list will still find the new one, and existing systems will still
keep using what they were using instead of preferring the new path
that may or may not be complete on some systems.

Fixes #18813.

Change-Id: I5275ab67424b95e7210e14938d3e986c8caee0ba
Reviewed-on: https://go-review.googlesource.com/36429
Run-TryBot: Russ Cox <rsc@golang.org>
Reviewed-by: Adam Langley <agl@golang.org>

cmd/link, crypto/tls: don't use append loops

Change-Id: Ib47e295e8646b769c30fd81e5c7f20f964df163e
Reviewed-on: https://go-review.googlesource.com/36335
Reviewed-by: Filippo Valsorda <hi@filippo.io>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

spec: clarify alignment of arrays

Fixes #18950.

Change-Id: I9f94748f36a896bcadc96f0642eb1f3bff387950
Reviewed-on: https://go-review.googlesource.com/36481
Reviewed-by: Russ Cox <rsc@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

testing: clarify T.Parallel() godoc wording

Fixes #18914.

Change-Id: Iec90d6aaa62595983db28b17794429f3c9a3dc36
Reviewed-on: https://go-review.googlesource.com/36272
Reviewed-by: Ian Lance Taylor <iant@golang.org>

Revert "image: fix the overlap check in Rectangle.Intersect."

This reverts commit a855da29dbd7a80c4d87a421c1f88a8603c020fa.

Change-Id: I23c0351b0708877e0b3d1b44a2bc2799cee52cd1
Reviewed-on: https://go-review.googlesource.com/36426
Reviewed-by: Russ Cox <rsc@golang.org>

text/template: remove duplicate logic in conditional

It looks like this conditional may have been refactored at some point,
but the logic was still very confusing. The outer conditional checks if
the function is variadic, so there's no need to verify that in the
result. Additionally, since the function isn't variadic, there is no
reason to permit the function call if the number of input arguments is
less than the function signature requires.

Change-Id: Ia957cf83d1c900c08dd66384efcb74f0c368422e
Reviewed-on: https://go-review.googlesource.com/35491
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

cmd/internal/obj: remove Follow pass

The Follow pass in the assembler backend reorders and copies
instructions. This even applies to hand-written assembly code,
which in many cases don't want to be reordered. Now that the
SSA compiler does a good job for laying out instructions, the
benefit of this pass is very little:

AMD64: (old = with Follow, new = without Follow)
name                      old time/op    new time/op    delta
BinaryTree17-12              2.78s ± 1%     2.79s ± 1%  +0.44%  (p=0.000 n=20+19)
Fannkuch11-12                3.11s ± 0%     3.31s ± 1%  +6.16%  (p=0.000 n=19+19)
FmtFprintfEmpty-12          50.9ns ± 1%    51.6ns ± 3%  +1.40%  (p=0.000 n=17+20)
FmtFprintfString-12          127ns ± 0%     128ns ± 1%  +0.88%  (p=0.000 n=17+17)
FmtFprintfInt-12             122ns ± 0%     123ns ± 1%  +0.76%  (p=0.000 n=20+19)
FmtFprintfIntInt-12          185ns ± 1%     186ns ± 1%  +0.65%  (p=0.000 n=20+19)
FmtFprintfPrefixedInt-12     192ns ± 1%     202ns ± 1%  +4.99%  (p=0.000 n=20+19)
FmtFprintfFloat-12           284ns ± 0%     288ns ± 0%  +1.33%  (p=0.000 n=15+19)
FmtManyArgs-12               807ns ± 0%     804ns ± 0%  -0.44%  (p=0.000 n=16+18)
GobDecode-12                7.23ms ± 1%    7.21ms ± 1%    ~     (p=0.052 n=20+20)
GobEncode-12                6.09ms ± 1%    6.12ms ± 1%  +0.41%  (p=0.002 n=19+19)
Gzip-12                      253ms ± 1%     255ms ± 1%  +0.95%  (p=0.000 n=18+20)
Gunzip-12                   38.4ms ± 0%    38.5ms ± 0%  +0.34%  (p=0.000 n=17+17)
HTTPClientServer-12         95.4µs ± 2%    96.1µs ± 1%  +0.78%  (p=0.002 n=19+19)
JSONEncode-12               16.5ms ± 1%    16.6ms ± 1%  +1.17%  (p=0.000 n=19+19)
JSONDecode-12               54.6ms ± 1%    55.3ms ± 1%  +1.23%  (p=0.000 n=18+18)
Mandelbrot200-12            4.47ms ± 0%    4.47ms ± 0%  +0.06%  (p=0.000 n=18+18)
GoParse-12                  3.47ms ± 1%    3.47ms ± 1%    ~     (p=0.583 n=20+20)
RegexpMatchEasy0_32-12      84.8ns ± 1%    85.2ns ± 2%  +0.51%  (p=0.022 n=20+20)
RegexpMatchEasy0_1K-12       206ns ± 1%     206ns ± 1%    ~     (p=0.770 n=20+20)
RegexpMatchEasy1_32-12      82.8ns ± 1%    83.4ns ± 1%  +0.64%  (p=0.000 n=20+19)
RegexpMatchEasy1_1K-12       363ns ± 1%     361ns ± 1%  -0.48%  (p=0.007 n=20+20)
RegexpMatchMedium_32-12      126ns ± 1%     126ns ± 0%  +0.72%  (p=0.000 n=20+20)
RegexpMatchMedium_1K-12     39.1µs ± 1%    39.8µs ± 0%  +1.73%  (p=0.000 n=19+19)
RegexpMatchHard_32-12       1.97µs ± 0%    1.98µs ± 1%  +0.29%  (p=0.005 n=18+20)
RegexpMatchHard_1K-12       59.5µs ± 1%    59.8µs ± 1%  +0.36%  (p=0.000 n=18+20)
Revcomp-12                   442ms ± 1%     445ms ± 2%  +0.67%  (p=0.000 n=19+20)
Template-12                 58.0ms ± 1%    57.5ms ± 1%  -0.85%  (p=0.000 n=19+19)
TimeParse-12                 311ns ± 0%     314ns ± 0%  +0.94%  (p=0.000 n=20+18)
TimeFormat-12                350ns ± 3%     346ns ± 0%    ~     (p=0.076 n=20+19)
[Geo mean]                  55.9µs         56.4µs       +0.80%

