]> Cypherpunks repositories - gostls13.git/log
gostls13.git
9 years agomisc/cgo/testsanitizers: fix test of whether compiler option works
Ian Lance Taylor [Tue, 3 Nov 2015 15:14:22 +0000 (07:14 -0800)]
misc/cgo/testsanitizers: fix test of whether compiler option works

On older versions of GCC we need to pass a file name before GCC will
report an unrecognized option.

Fixes #13065.

Change-Id: I7ed34c01a006966a446059025f7d10235c649072
Reviewed-on: https://go-review.googlesource.com/16589
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Crawshaw <crawshaw@golang.org>
9 years agoruntime: remove dead code
Dmitry Vyukov [Tue, 3 Nov 2015 19:00:21 +0000 (20:00 +0100)]
runtime: remove dead code

runtime.free has long gone.

Change-Id: I058f69e6481b8fa008e1951c29724731a8a3d081
Reviewed-on: https://go-review.googlesource.com/16593
Reviewed-by: Austin Clements <austin@google.com>
Run-TryBot: Austin Clements <austin@google.com>

9 years agoruntime: cache two workbufs to reduce contention
Austin Clements [Fri, 16 Oct 2015 13:53:11 +0000 (09:53 -0400)]
runtime: cache two workbufs to reduce contention

Currently the gcWork abstraction caches a single work buffer. As a
result, if a worker is putting and getting pointers right at the
boundary of a work buffer, it can flap between work buffers and
(potentially significantly) increase contention on the global work
buffer lists.

This change modifies gcWork to instead cache two work buffers and
switch off between them. This introduces one buffers' worth of
hysteresis and eliminates the above performance worst case by
amortizing the cost of getting or putting a work buffer over at least
one buffers' worth of work.

In practice, it's difficult to trigger this worst case with reasonably
large work buffers. On the garbage benchmark, this reduces the max
writes/sec to the global work list from 32K to 25K and the median from
6K to 5K. However, if a workload were to trigger this worst case
behavior, it could significantly drive up this contention.

This has negligible effects on the go1 benchmarks and slightly speeds
up the garbage benchmark.

name              old time/op  new time/op  delta
XBenchGarbage-12  5.90ms ± 3%  5.83ms ± 4%  -1.18%  (p=0.011 n=18+18)

name                      old time/op    new time/op    delta
BinaryTree17-12              3.22s ± 4%     3.17s ± 3%  -1.57%  (p=0.009 n=19+20)
Fannkuch11-12                2.44s ± 1%     2.53s ± 4%  +3.78%  (p=0.000 n=18+19)
FmtFprintfEmpty-12          50.2ns ± 2%    50.5ns ± 5%    ~     (p=0.631 n=19+20)
FmtFprintfString-12          167ns ± 1%     166ns ± 1%    ~     (p=0.141 n=20+20)
FmtFprintfInt-12             162ns ± 1%     159ns ± 1%  -1.80%  (p=0.000 n=20+20)
FmtFprintfIntInt-12          277ns ± 2%     263ns ± 1%  -4.78%  (p=0.000 n=20+18)
FmtFprintfPrefixedInt-12     240ns ± 1%     232ns ± 2%  -3.25%  (p=0.000 n=20+20)
FmtFprintfFloat-12           311ns ± 1%     315ns ± 2%  +1.17%  (p=0.000 n=20+20)
FmtManyArgs-12              1.05µs ± 2%    1.03µs ± 2%  -1.72%  (p=0.000 n=20+20)
GobDecode-12                8.65ms ± 1%    8.71ms ± 2%  +0.68%  (p=0.001 n=19+20)
GobEncode-12                6.51ms ± 1%    6.54ms ± 1%  +0.42%  (p=0.047 n=20+19)
Gzip-12                      318ms ± 2%     315ms ± 2%  -1.20%  (p=0.000 n=19+19)
Gunzip-12                   42.2ms ± 2%    42.1ms ± 1%    ~     (p=0.667 n=20+19)
HTTPClientServer-12         62.5µs ± 1%    62.4µs ± 1%    ~     (p=0.110 n=20+18)
JSONEncode-12               16.8ms ± 1%    16.8ms ± 2%    ~     (p=0.569 n=19+20)
JSONDecode-12               60.8ms ± 2%    59.8ms ± 1%  -1.69%  (p=0.000 n=19+19)
Mandelbrot200-12            3.87ms ± 1%    3.85ms ± 0%  -0.61%  (p=0.001 n=20+17)
GoParse-12                  3.76ms ± 2%    3.76ms ± 1%    ~     (p=0.698 n=20+20)
RegexpMatchEasy0_32-12       100ns ± 2%     101ns ± 2%    ~     (p=0.065 n=19+20)
RegexpMatchEasy0_1K-12       342ns ± 2%     333ns ± 1%  -2.82%  (p=0.000 n=20+19)
RegexpMatchEasy1_32-12      83.3ns ± 2%    83.2ns ± 2%    ~     (p=0.692 n=20+19)
RegexpMatchEasy1_1K-12       498ns ± 2%     490ns ± 1%  -1.52%  (p=0.000 n=18+20)
RegexpMatchMedium_32-12      131ns ± 2%     131ns ± 2%    ~     (p=0.464 n=20+18)
RegexpMatchMedium_1K-12     39.3µs ± 2%    39.6µs ± 1%  +0.77%  (p=0.000 n=18+19)
RegexpMatchHard_32-12       2.04µs ± 2%    2.06µs ± 1%  +0.69%  (p=0.009 n=19+20)
RegexpMatchHard_1K-12       61.4µs ± 2%    62.1µs ± 1%  +1.21%  (p=0.000 n=19+20)
Revcomp-12                   534ms ± 1%     529ms ± 1%  -0.97%  (p=0.000 n=19+16)
Template-12                 70.4ms ± 2%    70.0ms ± 1%    ~     (p=0.070 n=19+19)
TimeParse-12                 359ns ± 3%     344ns ± 1%  -4.15%  (p=0.000 n=19+19)
TimeFormat-12                357ns ± 1%     361ns ± 2%  +1.05%  (p=0.002 n=20+20)
[Geo mean]                  62.4µs         62.0µs       -0.56%

