]> Cypherpunks repositories - gostls13.git/log
gostls13.git
9 years agocmd/compile: a rule's line number is at its ->
Keith Randall [Tue, 26 Apr 2016 21:09:58 +0000 (14:09 -0700)]
cmd/compile: a rule's line number is at its ->

Let's define the line number of a multiline rule as the line
number on which the -> appears.  This helps make the rule
cover analysis look a bit nicer.

Change-Id: I4ac4c09f2240285976590ecfd416bc4c05e78946
Reviewed-on: https://go-review.googlesource.com/22473
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
9 years agocmd/compile: lazily initialize litbuf
Matthew Dempsky [Tue, 26 Apr 2016 17:55:32 +0000 (10:55 -0700)]
cmd/compile: lazily initialize litbuf

Instead of eagerly creating strings like "literal 2.01" for every
lexed number in case we need to mention it in an error message, defer
this work to (*parser).syntax_error.

name      old allocs/op  new allocs/op  delta
Template      482k ± 0%      482k ± 0%  -0.12%   (p=0.000 n=9+10)
GoTypes      1.35M ± 0%     1.35M ± 0%  -0.04%  (p=0.015 n=10+10)
Compiler     5.45M ± 0%     5.44M ± 0%  -0.12%    (p=0.000 n=9+8)

Change-Id: I333b3c80e583864914412fb38f8c0b7f1d8c8821
Reviewed-on: https://go-review.googlesource.com/22480
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
9 years agocmd/dist: sort entries in zcgo.go generated file for deterministic build
Robert Griesemer [Tue, 26 Apr 2016 22:18:48 +0000 (15:18 -0700)]
cmd/dist: sort entries in zcgo.go generated file for deterministic build

This simplifies comparison of object files across different builds
by ensuring that the strings in the zcgo.go always appear in the
same order.

Change-Id: I3639ea4fd10e0d645b838d1bbb03cd33deca340e
Reviewed-on: https://go-review.googlesource.com/22478
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
9 years agounicode: improve SimpleFold performance for ascii
Egon Elbre [Tue, 17 Nov 2015 14:51:23 +0000 (16:51 +0200)]
unicode: improve SimpleFold performance for ascii

This change significantly speeds up case-insensitive regexp matching.

benchmark                      old ns/op      new ns/op      delta
BenchmarkMatchEasy0i_32-8      2690           1473           -45.24%
BenchmarkMatchEasy0i_1K-8      80404          42269          -47.43%
BenchmarkMatchEasy0i_32K-8     3272187        2076118        -36.55%
BenchmarkMatchEasy0i_1M-8      104805990      66503805       -36.55%
BenchmarkMatchEasy0i_32M-8     3360192200     2126121600     -36.73%

benchmark                      old MB/s     new MB/s     speedup
BenchmarkMatchEasy0i_32-8      11.90        21.72        1.83x
BenchmarkMatchEasy0i_1K-8      12.74        24.23        1.90x
BenchmarkMatchEasy0i_32K-8     10.01        15.78        1.58x
BenchmarkMatchEasy0i_1M-8      10.00        15.77        1.58x
BenchmarkMatchEasy0i_32M-8     9.99         15.78        1.58x

Issue #13288

Change-Id: I94af7bb29e75d60b4f6ee760124867ab271b9642
Reviewed-on: https://go-review.googlesource.com/16943
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agogc: use AbsFileLine for deterministic binary export data
Alan Donovan [Mon, 25 Apr 2016 22:31:36 +0000 (18:31 -0400)]
gc: use AbsFileLine for deterministic binary export data

This version of the file name honors the -trimprefix flag,
which strips off variable parts like $WORK or $PWD.
The TestCgoConsistentResults test now passes.

Change-Id: If93980b054f9b13582dd314f9d082c26eaac4f41
Reviewed-on: https://go-review.googlesource.com/22444
Reviewed-by: Robert Griesemer <gri@golang.org>
9 years agocmd/compile: don't discard inlineable but empty functions with binary export format
Robert Griesemer [Tue, 26 Apr 2016 21:11:38 +0000 (14:11 -0700)]
cmd/compile: don't discard inlineable but empty functions with binary export format

Change-Id: I0f016fa000f949d27847d645b4cdebe68a8abf20
Reviewed-on: https://go-review.googlesource.com/22474
Run-TryBot: Robert Griesemer <gri@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/link: pass -no-pie (if supported) when creating a race-enabled executable.
Michael Hudson-Doyle [Tue, 26 Apr 2016 09:37:49 +0000 (21:37 +1200)]
cmd/link: pass -no-pie (if supported) when creating a race-enabled executable.

Fixes #15443

Change-Id: Ia3593104fc1a4255926ae5675c25390604b44b7b
Reviewed-on: https://go-review.googlesource.com/22453
Run-TryBot: Michael Hudson-Doyle <michael.hudson@canonical.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Dmitry Vyukov <dvyukov@google.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
9 years agocmd/link: fix gdb backtrace on architectures using a link register
Michael Munday [Wed, 13 Apr 2016 17:34:41 +0000 (13:34 -0400)]
cmd/link: fix gdb backtrace on architectures using a link register

Also adds TestGdbBacktrace to the runtime package.

Dwarf modifications written by Bryan Chan (@bryanpkc) who is also
at IBM and covered by the same CLA.

Fixes #14628

Change-Id: I106a1f704c3745a31f29cdadb0032e3905829850
Reviewed-on: https://go-review.googlesource.com/20193
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/compile/internal/gc: rewrite comment to avoid automated meaning
Russ Cox [Tue, 26 Apr 2016 17:18:14 +0000 (13:18 -0400)]
cmd/compile/internal/gc: rewrite comment to avoid automated meaning

The comment says 'DΟ NΟT SUBMIT', and that text being in a file can cause
automated errors or warnings when trying to check the Go sources into other
source control systems.

(We reject that string in CL commit messages, which I've avoided here
by changing the O's to Ο's above.)

Change-Id: I6cdd57a8612ded5208f05a8bd6b137f44424a030
Reviewed-on: https://go-review.googlesource.com/22434
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
9 years agocmd/compile: more sanity checks on rewrite rules
Keith Randall [Wed, 20 Apr 2016 18:17:41 +0000 (11:17 -0700)]
cmd/compile: more sanity checks on rewrite rules

Make sure ops have the right number of args, set
aux and auxint only if allowed, etc.

Normalize error reporting format.

Change-Id: Ie545fcc5990c8c7d62d40d9a0a55885f941eb645
Reviewed-on: https://go-review.googlesource.com/22320
Reviewed-by: David Chase <drchase@google.com>
9 years agocrypto/sha512: add s390x assembly implementation
Michael Munday [Mon, 25 Apr 2016 20:17:42 +0000 (16:17 -0400)]
crypto/sha512: add s390x assembly implementation

Renames block to blockGeneric so that it can be called when the
assembly feature check fails. This means making block a var on
platforms without an assembly implementation (similar to the sha1
package).

Also adds a test to check that the fallback path works correctly
when the feature check fails.

name        old speed      new speed       delta
Hash8Bytes  7.13MB/s ± 2%  19.89MB/s ± 1%  +178.82%   (p=0.000 n=9+10)
Hash1K       121MB/s ± 1%    661MB/s ± 1%  +444.54%   (p=0.000 n=10+9)
Hash8K       137MB/s ± 0%    918MB/s ± 1%  +569.29%  (p=0.000 n=10+10)

Change-Id: Id65dd6e943f14eeffe39a904dc88065fc6a60179
Reviewed-on: https://go-review.googlesource.com/22402
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Michael Munday <munday@ca.ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agonet: ignore lame referral responses like libresolv
Matthew Dempsky [Mon, 25 Apr 2016 20:09:11 +0000 (13:09 -0700)]
net: ignore lame referral responses like libresolv

Fixes #15434.

Change-Id: Ia88b740df5418a6d3af1c29a03756f4234f388b0
Reviewed-on: https://go-review.googlesource.com/22428
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/link: correctly decode name length
David Crawshaw [Tue, 26 Apr 2016 14:53:25 +0000 (10:53 -0400)]
cmd/link: correctly decode name length

The linker was incorrectly decoding type name lengths, causing
typelinks to be sorted out of order and in cases where the name was
the exact right length, linker panics.

Added a test to the reflect package that causes TestTypelinksSorted
to fail before this CL. It's not the exact failure seen in #15448
but it has the same cause: decodetype_name calculating the wrong
length.

The equivalent decoders in reflect/type.go and runtime/type.go
have the parenthesis in the right place.

Fixes #15448

Change-Id: I33257633d812b7d2091393cb9d6cc8a73e0138c8
Reviewed-on: https://go-review.googlesource.com/22403
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Russ Cox <rsc@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/compile: fix another bug in dominator computation
David Chase [Mon, 25 Apr 2016 20:24:11 +0000 (16:24 -0400)]
cmd/compile: fix another bug in dominator computation

Here, "fix" means "replace".  The new dominator computation
is the "simple" algorithm from Lengauer and Tarjan's TOPLAS
paper, with minimal changes.