ARM32:
name                     old time/op    new time/op    delta
BinaryTree17-4              30.4s ± 0%     30.1s ± 0%  -1.14%  (p=0.000 n=10+8)
Fannkuch11-4                13.7s ± 0%     13.6s ± 0%  -0.75%  (p=0.000 n=10+10)
FmtFprintfEmpty-4           664ns ± 1%     651ns ± 1%  -1.96%  (p=0.000 n=7+8)
FmtFprintfString-4         1.83µs ± 2%    1.77µs ± 2%  -3.21%  (p=0.000 n=10+10)
FmtFprintfInt-4            1.57µs ± 2%    1.54µs ± 2%  -2.25%  (p=0.007 n=10+10)
FmtFprintfIntInt-4         2.37µs ± 2%    2.31µs ± 1%  -2.68%  (p=0.000 n=10+10)
FmtFprintfPrefixedInt-4    2.14µs ± 2%    2.10µs ± 1%  -1.83%  (p=0.006 n=10+10)
FmtFprintfFloat-4          3.69µs ± 2%    3.74µs ± 1%  +1.60%  (p=0.000 n=10+10)
FmtManyArgs-4              9.43µs ± 1%    9.17µs ± 1%  -2.70%  (p=0.000 n=10+10)
GobDecode-4                76.3ms ± 1%    75.5ms ± 1%  -1.14%  (p=0.003 n=10+10)
GobEncode-4                70.7ms ± 2%    69.0ms ± 1%  -2.36%  (p=0.000 n=10+10)
Gzip-4                      2.64s ± 1%     2.65s ± 0%  +0.59%  (p=0.002 n=10+10)
Gunzip-4                    402ms ± 0%     398ms ± 0%  -1.11%  (p=0.000 n=10+9)
HTTPClientServer-4          458µs ± 0%     457µs ± 0%    ~     (p=0.247 n=10+10)
JSONEncode-4                171ms ± 0%     172ms ± 0%  +0.56%  (p=0.000 n=10+10)
JSONDecode-4                672ms ± 1%     668ms ± 1%    ~     (p=0.105 n=10+10)
Mandelbrot200-4            33.5ms ± 0%    33.5ms ± 0%    ~     (p=0.156 n=9+10)
GoParse-4                  33.9ms ± 0%    34.0ms ± 0%  +0.36%  (p=0.031 n=9+9)
RegexpMatchEasy0_32-4       823ns ± 1%     835ns ± 1%  +1.49%  (p=0.000 n=8+8)
RegexpMatchEasy0_1K-4      3.99µs ± 0%    4.02µs ± 1%  +0.92%  (p=0.000 n=8+10)
RegexpMatchEasy1_32-4       877ns ± 3%     904ns ± 2%  +3.07%  (p=0.012 n=10+10)
RegexpMatchEasy1_1K-4      5.99µs ± 0%    5.97µs ± 1%  -0.38%  (p=0.023 n=8+8)
RegexpMatchMedium_32-4     1.40µs ± 2%    1.40µs ± 2%    ~     (p=0.590 n=10+9)
RegexpMatchMedium_1K-4      357µs ± 0%     355µs ± 1%  -0.72%  (p=0.000 n=7+8)
RegexpMatchHard_32-4       22.3µs ± 0%    22.1µs ± 0%  -0.49%  (p=0.000 n=8+7)
RegexpMatchHard_1K-4        661µs ± 0%     658µs ± 0%  -0.42%  (p=0.000 n=8+7)
Revcomp-4                  46.3ms ± 0%    46.3ms ± 0%    ~     (p=0.393 n=10+10)
Template-4                  753ms ± 1%     750ms ± 0%    ~     (p=0.211 n=10+9)
TimeParse-4                4.28µs ± 1%    4.22µs ± 1%  -1.34%  (p=0.000 n=8+10)
TimeFormat-4               9.00µs ± 0%    9.05µs ± 0%  +0.59%  (p=0.000 n=10+10)
[Geo mean]                  538µs          535µs       -0.55%