name                      old speed      new speed      delta
GobDecode-12              88.7MB/s ± 1%  88.1MB/s ± 2%  -0.68%  (p=0.001 n=19+20)
GobEncode-12               118MB/s ± 1%   117MB/s ± 1%  -0.42%  (p=0.046 n=20+19)
Gzip-12                   60.9MB/s ± 2%  61.7MB/s ± 2%  +1.21%  (p=0.000 n=19+19)
Gunzip-12                  460MB/s ± 2%   461MB/s ± 1%    ~     (p=0.661 n=20+19)
JSONEncode-12              116MB/s ± 1%   115MB/s ± 2%    ~     (p=0.555 n=19+20)
JSONDecode-12             31.9MB/s ± 2%  32.5MB/s ± 1%  +1.72%  (p=0.000 n=19+19)
GoParse-12                15.4MB/s ± 2%  15.4MB/s ± 1%    ~     (p=0.653 n=20+20)
RegexpMatchEasy0_32-12     317MB/s ± 2%   315MB/s ± 2%    ~     (p=0.141 n=19+20)
RegexpMatchEasy0_1K-12    2.99GB/s ± 2%  3.07GB/s ± 1%  +2.86%  (p=0.000 n=20+19)
RegexpMatchEasy1_32-12     384MB/s ± 2%   385MB/s ± 2%    ~     (p=0.672 n=20+19)
RegexpMatchEasy1_1K-12    2.06GB/s ± 2%  2.09GB/s ± 1%  +1.54%  (p=0.000 n=18+20)
RegexpMatchMedium_32-12   7.62MB/s ± 2%  7.63MB/s ± 2%    ~     (p=0.800 n=20+18)
RegexpMatchMedium_1K-12   26.0MB/s ± 1%  25.8MB/s ± 1%  -0.77%  (p=0.000 n=18+19)
RegexpMatchHard_32-12     15.7MB/s ± 2%  15.6MB/s ± 1%  -0.69%  (p=0.010 n=19+20)
RegexpMatchHard_1K-12     16.7MB/s ± 2%  16.5MB/s ± 1%  -1.19%  (p=0.000 n=19+20)
Revcomp-12                 476MB/s ± 1%   481MB/s ± 1%  +0.97%  (p=0.000 n=19+16)
Template-12               27.6MB/s ± 2%  27.7MB/s ± 1%    ~     (p=0.071 n=19+19)
[Geo mean]                99.1MB/s       99.3MB/s       +0.27%

Change-Id: I68bcbf74ccb716cd5e844a554f67b679135105e6
Reviewed-on: https://go-review.googlesource.com/16042
Reviewed-by: Rick Hudson <rlh@golang.org>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agoruntime: fix finalization and profiling of tiny allocations
Dmitry Vyukov [Tue, 3 Nov 2015 11:19:15 +0000 (12:19 +0100)]
runtime: fix finalization and profiling of tiny allocations

Handling of special records for tiny allocations has two problems:
1. Once we queue a finalizer we mark the object. As the result any
   subsequent finalizers for the same object will not be queued
   during this GC cycle. If we have 16 finalizers setup (the worst case),
   finalization will take 16 GC cycles. This is what caused misbehave
   of tinyfin.go. The actual flakiness was caused by the fact that fing
   is asynchronous and don't always run before the check.
2. If a tiny block has both finalizer and profile specials,
   it is possible that we both queue finalizer, preserve the object live
   and free the profile record. As the result heap profile can be skewed.

Fix both issues by analyzing all special records for a single object at once.

Also, make tinyfin test stricter and remove reliance on real time.

Also, add a test for the problem 2. Currently heap profile missed about
a half of live memory.

Fixes #13100

Change-Id: I9ae4dc1c44893724138a4565ca5cae29f2e97544
Reviewed-on: https://go-review.googlesource.com/16591
Reviewed-by: Austin Clements <austin@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
Run-TryBot: Dmitry Vyukov <dvyukov@google.com>

9 years agostrings: add asm version of Index() for short strings on amd64
Ilya Tocar [Wed, 28 Oct 2015 15:05:05 +0000 (18:05 +0300)]
strings: add asm version of Index() for short strings on amd64

Currently we have special case for 1-byte strings,
This extends this to strings shorter than 32 bytes on amd64.
Results (broadwell):

name                 old time/op  new time/op  delta
IndexRune-4          57.4ns ± 0%  57.5ns ± 0%   +0.10%        (p=0.000 n=20+19)
IndexRuneFastPath-4  20.4ns ± 0%  20.4ns ± 0%     ~     (all samples are equal)
Index-4              21.0ns ± 0%  21.8ns ± 0%   +3.81%        (p=0.000 n=20+20)
LastIndex-4          7.07ns ± 1%  6.98ns ± 0%   -1.21%        (p=0.000 n=20+16)
IndexByte-4          18.3ns ± 0%  18.3ns ± 0%     ~     (all samples are equal)
IndexHard1-4         1.46ms ± 0%  0.39ms ± 0%  -73.06%        (p=0.000 n=16+16)
IndexHard2-4         1.46ms ± 0%  0.30ms ± 0%  -79.55%        (p=0.000 n=18+18)
IndexHard3-4         1.46ms ± 0%  0.66ms ± 0%  -54.68%        (p=0.000 n=19+19)
LastIndexHard1-4     1.46ms ± 0%  1.46ms ± 0%   -0.01%        (p=0.036 n=18+20)
LastIndexHard2-4     1.46ms ± 0%  1.46ms ± 0%     ~           (p=0.588 n=19+19)
LastIndexHard3-4     1.46ms ± 0%  1.46ms ± 0%     ~           (p=0.283 n=17+20)
IndexTorture-4       11.1µs ± 0%  11.1µs ± 0%   +0.01%        (p=0.000 n=18+17)

Change-Id: I892781549f558f698be4e41f9f568e3d0611efb5
Reviewed-on: https://go-review.googlesource.com/16430
Reviewed-by: Keith Randall <khr@golang.org>
Run-TryBot: Ilya Tocar <ilya.tocar@intel.com>

9 years agoruntime: enlarge GC work buffer size
Austin Clements [Fri, 2 Oct 2015 13:51:12 +0000 (09:51 -0400)]
runtime: enlarge GC work buffer size

Currently the GC work buffers are only 256 bytes and hence can record
only 24 64-bit pointer. They were reduced from 4K in commits db7fd1c
and a15818f as a way to minimize the amount of work the per-P workbuf
caches could "hide" from the mark phase and carry in to the mark
termination phase. However, this approach wasn't very robust and we
later added a "mark 2" phase to address this problem head-on.

Because of mark 2, there's now no benefit to having very small work
buffers. But there are plenty of downsides: small work buffers
increase contention on the work lists, increase the frequency and
hence net overhead of acquiring and releasing work buffers, and
somewhat increase memory overhead of the GC.

This commit expands work buffers back to 4K (504 64-bit pointers).
This reduces the rate of writes to work.full in the garbage benchmark
from a peak of ~780,000 writes/sec to a peak of ~32,000 writes/sec.

This has negligible effect on the go1 benchmarks. It slightly slows
down the garbage benchmark.

name              old time/op  new time/op  delta
XBenchGarbage-12  5.37ms ± 5%  5.60ms ± 2%  +4.37%  (p=0.000 n=20+20)

Change-Id: Ic9cc28e7a125d23d9faf4f5e690fb8aa9bcdfb28
Reviewed-on: https://go-review.googlesource.com/15893
Reviewed-by: Rick Hudson <rlh@golang.org>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agoruntime: make assists preemptible
Austin Clements [Thu, 15 Oct 2015 21:58:17 +0000 (17:58 -0400)]
runtime: make assists preemptible

Currently, assists are non-preemptible, which means a heavily
assisting G can block other Gs from running. At the beginning of a GC
cycle, it can also delay scang, which will spin until the assist is
done. Since scanning is currently done sequentially, this can
seriously extend the length of the scan phase.

Fix this by making assists preemptible. Since the assist holds work
buffers and runs on the system stack, this must be done cooperatively:
we make gcDrainN return on preemption, and make the assist return from
the system stack and voluntarily Gosched.