Also included is a test that tweaks the fixed error.

Change-Id: I0abdf53d5d64df1e67e4e62f55e88957045cd63b
Reviewed-on: https://go-review.googlesource.com/22401
Run-TryBot: David Chase <drchase@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
9 years agostrings: use SSE4.2 in strings.Index on AMD64
Ilya Tocar [Thu, 21 Apr 2016 15:24:12 +0000 (18:24 +0300)]
strings: use SSE4.2 in strings.Index on AMD64

Use PCMPESTRI instruction if available.

Index-4              21.1ns ± 0%  21.1ns ± 0%     ~     (all samples are equal)
IndexHard1-4          395µs ± 0%   105µs ± 0%  -73.53%        (p=0.000 n=19+20)
IndexHard2-4          300µs ± 0%   147µs ± 0%  -51.11%        (p=0.000 n=19+20)
IndexHard3-4          665µs ± 0%   665µs ± 0%     ~           (p=0.942 n=16+19)

Change-Id: I4f66794164740a2b939eb1c78934e2390b489064
Reviewed-on: https://go-review.googlesource.com/22337
Run-TryBot: Ilya Tocar <ilya.tocar@intel.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Russ Cox <rsc@golang.org>
9 years agocmd/compile: sort import strings for canonical obj files
Robert Griesemer [Mon, 25 Apr 2016 22:59:42 +0000 (15:59 -0700)]
cmd/compile: sort import strings for canonical obj files

This is not necessary for reproduceability but it removes
differences due to imported package order between compiles
using textual vs binary export format. The packages list
tends to be very short, so it's ok doing it always for now.

Guarded with a documented (const) flag so it's trivial to
disable and remove eventually.

Also, use the same flag now to enforce parameter numbering.

Change-Id: Ie05d2490df770239696ecbecc07532ed62ccd5c0
Reviewed-on: https://go-review.googlesource.com/22445
Run-TryBot: Robert Griesemer <gri@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
9 years agoruntime: arm5, fix large-offset floating-point stores
Keith Randall [Mon, 25 Apr 2016 21:12:26 +0000 (14:12 -0700)]
runtime: arm5, fix large-offset floating-point stores

The code sequence for large-offset floating-point stores
includes adding the base pointer to r11.  Make sure we
can interpret that instruction correctly.

Fixes build.

Fixes #15440

Change-Id: I7fe5a4a57e08682967052bf77c54e0ec47fcb53e
Reviewed-on: https://go-review.googlesource.com/22440
Reviewed-by: Michael Hudson-Doyle <michael.hudson@canonical.com>
9 years agocmd/compile: for now, keep parameter numbering in binary export format
Robert Griesemer [Mon, 25 Apr 2016 21:39:51 +0000 (14:39 -0700)]
cmd/compile: for now, keep parameter numbering in binary export format

The numbering is only required for parameters of functions/methods
with exported inlineable bodies. For now, always export parameter names
with internal numbering to minimize the diffs between assembly code
dumps of code compiled with the textual vs the binary format.

To be disabled again once the new export format is default.

Change-Id: I6d14c564e734cc5596c7e995d8851e06d5a35013
Reviewed-on: https://go-review.googlesource.com/22441
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
9 years agocmd/compile: treat empty and absent struct field tags as identical
Matthew Dempsky [Mon, 25 Apr 2016 20:24:48 +0000 (13:24 -0700)]
cmd/compile: treat empty and absent struct field tags as identical

Fixes #15439.

Change-Id: I5a32384c46e20f8db6968e5a9e854c45ab262fe4
Reviewed-on: https://go-review.googlesource.com/22429
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agoruntime: zero tmpbuf between len and cap
Keith Randall [Mon, 25 Apr 2016 00:04:32 +0000 (17:04 -0700)]
runtime: zero tmpbuf between len and cap

Zero the entire buffer so we don't need to
lower its capacity upon return.  This lets callers
do some appending without allocation.

Zeroing is cheap, the byte buffer requires only
4 extra instructions.

Fixes #14235

Change-Id: I970d7badcef047dafac75ac17130030181f18fe2
Reviewed-on: https://go-review.googlesource.com/22424
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
9 years agocmd/compile: introduce bool operations.
Alexandru Moșoi [Sun, 24 Apr 2016 19:21:07 +0000 (21:21 +0200)]
cmd/compile: introduce bool operations.

Introduce OrB, EqB, NeqB, AndB to handle bool operations.

Change-Id: I53e4d5125a8090d5eeb4576db619103f19fff58d
Reviewed-on: https://go-review.googlesource.com/22412
Reviewed-by: Keith Randall <khr@golang.org>
9 years agospec: be more explicit about equivalence of empty string and absent field tags
Robert Griesemer [Fri, 22 Apr 2016 23:35:29 +0000 (16:35 -0700)]
spec: be more explicit about equivalence of empty string and absent field tags

Note that the spec already makes that point with a comment in the very first
example for struct field tags. This change is simply stating this explicitly
in the actual spec prose.

- gccgo and go/types already follow this rule
- the current reflect package API doesn't distinguish between absent tags
  and empty tags (i.e., there is no discoverable difference)

Fixes #15412.

Change-Id: I92f9c283064137b4c8651630cee0343720717a02
Reviewed-on: https://go-review.googlesource.com/22391
Reviewed-by: Russ Cox <rsc@golang.org>
Reviewed-by: Rob Pike <r@golang.org>
9 years agotest: add test for issue 15084
Josh Bleecher Snyder [Mon, 25 Apr 2016 17:55:22 +0000 (10:55 -0700)]
test: add test for issue 15084

The underlying issues have been fixed.
All the individual fixes have their own tests,
but it's still useful to have a plain source test.

Fixes #15084

Change-Id: I06c485a7d0716201bd57d1f3be53668dddd7ec14
Reviewed-on: https://go-review.googlesource.com/22426
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Chase <drchase@google.com>
9 years agocmd/compile: encapsulate OSLICE* representation
Josh Bleecher Snyder [Thu, 21 Apr 2016 18:55:33 +0000 (11:55 -0700)]
cmd/compile: encapsulate OSLICE* representation

As a nice side-effect, this allows us to
unify several code paths.

The terminology (low, high, max, simple slice expr,
full slice expr) is taken from the spec and
the examples in the spec.

This is a trial run. The plan, probably for Go 1.8,
is to change slice expressions to use Node.List
instead of OKEY, and to do some similar
tree structure changes for other ops.

Passes toolstash -cmp. No performance change.
all.bash passes with GO_GCFLAGS=-newexport.

Updates #15350

Change-Id: Ic1efdc36e79cdb95ae1636e9817a3ac8f83ab1ac
Reviewed-on: https://go-review.googlesource.com/22425
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
9 years agocmd/go: disable failing TestGoGetInsecure
Russ Cox [Mon, 25 Apr 2016 14:48:45 +0000 (10:48 -0400)]
cmd/go: disable failing TestGoGetInsecure

Update #15410

Change-Id: Iad3f2639aa7a67b11efc35a629e1893f7d87b957
Reviewed-on: https://go-review.googlesource.com/22430
Run-TryBot: Russ Cox <rsc@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
9 years agodebug/pe: better error messages
Alex Brainman [Sun, 24 Apr 2016 05:09:00 +0000 (15:09 +1000)]
debug/pe: better error messages

Updates #15345

Change-Id: Iae35d3e378cbc8157ba1ff91e4971ed4515a5e5c
Reviewed-on: https://go-review.googlesource.com/22394
Reviewed-by: David Crawshaw <crawshaw@golang.org>
Run-TryBot: Alex Brainman <alex.brainman@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/compile: minor cleanup in inl
Josh Bleecher Snyder [Sun, 24 Apr 2016 21:09:03 +0000 (14:09 -0700)]
cmd/compile: minor cleanup in inl

* Make budget an int32 to avoid needless conversions.
* Introduce some temporary variables to reduce repetition.
* If ... args are present, they will be the last argument
  to the function. No need to scan all arguments.

Passes toolstash -cmp.

Change-Id: I55203609f5d2f25a4e238cd48c63214651120cfc
Reviewed-on: https://go-review.googlesource.com/22421
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
9 years agocmd/compile: replace len(Nodes.Slice()) with Nodes.Len()
Josh Bleecher Snyder [Sun, 24 Apr 2016 20:50:26 +0000 (13:50 -0700)]
cmd/compile: replace len(Nodes.Slice()) with Nodes.Len()

Generated with eg:

func before(n gc.Nodes) int { return len(n.Slice()) }
func after(n gc.Nodes) int  { return n.Len() }

Change-Id: Ifdf01915e60069166afe96aa7b1d08720bf62fc5
Reviewed-on: https://go-review.googlesource.com/22420
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
9 years agotest: add test that required algs get generated
Josh Bleecher Snyder [Sun, 24 Apr 2016 20:31:55 +0000 (13:31 -0700)]
test: add test that required algs get generated

This is a follow-up to CLs 19769 and 19770.