ARM64:
name                     old time/op    new time/op    delta
BinaryTree17-8              8.39s ± 0%     8.39s ± 0%    ~     (p=0.684 n=10+10)
Fannkuch11-8                5.95s ± 0%     5.99s ± 0%  +0.63%  (p=0.000 n=10+10)
FmtFprintfEmpty-8           116ns ± 0%     116ns ± 0%    ~     (all equal)
FmtFprintfString-8          361ns ± 0%     360ns ± 0%  -0.31%  (p=0.003 n=8+6)
FmtFprintfInt-8             290ns ± 0%     290ns ± 0%    ~     (p=0.620 n=9+9)
FmtFprintfIntInt-8          476ns ± 1%     469ns ± 0%  -1.47%  (p=0.000 n=10+6)
FmtFprintfPrefixedInt-8     412ns ± 2%     417ns ± 2%  +1.39%  (p=0.006 n=9+10)
FmtFprintfFloat-8           652ns ± 1%     652ns ± 0%    ~     (p=0.161 n=10+8)
FmtManyArgs-8              1.94µs ± 0%    1.94µs ± 2%    ~     (p=0.781 n=10+10)
GobDecode-8                17.7ms ± 1%    17.7ms ± 0%    ~     (p=0.962 n=10+7)
GobEncode-8                15.6ms ± 0%    15.6ms ± 1%    ~     (p=0.063 n=10+10)
Gzip-8                      786ms ± 0%     787ms ± 0%    ~     (p=0.356 n=10+9)
Gunzip-8                    127ms ± 0%     127ms ± 0%  +0.08%  (p=0.028 n=10+9)
HTTPClientServer-8          198µs ± 6%     198µs ± 7%    ~     (p=0.796 n=10+10)
JSONEncode-8               42.5ms ± 0%    42.2ms ± 0%  -0.73%  (p=0.000 n=9+8)
JSONDecode-8                158ms ± 1%     162ms ± 0%  +2.28%  (p=0.000 n=10+9)
Mandelbrot200-8            10.1ms ± 0%    10.1ms ± 0%  -0.01%  (p=0.000 n=10+9)
GoParse-8                  8.54ms ± 1%    8.63ms ± 1%  +1.06%  (p=0.000 n=10+9)
RegexpMatchEasy0_32-8       231ns ± 1%     225ns ± 0%  -2.52%  (p=0.000 n=9+10)
RegexpMatchEasy0_1K-8      1.63µs ± 0%    1.63µs ± 0%    ~     (p=0.170 n=10+10)
RegexpMatchEasy1_32-8       253ns ± 0%     249ns ± 0%  -1.41%  (p=0.000 n=9+10)
RegexpMatchEasy1_1K-8      2.08µs ± 0%    2.08µs ± 0%  -0.32%  (p=0.000 n=9+10)
RegexpMatchMedium_32-8      355ns ± 1%     351ns ± 0%  -1.04%  (p=0.007 n=10+7)
RegexpMatchMedium_1K-8      104µs ± 0%     104µs ± 0%    ~     (p=0.148 n=10+10)
RegexpMatchHard_32-8       5.79µs ± 0%    5.79µs ± 0%    ~     (p=0.578 n=10+10)
RegexpMatchHard_1K-8        176µs ± 0%     176µs ± 0%    ~     (p=0.137 n=10+10)
Revcomp-8                   1.37s ± 1%     1.36s ± 1%  -0.26%  (p=0.023 n=10+10)
Template-8                  151ms ± 1%     154ms ± 1%  +2.14%  (p=0.000 n=9+10)
TimeParse-8                 723ns ± 2%     721ns ± 1%    ~     (p=0.592 n=10+10)
TimeFormat-8                804ns ± 2%     798ns ± 3%    ~     (p=0.344 n=10+10)
[Geo mean]                  154µs          154µs       -0.02%

Therefore remove this pass. Also reduce text size by 0.5~2%.

