net: deduplicate Unix socket code

This change consolidates functions and methods related to UnixAddr,
UnixConn and UnixListener for maintenance purpose, especially for
documentation.

The followup changes will update comments and examples.

Updates #10624.

Change-Id: I372d152099ac10956284e6b3863d7e4d9fe5c8e9
Reviewed-on: https://go-review.googlesource.com/20125
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

net: deduplicate raw IP socket code

This change consolidates functions and methods related to IPAddr and
IPConn for maintenance purpose, especially for documentation.

The followup changes will update comments and examples.

Updates #10624.

Change-Id: Ia5146f234225704a3c0b6459e1903e56a7b68134
Reviewed-on: https://go-review.googlesource.com/20124
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

net: deduplicate UDP socket code

This change consolidates functions and methods related to UDPAddr and
UDPConn for maintenance purpose, especially for documentation.

The followup changes will update comments and examples.

Updates #10624.

Change-Id: Idfe9be8ea46ade1111b0ae176862b2048eafc7be
Reviewed-on: https://go-review.googlesource.com/20120
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

net: simplify ipToSockaddr

Change-Id: I5dbcdf0ee0b46b760b2a7decb1d937aac2a6fa8d
Reviewed-on: https://go-review.googlesource.com/20585
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

cmd/compile: change logging of spills for regalloc to Warnl format

Change-Id: I01c000ff3f6dc6b0ed691e289eeef0fa61500337
Reviewed-on: https://go-review.googlesource.com/20744
Reviewed-by: Keith Randall <khr@golang.org>

fmt: replace variables for type bit sizes with constants

Use constants instead of dynamically computed values to determine
the bit sizes of types similar to how strconv and other packages
directly compute these sizes. Move these constants near the code
that uses them.

Change-Id: I78d113b7e697466097e32653975df5990380c2c1
Reviewed-on: https://go-review.googlesource.com/20514
Run-TryBot: Rob Pike <r@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rob Pike <r@golang.org>

net: make SplitHostPort return an empty host on error

This change also refactors SplitHostPort to avoid using gotos and
naked returns.

Fixes #14827

Change-Id: I4dca528936757fd06da76c23af8a0f6175bbedd1
Reviewed-on: https://go-review.googlesource.com/20726
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd/compile: add logging to critical and phielim

Change-Id: Ieefeceea40bd29657fd519368b0920dad8443844
Reviewed-on: https://go-review.googlesource.com/20712
Run-TryBot: Todd Neal <todd@tneal.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>

cmd/compile: turn off SSA internal consistency checks

They've been on for a few weeks of general use and nothing
has tripped up on them yet.

Makes the compiler ~18% faster.

Change-Id: I42d7bbc0581597f9cf4fb28989847814c81b08a2
Reviewed-on: https://go-review.googlesource.com/20741
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

runtime: update link to WriteHeapDump format

The new link is https://golang.org/s/go15heapdump.

Change-Id: Ifcaf8572bfe815ffaa78442a1991f6e20e990a50
Reviewed-on: https://go-review.googlesource.com/20740
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd: collapse internal/obj/fmt.go into compile/internal/gc/fmt.go

The obj.Fmt* values are only used by gc/fmt.go, so just move them
there. Also, add comments documenting the correspondance between
FmtFoo names and their flag characters to make understanding the
existing documentation slightly less confusing.

While here, add a new FmtFlag named type to represent these values.

Change-Id: I9631214b892557d094823f1ac575d0c43a84007b
Reviewed-on: https://go-review.googlesource.com/20717
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

sync: new Cond implementation

Change Cond implementation to use a notification list such that waiters
can first register for a notification, release the lock, then actually
wait. Signalers never have to park anymore.

This is intended to address an issue in the previous implementation
where Broadcast could fail to signal all waiters.

Results of the existing benchmark are below.

                                          Original          New  Diff
BenchmarkCond1-48        2000000               745 ns/op    755 +1.3%
BenchmarkCond2-48        1000000              1545 ns/op   1532 -0.8%
BenchmarkCond4-48         300000              3833 ns/op   3896 +1.6%
BenchmarkCond8-48         200000             10049 ns/op  10257 +2.1%
BenchmarkCond16-48        100000             21123 ns/op  21236 +0.5%
BenchmarkCond32-48         30000             40393 ns/op  41097 +1.7%

Fixes #14064

Change-Id: I083466d61593a791a034df61f5305adfb8f1c7f9
Reviewed-on: https://go-review.googlesource.com/18892
Reviewed-by: Dmitry Vyukov <dvyukov@google.com>
Reviewed-by: Austin Clements <austin@google.com>
Run-TryBot: Caleb Spare <cespare@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

encoding/base64: correct DecodedLen overestimate for unpadded encodings

While we're at it, add tests for EncodedLen and DecodedLen.

Fixes #14803.

Change-Id: I200c72cf11c51669b8d9f70c6e57ece359f7ae61
Reviewed-on: https://go-review.googlesource.com/20649
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

math/big: fix comment typos

Change-Id: I34cdc9cb3d32e86ff3a57db0012326c39cd55670
Reviewed-on: https://go-review.googlesource.com/20718
Reviewed-by: Robert Griesemer <gri@golang.org>

cmd/compile: compute second method type at runtime

The type information for a method includes two variants: a func
without the receiver, and a func with the receiver as the first
parameter. The former is used as part of the dynamic interface
checks, but the latter is only returned as a type in the
reflect.Method struct.

Instead of computing it at compile time, construct it at run time
with reflect.FuncOf.

Using cl/20701 as a baseline,

cmd/go: -480KB, (4.4%)
jujud:  -5.6MB, (7.8%)

For #6853.

Change-Id: I1b8c73f3ab894735f53d00cb9c0b506d84d54e92
Reviewed-on: https://go-review.googlesource.com/20709
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

cmd/compile: use int for field index

All of a struct's fields have to fit into memory anyway, so index them
with int instead of int64.  This also makes it nicer for
cmd/compile/internal/gc to reuse the same NumFields function.