Change-Id: Ia9b71055613b80df4ce62b34fcc4f479f04f72fd
Reviewed-on: https://go-review.googlesource.com/22399
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Crawshaw <crawshaw@golang.org>
9 years agocmd/compile: use gc.Etype's String method
Josh Bleecher Snyder [Fri, 22 Apr 2016 15:39:56 +0000 (08:39 -0700)]
cmd/compile: use gc.Etype's String method

Passes toolstash -cmp.

Change-Id: I42c962cc5a3ffec2969f223cf238c2fdadbf5857
Reviewed-on: https://go-review.googlesource.com/22381
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/compile: give gc.Op a String method, use it
Josh Bleecher Snyder [Fri, 22 Apr 2016 14:14:10 +0000 (07:14 -0700)]
cmd/compile: give gc.Op a String method, use it

Passes toolstash -cmp.

Change-Id: I915e76374fd64aa2597e6fa47e4fa95ca00ca643
Reviewed-on: https://go-review.googlesource.com/22380
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Crawshaw <crawshaw@golang.org>
9 years agocmd/compile: don't generate algs for map buckets
Josh Bleecher Snyder [Mon, 4 May 2015 22:01:03 +0000 (15:01 -0700)]
cmd/compile: don't generate algs for map buckets

Note that this is only safe because
the compiler generates multiple distinct
gc.Types. If we switch to having canonical
gc.Types, then this will need to be updated
to handle the case in which the user uses both
map[T]S and also map[[8]T]S. In that case,
the runtime needs algs for [8]T, but this could
mark the sole [8]T type as Noalg. This is a general
problem with having a single bool to represent
whether alg generation is needed for a type.

Cuts 5k off cmd/go and 22k off golang.org/x/tools/cmd/godoc,
approx 0.04% and 0.12% respectively.

For #6853 and #9930

Change-Id: I30a15ec72ecb62e2aa053260a7f0f75015fc0ade
Reviewed-on: https://go-review.googlesource.com/19769
Reviewed-by: David Crawshaw <crawshaw@golang.org>
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/compile: don't generate algs for ... args
Josh Bleecher Snyder [Mon, 4 May 2015 22:01:29 +0000 (15:01 -0700)]
cmd/compile: don't generate algs for ... args

Note that this is only safe because
the compiler generates multiple distinct
gc.Types. If we switch to having canonical
gc.Types, then this will need to be updated
to handle the case in which the user uses both
map[[n]T]S and also calls a function f(...T) with n arguments.
In that case, the runtime needs algs for [n]T, but this could
mark the sole [n]T type as Noalg. This is a general
problem with having a single bool to represent
whether alg generation is needed for a type.

Cuts 17k off cmd/go and 13k off golang.org/x/tools/cmd/godoc,
approx 0.14% and 0.07% respectively.

For #6853 and #9930

Change-Id: Iccb6b9fd88ade5497d7090528a903816d340bf0a
Reviewed-on: https://go-review.googlesource.com/19770
Reviewed-by: David Crawshaw <crawshaw@golang.org>
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/compile: reorder how slicelit initializes a slice
Keith Randall [Sun, 24 Apr 2016 05:59:01 +0000 (22:59 -0700)]
cmd/compile: reorder how slicelit initializes a slice

  func f(x, y, z *int) {
    a := []*int{x,y,z}
    ...
  }

We used to use:
  var tmp [3]*int
  a := tmp[:]
  a[0] = x
  a[1] = y
  a[2] = z

Now we do:
  var tmp [3]*int
  tmp[0] = x
  tmp[1] = y
  tmp[2] = z
  a := tmp[:]

Doesn't sound like a big deal, but the compiler has trouble
eliminating write barriers when using the former method because it
doesn't know that the slice points to the stack.  In the latter
method, the compiler knows the array is on the stack and as a result
doesn't emit any write barriers.

This turns out to be extremely common when building ... args, like
for calls fmt.Printf.

Makes go binaries ~1% smaller.

Doesn't have a measurable effect on the go1 fmt benchmarks,
unfortunately.

Fixes #14263
Update #6853

Change-Id: I9074a2788ec9e561a75f3b71c119b69f304d6ba2
Reviewed-on: https://go-review.googlesource.com/22395
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
9 years agointernal/trace: fix event ordering for coarse timestamps
Dmitry Vyukov [Sat, 23 Apr 2016 19:18:34 +0000 (21:18 +0200)]
internal/trace: fix event ordering for coarse timestamps

Arm arch uses coarse-grained kernel timer as cputicks.
As the result sort.Sort smashes trace entirely. Use sort.Stable instead.

Change-Id: Idfa017a86a489be58cf239f7fe56d7f4b66b52a9
Reviewed-on: https://go-review.googlesource.com/22317
Run-TryBot: Dmitry Vyukov <dvyukov@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
9 years agoruntime/trace: test detection of broken timestamps
Dmitry Vyukov [Thu, 7 Apr 2016 13:48:15 +0000 (15:48 +0200)]
runtime/trace: test detection of broken timestamps

On some processors cputicks (used to generate trace timestamps)
produce non-monotonic timestamps. It is important that the parser
distinguishes logically inconsistent traces (e.g. missing, excessive
or misordered events) from broken timestamps. The former is a bug
in tracer, the latter is a machine issue.

Test that (1) parser does not return a logical error in case of
broken timestamps and (2) broken timestamps are eventually detected
and reported.

Change-Id: Ib4b1eb43ce128b268e754400ed8b5e8def04bd78
Reviewed-on: https://go-review.googlesource.com/21608
Reviewed-by: Austin Clements <austin@google.com>
9 years agodebug/pe: introduce File.COFFSymbols and (*COFFSymbol).FullName
Alex Brainman [Thu, 21 Apr 2016 06:51:36 +0000 (16:51 +1000)]
debug/pe: introduce File.COFFSymbols and (*COFFSymbol).FullName

Reloc.SymbolTableIndex is an index into symbol table. But
Reloc.SymbolTableIndex cannot be used as index into File.Symbols,
because File.Symbols slice has Aux lines removed as it is built.

We cannot change the way File.Symbols works, so I propose we
introduce new File.COFFSymbols that does not have that limitation.

Also unlike File.Symbols, File.COFFSymbols will consist of
COFFSymbol. COFFSymbol matches PE COFF specification exactly,
and it is simpler to use.

Updates #15345

Change-Id: Icbc265853a472529cd6d64a76427b27e5459e373
Reviewed-on: https://go-review.googlesource.com/22336
Reviewed-by: David Crawshaw <crawshaw@golang.org>
Run-TryBot: Alex Brainman <alex.brainman@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/link: add -dumpdep flag to dump linker dependency graph
Brad Fitzpatrick [Sat, 23 Apr 2016 01:49:59 +0000 (18:49 -0700)]
cmd/link: add -dumpdep flag to dump linker dependency graph

This is what led to https://golang.org/cl/20763 and
https://golang.org/cl/20765 to shrink binary sizes.

Change-Id: Id360d474e6153cfe32a525b0a720810fd113195b
Reviewed-on: https://go-review.googlesource.com/22392
Reviewed-by: David Crawshaw <crawshaw@golang.org>
9 years agocmd/compile: get rid of most byte and word insns for amd64
Keith Randall [Fri, 22 Apr 2016 20:09:18 +0000 (13:09 -0700)]
cmd/compile: get rid of most byte and word insns for amd64

Now that we're using 32-bit ops for 8/16-bit logical operations
(to avoid partial register stalls), there's really no need to
keep track of the 8/16-bit ops at all.  Convert everything we
can to 32-bit ops.

This CL is the obvious stuff.  I might think a bit more about
whether we can get rid of weirder stuff like HMULWU.

The only downside to this CL is that we lose some information
about constants.  If we had source like:
  var a byte = ...
  a += 128
  a += 128
We will convert that to a += 256, when we could get rid of the
add altogether.  This seems like a fairly unusual scenario and
I'm happy with forgoing that optimization.

Change-Id: Ia7c1e5203d0d110807da69ed646535194a3efba1
Reviewed-on: https://go-review.googlesource.com/22382
Reviewed-by: Todd Neal <todd@tneal.org>
9 years agocmd/compile: combine stores into larger widths
Keith Randall [Wed, 20 Apr 2016 22:02:48 +0000 (15:02 -0700)]
cmd/compile: combine stores into larger widths

Combine stores into larger widths when it is safe to do so.

Add clobber() function so stray dead uses do not impede the
above rewrites.

Fix bug in loads where all intermediate values depending on
a small load (not just the load itself) must have no other uses.
We really need the small load to be dead after the rewrite..

Fixes #14267

Change-Id: Ib25666cb19777f65082c76238fba51a76beb5d74
Reviewed-on: https://go-review.googlesource.com/22326
Run-TryBot: Keith Randall <khr@golang.org>
Reviewed-by: Todd Neal <todd@tneal.org>
9 years agoruntime: use per-goroutine sequence numbers in tracer
Dmitry Vyukov [Tue, 5 Apr 2016 13:29:14 +0000 (15:29 +0200)]
runtime: use per-goroutine sequence numbers in tracer

Currently tracer uses global sequencer and it introduces
significant slowdown on parallel machines (up to 10x).
Replace the global sequencer with per-goroutine sequencer.