This is prerequisite to enlarging the work buffers. Without this
change, the delays and spinning in scang increase significantly.

This has no effect on the go1 benchmarks.

name              old time/op  new time/op  delta
XBenchGarbage-12  5.72ms ± 4%  5.37ms ± 5%  -6.11%  (p=0.000 n=20+20)

Change-Id: I829e732a0f23b126da633516a1a9ec1a508fdbf1
Reviewed-on: https://go-review.googlesource.com/15894
Reviewed-by: Rick Hudson <rlh@golang.org>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agoruntime: replace assist sleep loop with park/ready
Austin Clements [Thu, 15 Oct 2015 01:31:33 +0000 (21:31 -0400)]
runtime: replace assist sleep loop with park/ready

GC assists must block until the assist can be satisfied (either
through stealing credit or doing work) or the GC cycle ends.
Currently, this is implemented as a retry loop with a 100 µs delay.
This obviously isn't ideal, as it wastes CPU and delays mutator
execution. It also has the somewhat peculiar downside that sleeping a
G requires allocation, and this requires working around recursive
allocation.

Replace this timed delay with a proper scheduling queue. When an
assist can't be satisfied immediately, it adds the allocating G to a
queue and parks it. Any time background scan credit is flushed, it
consults this queue, directly satisfies the debt of queued assists,
and wakes up satisfied assists before flushing any remaining credit to
the background credit pool.

No effect on the go1 benchmarks. Slightly speeds up the garbage
benchmark.

name              old time/op  new time/op  delta
XBenchGarbage-12  5.81ms ± 1%  5.72ms ± 4%  -1.65%  (p=0.011 n=20+20)

Updates #12041.

Change-Id: I8ee3b6274dd097b12b10a8030796a958a4b0e7b7
Reviewed-on: https://go-review.googlesource.com/15890
Reviewed-by: Rick Hudson <rlh@golang.org>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agoruntime: change p.runq from []*g to []guintptr
Austin Clements [Mon, 2 Nov 2015 21:59:39 +0000 (16:59 -0500)]
runtime: change p.runq from []*g to []guintptr

This eliminates many write barriers in the scheduler code that are
unnecessary and will interfere with upcoming changes where the garbage
collector will have to invoke run queue functions in contexts that
must not have write barriers.

Change-Id: I702d0ac99cfd00ffff406e7362917db6a43e7e55
Reviewed-on: https://go-review.googlesource.com/16556
Reviewed-by: Russ Cox <rsc@golang.org>
Run-TryBot: Austin Clements <austin@google.com>

9 years agomisc/ios: fix an error when getenv encounters unset variable
Michael Matloob [Mon, 2 Nov 2015 22:37:31 +0000 (17:37 -0500)]
misc/ios: fix an error when getenv encounters unset variable

The error message should indicate the name of the unset variable,
rather than the value. The value will alwayse be empty.

Change-Id: I6f6c165074dfce857b6523703a890d205423cd28
Reviewed-on: https://go-review.googlesource.com/16555
Reviewed-by: David Crawshaw <crawshaw@golang.org>
9 years agotest: use go:noinline consistently
Todd Neal [Fri, 30 Oct 2015 02:45:19 +0000 (21:45 -0500)]
test: use go:noinline consistently

Replace various implementations of inlining prevention with
"go:noinline"

Change-Id: Iac90895c3a62d6f4b7a6c72e11e165d15a0abfa4
Reviewed-on: https://go-review.googlesource.com/16510
Reviewed-by: Keith Randall <khr@golang.org>
Run-TryBot: Todd Neal <todd@tneal.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/link: remove duplicate symtab entry for global functions
Michael Hudson-Doyle [Mon, 2 Nov 2015 20:36:53 +0000 (09:36 +1300)]
cmd/link: remove duplicate symtab entry for global functions

golang.org/cl/16436 added a local symbol for every global function, but also
added a duplicate entry for the global symbol. Surprisingly this hasn't caused
any noticeable problems, but it's still wrong.

Change-Id: Icd3906760f8aaf7bef31ffd4f2d866d73d36dc2c
Reviewed-on: https://go-review.googlesource.com/16581
Reviewed-by: Ian Lance Taylor <iant@golang.org>
9 years agobytes: speed up Compare() on amd64
Ilya Tocar [Wed, 28 Oct 2015 20:20:26 +0000 (23:20 +0300)]
bytes: speed up Compare() on amd64

Use AVX2 if available.
Results (haswell), below:

name                           old time/op    new time/op     delta
BytesCompare1-6                  11.4ns ± 0%     11.4ns ± 0%     ~     (all samples are equal)
BytesCompare2-6                  11.4ns ± 0%     11.4ns ± 0%     ~     (all samples are equal)
BytesCompare4-6                  11.4ns ± 0%     11.4ns ± 0%     ~     (all samples are equal)
BytesCompare8-6                  9.29ns ± 2%     8.76ns ± 0%   -5.72%        (p=0.000 n=16+17)
BytesCompare16-6                 9.29ns ± 2%     9.20ns ± 0%   -1.02%        (p=0.000 n=20+16)
BytesCompare32-6                 11.4ns ± 1%     11.4ns ± 0%     ~           (p=0.191 n=20+20)
BytesCompare64-6                 14.4ns ± 0%     13.1ns ± 0%   -8.68%        (p=0.000 n=20+20)
BytesCompare128-6                20.2ns ± 0%     18.5ns ± 0%   -8.27%        (p=0.000 n=16+20)
BytesCompare256-6                29.3ns ± 0%     24.5ns ± 0%  -16.38%        (p=0.000 n=16+16)
BytesCompare512-6                46.8ns ± 0%     37.1ns ± 0%  -20.78%        (p=0.000 n=18+16)
BytesCompare1024-6               82.9ns ± 0%     62.3ns ± 0%  -24.86%        (p=0.000 n=20+14)
BytesCompare2048-6                155ns ± 0%      112ns ± 0%  -27.74%        (p=0.000 n=20+20)
CompareBytesEqual-6              10.1ns ± 1%     10.0ns ± 1%     ~           (p=0.527 n=20+20)
CompareBytesToNil-6              10.0ns ± 2%      9.4ns ± 0%   -6.57%        (p=0.000 n=20+17)
CompareBytesEmpty-6              8.76ns ± 0%     8.76ns ± 0%     ~     (all samples are equal)
CompareBytesIdentical-6          8.76ns ± 0%     8.76ns ± 0%     ~     (all samples are equal)
CompareBytesSameLength-6         10.6ns ± 1%     10.6ns ± 1%     ~           (p=0.240 n=20+20)
CompareBytesDifferentLength-6    10.6ns ± 0%     10.6ns ± 1%     ~           (p=1.000 n=20+20)
CompareBytesBigUnaligned-6        132±s ± 1%      105±s ± 1%  -20.61%        (p=0.000 n=20+18)
CompareBytesBig-6                 125±s ± 1%      105±s ± 1%  -16.31%        (p=0.000 n=20+20)
CompareBytesBigIdentical-6       8.13ns ± 0%     8.13ns ± 0%     ~     (all samples are equal)

name                           old speed      new speed       delta
CompareBytesBigUnaligned-6     7.94GB/s ± 1%  10.01GB/s ± 1%  +25.96%        (p=0.000 n=20+18)
CompareBytesBig-6              8.38GB/s ± 1%  10.01GB/s ± 1%  +19.48%        (p=0.000 n=20+20)
CompareBytesBigIdentical-6      129TB/s ± 0%    129TB/s ± 0%   +0.01%        (p=0.003 n=17+19)