Change-Id: I210be804a0c33370ec9977414918c02c675b0fbe
Reviewed-on: https://go-review.googlesource.com/20691
Reviewed-by: Ian Lance Taylor <iant@golang.org>

cmd/link: when pruning methods also prune funcType

Remove method type information for pruned methods from any program
that does not reflect on methods. This can be a significant saving:

addr2line: -310KB (8.8%)

A future update might want to consider a more aggressive variant of
this: setting the Type and Func fields of reflect.Method to nil for
unexported methods. That would shrink cmd/go by 2% and jujud by 2.6%
but could be considered an API change. So this CL sticks to the
uncontroversial change.

For #6853.

Change-Id: I5d186d9f822dc118ee89dc572c4912a3b3c72577
Reviewed-on: https://go-review.googlesource.com/20701
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

cmd/compile: fix load-combining

Make sure symbol gets carried along by load-combining rule.
Add the new load into the right block where we know that
mem is live.

Use auxInt field to carry i along instead of an explicit ADDQ.

Incorporate LEA ops into MOVBQZX and friends.

Change-Id: I587f7c6120b98fd2a0d48ddd6ddd13345d4421b4
Reviewed-on: https://go-review.googlesource.com/20732
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Todd Neal <todd@tneal.org>

archive/zip: add missing argument to error message

Silence vet.

Change-Id: I987438847389500cf3b5bc545ef918c66917b51a
Reviewed-on: https://go-review.googlesource.com/20683
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd/compile: make Type.Field stricter about bounds checking

Turns out there were only two call sites that expected
t.Field(t.NumFields()) to return nil.

Change-Id: I4679988d38ee9d7c9d89883537a17046717b2a77
Reviewed-on: https://go-review.googlesource.com/20731
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Crawshaw <crawshaw@golang.org>

os/user: allow LookupGroupId to fail during test

On my Mac I am in group 5000 which apparently has no name
(I suspect because it is an LDAP group and I cannot reach the
LDAP server). Do not make the test fail in that case.

Fixes #14806

Change-Id: I56b11a8e86b048abfb00812eaad37802fd2adcc5
Reviewed-on: https://go-review.googlesource.com/20710
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd/compile/internal/gc: remove Thearch.Linkarchinit

Change-Id: I8e990204c7a1e65d6b8e4262f6b5b9bd0cb3540c
Reviewed-on: https://go-review.googlesource.com/20725
Run-TryBot: Dave Cheney <dave@cheney.net>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

cmd/compile/internal/gc: tidy plive.go

Make boolean looking things boolean.

Change-Id: I8d1c0a32b471412b25a72908c7da6458d7bbe65b
Reviewed-on: https://go-review.googlesource.com/20723
Run-TryBot: Dave Cheney <dave@cheney.net>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd/link, cmd/compile: Add symbol references to object file.

Symbols in the object file currently refer to each other using symbol name
and version. Referring to the same symbol many times in an object file takes
up space and causes redundant map lookups. Instead write out a list of unique
symbol references and have symbols refer to each other using indexes into this
list.

Credit to Michael Hudson-Doyle for kicking this off.

Reduces pkg/linux_amd64 size by 30% from 61MB to 43MB

name       old s/op   new s/op   delta
LinkCmdGo  0.74 ± 3%  0.63 ± 4%  -15.22%  (p=0.000 n=20+20)
LinkJuju   6.38 ± 6%  5.73 ± 6%  -10.16%  (p=0.000 n=20+19)

Change-Id: I7e101a0c80b8e673a3ba688295e6f80ea04e1cfb
Reviewed-on: https://go-review.googlesource.com/20099
Reviewed-by: David Crawshaw <crawshaw@golang.org>
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

runtime: preserve darwin/arm{,64} callee-save registers

CL 14603 attempted to preserve the callee-save registers for
the darwin/arm runtime initialization routine, but I believe it
wasn't sufficient and resulted in the crash reported in issue

Saving and restoring the registers on the stack the same way
linux/arm does seems more obvious and fixes #14778, so do that.

Even though #14778 is not reproducible on darwin/arm64, I applied
a similar change there, and to linux/arm64 which obeys the same
calling convention.

Finally, this CL is a candidate for a 1.6 minor release for the same
reason CL 14603 was in a 1.5 minor release (as CL 16968). It is
small and only touches the iOS platforms and gomobile on darwin/arm
is currently useless without it.

Fixes #14778
Fixes #12590 (again)

Change-Id: I7401daf0bbd7c579a7e84761384a7b763651752a
Reviewed-on: https://go-review.googlesource.com/20621
Reviewed-by: David Crawshaw <crawshaw@golang.org>
Run-TryBot: Elias Naur <elias.naur@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile/internal: clean up galign.go constants

Move the C header style architecture constants to the per arch Main
methods.

Change-Id: Ie7ff39baa275ceaa6680e7d16441ca9f0aa12597
Reviewed-on: https://go-review.googlesource.com/20722
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Dave Cheney <dave@cheney.net>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/internal/obj: remove dead code

Partial automatic cleanup driven by Dominik Honnef's unused tool.

As _lookup now only has one caller, merge it into the caller and remove
the conditional create logic.

Change-Id: I2ea354d9d4b32a19905271eca74725231b6d8a93
Reviewed-on: https://go-review.googlesource.com/20589
Run-TryBot: Dave Cheney <dave@cheney.net>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

cmd/compile: move testdclstack out of the xdcl_list loop

This is a band-aid, but it fixes the problem
until a deeper fix is in place.

Testing with genpkg -n 50000, I see:

Before:

      154.67 real       184.66 user         3.15 sys

After:

       61.82 real        96.99 user         2.17 sys

Fixes #14781.