If we assign per-goroutine sequence numbers to only 3 types
of events (start, unblock and syscall exit), it is enough to
restore consistent partial ordering of all events. Even these
events don't need sequence numbers all the time (if goroutine
starts on the same P where it was unblocked, then start does
not need sequence number).
The burden of restoring the order is put on trace parser.
Details of the algorithm are described in the comments.

On http benchmark with GOMAXPROCS=48:
no tracing: 5026 ns/op
tracing: 27803 ns/op (+453%)
with this change: 6369 ns/op (+26%, mostly for traceback)

Also trace size is reduced by ~22%. Average event size before: 4.63
bytes/event, after: 3.62 bytes/event.

Besides running trace tests, I've also tested with manually broken
cputicks (random skew for each event, per-P skew and episodic random skew).
In all cases broken timestamps were detected and no test failures.

Change-Id: I078bde421ccc386a66f6c2051ab207bcd5613efa
Reviewed-on: https://go-review.googlesource.com/21512
Run-TryBot: Dmitry Vyukov <dvyukov@google.com>
Reviewed-by: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agodoc: mention security from contribution guidelines
Francesc Campoy [Fri, 22 Apr 2016 23:27:34 +0000 (16:27 -0700)]
doc: mention security from contribution guidelines

Fixes #15413

Change-Id: I837a391276eed565cf66d3715ec68b7b959ce143
Reviewed-on: https://go-review.googlesource.com/22390
Reviewed-by: Andrew Gerrand <adg@golang.org>
9 years agocmd/compile: remove redundant "// fallthrough" comments
Matthew Dempsky [Fri, 22 Apr 2016 22:45:24 +0000 (15:45 -0700)]
cmd/compile: remove redundant "// fallthrough" comments

Change-Id: Ia3f262f06592b66447c213e2350402cd5e6e2ccd
Reviewed-on: https://go-review.googlesource.com/22389
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
9 years agocmd/compile: don't export pos info in new export format for now
Robert Griesemer [Fri, 22 Apr 2016 21:50:20 +0000 (14:50 -0700)]
cmd/compile: don't export pos info in new export format for now

Exporting filenames as part of the position information can lead
to different object files which breaks tests.

Change-Id: Ia678ab64293ebf04bf83601e6ba72919d05762a4
Reviewed-on: https://go-review.googlesource.com/22385
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
9 years agoapi: update next.txt
Brad Fitzpatrick [Fri, 22 Apr 2016 22:15:17 +0000 (22:15 +0000)]
api: update next.txt

Change-Id: I12d5e5d0e74b354f26898bab4ea30eb27ac45cd7
Reviewed-on: https://go-review.googlesource.com/22387
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/api: ignore vendored packages
Brad Fitzpatrick [Fri, 22 Apr 2016 22:13:50 +0000 (22:13 +0000)]
cmd/api: ignore vendored packages

Fixes #15404

Change-Id: I16f2a34a1e4c3457053a1fc2141f21747cfb22b4
Reviewed-on: https://go-review.googlesource.com/22386
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
9 years agonet: keep waiting for valid DNS response until timeout
Matthew Dempsky [Sat, 16 Apr 2016 02:19:58 +0000 (19:19 -0700)]
net: keep waiting for valid DNS response until timeout

Prevents denial of service attacks from bogus UDP packets.

Fixes #13281.

Change-Id: Ifb51b17a1b0807bfd27b144d6037431701184e7b
Reviewed-on: https://go-review.googlesource.com/22126
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agonet/url: validate ports in IPv4 addresses
Mohit Agarwal [Thu, 21 Apr 2016 19:17:04 +0000 (00:47 +0530)]
net/url: validate ports in IPv4 addresses

Fixes #14860

Change-Id: Id55ad942d45a104d560a879d6e8e1aa09671789b
Reviewed-on: https://go-review.googlesource.com/22351
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agoencoding/json: add Encoder.DisableHTMLEscaping
Caleb Spare [Sun, 10 Apr 2016 04:18:22 +0000 (21:18 -0700)]
encoding/json: add Encoder.DisableHTMLEscaping

This provides a way to disable the escaping of <, >, and & in JSON
strings.

Fixes #14749.

Change-Id: I1afeb0244455fc8b06c6cce920444532f229555b
Reviewed-on: https://go-review.googlesource.com/21796
Run-TryBot: Caleb Spare <cespare@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
9 years agocmd/compile: replace Ctype switches with type switches
Matthew Dempsky [Fri, 22 Apr 2016 19:27:29 +0000 (12:27 -0700)]
cmd/compile: replace Ctype switches with type switches

Instead of switching on Ctype (which internally uses a type switch)
and then scattering lots of type assertions throughout the CTFOO case
clauses, just use type switches directly on the underlying constant
value.

Passes toolstash/buildall.

Change-Id: I9bc172cc67e5f391cddc15539907883b4010689e
Reviewed-on: https://go-review.googlesource.com/22384
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agoruntime: fix description of trace events
Dmitry Vyukov [Fri, 22 Apr 2016 20:48:11 +0000 (22:48 +0200)]
runtime: fix description of trace events

Change-Id: I037101b1921fe151695d32e9874b50dd64982298
Reviewed-on: https://go-review.googlesource.com/22314
Reviewed-by: Austin Clements <austin@google.com>
9 years agocmd/compile: map TSLICE to obj.KindSlice directly
Matthew Dempsky [Fri, 22 Apr 2016 03:58:35 +0000 (20:58 -0700)]
cmd/compile: map TSLICE to obj.KindSlice directly

Change-Id: Idab5f603c1743895b8f4edbcc55f7be83419a099
Reviewed-on: https://go-review.googlesource.com/22383
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: David Crawshaw <crawshaw@golang.org>
9 years agocmd/dist: skip misc/cgo/test with internal linking on ppc64le
Ian Lance Taylor [Fri, 22 Apr 2016 17:17:06 +0000 (10:17 -0700)]
cmd/dist: skip misc/cgo/test with internal linking on ppc64le

CL 22372 changed ppc64le to use normal cgo initialization on ppc64le.
Doing this uncovered a cmd/link error using internal linking.
Opened issue 15409 for the problem.  This CL disables the test.

Update #15409.

Change-Id: Ia1bb6b874c1b5a4df1a0436c8841c145142c30f7
Reviewed-on: https://go-review.googlesource.com/22379
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
9 years agocmd/trace: generate new pprof profiles
Dmitry Vyukov [Tue, 12 Apr 2016 18:50:59 +0000 (20:50 +0200)]
cmd/trace: generate new pprof profiles

Generate new protobuf pprof profiles with embed symbol info.
This makes program binary unnecessary.

Change-Id: Ie628439c13c5e34199782031138102c83ea50621
Reviewed-on: https://go-review.googlesource.com/21873
Reviewed-by: Hyang-Ah Hana Kim <hyangah@gmail.com>
Run-TryBot: Dmitry Vyukov <dvyukov@google.com>

9 years agocmd/compile: in a Tarjan algorithm, DFS should really be DFS
David Chase [Fri, 22 Apr 2016 16:15:08 +0000 (12:15 -0400)]
cmd/compile: in a Tarjan algorithm, DFS should really be DFS

Replaced incorrect recursion-free rendering of DFS with
something that was correct.  Enhanced test with all
permutations of IF successors to ensure that all possible
DFS traversals are exercised.

Test is improved version of
https://go-review.googlesource.com/#/c/22334

Update 15084.

Change-Id: I6e944c41244e47fe5f568dfc2b360ff93b94079e
Reviewed-on: https://go-review.googlesource.com/22347
Reviewed-by: Keith Randall <khr@golang.org>
Run-TryBot: David Chase <drchase@google.com>

9 years agocrypto/aes: use asm for BenchmarkExpand on amd64
Michael Munday [Thu, 21 Apr 2016 12:00:07 +0000 (08:00 -0400)]
crypto/aes: use asm for BenchmarkExpand on amd64

This reverses the change to this benchmark made in 9b6bf20.

Change-Id: I79ab88286c3028d3be561957140375bbc413e7ab
Reviewed-on: https://go-review.googlesource.com/22340
Run-TryBot: Michael Munday <munday@ca.ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Minux Ma <minux@golang.org>
9 years agohash/crc32: use vector instructions on s390x
Chris Zou [Mon, 18 Apr 2016 23:30:17 +0000 (19:30 -0400)]
hash/crc32: use vector instructions on s390x

The input buffer is aligned to a doubleword boundary to
improve performance of the vector instructions. The pure
Go implementation is used to align the input data, and is
also used when the vector instructions are not available
or the data length is less than 64 bytes.