Change-Id: I820f31bab4582dd4204b146bb077c0d2f24cd8f5
Reviewed-on: https://go-review.googlesource.com/16434
Run-TryBot: Ilya Tocar <ilya.tocar@intel.com>
Reviewed-by: Klaus Post <klauspost@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
9 years agocmd/link: concurrent obj copy for external linking
David Crawshaw [Wed, 28 Oct 2015 18:41:58 +0000 (14:41 -0400)]
cmd/link: concurrent obj copy for external linking

Change-Id: I630ae29ecb39252642883398cc51d49133c6f3d7
Reviewed-on: https://go-review.googlesource.com/16451
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agotest: update tinyfin test
Brad Fitzpatrick [Mon, 2 Nov 2015 15:46:44 +0000 (07:46 -0800)]
test: update tinyfin test

* use new(int32) to be pedantic about documented SetFinalizer rules:
  "The argument x must be a pointer to an object allocated by calling
  new or by taking the address of a composite literal"

* remove the amd64-only restriction. The GC is fully precise everywhere
  now, even on 32-bit. (keep the gccgo restriction, though)

* remove a data race (perhaps the actual bug) and use atomic.LoadInt32
  for the final check. The race detector is now happy, too.

Updates #13100

Change-Id: I8d05c0ac4f046af9ba05701ad709c57984b34893
Reviewed-on: https://go-review.googlesource.com/16535
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agotest: add gcc6798, valid code that caused gccgo to crash
Ian Lance Taylor [Sun, 1 Nov 2015 20:30:54 +0000 (12:30 -0800)]
test: add gcc6798, valid code that caused gccgo to crash

Change-Id: I288bd3091ea81db7b616747cbec8958a31d98b7e
Reviewed-on: https://go-review.googlesource.com/16532
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
9 years agocmd/go, runtime: define GOBUILDMODE_shared rather than shared when dynamically linking
Michael Hudson-Doyle [Thu, 29 Oct 2015 22:48:43 +0000 (11:48 +1300)]
cmd/go, runtime: define GOBUILDMODE_shared rather than shared when dynamically linking

To avoid collisions with what existing code may already be doing.

Change-Id: Ice639440aafc0724714c25333d90a49954372230
Reviewed-on: https://go-review.googlesource.com/16503
Reviewed-by: Ian Lance Taylor <iant@golang.org>
9 years agonet: make Dial, Listen{,Packet} for TCP/UDP with invalid port fail
Mikio Hara [Tue, 21 Jul 2015 03:51:01 +0000 (12:51 +0900)]
net: make Dial, Listen{,Packet} for TCP/UDP with invalid port fail

This change makes Dial, Listen and ListenPacket with invalid port fail
whatever GODEBUG=netdns is.

Please be informed that cgoLookupPort with an out of range literal
number may return either the lower or upper bound value, 0 or 65535,
with no error on some platform.

Fixes #11715.

Change-Id: I43f9c4fb5526d1bf50b97698e0eb39d29fd74c35
Reviewed-on: https://go-review.googlesource.com/12447
Reviewed-by: Ian Lance Taylor <iant@golang.org>
9 years agopath/filepath: fix EvalSymlinks(".") on windows
Alex Brainman [Mon, 21 Sep 2015 06:12:24 +0000 (16:12 +1000)]
path/filepath: fix EvalSymlinks(".") on windows

Also new tests added. So, perhaps, this CL corrects
even more broken EvalSymlinks behaviour.

Fixes #12451

Change-Id: I81b9d92bab74bcb8eca6db6633546982fe5cec87
Reviewed-on: https://go-review.googlesource.com/16192
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/link: support new 386/amd64 relocations
Ian Lance Taylor [Sat, 31 Oct 2015 21:44:48 +0000 (14:44 -0700)]
cmd/link: support new 386/amd64 relocations

The GNU binutils recently picked up support for new 386/amd64
relocations.  Add support for them in the Go linker when doing an
internal link.

The 386 relocation R_386_GOT32X was proposed in
https://groups.google.com/forum/#!topic/ia32-abi/GbJJskkid4I .  It can
be treated as identical to the R_386_GOT32 relocation.

The amd64 relocations R_X86_64_GOTPCRELX and R_X86_64_REX_GOTPCRELX were
proposed in
https://groups.google.com/forum/#!topic/x86-64-abi/n9AWHogmVY0 .  They
can both be treated as identical to the R_X86_64_GOTPCREL relocation.

The purpose of the new relocations is to permit additional linker
relaxations in some cases.  We do not attempt to support those cases.

While we're at it, remove the unused and in some cases out of date
_COUNT names from ld/elf.go.

Fixes #13114.

Change-Id: I34ef07f6fcd00cdd2996038ecf46bb77a49e968b
Reviewed-on: https://go-review.googlesource.com/16529
Reviewed-by: Minux Ma <minux@golang.org>
9 years agogo/types: fix TypeString(nil, nil)
Matthew Dempsky [Thu, 29 Oct 2015 01:55:30 +0000 (18:55 -0700)]
go/types: fix TypeString(nil, nil)

The code is meant to return "<nil>", but because of a make([]Type, 8)
call that should be make([]Type, 0, 8), the nil Type happens to
already appear in the array.

Change-Id: I2db140046e52f27db1b0ac84bde2b6680677dd95
Reviewed-on: https://go-review.googlesource.com/16464
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
9 years agomath: fix bad shift in Expm1
Matthew Dempsky [Fri, 30 Oct 2015 02:16:20 +0000 (19:16 -0700)]
math: fix bad shift in Expm1

Noticed by cmd/vet.

Expected values array produced by Python instead of Keisan because:

1) Keisan's website calculator is painfully difficult to copy/paste
values into and out of, and

2) after tediously computing e^(vf[i] * 10) - 1 via Keisan I
discovered that Keisan computing vf[i]*10 in a higher precision was
giving substantially different output values.

Also, testing uses "close" instead of "veryclose" because 386's
assembly implementation produces values for some of the test cases
that fail "veryclose".  Curiously, Expm1(vf[i]*10) is identical to
Exp(vf[i]*10)-1 on 386, whereas with the portable implementation
they're only "veryclose".

Investigating these questions is left to someone else.  I just wanted
to fix the cmd/vet warning.

Fixes #13101.

Change-Id: Ica8f6c267d01aa4cc31f53593e95812746942fbc
Reviewed-on: https://go-review.googlesource.com/16505
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Klaus Post <klauspost@gmail.com>
Reviewed-by: Robert Griesemer <gri@golang.org>
9 years agoruntime: perform mark 2 root re-scanning in GC workers
Austin Clements [Mon, 19 Oct 2015 18:15:16 +0000 (14:15 -0400)]
runtime: perform mark 2 root re-scanning in GC workers

This moves another root scanning task out of the GC coordinator and
parallelizes it on the GC workers.