Change-Id: I24c7822d60c289bdd6a18a7840b984954c95f7d4
Reviewed-on: https://go-review.googlesource.com/20696
Reviewed-by: Robert Griesemer <gri@golang.org>

fmt: handle %X like %x for byte type arrays and slices

Treat the verb %X in the same special way as %q, %s and %x
are for arrays and slices with byte type elements.

Modify input for tests so the result of %x and %X is distinct.

Change-Id: I38d227755e98c7fad5e4adc2f603c6873aa910fd
Reviewed-on: https://go-review.googlesource.com/20516
Run-TryBot: Rob Pike <r@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rob Pike <r@golang.org>

cmd/compile: clean up C-style variable declarations in plive.go

Change-Id: I928f51a1fe4830a81d4f5d3eb572785e06a75b77
Reviewed-on: https://go-review.googlesource.com/20581
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: keep variable numbering for inlineable exported functions

Another step towards hooking up exported inlined function bodies.

Change-Id: Ib8094b03ac7970fee0e51b5826b5f8aa232e23fb
Reviewed-on: https://go-review.googlesource.com/20605
Reviewed-by: Alan Donovan <adonovan@google.com>
Run-TryBot: Robert Griesemer <gri@golang.org>

cmd/compile: replace TFIELD kind with separate Field type

Allows removing a bunch of unnecessary fields.

Passes toolstash/buildall.

Change-Id: Iec2492920e1c3ef352a9bf4296c74a55d9cc9ad6
Reviewed-on: https://go-review.googlesource.com/20677
Reviewed-by: Robert Griesemer <gri@golang.org>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/dist: build bootstrap toolchain with -l

Workaround Go 1.4 compiler bugs.  See discussion at:
https://groups.google.com/d/msg/golang-dev/Ss7mCKsvk8w/Gsq7VYI0AwAJ

Change-Id: I842335fddffb67a6e21c000fe5bef258ea61c77a
Reviewed-on: https://go-review.googlesource.com/20690
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

crypto/tls: minor refactors for readability

Change-Id: I93e73f16474b4b31f7097af2f9479822dfc34c5c
Reviewed-on: https://go-review.googlesource.com/20678
Reviewed-by: Adam Langley <agl@golang.org>

cmd/compile: stop storing TFIELD types in Node.Type

Currently, the only use for this is on the Left side of OKEY nodes
within struct literals.  esc and fmt only care so they can recognize
that the ONAME nodes are actually field names, which need special
handling.

sinit additionally needs to know the field's offset within the struct,
which we can provide via Xoffset.

Passes toolstash/buildall.

Change-Id: I362d965e161f4d80fcd9c9bae0dfacc657dc0b29
Reviewed-on: https://go-review.googlesource.com/20676
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile: minor cleanup in sinit.go

Passes toolstash -cmp.

Change-Id: Id16009ef3ef1173eafe0f0c578dbf325b61aab3c
Reviewed-on: https://go-review.googlesource.com/20674
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Dave Cheney <dave@cheney.net>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile: add Fields field to Type

Switch TSTRUCT and TINTER to use Fields instead of Type, which wrings
out the remaining few direct uses of the latter.

Preparation for converting fields to use a separate "Field" type.

Passes toolstash/buildall.

Change-Id: I5a2ea7e159d0dde1be2c9afafc10a8f739d95743
Reviewed-on: https://go-review.googlesource.com/20675
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>

cmd/compile: simplify InitPlan

Passes toolstash -cmp.

Change-Id: Iaa0d78c2552efb29e67f6c99c7287f8566027add
Reviewed-on: https://go-review.googlesource.com/20673
Reviewed-by: Matthew Dempsky <mdempsky@google.com>

cmd/compile: convert data siz to int

Follow-up to CL 20645.

Passes toolstash -cmp.

Change-Id: Idc63c41b2be2d52e3a6ac59b3a12eb41aa2efbed
Reviewed-on: https://go-review.googlesource.com/20670
Reviewed-by: Ian Lance Taylor <iant@golang.org>

cmd/compile: fix CTNIL static data comments

Change-Id: I434f881c10dec41a802641db8873abf1353beba2
Reviewed-on: https://go-review.googlesource.com/20671
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile: allow deduplication of long strings

String symbols' names used to appear in the final binary.
Using a string's contents as it's symbol's name
was a thus a bad idea if the string's name was long.
Recent improvements by crawshaw have changed that.

Instead of placing long strings behind opaque names
in local packages, place them in the global string
package and make them content-addressable.
Symbol names still occur in the object files,
so use a hash to avoid needless length there.

Reduces the size of cmd/go by 30k.

Change-Id: Ifdbbaf47bf44352418c90ddd903d5106e48db4f1
Reviewed-on: https://go-review.googlesource.com/20524
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Crawshaw <crawshaw@golang.org>

cmd/compile: fix no-opt build after moving decomposing user functions

decompose-builtin pass requires an opt pass, but -N disables
late-opt, the only opt pass (out of two) that happens
after decompose-builtin.  This CL enables both 'opt' and 'late opt'
passes. The extra compile time for 'late opt' in negligible
since most rewrites were already done in the first 'opt'
(also measured before). We should put some effort in splitting the
generic rules into required and optional.

Also update generic.rules comments about lowering
of StringMake and SliceMake.

Tested with GO_GCFLAGS=-N ./all.bash

Change-Id: I92999681aaa02587b6dc6e32ce997a91f1fc9499
Reviewed-on: https://go-review.googlesource.com/20682
Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd/compile: remove amd64 code from package gc and the core gen tool

Parts of the SSA compiler in package gc contain amd64-specific code,
most notably Prog generation. Move this code into package amd64, so that
other architectures can be added more easily.

In package gc, this change is just moving code. There are no functional
changes or even any larger structural changes beyond changing function
names (mostly for export).

In the cmd/compile/internal/ssa/gen tool, more information is included
in arch to remove the AMD64-specific behavior in the main portion of the
tool. The generated opGen.go is identical.