Change-Id: Ie259a5f2f1562bcc17961c99e5776c99091d6bed
Reviewed-on: https://go-review.googlesource.com/22201
Reviewed-by: Michael Munday <munday@ca.ibm.com>
Run-TryBot: Michael Munday <munday@ca.ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bill O'Farrell <billotosyr@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
9 years agocmd/compile: reenable phielim during rewrite
Alexandru Moșoi [Thu, 21 Apr 2016 08:11:33 +0000 (10:11 +0200)]
cmd/compile: reenable phielim during rewrite

Remove the "optimization" that was causing the issue.

For the following code the "optimization" was
converting v to (OpCopy x) which is wrong because
x doesn't dominate v.

b1:
    y = ...
    First .. b3
b2:
   x = ...
   Goto b3
b3:
   v = phi x y
   ... use v ...

That "optimization" is likely no longer needed because
we now have a second opt pass with a dce in between
which removes blocks of type First.

For pkg/tools/linux_amd64/* the binary size drops
from 82142886 to 82060034.

Change-Id: I10428abbd8b32c5ca66fec3da2e6f3686dddbe31
Reviewed-on: https://go-review.googlesource.com/22312
Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
9 years agocmd/compile: convert some Phis into And8.
Alexandru Moșoi [Fri, 22 Apr 2016 10:44:31 +0000 (12:44 +0200)]
cmd/compile: convert some Phis into And8.

See discussion at [1]. True value must have a fixed non-zero
representation meaning that a && b can be implemented as a & b.

[1] https://groups.google.com/forum/#!topic/golang-dev/xV0vPuFP9Vg

This change helps with m := a && b, but it's more common to see
if a && b { do something } which is not handled.

Change-Id: Ib6f9ff898a0a8c05d12466e2464e4fe781035394
Reviewed-on: https://go-review.googlesource.com/22313
Run-TryBot: Alexandru Moșoi <alexandru@mosoi.ro>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
9 years agocmd/compile: don't nilcheck newobject and return values from mapaccess{1,2}
Keith Randall [Wed, 20 Apr 2016 04:06:53 +0000 (21:06 -0700)]
cmd/compile: don't nilcheck newobject and return values from mapaccess{1,2}

They are guaranteed to be non-nil, no point in inserting
nil checks for them.

Fixes #15390

Change-Id: I3b9a0f2319affc2139dcc446d0a56c6785ae5a86
Reviewed-on: https://go-review.googlesource.com/22291
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
9 years agoruntime/cgo: use normal libinit on PPC GNU/Linux
Ian Lance Taylor [Fri, 22 Apr 2016 14:08:13 +0000 (07:08 -0700)]
runtime/cgo: use normal libinit on PPC GNU/Linux

The special case was because PPC did not support external linking, but
now it does.

Fixes #10410.

Change-Id: I9b024686e0f03da7a44c1c59b41c529802f16ab0
Reviewed-on: https://go-review.googlesource.com/22372
Run-TryBot: Ian Lance Taylor <iant@golang.org>
Reviewed-by: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/compile, etc: use nameOff in uncommonType
David Crawshaw [Fri, 8 Apr 2016 01:37:45 +0000 (21:37 -0400)]
cmd/compile, etc: use nameOff in uncommonType

linux/amd64 PIE:
cmd/go:  -62KB (0.5%)
jujud:  -550KB (0.7%)

For #6853.

Change-Id: Ieb67982abce5832e24b997506f0ae7108f747108
Reviewed-on: https://go-review.googlesource.com/22371
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
9 years agocmd/compile, etc: use nameOff for rtype string
David Crawshaw [Thu, 7 Apr 2016 20:29:16 +0000 (16:29 -0400)]
cmd/compile, etc: use nameOff for rtype string

linux/amd64:
cmd/go:   -8KB (basically nothing)

linux/amd64 PIE:
cmd/go: -191KB (1.6%)
jujud:  -1.5MB (1.9%)

Updates #6853
Fixes #15064

Change-Id: I0adbb95685e28be92e8548741df0e11daa0a9b5f
Reviewed-on: https://go-review.googlesource.com/21777
Reviewed-by: Ian Lance Taylor <iant@golang.org>
9 years agocmd/link: do not align type name data
David Crawshaw [Fri, 8 Apr 2016 00:07:09 +0000 (20:07 -0400)]
cmd/link: do not align type name data

Now that reflect.name objects contain an offset to pkgPath instead of a
pointer, there is no need to align the symbol data.

Removes approx. 10KB from the cmd/go binary. The effect becomes more
important later as more type data is moved into name objects.

For #6853

Change-Id: Idb507fdbdad04f16fc224378f82272cb5c236ab7
Reviewed-on: https://go-review.googlesource.com/21776
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
9 years agocrypto/sha1: add s390x assembly implementation
Michael Munday [Mon, 18 Apr 2016 21:41:50 +0000 (17:41 -0400)]
crypto/sha1: add s390x assembly implementation

Use the compute intermediate message digest (KIMD) instruction
when possible. Adds test to check fallback code path in case
KIMD is not available.

Benchmark changes:
Hash8Bytes  3.4x
Hash1K      9.3x
Hash8K      10.9x

Change-Id: Ibcd71a886dfd7b3822042235b4f4eaa7a148036b
Reviewed-on: https://go-review.googlesource.com/22350
Run-TryBot: Michael Munday <munday@ca.ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
9 years agohtml/template, text/template: clarify Parse{Files,Glob} semantics
Andrew Gerrand [Mon, 11 Apr 2016 01:14:53 +0000 (11:14 +1000)]
html/template, text/template: clarify Parse{Files,Glob} semantics

Document the subtle property that files with equivalent base names
will overwrite extant templates with those same names.

Fixes golang/go#14320

Change-Id: Ie9ace1b08e6896ea599836e31582123169aa7a25
Reviewed-on: https://go-review.googlesource.com/21824
Reviewed-by: Rob Pike <r@golang.org>
9 years agocmd/go: write test file to temporary directory
Andrew Gerrand [Mon, 18 Apr 2016 03:25:51 +0000 (13:25 +1000)]
cmd/go: write test file to temporary directory

Before this change, a go-vendor-issue-14613 file would be left in the
working directory after tests run.

Change-Id: If1858421bb287215ab4a19163f489131b2e8912c
Reviewed-on: https://go-review.googlesource.com/22169
Reviewed-by: Chris Broadfoot <cbro@golang.org>
9 years agoencoding/gob: document compatibility
Rob Pike [Thu, 21 Apr 2016 19:43:22 +0000 (12:43 -0700)]
encoding/gob: document compatibility

Fixes #13808.

Change-Id: Ifbd5644da995a812438a405485c9e08b4503a313
Reviewed-on: https://go-review.googlesource.com/22352
Reviewed-by: Ian Lance Taylor <iant@golang.org>
9 years agoflag: update test case (fix build)
Robert Griesemer [Thu, 21 Apr 2016 22:55:50 +0000 (15:55 -0700)]
flag: update test case (fix build)

Change-Id: I2275dc703be4fda3feedf76483148eab853b43b8
Reviewed-on: https://go-review.googlesource.com/22360
Run-TryBot: Robert Griesemer <gri@golang.org>
Reviewed-by: Rob Pike <r@golang.org>
9 years agocmd/link: convert Link.Filesyms into a slice
Michael Hudson-Doyle [Thu, 21 Apr 2016 22:10:08 +0000 (10:10 +1200)]
cmd/link: convert Link.Filesyms into a slice

Change-Id: I6490de325b0f4ba962c679503102d30d41dcc384
Reviewed-on: https://go-review.googlesource.com/22359
Run-TryBot: Michael Hudson-Doyle <michael.hudson@canonical.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/link: fix Codeblk printing when -a to use Textp as a slice
Michael Hudson-Doyle [Thu, 21 Apr 2016 21:38:41 +0000 (09:38 +1200)]
cmd/link: fix Codeblk printing when -a to use Textp as a slice

Does anyone actually pass -a to the linker?

Change-Id: I1d31ea66aa5604b7fd42adf15bdab71e9f52d0ed
Reviewed-on: https://go-review.googlesource.com/22356
Run-TryBot: Michael Hudson-Doyle <michael.hudson@canonical.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Crawshaw <crawshaw@golang.org>
9 years agodoc/go1.7.txt: 0s for zero duration, go doc groups constructors with types
Rob Pike [Thu, 21 Apr 2016 22:09:49 +0000 (15:09 -0700)]
doc/go1.7.txt: 0s for zero duration, go doc groups constructors with types

Change-Id: I4fc35649ff5a3510f5667b62e7e84e113e95dffe
Reviewed-on: https://go-review.googlesource.com/22358
Reviewed-by: Rob Pike <r@golang.org>
9 years agotime: print zero duration as 0s, not 0
Rob Pike [Thu, 21 Apr 2016 21:53:19 +0000 (14:53 -0700)]
time: print zero duration as 0s, not 0

There should be a unit, and s is the SI unit name, so use that.
The other obvious possibility is ns (nanosecond), but the fact
that durations are measured in nanoseconds is an internal detail.

Fixes #14058.