This has negligible effect on the go1 benchmarks and the garbage
benchmark.

name              old time/op  new time/op  delta
XBenchGarbage-12  5.24ms ± 1%  5.26ms ± 1%  +0.30%  (p=0.007 n=18+17)

name                      old time/op    new time/op    delta
BinaryTree17-12              3.20s ± 5%     3.21s ± 5%    ~     (p=0.264 n=20+18)
Fannkuch11-12                2.46s ± 1%     2.54s ± 2%  +3.09%  (p=0.000 n=18+20)
FmtFprintfEmpty-12          49.9ns ± 4%    50.0ns ± 5%    ~     (p=0.356 n=20+20)
FmtFprintfString-12          170ns ± 1%     170ns ± 2%    ~     (p=0.815 n=19+20)
FmtFprintfInt-12             160ns ± 1%     159ns ± 1%  -0.63%  (p=0.003 n=18+19)
FmtFprintfIntInt-12          270ns ± 1%     267ns ± 1%  -1.00%  (p=0.000 n=19+18)
FmtFprintfPrefixedInt-12     238ns ± 1%     232ns ± 1%  -2.28%  (p=0.000 n=19+19)
FmtFprintfFloat-12           310ns ± 2%     313ns ± 2%  +0.93%  (p=0.000 n=19+19)
FmtManyArgs-12              1.06µs ± 1%    1.04µs ± 1%  -1.93%  (p=0.000 n=20+19)
GobDecode-12                8.63ms ± 1%    8.70ms ± 1%  +0.81%  (p=0.001 n=20+19)
GobEncode-12                6.52ms ± 1%    6.56ms ± 1%  +0.66%  (p=0.000 n=20+19)
Gzip-12                      318ms ± 1%     319ms ± 1%    ~     (p=0.405 n=17+18)
Gunzip-12                   42.1ms ± 2%    42.0ms ± 1%    ~     (p=0.771 n=20+19)
HTTPClientServer-12         62.6µs ± 1%    62.9µs ± 1%  +0.41%  (p=0.038 n=20+20)
JSONEncode-12               16.9ms ± 1%    16.9ms ± 1%    ~     (p=0.077 n=18+20)
JSONDecode-12               60.7ms ± 1%    62.3ms ± 1%  +2.73%  (p=0.000 n=20+20)
Mandelbrot200-12            3.86ms ± 1%    3.85ms ± 1%    ~     (p=0.084 n=19+20)
GoParse-12                  3.75ms ± 2%    3.73ms ± 1%    ~     (p=0.107 n=20+19)
RegexpMatchEasy0_32-12       100ns ± 2%     101ns ± 2%  +0.97%  (p=0.001 n=20+19)
RegexpMatchEasy0_1K-12       342ns ± 2%     332ns ± 2%  -2.86%  (p=0.000 n=19+19)
RegexpMatchEasy1_32-12      83.2ns ± 2%    82.8ns ± 2%    ~     (p=0.108 n=19+20)
RegexpMatchEasy1_1K-12       495ns ± 2%     490ns ± 2%  -1.04%  (p=0.000 n=18+19)
RegexpMatchMedium_32-12      130ns ± 2%     131ns ± 2%    ~     (p=0.291 n=20+20)
RegexpMatchMedium_1K-12     39.3µs ± 1%    39.9µs ± 1%  +1.54%  (p=0.000 n=18+20)
RegexpMatchHard_32-12       2.02µs ± 1%    2.05µs ± 2%  +1.19%  (p=0.000 n=19+19)
RegexpMatchHard_1K-12       60.9µs ± 1%    61.5µs ± 1%  +0.99%  (p=0.000 n=18+18)
Revcomp-12                   535ms ± 1%     531ms ± 1%  -0.82%  (p=0.000 n=17+17)
Template-12                 73.0ms ± 1%    74.1ms ± 1%  +1.47%  (p=0.000 n=20+20)
TimeParse-12                 356ns ± 2%     348ns ± 1%  -2.30%  (p=0.000 n=20+20)
TimeFormat-12                347ns ± 1%     353ns ± 1%  +1.68%  (p=0.000 n=19+20)
[Geo mean]                  62.3µs         62.4µs       +0.12%

name                      old speed      new speed      delta
GobDecode-12              88.9MB/s ± 1%  88.2MB/s ± 1%  -0.81%  (p=0.001 n=20+19)
GobEncode-12               118MB/s ± 1%   117MB/s ± 1%  -0.66%  (p=0.000 n=20+19)
Gzip-12                   60.9MB/s ± 1%  60.8MB/s ± 1%    ~     (p=0.409 n=17+18)
Gunzip-12                  461MB/s ± 2%   462MB/s ± 1%    ~     (p=0.765 n=20+19)
JSONEncode-12              115MB/s ± 1%   115MB/s ± 1%    ~     (p=0.078 n=18+20)
JSONDecode-12             32.0MB/s ± 1%  31.1MB/s ± 1%  -2.65%  (p=0.000 n=20+20)
GoParse-12                15.5MB/s ± 2%  15.5MB/s ± 1%    ~     (p=0.111 n=20+19)
RegexpMatchEasy0_32-12     318MB/s ± 2%   314MB/s ± 2%  -1.27%  (p=0.000 n=20+19)
RegexpMatchEasy0_1K-12    2.99GB/s ± 1%  3.08GB/s ± 2%  +2.94%  (p=0.000 n=19+19)
RegexpMatchEasy1_32-12     385MB/s ± 2%   386MB/s ± 2%    ~     (p=0.105 n=19+20)
RegexpMatchEasy1_1K-12    2.07GB/s ± 1%  2.09GB/s ± 2%  +1.06%  (p=0.000 n=18+19)
RegexpMatchMedium_32-12   7.64MB/s ± 2%  7.61MB/s ± 1%    ~     (p=0.179 n=20+20)
RegexpMatchMedium_1K-12   26.1MB/s ± 1%  25.7MB/s ± 1%  -1.52%  (p=0.000 n=18+20)
RegexpMatchHard_32-12     15.8MB/s ± 1%  15.6MB/s ± 2%  -1.18%  (p=0.000 n=19+19)
RegexpMatchHard_1K-12     16.8MB/s ± 2%  16.6MB/s ± 1%  -0.90%  (p=0.000 n=19+18)
Revcomp-12                 475MB/s ± 1%   479MB/s ± 1%  +0.83%  (p=0.000 n=17+17)
Template-12               26.6MB/s ± 1%  26.2MB/s ± 1%  -1.45%  (p=0.000 n=20+20)
[Geo mean]                99.0MB/s       98.7MB/s       -0.32%

Change-Id: I6ea44d7a59aaa6851c64695277ab65645ff9d32e
Reviewed-on: https://go-review.googlesource.com/16070
Reviewed-by: Rick Hudson <rlh@golang.org>
Run-TryBot: Austin Clements <austin@google.com>

9 years agoruntime: perform concurrent scan in GC workers
Austin Clements [Mon, 19 Oct 2015 17:46:32 +0000 (13:46 -0400)]
runtime: perform concurrent scan in GC workers

Currently the concurrent root scan is performed in its entirety by the
GC coordinator before entering concurrent mark (which enables GC
workers). This scan is done sequentially, which can prolong the scan
phase, delay the mark phase, and means that the scan phase does not
obey the 25% CPU goal. Furthermore, there's no need to complete the
root scan before starting marking (in fact, we already allow GC
assists to happen during the scan phase), so this acts as an
unnecessary barrier between root scanning and marking.