Change-Id: I8eb37c6e6df6de1b65fa7dab6f3bc32c29daf643
Reviewed-on: https://go-review.googlesource.com/20609
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

runtime: document the G states

In particular, write down the rules for stack ownership because the
details of this are about to get very important with concurrent stack
shrinking. (Interestingly, the details don't actually change, but
anything that's currently skating on thin ice is likely to fall
through.)

Fox #12967.

Change-Id: I561e2610e864295e9faba07717a934aabefcaab9
Reviewed-on: https://go-review.googlesource.com/20034
Reviewed-by: Rick Hudson <rlh@golang.org>

runtime: copy stack before adjusting

Currently copystack adjusts pointers in the old stack and then copies
the adjusted stack to the new stack. In addition to being generally
confusing, this is going to make concurrent stack shrinking harder.

Switch this around so that we first copy the stack and then adjust
pointers on the new stack (never writing to the old stack).

This reprises CL 15996, but takes a different and simpler approach. CL
15996 still walked the old stack while adjusting pointers on the new
stack. In this CL, we adjust auxiliary structures before walking the
stack, so we can just walk the new stack.

For #12967.

Change-Id: I94fa86f823ba9ee478e73b2ba509eed3361c43df
Reviewed-on: https://go-review.googlesource.com/20033
Reviewed-by: Rick Hudson <rlh@golang.org>

runtime: improve comment on selectgo

In particular, document that *sel is on the stack no matter what.

Change-Id: I1c264215e026c27721b13eedae73ec845066cdec
Reviewed-on: https://go-review.googlesource.com/20032
Reviewed-by: Rick Hudson <rlh@golang.org>

runtime: fix a typo in asssembly macro GO_RESULTS_INITIALIZED

Fixes #14772

Change-Id: I32f2b6b74de28be406b1306364bc07620a453962
Reviewed-on: https://go-review.googlesource.com/20680
Reviewed-by: David du Colombier <0intro@gmail.com>
Reviewed-by: Minux Ma <minux@golang.org>
Run-TryBot: Minux Ma <minux@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile: change the type of ssa Warnl line number

Line numbers are always int32, so the Warnl function should take the
line number as an int32 as well.  This matches gc.Warnl and removes
a cast every place it's used.

Change-Id: I5d6201e640d52ec390eb7174f8fd8c438d4efe58
Reviewed-on: https://go-review.googlesource.com/20662
Run-TryBot: Todd Neal <todd@tneal.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>

cmd/internal/obj, cmd/asm: get rid of obj.ADATA

Just recognize "DATA" as a special pseudo op word in the assembler
directly.

Change-Id: I508e111fd71f561efa600ad69567a7089a57adb2
Reviewed-on: https://go-review.googlesource.com/20648
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rob Pike <r@golang.org>

cmd/internal/obj: support arbitrarily sized string data

Updates #14786.

Change-Id: I5fe889886f772167386cd10390ac50abc1383937
Reviewed-on: https://go-review.googlesource.com/20607
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>

cmd/all: stop generating ADATA Progs

The only remaining place that generated ADATA
Prog was the assembler. Stop, and delete some
now-dead code.

Passes toolstash -cmp.

Change-Id: I26578ff1b4868e98562b44f69d909c083e96f8d5
Reviewed-on: https://go-review.googlesource.com/20646
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

cmd/compile: write some static data directly

Instead of generating ADATA instructions for
static data, write that static data directly
into the linker sym.
This is considerably more efficient.
The assembler still generates
ADATA instructions, so the ADATA machinery
cannot be dismantled yet. (Future work.)
Skipping ADATA has a significant impact
compiling the unicode package, which has lots
of static data.

name     old time/op    new time/op    delta
Unicode     227ms ±10%     192ms ± 4%  -15.61%  (p=0.000 n=29+30)

name     old alloc/op   new alloc/op   delta
Unicode    51.0MB ± 0%    45.8MB ± 0%  -10.29%  (p=0.000 n=30+30)

name     old allocs/op  new allocs/op  delta
Unicode      610k ± 0%      578k ± 0%   -5.29%  (p=0.000 n=30+30)

This does not pass toolstash -cmp, because
this changes the order in which some relocations
get added, and thus it changes the output from
the compiler. It is not worth the execution time
to sort the relocs in the normal case.

However, compiling with -S -v generates identical
output if (1) you suppress printing of ADATA progs
in flushplist and (2) you suppress printing of
cpu timing. It is reasonable to suppress printing
the ADATA progs, since the data itself is dumped
later. I am therefore fairly confident that all
changes are superficial and non-functional.

Fixes #14786, although there's more to do
in general.

Change-Id: I8dfabe7b423b31a30e516cfdf005b62a2e9ccd82
Reviewed-on: https://go-review.googlesource.com/20645
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

reflect: use SelectDir instead of uintptr in runtimeSelect

And fix the wrong comment.

Initially found this because the comment was wrong about the possible
values. Then noticed that there doesn't seem to be any reason to use
uintptr over SelectDir.

Change-Id: I4f9f9640e49d89e558ed00bd99e57dab890785f5
Reviewed-on: https://go-review.googlesource.com/20655
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

runtime: speed up growslice by avoiding divisions

Only compute the number of maximum allowed elements per slice once.
Special case newcap computation for slices with byte sized elements.

name              old time/op  new time/op  delta
GrowSliceBytes-2  61.1ns ± 1%  43.4ns ± 1%  -29.00%  (p=0.000 n=20+20)
GrowSliceInts-2   85.9ns ± 1%  75.7ns ± 1%  -11.80%  (p=0.000 n=20+20)

Change-Id: I5d9c0d5987cdd108ac29dc32e31912dcefa2324d
Reviewed-on: https://go-review.googlesource.com/20653
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

os/user: fix formatting of error group lookup message

It was failing like "unknown groupid ᎈ|" instead of "unknown groupid
5000" due to the conversion from int to string.