Change-Id: Id1f8f3c77088224d9f7cd643778713d5cc3be5d9
Reviewed-on: https://go-review.googlesource.com/22357
Reviewed-by: Robert Griesemer <gri@golang.org>
9 years agocmd/doc: group constructors with type in package presentation
Rob Pike [Thu, 21 Apr 2016 20:29:07 +0000 (13:29 -0700)]
cmd/doc: group constructors with type in package presentation

Fixes #14004.

$ go doc encoding.gob
Before:
func Register(value interface{})
func RegisterName(name string, value interface{})
func NewDecoder(r io.Reader) *Decoder
func NewEncoder(w io.Writer) *Encoder
type CommonType struct { ... }
type Decoder struct { ... }
type Encoder struct { ... }
type GobDecoder interface { ... }
type GobEncoder interface { ... }

After:
func Register(value interface{})
func RegisterName(name string, value interface{})
type CommonType struct { ... }
type Decoder struct { ... }
    func NewDecoder(r io.Reader) *Decoder
type Encoder struct { ... }
    func NewEncoder(w io.Writer) *Encoder
type GobDecoder interface { ... }
type GobEncoder interface { ... }

Change-Id: I021db25bce4a16b3dfa22ab323ca1f4e68d50111
Reviewed-on: https://go-review.googlesource.com/22354
Reviewed-by: Robert Griesemer <gri@golang.org>
9 years agocmd/compile: Use pre-regalloc value ID in lateSpillUse
Keith Randall [Thu, 21 Apr 2016 20:58:22 +0000 (13:58 -0700)]
cmd/compile: Use pre-regalloc value ID in lateSpillUse

The cached copy's ID is sometimes outside the bounds of the orig array.

There's no reason to start at the cached copy and work backwards
to the original value. We already have the original value ID at
all the callsites.

Fixes noopt build

Change-Id: I313508a1917e838a87e8cc83b2ef3c2e4a8db304
Reviewed-on: https://go-review.googlesource.com/22355
Run-TryBot: Keith Randall <khr@golang.org>
Reviewed-by: David Chase <drchase@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/compile: split TSLICE into separate Type kind
Matthew Dempsky [Mon, 18 Apr 2016 21:02:08 +0000 (14:02 -0700)]
cmd/compile: split TSLICE into separate Type kind

Instead of using TARRAY for both arrays and slices, create a new
TSLICE kind to handle slices.

Also, get rid of the "DDDArray" distinction. While kinda ugly, it
seems likely we'll need to defer evaluating the constant bounds
expressions for golang.org/issue/13890.

Passes toolstash/buildall.

Change-Id: I8e45d4900e7df3a04cce59428ec8b38035d3cc3a
Reviewed-on: https://go-review.googlesource.com/22329
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agospec: fix incorrect comment in shift example
Robert Griesemer [Thu, 21 Apr 2016 19:47:24 +0000 (12:47 -0700)]
spec: fix incorrect comment in shift example

- adjusted example code
- fixed comments

Fixes #14785.

Change-Id: Ia757dc93b0a69b8408559885ece7f3685a37daaa
Reviewed-on: https://go-review.googlesource.com/22353
Reviewed-by: Rob Pike <r@golang.org>
9 years agoruntime: eliminate floating garbage estimate
Austin Clements [Wed, 30 Mar 2016 21:15:15 +0000 (17:15 -0400)]
runtime: eliminate floating garbage estimate

Currently when we compute the trigger for the next GC, we do it based
on an estimate of the reachable heap size at the start of the GC
cycle, which is itself based on an estimate of the floating garbage.
This was introduced by 4655aad to fix a bad feedback loop that allowed
the heap to grow to many times the true reachable size.

However, this estimate gets easily confused by rapidly allocating
applications, and, worse it's different than the heap size the trigger
controller uses to compute the trigger itself. This results in the
trigger controller often thinking that GC finished before it started.
Since this would be a pretty great outcome from it's perspective, it
sets the trigger for the next cycle as close to the next goal as
possible (which is limited to 95% of the goal).

Furthermore, the bad feedback loop this estimate originally fixed
seems not to happen any more, suggesting it was fixed more correctly
by some other change in the mean time. Finally, with the change to
allocate black, it shouldn't even be theoretically possible for this
bad feedback loop to occur.

Hence, eliminate the floating garbage estimate and simply consider the
reachable heap to be the marked heap. This harms overall throughput
slightly for allocation-heavy benchmarks, but significantly improves
mutator availability.

Fixes #12204. This brings the average trigger in this benchmark from
0.95 (the cap) to 0.7 and the active GC utilization from ~90% to ~45%.

Updates #14951. This makes the trigger controller much better behaved,
so it pulls the trigger lower if assists are consuming a lot of CPU
like it's supposed to, increasing mutator availability.

name              old time/op  new time/op  delta
XBenchGarbage-12  2.21ms ± 1%  2.28ms ± 3%  +3.29%  (p=0.000 n=17+17)

Some of this slow down we paid for in earlier commits. Relative to the
start of the series to switch to allocate-black (the parent of "count
black allocations toward scan work"), the garbage benchmark is 2.62%
slower.

name                      old time/op    new time/op    delta
BinaryTree17-12              2.53s ± 3%     2.53s ± 3%    ~     (p=0.708 n=20+19)
Fannkuch11-12                2.08s ± 0%     2.08s ± 0%  -0.22%  (p=0.002 n=19+18)
FmtFprintfEmpty-12          45.3ns ± 2%    45.2ns ± 3%    ~     (p=0.505 n=20+20)
FmtFprintfString-12          129ns ± 0%     131ns ± 2%  +1.80%  (p=0.000 n=16+19)
FmtFprintfInt-12             121ns ± 2%     121ns ± 2%    ~     (p=0.768 n=19+19)
FmtFprintfIntInt-12          186ns ± 1%     188ns ± 3%  +0.99%  (p=0.000 n=19+19)
FmtFprintfPrefixedInt-12     188ns ± 1%     188ns ± 1%    ~     (p=0.947 n=18+16)
FmtFprintfFloat-12           254ns ± 1%     255ns ± 1%  +0.30%  (p=0.002 n=19+17)
FmtManyArgs-12               763ns ± 0%     770ns ± 0%  +0.92%  (p=0.000 n=18+18)
GobDecode-12                7.00ms ± 1%    7.04ms ± 1%  +0.61%  (p=0.049 n=20+20)
GobEncode-12                5.88ms ± 1%    5.88ms ± 0%    ~     (p=0.641 n=18+19)
Gzip-12                      214ms ± 1%     215ms ± 1%  +0.43%  (p=0.002 n=18+19)
Gunzip-12                   37.6ms ± 0%    37.6ms ± 0%  +0.11%  (p=0.015 n=17+18)
HTTPClientServer-12         76.9µs ± 2%    78.1µs ± 2%  +1.44%  (p=0.000 n=20+18)
JSONEncode-12               15.2ms ± 2%    15.1ms ± 1%    ~     (p=0.271 n=19+18)
JSONDecode-12               53.1ms ± 1%    53.3ms ± 0%  +0.49%  (p=0.000 n=18+19)
Mandelbrot200-12            4.04ms ± 1%    4.03ms ± 0%  -0.33%  (p=0.005 n=18+18)
GoParse-12                  3.29ms ± 1%    3.28ms ± 1%    ~     (p=0.146 n=16+17)
RegexpMatchEasy0_32-12      69.9ns ± 3%    69.5ns ± 1%    ~     (p=0.785 n=20+19)
RegexpMatchEasy0_1K-12       237ns ± 0%     237ns ± 0%    ~     (p=1.000 n=18+18)
RegexpMatchEasy1_32-12      69.5ns ± 1%    69.2ns ± 1%  -0.44%  (p=0.020 n=16+19)
RegexpMatchEasy1_1K-12       372ns ± 1%     371ns ± 2%    ~     (p=0.086 n=20+19)
RegexpMatchMedium_32-12      108ns ± 3%     107ns ± 1%  -1.00%  (p=0.004 n=19+14)
RegexpMatchMedium_1K-12     34.2µs ± 4%    34.0µs ± 2%    ~     (p=0.380 n=19+20)
RegexpMatchHard_32-12       1.77µs ± 4%    1.76µs ± 3%    ~     (p=0.558 n=18+20)
RegexpMatchHard_1K-12       53.4µs ± 4%    52.8µs ± 2%  -1.10%  (p=0.020 n=18+20)
Revcomp-12                   359ms ± 4%     377ms ± 0%  +5.19%  (p=0.000 n=20+18)
Template-12                 63.7ms ± 2%    62.9ms ± 2%  -1.27%  (p=0.005 n=18+20)
TimeParse-12                 316ns ± 2%     313ns ± 1%    ~     (p=0.059 n=20+16)
TimeFormat-12                329ns ± 0%     331ns ± 0%  +0.39%  (p=0.000 n=16+18)
[Geo mean]                  51.6µs         51.7µs       +0.18%

Change-Id: I1dce4640c8205d41717943b021039fffea863c57
Reviewed-on: https://go-review.googlesource.com/21324
Reviewed-by: Rick Hudson <rlh@golang.org>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agoruntime: allocate black during GC
Austin Clements [Wed, 30 Mar 2016 21:02:23 +0000 (17:02 -0400)]
runtime: allocate black during GC

Currently we allocate white for most of concurrent marking. This is
based on the classical argument that it produces less floating
garbage, since allocations during GC may not get linked into the heap
and allocating white lets us reclaim these. However, it's not clear
how often this actually happens, especially since our write barrier
shades any pointer as soon as it's installed in the heap regardless of
the color of the slot.