This change shifts the root scan work out of the GC coordinator and in
to the GC workers. The coordinator simply sets up the scan state and
enqueues the right number of root scan jobs. The GC workers then drain
the root scan jobs prior to draining heap scan jobs.

This parallelizes the root scan process, makes it obey the 25% CPU
goal, and effectively eliminates root scanning as an isolated phase,
allowing the system to smoothly transition from root scanning to heap
marking. This also eliminates a major non-STW responsibility of the GC
coordinator, which will make it easier to switch to a decentralized
state machine. Finally, it puts us in a good position to perform root
scanning in assists as well, which will help satisfy assists at the
beginning of the GC cycle.

This is mostly straightforward. One tricky aspect is that we have to
deal with preemption deadlock: where two non-preemptible gorountines
are trying to preempt each other to perform a stack scan. Given the
context where this happens, the only instance of this is two
background workers trying to scan each other. We avoid this by simply
not scanning the stacks of background workers during the concurrent
phase; this is safe because we'll scan them during mark termination
(and their stacks are *very* small and should not contain any new
pointers).

This change also switches the root marking during mark termination to
use the same gcDrain-based code path as concurrent mark. This
shouldn't affect performance because STW root marking was already
parallel and tasks switched to heap marking immediately when no more
root marking tasks were available. However, it simplifies the code and
unifies these code paths.

This has negligible effect on the go1 benchmarks. It slightly slows
down the garbage benchmark, possibly by making GC run slightly more
frequently.

name              old time/op  new time/op  delta
XBenchGarbage-12  5.10ms ± 1%  5.24ms ± 1%  +2.87%  (p=0.000 n=18+18)

name                      old time/op    new time/op    delta
BinaryTree17-12              3.25s ± 3%     3.20s ± 5%  -1.57%  (p=0.013 n=20+20)
Fannkuch11-12                2.45s ± 1%     2.46s ± 1%  +0.38%  (p=0.019 n=20+18)
FmtFprintfEmpty-12          49.7ns ± 3%    49.9ns ± 4%    ~     (p=0.851 n=19+20)
FmtFprintfString-12          170ns ± 2%     170ns ± 1%    ~     (p=0.775 n=20+19)
FmtFprintfInt-12             161ns ± 1%     160ns ± 1%  -0.78%  (p=0.000 n=19+18)
FmtFprintfIntInt-12          267ns ± 1%     270ns ± 1%  +1.04%  (p=0.000 n=19+19)
FmtFprintfPrefixedInt-12     238ns ± 2%     238ns ± 1%    ~     (p=0.133 n=18+19)
FmtFprintfFloat-12           311ns ± 1%     310ns ± 2%  -0.35%  (p=0.023 n=20+19)
FmtManyArgs-12              1.08µs ± 1%    1.06µs ± 1%  -2.31%  (p=0.000 n=20+20)
GobDecode-12                8.65ms ± 1%    8.63ms ± 1%    ~     (p=0.377 n=18+20)
GobEncode-12                6.49ms ± 1%    6.52ms ± 1%  +0.37%  (p=0.015 n=20+20)
Gzip-12                      319ms ± 3%     318ms ± 1%    ~     (p=0.975 n=19+17)
Gunzip-12                   41.9ms ± 1%    42.1ms ± 2%  +0.65%  (p=0.004 n=19+20)
HTTPClientServer-12         61.7µs ± 1%    62.6µs ± 1%  +1.40%  (p=0.000 n=18+20)
JSONEncode-12               16.8ms ± 1%    16.9ms ± 1%    ~     (p=0.239 n=20+18)
JSONDecode-12               58.4ms ± 1%    60.7ms ± 1%  +3.85%  (p=0.000 n=19+20)
Mandelbrot200-12            3.86ms ± 0%    3.86ms ± 1%    ~     (p=0.092 n=18+19)
GoParse-12                  3.75ms ± 2%    3.75ms ± 2%    ~     (p=0.708 n=19+20)
RegexpMatchEasy0_32-12       100ns ± 1%     100ns ± 2%  +0.60%  (p=0.010 n=17+20)
RegexpMatchEasy0_1K-12       341ns ± 1%     342ns ± 2%    ~     (p=0.203 n=20+19)
RegexpMatchEasy1_32-12      82.5ns ± 2%    83.2ns ± 2%  +0.83%  (p=0.007 n=19+19)
RegexpMatchEasy1_1K-12       495ns ± 1%     495ns ± 2%    ~     (p=0.970 n=19+18)
RegexpMatchMedium_32-12      130ns ± 2%     130ns ± 2%  +0.59%  (p=0.039 n=19+20)
RegexpMatchMedium_1K-12     39.2µs ± 1%    39.3µs ± 1%    ~     (p=0.214 n=18+18)
RegexpMatchHard_32-12       2.03µs ± 2%    2.02µs ± 1%    ~     (p=0.166 n=18+19)
RegexpMatchHard_1K-12       61.0µs ± 1%    60.9µs ± 1%    ~     (p=0.169 n=20+18)
Revcomp-12                   533ms ± 1%     535ms ± 1%    ~     (p=0.071 n=19+17)
Template-12                 68.1ms ± 2%    73.0ms ± 1%  +7.26%  (p=0.000 n=19+20)
TimeParse-12                 355ns ± 2%     356ns ± 2%    ~     (p=0.530 n=19+20)
TimeFormat-12                357ns ± 2%     347ns ± 1%  -2.59%  (p=0.000 n=20+19)
[Geo mean]                  62.1µs         62.3µs       +0.31%

name                      old speed      new speed      delta
GobDecode-12              88.7MB/s ± 1%  88.9MB/s ± 1%    ~     (p=0.377 n=18+20)
GobEncode-12               118MB/s ± 1%   118MB/s ± 1%  -0.37%  (p=0.015 n=20+20)
Gzip-12                   60.9MB/s ± 3%  60.9MB/s ± 1%    ~     (p=0.944 n=19+17)
Gunzip-12                  464MB/s ± 1%   461MB/s ± 2%  -0.64%  (p=0.004 n=19+20)
JSONEncode-12              115MB/s ± 1%   115MB/s ± 1%    ~     (p=0.236 n=20+18)
JSONDecode-12             33.2MB/s ± 1%  32.0MB/s ± 1%  -3.71%  (p=0.000 n=19+20)
GoParse-12                15.5MB/s ± 2%  15.5MB/s ± 2%    ~     (p=0.702 n=19+20)
RegexpMatchEasy0_32-12     320MB/s ± 1%   318MB/s ± 2%    ~     (p=0.094 n=18+20)
RegexpMatchEasy0_1K-12    3.00GB/s ± 1%  2.99GB/s ± 1%    ~     (p=0.194 n=20+19)
RegexpMatchEasy1_32-12     388MB/s ± 2%   385MB/s ± 2%  -0.83%  (p=0.008 n=19+19)
RegexpMatchEasy1_1K-12    2.07GB/s ± 1%  2.07GB/s ± 1%    ~     (p=0.964 n=19+18)
RegexpMatchMedium_32-12   7.68MB/s ± 1%  7.64MB/s ± 2%  -0.57%  (p=0.020 n=19+20)
RegexpMatchMedium_1K-12   26.1MB/s ± 1%  26.1MB/s ± 1%    ~     (p=0.211 n=18+18)
RegexpMatchHard_32-12     15.8MB/s ± 1%  15.8MB/s ± 1%    ~     (p=0.180 n=18+19)
RegexpMatchHard_1K-12     16.8MB/s ± 1%  16.8MB/s ± 2%    ~     (p=0.236 n=20+19)
Revcomp-12                 477MB/s ± 1%   475MB/s ± 1%    ~     (p=0.071 n=19+17)
Template-12               28.5MB/s ± 2%  26.6MB/s ± 1%  -6.77%  (p=0.000 n=19+20)
[Geo mean]                 100MB/s       99.0MB/s       -0.82%