Comment out some dead code in runtime/sys_nacl_amd64p32.s
which contains undefined symbols.

Change-Id: I1473986fe5b18b3d2554ce96cdc6f0999b8d955d
Reviewed-on: https://go-review.googlesource.com/36205
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Russ Cox <rsc@golang.org>

crypto/des: improve the throughput of DES and 3DES

For detailed explanation of the adopted (Eric Young's) algorithm,
see http://ftp.nluug.nl/security/coast/libs/libdes/ALGORITHM

benchmark                   old ns/op     new ns/op     delta
BenchmarkEncrypt-16         649           164           -74.73%
BenchmarkDecrypt-16         546           156           -71.43%
BenchmarkTDESEncrypt-16     1651          385           -76.68%
BenchmarkTDESDecrypt-16     1645          378           -77.02%

benchmark                   old MB/s     new MB/s     speedup
BenchmarkEncrypt-16         12.31        48.76        3.96x
BenchmarkDecrypt-16         14.64        51.03        3.49x
BenchmarkTDESEncrypt-16     4.84         20.74        4.29x
BenchmarkTDESDecrypt-16     4.86         21.16        4.35x

Change-Id: Ic3e1fe3340419ec5a0e6379434911eb41e0246f6
Reviewed-on: https://go-review.googlesource.com/36490
Run-TryBot: Minux Ma <minux@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Russ Cox <rsc@golang.org>

go/types: permit f(nil...) for variadic arguments

This code may be pointless, but it is legal.

Fixes golang/go#18268

Change-Id: Ibacae583606e1a6fdf0c0f01abe2e22e9e608393
Reviewed-on: https://go-review.googlesource.com/34194
Reviewed-by: Robert Griesemer <gri@golang.org>

image: fix the overlap check in Rectangle.Intersect.

The doc comment for Rectangle.Intersect clearly states, "If the two
rectangles do not overlap then the zero rectangle will be returned."
Prior to this fix, calling Intersect on adjacent but non-overlapping
rectangles would return an empty but non-zero rectangle.

The fix essentially changes
if r.Min.X > r.Max.X || r.Min.Y > r.Max.Y { etc }
to
if r.Min.X >= r.Max.X || r.Min.Y >= r.Max.Y { etc }
(note that the > signs have become >= signs), but changing that line to:
if r.Empty() { etc }
seems clearer (and equivalent).

Change-Id: Ia654e4b9dc805978db3e94d7a9718b6366005360
Reviewed-on: https://go-review.googlesource.com/34853
Reviewed-by: David Crawshaw <crawshaw@golang.org>

runtime/pprof: symbolize proto profiles

When generating pprof profiles in proto format, symbolize the profiles.

Change-Id: I2471ed7f919483e5828868306418a63e41aff5c5
Reviewed-on: https://go-review.googlesource.com/34192
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Russ Cox <rsc@golang.org>

test: improve output format of issue10607a.go test

Change-Id: Iad5ff820a95f5082b75aa5260e40c33c7b0ecf22
Reviewed-on: https://go-review.googlesource.com/35990
Reviewed-by: Russ Cox <rsc@golang.org>

cmd/compile/internal/syntax: avoid follow-up error for incorrect if statement

This is a follow-up on https://go-review.googlesource.com/36470
and leads to a more stable fix. The above CL relied on filtering
of multiple errors on the same line to avoid more than one error
for an `if` statement of the form `if a := 10 {}`. This CL avoids
the secondary error ("missing condition in if statement") in the
first place.

For #18915.

Change-Id: I8517f485cc2305965276c17d8f8797d61ef9e999
Reviewed-on: https://go-review.googlesource.com/36479
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>

testing: print extra labels on benchmarks

When running benchmarks, print "goos", "goarch", and "pkg"
labels. This makes it easier to refer to benchmark logs and understand
how they were generated. "pkg" is printed only for benchmarks located
in GOPATH.