On the other hand, allocating black has several advantages that seem
to significantly outweigh this downside.

1) It naturally bounds the total scan work to the live heap size at
the start of a GC cycle. Allocating white does not, and thus depends
entirely on assists to prevent the heap from growing faster than it
can be scanned.

2) It reduces the total amount of scan work per GC cycle by the size
of newly allocated objects that are linked into the heap graph, since
objects allocated black never need to be scanned.

3) It reduces total write barrier work since more objects will already
be black when they are linked into the heap graph.

This gives a slight overall improvement in benchmarks.

name              old time/op  new time/op  delta
XBenchGarbage-12  2.24ms ± 0%  2.21ms ± 1%  -1.32%  (p=0.000 n=18+17)

name                      old time/op    new time/op    delta
BinaryTree17-12              2.60s ± 3%     2.53s ± 3%  -2.56%  (p=0.000 n=20+20)
Fannkuch11-12                2.08s ± 1%     2.08s ± 0%    ~     (p=0.452 n=19+19)
FmtFprintfEmpty-12          45.1ns ± 2%    45.3ns ± 2%    ~     (p=0.367 n=19+20)
FmtFprintfString-12          131ns ± 3%     129ns ± 0%  -1.60%  (p=0.000 n=20+16)
FmtFprintfInt-12             122ns ± 0%     121ns ± 2%  -0.86%  (p=0.000 n=16+19)
FmtFprintfIntInt-12          187ns ± 1%     186ns ± 1%    ~     (p=0.514 n=18+19)
FmtFprintfPrefixedInt-12     189ns ± 0%     188ns ± 1%  -0.54%  (p=0.000 n=16+18)
FmtFprintfFloat-12           256ns ± 0%     254ns ± 1%  -0.43%  (p=0.000 n=17+19)
FmtManyArgs-12               769ns ± 0%     763ns ± 0%  -0.72%  (p=0.000 n=18+18)
GobDecode-12                7.08ms ± 2%    7.00ms ± 1%  -1.22%  (p=0.000 n=20+20)
GobEncode-12                5.88ms ± 0%    5.88ms ± 1%    ~     (p=0.406 n=18+18)
Gzip-12                      214ms ± 0%     214ms ± 1%    ~     (p=0.103 n=17+18)
Gunzip-12                   37.6ms ± 0%    37.6ms ± 0%    ~     (p=0.563 n=17+17)
HTTPClientServer-12         77.2µs ± 3%    76.9µs ± 2%    ~     (p=0.606 n=20+20)
JSONEncode-12               15.1ms ± 1%    15.2ms ± 2%    ~     (p=0.138 n=19+19)
JSONDecode-12               53.3ms ± 1%    53.1ms ± 1%  -0.33%  (p=0.000 n=19+18)
Mandelbrot200-12            4.04ms ± 1%    4.04ms ± 1%    ~     (p=0.075 n=19+18)
GoParse-12                  3.30ms ± 1%    3.29ms ± 1%  -0.57%  (p=0.000 n=18+16)
RegexpMatchEasy0_32-12      69.5ns ± 1%    69.9ns ± 3%    ~     (p=0.822 n=18+20)
RegexpMatchEasy0_1K-12       237ns ± 1%     237ns ± 0%    ~     (p=0.398 n=19+18)
RegexpMatchEasy1_32-12      69.8ns ± 2%    69.5ns ± 1%    ~     (p=0.090 n=20+16)
RegexpMatchEasy1_1K-12       371ns ± 1%     372ns ± 1%    ~     (p=0.178 n=19+20)
RegexpMatchMedium_32-12      108ns ± 2%     108ns ± 3%    ~     (p=0.124 n=20+19)
RegexpMatchMedium_1K-12     33.9µs ± 2%    34.2µs ± 4%    ~     (p=0.309 n=20+19)
RegexpMatchHard_32-12       1.75µs ± 2%    1.77µs ± 4%  +1.28%  (p=0.018 n=19+18)
RegexpMatchHard_1K-12       52.7µs ± 1%    53.4µs ± 4%  +1.23%  (p=0.013 n=15+18)
Revcomp-12                   354ms ± 1%     359ms ± 4%  +1.27%  (p=0.043 n=20+20)
Template-12                 63.6ms ± 2%    63.7ms ± 2%    ~     (p=0.654 n=20+18)
TimeParse-12                 313ns ± 1%     316ns ± 2%  +0.80%  (p=0.014 n=17+20)
TimeFormat-12                332ns ± 0%     329ns ± 0%  -0.66%  (p=0.000 n=16+16)
[Geo mean]                  51.7µs         51.6µs       -0.09%

Change-Id: I2214a6a0e4f544699ea166073249a8efdf080dc0
Reviewed-on: https://go-review.googlesource.com/21323
Reviewed-by: Rick Hudson <rlh@golang.org>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agoruntime: simplify/optimize allocate-black a bit
Austin Clements [Sun, 17 Apr 2016 15:42:37 +0000 (11:42 -0400)]
runtime: simplify/optimize allocate-black a bit

Currently allocating black switches to the system stack (which is
probably a historical accident) and atomically updates the global
bytes marked stat. Since we're about to depend on this much more,
optimize it a bit by putting it back on the regular stack and updating
the per-P bytes marked stat, which gets lazily folded into the global
bytes marked stat.

Change-Id: Ibbe16e5382d3fd2256e4381f88af342bf7020b04
Reviewed-on: https://go-review.googlesource.com/22170
Reviewed-by: Rick Hudson <rlh@golang.org>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agoruntime: count black allocations toward scan work
Austin Clements [Sat, 16 Apr 2016 22:27:38 +0000 (18:27 -0400)]
runtime: count black allocations toward scan work

Currently we count black allocations toward the scannable heap size,
but not toward the scan work we've done so far. This is clearly
inconsistent (we have, in effect, scanned these allocations and since
they're already black, we're not going to scan them again). Worse, it
means we don't count black allocations toward the scannable heap size
as of the *next* GC because this is based on the amount of scan work
we did in this cycle.

Fix this by counting black allocations as scan work. Currently the GC
spends very little time in allocate-black mode, so this probably
hasn't been a problem, but this will become important when we switch
to always allocating black.

Change-Id: If6ff693b070c385b65b6ecbbbbf76283a0f9d990
Reviewed-on: https://go-review.googlesource.com/22119
Reviewed-by: Rick Hudson <rlh@golang.org>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agotesting: add matching of subtest
Marcel van Lohuizen [Fri, 29 Jan 2016 15:16:03 +0000 (16:16 +0100)]
testing: add matching of subtest

Allows passing regexps per subtest to --test.run and --test.bench

Note that the documentation explicitly states that the split regular
expressions match the correpsonding parts (path components) of
the bench/test identifier. This is intended and slightly different
from the i'th RE matching the subtest/subbench at the respective
level.  Picking this semantics allows guaranteeing that a test or
benchmark identifier as printed by go test can be passed verbatim
(possibly quoted) to, respectively, -run or -bench: subtests and
subbenches might have a '/' in their name, causing a misaligment if
their ID is passed to -run or -bench as is.
This semantics has other benefits, but this is the main motivation.

Fixes golang.go#15126

Change-Id: If72e6d3f54db1df6bc2729ac6edc7ab3c740e7c3
Reviewed-on: https://go-review.googlesource.com/19122
Reviewed-by: Russ Cox <rsc@golang.org>
Run-TryBot: Marcel van Lohuizen <mpvl@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/compile: handle mem copies in amd64 backend
Keith Randall [Thu, 21 Apr 2016 17:02:36 +0000 (10:02 -0700)]
cmd/compile: handle mem copies in amd64 backend

Fixes noopt builder.

Change-Id: If13373b2597f0fcc9b1b2f9c860f2bd043e43c6c
Reviewed-on: https://go-review.googlesource.com/22338
Reviewed-by: Keith Randall <khr@golang.org>
9 years agocmd/compile/internal/gc: fix return value offset for SSA backend on ARM
Cherry Zhang [Mon, 18 Apr 2016 14:30:20 +0000 (10:30 -0400)]
cmd/compile/internal/gc: fix return value offset for SSA backend on ARM

Progress on SSA backend for ARM. Still not complete. It compiles a
Fibonacci function, but the caller picked the return value from an
incorrect offset. This CL adjusts it to match the stack frame layout
for architectures with link register.

Updates #15365.