Change-Id: I875bf6ceb306d1ee2f470cabf88aa6ede27c47a0
Reviewed-on: https://go-review.googlesource.com/16059
Reviewed-by: Rick Hudson <rlh@golang.org>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agoruntime: consolidate "out of GC work" checks
Austin Clements [Mon, 19 Oct 2015 17:35:25 +0000 (13:35 -0400)]
runtime: consolidate "out of GC work" checks

We already have gcMarkWorkAvailable, but the check for GC mark work is
open-coded in several places. Generalize gcMarkWorkAvailable slightly
and replace these open-coded checks with calls to gcMarkWorkAvailable.

In addition to cleaning up the code, this puts us in a better position
to make this check slightly more complicated.

Change-Id: I1b29883300ecd82a1bf6be193e9b4ee96582a860
Reviewed-on: https://go-review.googlesource.com/16058
Reviewed-by: Rick Hudson <rlh@golang.org>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agotest: disable flaky heapsampling test for now
Brad Fitzpatrick [Fri, 30 Oct 2015 20:33:28 +0000 (13:33 -0700)]
test: disable flaky heapsampling test for now

Flaky tests do more harm than good.

Updates #13098

Change-Id: I179ed810b49bbb96c8df462bfa20b70231b26772
Reviewed-on: https://go-review.googlesource.com/16521
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/compile/internal: named types for Etype and Op in struct Node
Marvin Stenger [Thu, 24 Sep 2015 21:21:18 +0000 (23:21 +0200)]
cmd/compile/internal: named types for Etype and Op in struct Node

Type Op is enfored now.
Type EType will need further CLs.
Added TODOs where Node.EType is used as a union type.
The TODOs have the format `TODO(marvin): Fix Node.EType union type.`.

Furthermore:
-The flag of Econv function in fmt.go is removed, since unused.
-Some cleaning along the way, e.g. declare vars first when getting initialized.

Passes go build -toolexec 'toolstash -cmp' -a std.

Fixes #11846

Change-Id: I908b955d5a78a195604970983fb9194bd9e9260b
Reviewed-on: https://go-review.googlesource.com/14956
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Marvin Stenger <marvin.stenger94@gmail.com>
9 years agoos: reduce allocations in Readdir on unix
Taru Karttunen [Wed, 28 Oct 2015 10:58:58 +0000 (12:58 +0200)]
os: reduce allocations in Readdir on unix

Include syscall.Stat_t on unix to the
unexported fileStat structure rather than
accessing it though an interface.

Additionally add a benchmark for Readdir
(and Readdirnames).

Tested on linux, freebsd, netbsd, openbsd
darwin, solaris, does not touch windows
stuff. Does not change the API, as
discussed on golang-dev.

E.g. on linux/amd64 with a directory of 65 files:

benchmark              old ns/op     new ns/op     delta
BenchmarkReaddir-4     67774         66225         -2.29%

benchmark              old allocs     new allocs     delta
BenchmarkReaddir-4     334            269            -19.46%

benchmark              old bytes     new bytes     delta
BenchmarkReaddir-4     25208         24168         -4.13%

Change-Id: I44ef72a04ad7055523a980f29aa11122040ae8fe
Reviewed-on: https://go-review.googlesource.com/16423
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agodoc/go1.6.txt: mention possible GOTRACEBACK change
Russ Cox [Fri, 30 Oct 2015 15:08:25 +0000 (11:08 -0400)]
doc/go1.6.txt: mention possible GOTRACEBACK change

For CL 16512, #12366, #13107.

Change-Id: I0ed1bb9597ac3db3fa35a037c304060d5a7e6d51
Reviewed-on: https://go-review.googlesource.com/16513
Reviewed-by: Austin Clements <austin@google.com>
9 years agoruntime: introduce GOTRACEBACK=single, now the default
Russ Cox [Fri, 30 Oct 2015 15:03:02 +0000 (11:03 -0400)]
runtime: introduce GOTRACEBACK=single, now the default

Abandon (but still support) the old numbering system.

GOTRACEBACK=none is old 0
GOTRACEBACK=single is the new behavior
GOTRACEBACK=all is old 1
GOTRACEBACK=system is old 2
GOTRACEBACK=crash is unchanged

See doc comment change in runtime1.go for details.

Filed #13107 to decide whether to change default back to GOTRACEBACK=all for Go 1.6 release.
If you run into programs where printing only the current goroutine omits
needed information, please add details in a comment on that issue.

Fixes #12366.

Change-Id: I82ca8b99b5d86dceb3f7102d38d2659d45dbe0db
Reviewed-on: https://go-review.googlesource.com/16512
Reviewed-by: Austin Clements <austin@google.com>
9 years agonet/http: do not listen on public network during HTTP/2 test
Russ Cox [Fri, 30 Oct 2015 17:50:19 +0000 (13:50 -0400)]
net/http: do not listen on public network during HTTP/2 test

Avoids Mac firewall box.

Change-Id: I000e421fa9639612d636b6fa4baf905459c5aeb2
Reviewed-on: https://go-review.googlesource.com/16514
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/compile/internal/gc: fix complex literal overflow
Robert Griesemer [Thu, 29 Oct 2015 21:46:47 +0000 (14:46 -0700)]
cmd/compile/internal/gc: fix complex literal overflow

Change-Id: Ibb15282a2baeb58439b085d70b82797d8c71de36
Reviewed-on: https://go-review.googlesource.com/16502
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
9 years agosyscall: correct parameter name of Setgid on linux
Marvin Stenger [Fri, 30 Oct 2015 17:16:47 +0000 (18:16 +0100)]
syscall: correct parameter name of Setgid on linux

Change parameter name from uid to gid, to fix an obvious copy-paste
error.