Updates #14806

Change-Id: I83e4b478ff628ad4053573a9f32b3fadce22e847
Reviewed-on: https://go-review.googlesource.com/20642
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>

cmd/compile, runtime: eliminate growslice_n

Fixes #11419.

Change-Id: I7935a253e3e96191a33f5041bab203ecc5f0c976
Reviewed-on: https://go-review.googlesource.com/20647
Reviewed-by: Keith Randall <khr@golang.org>

cmd/compile: move decompose builtin closer to late opt

* Shaves about 10k from pkg/tools/linux_amd64.
* Was suggested by drchase before
* Found by looking at ssa output of #14758

Change-Id: If2c4ddf3b2603d4dfd8fb4d9199b9a3dcb05b17d
Reviewed-on: https://go-review.googlesource.com/20570
Reviewed-by: David Chase <drchase@google.com>
Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile: clean up savedata

Simplify function signature.
Remove unused Ctxt field Enforce_data_order.

Passes toolstash -cmp.

Change-Id: I5caa270ca9ae725708fe415b2c978b5f40780255
Reviewed-on: https://go-review.googlesource.com/20644
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd/compile: minor cleanup

Passes toolstash -cmp.

Change-Id: Ib8c76081956e1d4eecef16b82bc8d527d7849d7e
Reviewed-on: https://go-review.googlesource.com/20643
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/pprof: fix crash with invalid source

pprof crashed when running with:

     go tool pprof %INVALID

Change-Id: I47c2a4da7273e3a97a833f320a650b8c6b59d6b9
Reviewed-on: https://go-review.googlesource.com/20632
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd/internal/obj: sort relocs by off when printing

This makes the output of compiling with -S more
stable in the face of unimportant variation in the
order in which relocs are generated.
It is also more pleasant to read the relocs when
they are sorted.

Also, do some minor cleanup.

For #14786

Change-Id: Id92020b13fd21777dfb5b29c2722c3b2eb27001b
Reviewed-on: https://go-review.googlesource.com/20641
Reviewed-by: Matthew Dempsky <mdempsky@google.com>

cmd/compile: rework how fieldtrack is implemented

Shrinks gc.Type and gc.Func slightly.

Passes "GOEXPERIMENT=fieldtrack ./all.bash" and "go test -a
-toolexec='toolstash -cmp' -ldflags=-k=rsc.io/tmp/fieldtrack.tracked
rsc.io/tmp/fieldtrack".

Change-Id: I785fe8a18eb830d9867d34247e4cd41a6a7921d4
Reviewed-on: https://go-review.googlesource.com/20557
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd/compile: const folding for float32/64

Split the auxFloat type into 32/64 bit versions and perform checking for
exactly representable float32 values.  Perform const folding on
float32/64.  Comment out some const negation rules that the frontend
already performs.

Change-Id: Ib3f8d59fa8b30e50fe0267786cfb3c50a06169d2
Reviewed-on: https://go-review.googlesource.com/20568
Run-TryBot: Todd Neal <todd@tneal.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>

cmd/compile/internal/ssa: generalize prove to all booleans

* Refacts a bit saving and restoring parents restrictions
* Shaves ~100k from pkg/tools/linux_amd64,
but most of the savings come from the rewrite rules.
* Improves on the following artificial test case:
func f1(a4 bool, a6 bool) bool {
  return a6 || (a6 || (a6 || a4)) || (a6 || (a4 || a6 || (false || a6)))
}

Change-Id: I714000f75a37a3a6617c6e6834c75bd23674215f
Reviewed-on: https://go-review.googlesource.com/20306
Reviewed-by: Keith Randall <khr@golang.org>
Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro>
TryBot-Result: Gobot Gobot <gobot@golang.org>

runtime: move testSchedLocalQueue* to export_test

Move functions testSchedLocalQueueLocal and testSchedLocalQueueSteal
from proc.go to export_test.go, the only site that they are used.

Fixes #14796

Change-Id: I16b6fa4a13835eab33f66a2c2e87a5f5c79b7bd3
Reviewed-on: https://go-review.googlesource.com/20640
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

net/http: delete unused type readClose

readClose hasn't been used in years.

Change-Id: Ia963c082e049ef856c2a637fec15204efe48b3b6
Reviewed-on: https://go-review.googlesource.com/20631
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

net/http: make ParseMultipartForm also populate Request.PostForm

Ensures that after request.ParseMultipartForm has been invoked,
Request.PostForm and Request.Form are both populated with the
same formValues read in, instead of only populating Request.Form.

Fixes #9305

Change-Id: I3d4a11b006fc7dffaa35360014fe15b8c74d00a3
Reviewed-on: https://go-review.googlesource.com/19986
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile: reuse auto symbols during ssa construction

Reuse auto symbols so cse can eliminate OpAddrs that refer to
them.

Change-Id: I69e6a3f77a3a33946459cf8c6eccf223f6125048
Reviewed-on: https://go-review.googlesource.com/20569
Run-TryBot: Todd Neal <todd@tneal.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>

cmd/link: use string map for symbols with single version

Reduces link time by ~3%

Results with gc on:
name       old s/op   new s/op   delta
LinkCmdGo  0.82 ± 2%  0.78 ± 2%  -3.90%  (p=0.000 n=17+17)
LinkJuju   7.11 ± 7%  6.87 ± 6%  -3.41%  (p=0.012 n=20+19)

Less noisy results with gc turned off:
name       old s/op   new s/op   delta
LinkCmdGo  0.66 ± 2%  0.64 ± 2%  -3.14%  (p=0.000 n=18+20)
LinkJuju   5.91 ± 1%  5.72 ± 2%  -3.17%  (p=0.000 n=20+20)

Change-Id: I4cac7933b0b22d0aee18255e1ab54550ad364593
Reviewed-on: https://go-review.googlesource.com/20478
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: David Crawshaw <crawshaw@golang.org>
Run-TryBot: David Crawshaw <crawshaw@golang.org>

cmd/link: rewrite pe symbol table generating code

Every go executable has COFF symbol table appended at the end. The table is
used by nm and addr2line and contains all symbols present in the executable.
The table is quite large. For example, my go.exe has 11736 records.