Change-Id: I397cbdd57b9fe8cbabbb354ec7bfba59f5625c42
Reviewed-on: https://go-review.googlesource.com/36356
Run-TryBot: Quentin Smith <quentin@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Russ Cox <rsc@golang.org>

spec: pick up a few corrections missed in prior commit

This CL picks up a couple of minor fixes that were present
in https://go-review.googlesource.com/#/c/36213/6..5 but
accidentally got dropped in https://go-review.googlesource.com/#/c/36213/
because I submitted from the wrong client.

Change-Id: I3ad0d20457152ea9a116cbb65a23eb0dc3a8525e
Reviewed-on: https://go-review.googlesource.com/36471
Reviewed-by: Matthew Dempsky <mdempsky@google.com>

spec: introduce alias declarations and type definitions

To avoid confusion caused by the term "named type" (which now just
means a type with a name, but formerly meant a type declared with
a non-alias type declaration), a type declaration now comes in two
forms: alias declarations and type definitions. Both declare a type
name, but type definitions also define new types.

Replace the use of "named type" with "defined type" elsewhere in
the spec.

For #18130.

Change-Id: I49f5ddacefce90354eb65ee5fbf10ba737221995
Reviewed-on: https://go-review.googlesource.com/36213
Reviewed-by: Rob Pike <r@golang.org>

cmd/compile/internal/syntax: make a parser error "1.7 compliant"

For code such as

if a := 10 { ...

the 1.7 compiler reported

a := 10 used as value

while the 1.8 compiler reported

invalid condition, tag, or type switch guard

Changed the error message to match the 1.7 compiler.

Fixes #18915.

Change-Id: I01308862e461922e717f9f8295a9db53d5a914eb
Reviewed-on: https://go-review.googlesource.com/36470
Run-TryBot: Robert Griesemer <gri@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile/internal/gc: remove more backend Sym uses

Removes all external uses of Linksym and Pkglookup, which are the only
two exported functions that return Syms.

Also add Duffcopy and Duffzero since they're used often enough across
SSA backends.

Passes toolstash -cmp.

Change-Id: I8d3fd048ad5cd676fc46378f09a917569ffc9b2c
Reviewed-on: https://go-review.googlesource.com/36418
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>

cmd/link: use external linking for PIE by default

Now `go test -buildmode=pie std -short` passes on linux/amd64.

Updates #18968

Change-Id: Ide21877713e00edc64c1700c950016d6bff8de0e
Reviewed-on: https://go-review.googlesource.com/36417
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile/internal/gc: stop exporting *gc.Sym-typed globals

The arch-specific SSA backends now no longer use gc.Sym either.

Passes toolstash -cmp.

Change-Id: Ic13b934b92a1b89b4b79c6c4796ab0a137608163
Reviewed-on: https://go-review.googlesource.com/36416
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>

cmd/compile/internal/ssa: use obj.LSym instead of gc.Sym

Gc's Sym type represents a package-qualified identifier, which is a
frontend concept and doesn't belong in SSA. Bonus: we can replace some
interface{} types with *obj.LSym.

Passes toolstash -cmp.

Change-Id: I456eb9957207d80f99f6eb9b8eab4a1f3263e9ed
Reviewed-on: https://go-review.googlesource.com/36415
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>

doc: remove inactive members of the CoC working group

Dave and Jason have moved on to other things.

Change-Id: I702d11bedfab1f47a33679a48c2309f49021229e
Reviewed-on: https://go-review.googlesource.com/36450
Reviewed-by: Dave Cheney <dave@cheney.net>

runtime: add definitions for SetGoroutineLabels and Do

This change defines runtime/pprof.SetGoroutineLabels and runtime/pprof.Do, which
are used to set profiler labels on goroutines. The change defines functions
in the runtime for setting and getting profile labels, and sets and unsets
profile labels when goroutines are created and deleted. The change also adds
the package runtime/internal/proflabel, which defines the structure the runtime
uses to store profile labels.

Change-Id: I747a4400141f89b6e8160dab6aa94ca9f0d4c94d
Reviewed-on: https://go-review.googlesource.com/34198
Run-TryBot: Michael Matloob <matloob@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Russ Cox <rsc@golang.org>
Reviewed-on: https://go-review.googlesource.com/35010

vendor/golang.org/x/crypto/curve25519: avoid loss of R15 in -dynlink mode

Original code fixed in https://go-review.googlesource.com/#/c/36359/.

Fixes #18820.

Change-Id: I060e6c9d0e312b4fd5d0674aff131055bf5cf61d
Reviewed-on: https://go-review.googlesource.com/36412
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Adam Langley <agl@golang.org>