Change-Id: I01e03c3e95f5503a185e8ac2b6d9caf4faf3d014
Reviewed-on: https://go-review.googlesource.com/22186
Reviewed-by: Keith Randall <khr@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/compile/internal/arm: fix comparison & conditional branch for SSA on ARM
Cherry Zhang [Mon, 18 Apr 2016 16:21:51 +0000 (12:21 -0400)]
cmd/compile/internal/arm: fix comparison & conditional branch for SSA on ARM

Progress on SSA for ARM. Still not complete. Now Fibonacci function compiles
and runs correctly.

The old backend swaps the operands for CMP instruction. This CL does the same
on SSA backend, and uses conditional branch accordingly.

Updates #15365.

Change-Id: I117e17feb22f03d936608bd232f76970e4bbe21a
Reviewed-on: https://go-review.googlesource.com/22187
Reviewed-by: Keith Randall <khr@golang.org>
9 years agocmd/link: convert textp into a slice
David Crawshaw [Tue, 19 Apr 2016 18:02:21 +0000 (14:02 -0400)]
cmd/link: convert textp into a slice

Updates #15374

Change-Id: I3ea715735862fe9550b88d7a29def6cb9d4419a6
Reviewed-on: https://go-review.googlesource.com/22243
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Michael Hudson-Doyle <michael.hudson@canonical.com>
9 years agocmd/link: calculate section alignment concurrently
David Crawshaw [Tue, 19 Apr 2016 12:59:56 +0000 (08:59 -0400)]
cmd/link: calculate section alignment concurrently

Reduces link time for cmd/go by 1%.

Change-Id: Iad4a16db0aedc56f81ddf73ba9b632e418dc1b19
Reviewed-on: https://go-review.googlesource.com/22242
Reviewed-by: Michael Hudson-Doyle <michael.hudson@canonical.com>
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agodebug/pe: introduce Section.Relocs
Alex Brainman [Thu, 21 Apr 2016 01:44:05 +0000 (11:44 +1000)]
debug/pe: introduce Section.Relocs

cmd/link reads PE object files when building programs with cgo.
cmd/link accesses object relocations. Add new Section.Relocs that
provides similar functionality in debug/pe.

Updates #15345

Change-Id: I34de91b7f18cf1c9e4cdb3aedd685486a625ac92
Reviewed-on: https://go-review.googlesource.com/22332
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
9 years agocmd/compile: fix ssa/check/on build
Keith Randall [Thu, 21 Apr 2016 04:46:39 +0000 (21:46 -0700)]
cmd/compile: fix ssa/check/on build

Disable phielimValue from rewrite pass for now.

Change-Id: I9f3bb1f527b50bc7a21cc6b7cb89f6136efd81e8
Reviewed-on: https://go-review.googlesource.com/22335
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
9 years agoruntime: use type int to specify size for newarray
Martin Möhrmann [Wed, 20 Apr 2016 16:00:52 +0000 (18:00 +0200)]
runtime: use type int to specify size for newarray

Consistently use type int for the size argument of
runtime.newarray, runtime.reflect_unsafe_NewArray
and reflect.unsafe_NewArray.

Change-Id: Ic77bf2dde216c92ca8c49462f8eedc0385b6314e
Reviewed-on: https://go-review.googlesource.com/22311
Reviewed-by: Keith Randall <khr@golang.org>
Run-TryBot: Martin Möhrmann <martisch@uos.de>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/compile: fix dominator check in check()
Keith Randall [Thu, 21 Apr 2016 00:29:50 +0000 (17:29 -0700)]
cmd/compile: fix dominator check in check()

Ancestor comparison was the wrong way around, effectively
disabling the def-must-dominate-use check.

Update #15084

Change-Id: Ic56d674c5000569d2cc855bbb000a60eae517c7c
Reviewed-on: https://go-review.googlesource.com/22330
Run-TryBot: Keith Randall <khr@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
9 years agocmd/link: fix reordering of plt/rel
David Crawshaw [Wed, 20 Apr 2016 23:10:20 +0000 (19:10 -0400)]
cmd/link: fix reordering of plt/rel

For the Solaris and S/390 builders.

Change-Id: Id9a83e0df91e6d0df8488ec5e2a546ba8e2d800e
Reviewed-on: https://go-review.googlesource.com/22327
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Michael Hudson-Doyle <michael.hudson@canonical.com>
Reviewed-by: Michael Munday <munday@ca.ibm.com>
9 years agodebug/pe: update cstring documentation
Alex Brainman [Thu, 21 Apr 2016 00:42:25 +0000 (10:42 +1000)]
debug/pe: update cstring documentation

Updates #15345

Change-Id: If1fca1f6042571cb0ac689bbb3c294309dd6e7b4
Reviewed-on: https://go-review.googlesource.com/22331
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Alex Brainman <alex.brainman@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agodebug/pe: pretty section.go code
Alex Brainman [Wed, 20 Apr 2016 05:58:55 +0000 (15:58 +1000)]
debug/pe: pretty section.go code

Introduce (*SectionHeader32).fullName and add documentation comments.

Updates #15345

Change-Id: I8f3b8ab9492642d62e7aad010c91c68daea3f14b
Reviewed-on: https://go-review.googlesource.com/22301
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Alex Brainman <alex.brainman@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/compile: reject embedded unsafe.Pointer values
Tal Shprecher [Wed, 20 Apr 2016 21:05:48 +0000 (14:05 -0700)]
cmd/compile: reject embedded unsafe.Pointer values

Fixes #14729

Change-Id: Ied819aa7b23e25de30aa8cde049c97297b4cab11
Reviewed-on: https://go-review.googlesource.com/22325
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/compile: accept old and new import format for builtin declarations
Robert Griesemer [Wed, 20 Apr 2016 23:41:43 +0000 (16:41 -0700)]
cmd/compile: accept old and new import format for builtin declarations

Test with forceNewExport set to true (but continues to be disabled by
default for now).

Fixes #15322.

Change-Id: I3b893db2206cbb79e66339284f22f4a0b20bf137
Reviewed-on: https://go-review.googlesource.com/22328
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
9 years agocmd/link: process data symbols with slices
David Crawshaw [Mon, 18 Apr 2016 18:50:14 +0000 (14:50 -0400)]
cmd/link: process data symbols with slices

First (and largest single) step to switching cmd/link from linked
lists of symbols to slices.

Sort sections independently and concurrently.
This reduces jujud link times on linux/amd64 by ~4%.

Updates #15374

Change-Id: I452bc8f33081039468636502fe3c1cc8d6ed9efa
Reviewed-on: https://go-review.googlesource.com/22205
Reviewed-by: Michael Hudson-Doyle <michael.hudson@canonical.com>
9 years agocmd/link: cleanup Datablk debug printing
David Crawshaw [Wed, 20 Apr 2016 14:36:49 +0000 (10:36 -0400)]
cmd/link: cleanup Datablk debug printing

(Split out from CL 22205.)

Change-Id: I45838dda8ea8c451b4388b8aade2c209cde2c0e1
Reviewed-on: https://go-review.googlesource.com/22299
Reviewed-by: Michael Hudson-Doyle <michael.hudson@canonical.com>
Run-TryBot: Michael Hudson-Doyle <michael.hudson@canonical.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocrypto/md5: add assembly implementation on ppc64le
Lynn Boger [Wed, 20 Apr 2016 13:15:07 +0000 (08:15 -0500)]
crypto/md5: add assembly implementation on ppc64le

This change improves the performance of the block
function used within crypto/md5 on ppc64le.  The following
improvement was seen:

BenchmarkHash8Bytes              8.39         26.04        3.10x
BenchmarkHash1K                  99.41        407.84       4.10x
BenchmarkHash8K                  108.87       460.00       4.23x
BenchmarkHash8BytesUnaligned     8.39         25.80        3.08x
BenchmarkHash1KUnaligned         89.94        407.81       4.53x
BenchmarkHash8KUnaligned         96.57        459.22       4.76x

Fixes #15385

Change-Id: I8af5af089cc3e3740c33c662003d104de5fe1d1b
Reviewed-on: https://go-review.googlesource.com/22294
Run-TryBot: Michael Munday <munday@ca.ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
9 years agomath/big: more tests, documentation for Flot gob marshalling
Robert Griesemer [Wed, 20 Apr 2016 20:40:55 +0000 (13:40 -0700)]
math/big: more tests, documentation for Flot gob marshalling

Follow-up to https://golang.org/cl/21755.
This turned out to be a bit more than just a few nits
as originally expected in that CL.

1) The actual mantissa may be shorter than required for the
   given precision (because of trailing 0's): no need to
   allocate space for it (and transmit 0's). This can save
   a lot of space when the precision is high: E.g., for
   prec == 1000, 16 words or 128 bytes are required at the
   most, but if the actual number is short, it may be much
   less (for the test cases present, it's significantly less).

2) The actual mantissa may be longer than the number of
   words required for the given precision: make sure to
   not overflow when encoding in bytes.

3) Add more documentation.

4) Add more tests.

Change-Id: I9f40c408cfdd9183a8e81076d2f7d6c75e7a00e9
Reviewed-on: https://go-review.googlesource.com/22324
Reviewed-by: Alan Donovan <adonovan@google.com>