To generate symbol table:
1) we walk "all symbols" list to count symbols we want for the table;
2) we allocate large global array of COFFSym structs (32 bytes each)
   to fit our symbols;
3) we walk "all symbols" list again to fill our array with contents;
4) we iterate over our global array to write all records to the file.

This CL changes all these steps with single step:
- walk "all symbols" list and write each COFF symbol table record to
  the file as we go.

I hope new version is faster and uses less garbage, but I don't know
how to benchmark this.

Change-Id: Ie4870583250131ea4428e0e83a0696c9df1794e0
Reviewed-on: https://go-review.googlesource.com/20580
Reviewed-by: David Crawshaw <crawshaw@golang.org>
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Run-TryBot: Alex Brainman <alex.brainman@gmail.com>

cmd/compile: strength reduce *24

We use *24 a lot for pointer arithmetic when accessing slices
of slices ([][]T).  Rewrite to use an LEA and a shift.
The shift will likely be free, as it often gets folded into
an indexed load/store.

Update #14606

Change-Id: Ie0bf6dc1093876efd57e88ce5f62c26a9bf21cec
Reviewed-on: https://go-review.googlesource.com/20567
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Todd Neal <todd@tneal.org>

cmd/compile: remove structpkg global variable

The structpkg global variable was only used to verify internal
consistency when declaring methods during import. Track the
value in the parser and binary importer directly and pass it
to the relevant function as an argument.

Change-Id: I7e5e006f9046d84f9a3959616f073798fda36c97
Reviewed-on: https://go-review.googlesource.com/20606
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd/compile: compute number of arguments correctly

The outCount value includes a flag bit for dotdotdot.

If we have this count incorrect, then the offset for the
methodset *rtype are in the wrong place.

Fixes #14783

Change-Id: If5acb16af08d4ffe36c8c9ee389c32f2712ce757
Reviewed-on: https://go-review.googlesource.com/20566
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

crypto/tls: implement dynamic record sizing

Currently, if a client of crypto/tls (e.g., net/http, http2) calls
tls.Conn.Write with a 33KB buffer, that ends up writing three TLS
records: 16KB, 16KB, and 1KB. Slow clients (such as 2G phones) must
download the first 16KB record before they can decrypt the first byte.
To improve latency, it's better to send smaller TLS records. However,
sending smaller records adds overhead (more overhead bytes and more
crypto calls), which slightly hurts throughput.

A simple heuristic, implemented in this change, is to send small
records for new connections, then boost to large records after the
first 1MB has been written on the connection.

Fixes #14376

Change-Id: Ice0f6279325be6775aa55351809f88e07dd700cd
Reviewed-on: https://go-review.googlesource.com/19591
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Tom Bergan <tombergan@google.com>
Reviewed-by: Adam Langley <agl@golang.org>

cmd/compile: don't copy all type nodes for builtin functions

Only copy the ones that actually change.  Also combine deep and substAny
functions into one.  The Type.Copyany field is now unused, so remove it.

Passes toolstash -cmp.

Change-Id: Id28a9bf144ecf3e522aad00496f8a21ae2b74680
Reviewed-on: https://go-review.googlesource.com/20600
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Run-TryBot: Ian Lance Taylor <iant@golang.org>

crypto/tls: document that the cipher suite list is non-exhaustive.

Fixes #14370.

Change-Id: Ieb95ee3494f592fb5fc74aa4b803479671816927
Reviewed-on: https://go-review.googlesource.com/20551
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

crypto/x509: correct OID for DSA-with-SHA-256.

I copied this down incorrectly. See
https://tools.ietf.org/html/rfc5758#section-3.1.

Thankfully it's incredibly obscure.

Fixes #14663.

Change-Id: I4792ee979916cc15aa2e300c8bc989bc4f9c63ac
Reviewed-on: https://go-review.googlesource.com/20549
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Adam Langley <agl@golang.org>

crypto/tls: better error for oversized handshake messages.

This change improves the error message when encountering a TLS handshake
message that is larger than our limit (64KB). Previously the error was
just “local error: internal error”.

Updates #13401.

Change-Id: I86127112045ae33e51079e3bc047dd7386ddc71a
Reviewed-on: https://go-review.googlesource.com/20547
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Adam Langley <agl@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile: move universe block code into new universe.go

Move lexinit, typeinit, lexinit1, and lexfini into new universe.go
file, and give them a more idiomatic and descriptive API. No code
changes.

Change-Id: I0e9b25dcc86ad10f4b990dc02bd33477b488cc85
Reviewed-on: https://go-review.googlesource.com/20604
Reviewed-by: Robert Griesemer <gri@golang.org>

cmd/compile: move lexn and lexname into lex.go (cleanup)

Missed these two declarations in the previous cleanup.

Change-Id: I54ff3accd387dd90e12847daccf4477169797f81
Reviewed-on: https://go-review.googlesource.com/20603
Run-TryBot: Robert Griesemer <gri@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd/compile: move lexer into separate file (cleanup)

This is really moving all the non-lexer pieces out of lex.go
into main.go. It's always been confusing that the top-most
compiler entry point (Main) is in the same file with the
lexer. Both files remain of substantial size (> 1000 lines),
which justifies this even more.

No other changes.

Change-Id: I03895589d5e3cc2340580350bbc1420539893dfc
Reviewed-on: https://go-review.googlesource.com/20601
Run-TryBot: Robert Griesemer <gri@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile: simplify lexer.ungetr

Change-Id: Id041c1b2d364aecd7a6613a53237f7de2c650a7e
Reviewed-on: https://go-review.googlesource.com/20599
Run-TryBot: Robert Griesemer <gri@golang.org>
Reviewed-by: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/link: treat reflect.Value.Method like Call

Fixes #14740

Change-Id: Iad8d971c21977b0a1f4ef55a08bb180a8125e976
Reviewed-on: https://go-review.googlesource.com/20562
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile: use bufio.Reader directly in lexer

Removes an intermediate layer of functions that was clogging up a
corner of the compiler's profile graph.

I can't measure a performance improvement running a large build
like jujud, but the profile reports less total time spent in
gc.(*lexer).getr.

Change-Id: I3000585cfcb0f9729d3a3859e9023690a6528591
Reviewed-on: https://go-review.googlesource.com/20565
Reviewed-by: Robert Griesemer <gri@golang.org>
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile: track reflect.Type.Method in deadcode

In addition to reflect.Value.Call, exported methods can be invoked
by the Func value in the reflect.Method struct. This CL has the
compiler track what functions get access to a legitimate reflect.Method
struct by looking for interface calls to either of:

Method(int) reflect.Method
MethodByName(string) (reflect.Method, bool)

This is a little overly conservative. If a user implements a type
with one of these methods without using the underlying calls on
reflect.Type, the linker will assume the worst and include all
exported methods. But it's cheap.

No change to any of the binary sizes reported in cl/20483.

For #14740

Change-Id: Ie17786395d0453ce0384d8b240ecb043b7726137
Reviewed-on: https://go-review.googlesource.com/20489
Reviewed-by: Matthew Dempsky <mdempsky@google.com>

cmd/compile: improve go.importpath.* docs

While we're here, remove a bit of dead code.

Change-Id: I1344bb41e5d368825ca5748189382f9e7023c59a
Reviewed-on: https://go-review.googlesource.com/20596
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd/link: tidy up rdsym

Use an early return. Check errors. Deduplicate.

Change-Id: Iabefd563b5ef82a16fab4791277630804fd09003
Reviewed-on: https://go-review.googlesource.com/20597
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd/compile: eliminate uses of Type.Down in bexport.go

Change-Id: I298b9d389ac33a24365f4c06a122c3af989906c0
Reviewed-on: https://go-review.googlesource.com/20594
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>

cmd/compile: sort interface methods in tointerface0

Might as well sort them while they're still in a slice.

Change-Id: I40c25ddc5c054dcb4da2aeefa79947967609d599
Reviewed-on: https://go-review.googlesource.com/20591
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

compress/flate: optimize huffman bit encoder

Part 1 of optimizing the deflater. This optimizes the bitwriter by:

* Removing allocations.
* Storing compound values for bit codes instead of 2 separate tables.
* Accumulate 48 bits between writes instead of 24.
* Inline bit flushing.

This also contains code that will be used in later CL's
(writeBlockDynamic, writeBlockHuff).

Tests for Huffman bit writer encoding regressions has been added.

name                       old speed      new speed      delta
EncodeDigitsSpeed1e4-4     19.3MB/s ± 1%  21.6MB/s ± 1%  +11.77%
EncodeDigitsSpeed1e5-4     25.0MB/s ± 6%  30.7MB/s ± 1%  +22.70%
EncodeDigitsSpeed1e6-4     28.2MB/s ± 1%  32.3MB/s ± 1%  +14.64%
EncodeDigitsDefault1e4-4   13.3MB/s ± 0%  14.2MB/s ± 1%   +7.07%
EncodeDigitsDefault1e5-4   6.43MB/s ± 1%  6.64MB/s ± 1%   +3.27%
EncodeDigitsDefault1e6-4   5.81MB/s ± 0%  5.85MB/s ± 1%   +0.69%
EncodeDigitsCompress1e4-4  13.2MB/s ± 0%  14.4MB/s ± 0%   +9.10%
EncodeDigitsCompress1e5-4  6.40MB/s ± 1%  6.61MB/s ± 0%   +3.20%
EncodeDigitsCompress1e6-4  5.80MB/s ± 1%  5.90MB/s ± 1%   +1.64%
EncodeTwainSpeed1e4-4      18.4MB/s ± 1%  20.7MB/s ± 1%  +12.72%
EncodeTwainSpeed1e5-4      27.7MB/s ± 1%  31.0MB/s ± 1%  +11.78%
EncodeTwainSpeed1e6-4      29.1MB/s ± 0%  32.9MB/s ± 2%  +13.25%
EncodeTwainDefault1e4-4    12.4MB/s ± 0%  13.1MB/s ± 1%   +5.88%
EncodeTwainDefault1e5-4    7.52MB/s ± 1%  7.83MB/s ± 0%   +4.19%
EncodeTwainDefault1e6-4    7.08MB/s ± 1%  7.26MB/s ± 0%   +2.54%
EncodeTwainCompress1e4-4   12.0MB/s ± 1%  12.8MB/s ± 1%   +6.70%
EncodeTwainCompress1e5-4   5.96MB/s ± 1%  6.16MB/s ± 0%   +3.27%
EncodeTwainCompress1e6-4   5.37MB/s ± 0%  5.39MB/s ± 1%   +0.47%

>Allocations:

benchmark                              old allocs     new allocs     delta
BenchmarkEncodeDigitsSpeed1e4-4        50             0              -100.00%
BenchmarkEncodeDigitsSpeed1e5-4        110            0              -100.00%
BenchmarkEncodeDigitsSpeed1e6-4        1032           0              -100.00%
BenchmarkEncodeDigitsDefault1e4-4      56             0              -100.00%
BenchmarkEncodeDigitsDefault1e5-4      120            0              -100.00%
BenchmarkEncodeDigitsDefault1e6-4      966            0              -100.00%
BenchmarkEncodeDigitsCompress1e4-4     56             0              -100.00%
BenchmarkEncodeDigitsCompress1e5-4     120            0              -100.00%
BenchmarkEncodeDigitsCompress1e6-4     966            0              -100.00%
BenchmarkEncodeTwainSpeed1e4-4         58             0              -100.00%
BenchmarkEncodeTwainSpeed1e5-4         132            0              -100.00%
BenchmarkEncodeTwainSpeed1e6-4         1082           0              -100.00%
BenchmarkEncodeTwainDefault1e4-4       52             0              -100.00%
BenchmarkEncodeTwainDefault1e5-4       126            0              -100.00%
BenchmarkEncodeTwainDefault1e6-4       886            0              -100.00%
BenchmarkEncodeTwainCompress1e4-4      52             0              -100.00%
BenchmarkEncodeTwainCompress1e5-4      120            0              -100.00%
BenchmarkEncodeTwainCompress1e6-4      880            0              -100.00%

benchmark                              old bytes     new bytes     delta
BenchmarkEncodeDigitsSpeed1e4-4        4288          2             -99.95%
BenchmarkEncodeDigitsSpeed1e5-4        8896          15            -99.83%
BenchmarkEncodeDigitsSpeed1e6-4        84098         153           -99.82%
BenchmarkEncodeDigitsDefault1e4-4      4480          3             -99.93%
BenchmarkEncodeDigitsDefault1e5-4      9216          76            -99.18%
BenchmarkEncodeDigitsDefault1e6-4      73920         768           -98.96%
BenchmarkEncodeDigitsCompress1e4-4     4480          3             -99.93%
BenchmarkEncodeDigitsCompress1e5-4     9216          76            -99.18%
BenchmarkEncodeDigitsCompress1e6-4     73920         768           -98.96%
BenchmarkEncodeTwainSpeed1e4-4         4544          2             -99.96%
BenchmarkEncodeTwainSpeed1e5-4         9600          15            -99.84%
BenchmarkEncodeTwainSpeed1e6-4         77633         153           -99.80%
BenchmarkEncodeTwainDefault1e4-4       4352          3             -99.93%
BenchmarkEncodeTwainDefault1e5-4       9408          76            -99.19%
BenchmarkEncodeTwainDefault1e6-4       65984         768           -98.84%
BenchmarkEncodeTwainCompress1e4-4      4352          3             -99.93%
BenchmarkEncodeTwainCompress1e5-4      9216          76            -99.18%
BenchmarkEncodeTwainCompress1e6-4      65792         768           -98.83%

Updates #14258

Change-Id: Ibaa97b9619743ad623094727228eb2ada1ec7f1f
Reviewed-on: https://go-review.googlesource.com/19336
Reviewed-by: Nigel Tao <nigeltao@golang.org>
Reviewed-by: Joe Tsai <joetsai@digital-static.net>
Run-TryBot: Joe Tsai <joetsai@digital-static.net>
TryBot-Result: Gobot Gobot <gobot@golang.org>

compress/flate: test if results are deterministic

This will test if deflate output is deterministic between two runs
of the deflater, when write sizes differ.

The deflater makes no official promises that results are
deterministic between runs, but this is a good test to determine
unintentional randomness.

Note that this does not guarantee that results are deterministic
across platforms nor that results will be deterministic between
Go versions. This is also not guarantees we should imply.

Change-Id: Id7dd89fe276060fd83a43d0b34ac35d50fcd32d9
Reviewed-on: https://go-review.googlesource.com/20573
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>

cmd/internal/obj: fix nil pointer dereference in Dconv

p can be nil in Dconv so we need to do a check before dereferencing
it. Fixes a problem I was having running toolstash.

Change-Id: I34d6d278b319583d8454c2342ac88e054fc4b641
Reviewed-on: https://go-review.googlesource.com/20595
Reviewed-by: David Crawshaw <crawshaw@golang.org>
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile: move InitPlan and InitEntry into sinit.go

I don't know what they're used for, but that's the only file they're
referenced in.

Change-Id: Ie39d7d4621e2d5224408243b5789597ca0dc14be
Reviewed-on: https://go-review.googlesource.com/20593
Reviewed-by: Dave Cheney <dave@cheney.net>

cmd/compile: give "magic" code its own source file

This code is an eye sore to keep scrolling past in subr.go, so move it
out of the way.

Change-Id: I8eafc1725d868a4924ee7ca9b7738cce309f9eff
Reviewed-on: https://go-review.googlesource.com/20592
Reviewed-by: Dave Cheney <dave@cheney.net>

runtime: limit TestCgoCCodeSIGPROF test to 1 second

Still fails about 20% of the time on my laptop.

Fixes #14766.

Change-Id: I169ab728c6022dceeb91188f5ad466ed6413c062
Reviewed-on: https://go-review.googlesource.com/20590
Run-TryBot: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile: remove Label.Link field and lastlabel global var

Change-Id: If2a174f482ecd56dee43f921d13fef98439872fc
Reviewed-on: https://go-review.googlesource.com/20559
Reviewed-by: Matthew Dempsky <mdempsky@google.com>

net/http: defer idleMu.Unlock() in tryPutIdleConn + minor typo fixes.

Change-Id: Ia2273c3a9f0001d16b0c767fea91498a9acb0af5
Reviewed-on: https://go-review.googlesource.com/19963
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

cmd/compile: simplify transformclosure

Use idiomatic slicing operations instead of incrementally building a
linked list.

Passes toolstash -cmp.

Change-Id: Idb0e40c7b4d7d1110d23828afa8ae1d157ba905f
Reviewed-on: https://go-review.googlesource.com/20556
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: cleanup unsafenmagic

In particular, make Alignof work more like Sizeof. Other idiomatic
cleanups while here.

Passes toolstash -cmp.

Change-Id: I4def20894f3d95e49ab6a50ddba189be36fdd258
Reviewed-on: https://go-review.googlesource.com/20555
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>

cmd/compile: eliminate uses of Type.Down in alg.go

This could be done by threading the Iter value down through memrun and
ispaddedfield, but that ends up a bit clunky. This way is also closer
to how we'll want the code to look once fields are kept in slices.

Passes toolstash -cmp.

Change-Id: I8a44445c85f921eb18d97199df2026c5ce0f4f67
Reviewed-on: https://go-review.googlesource.com/20558
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>