Change-Id: Iba13a45c87fde9625b82976a7d7901af4b705230
Reviewed-on: https://go-review.googlesource.com/16474
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
9 years agocmd/link/internal/ld: fix extra arguments Errorf
Quentin Perez [Fri, 30 Oct 2015 15:25:29 +0000 (16:25 +0100)]
cmd/link/internal/ld: fix extra arguments Errorf

lib.go:1309
lib.go:1313
lib.go:1317
lib.go:1325

Change-Id: Ic3d3282d836c28f8bdeb5ce52fb7c018450871b9
Reviewed-on: https://go-review.googlesource.com/16473
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agonet/http: update x/net/http2
Brad Fitzpatrick [Fri, 30 Oct 2015 02:31:15 +0000 (02:31 +0000)]
net/http: update x/net/http2

Updated bundle of golang.org/x/net/http2 to rev c95266

Change-Id: Id6df66a8f4eadd6a100f359fd6cac981653dafa6
Reviewed-on: https://go-review.googlesource.com/16448
Reviewed-by: Andrew Gerrand <adg@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agodoc: some go1.6.txt additions
Brad Fitzpatrick [Fri, 30 Oct 2015 02:20:29 +0000 (02:20 +0000)]
doc: some go1.6.txt additions

Many remain.

Change-Id: Icfe190c145a34f8aaa4d78e853dbc708962ba5ce
Reviewed-on: https://go-review.googlesource.com/16447
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
9 years agocmd/compile: add support for a go:noinline directive
Todd Neal [Tue, 25 Aug 2015 00:45:59 +0000 (19:45 -0500)]
cmd/compile: add support for a go:noinline directive

Some tests need to disable inlining of a function.  It's currently done
in one of a few ways (adding a function call, an empty switch, or a
defer).  Add support for a less fragile 'go:noinline' directive that
prevents inlining.

Fixes #12312

Change-Id: Ife444e13361b4a927709d81aa41e448f32eec8d4
Reviewed-on: https://go-review.googlesource.com/13911
Run-TryBot: Todd Neal <todd@tneal.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Josh Bleecher Snyder <josharian@gmail.com>
9 years agocmd/internal/obj, cmd/link, runtime: handle TLS more like a platform linker on ppc64
Michael Hudson-Doyle [Thu, 29 Oct 2015 21:13:13 +0000 (10:13 +1300)]
cmd/internal/obj, cmd/link, runtime: handle TLS more like a platform linker on ppc64

On ppc64x, the thread pointer, held in R13, points 0x7000 bytes past where
thread-local storage begins (presumably to maximize the amount of storage that
can be accessed with a 16-bit signed displacement). The relocations used to
indicate thread-local storage to the platform linker account for this, so to be
able to support external linking we need to change things so the linker applies
this offset instead of the runtime assembly.

Change-Id: I2556c249ab2d802cae62c44b2b4c5b44787d7059
Reviewed-on: https://go-review.googlesource.com/14233
Reviewed-by: Russ Cox <rsc@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
9 years agomath: fix typo and braino in my earlier commit
Brad Fitzpatrick [Thu, 29 Oct 2015 20:02:28 +0000 (13:02 -0700)]
math: fix typo and braino in my earlier commit

The bug number was a typo, and I forgot to switch the implementation
back to if statements after the change from Float64bits in the first
patchset back to branching.

if statements can currently be inlined, but switch cannot (#13071)

Change-Id: I81d0cf64bda69186c3d747a07047f6a694f8fa70
Reviewed-on: https://go-review.googlesource.com/16446
Reviewed-by: Robert Griesemer <gri@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/internal/obj, cmd/link: simplify ppc64 archreloc now that the original value...
Michael Hudson-Doyle [Mon, 31 Aug 2015 01:51:37 +0000 (13:51 +1200)]
cmd/internal/obj, cmd/link: simplify ppc64 archreloc now that the original value is passed to it

And get rid of the stupid game of encoding the instruction in the addend.

Change-Id: Ib4de7515196cbc1e63b4261b01931cf02a44c1e6
Reviewed-on: https://go-review.googlesource.com/14055
Reviewed-by: Russ Cox <rsc@golang.org>
9 years agonet/http: extra documentation for Redirect and RedirectHandler
Taru Karttunen [Fri, 16 Oct 2015 10:26:20 +0000 (13:26 +0300)]
net/http: extra documentation for Redirect and RedirectHandler

Errors with http.Redirect and http.StatusOk seem
to occur from time to time on the irc channel.
This change adds documentation suggesting
to use one of the 3xx codes and not StatusOk
with Redirect.

Change-Id: I6b900a8eb868265fbbb846ee6a53e426d90a727d
Reviewed-on: https://go-review.googlesource.com/15980
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
9 years agomath: replace assembly implementations of Abs with pure Go version
Brad Fitzpatrick [Thu, 29 Oct 2015 15:13:58 +0000 (15:13 +0000)]
math: replace assembly implementations of Abs with pure Go version

The compiler can do a fine job, and can also inline it.

From Jeremy Jackins's observation and rsc's recommendation in thread:

"Pure Go math.Abs outperforms assembly version"
https://groups.google.com/forum/#!topic/golang-dev/nP5mWvwAXZo

Updates #13095

Change-Id: I3066f8eaa327bb403173b29791cc8661d7c0532c
Reviewed-on: https://go-review.googlesource.com/16444
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/compile, cmd/go, cmd/link: enable -buildmode=shared and related flags on linux...
Michael Hudson-Doyle [Thu, 29 Oct 2015 07:24:29 +0000 (20:24 +1300)]
cmd/compile, cmd/go, cmd/link: enable -buildmode=shared and related flags on linux/arm64

Change-Id: Ibddbbf6f4a5bd336a8b234d40fad0fcea574cd6e
Reviewed-on: https://go-review.googlesource.com/13994
Reviewed-by: Ian Lance Taylor <iant@golang.org>
9 years agocmd/link: always resolve functions locally when linking dynamically
Michael Hudson-Doyle [Wed, 28 Oct 2015 23:17:43 +0000 (12:17 +1300)]
cmd/link: always resolve functions locally when linking dynamically

When dynamically linking, we want references to functions defined
in this module to always be to the function object, not to the
PLT. We force this by writing an additional local symbol for
every global function symbol and making all relocations against
the global symbol refer to this local symbol instead. This is
approximately equivalent to the ELF linker -Bsymbolic-functions
option, but that is buggy on several platforms.

Change-Id: Ie6983eb4d1947f8543736fd349f9a90df3cce91a
Reviewed-on: https://go-review.googlesource.com/16436
Reviewed-by: Ian Lance Taylor <iant@golang.org>
9 years agocmd/go: buildmode=c-shared support for linux/arm64
David Crawshaw [Wed, 28 Oct 2015 21:57:30 +0000 (17:57 -0400)]
cmd/go: buildmode=c-shared support for linux/arm64

Change-Id: Ic826dc25b5203b2e9fc253d6fe997e4b41de3789
Reviewed-on: https://go-review.googlesource.com/16453
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agoos/signal: don't expect to be notified for os.Kill
Burcu Dogan [Thu, 29 Oct 2015 07:02:02 +0000 (00:02 -0700)]
os/signal: don't expect to be notified for os.Kill

os.Kill cannot be caught on Unix systems.
The example gives the false impression that it can.

Fixes #13080.

Change-Id: I3b9e6f38a38f437a463c5b869ae84a0d3fd23f72
Reviewed-on: https://go-review.googlesource.com/16467
Reviewed-by: Andrew Gerrand <adg@golang.org>
9 years agocmd/go: use t.Fatalf for formatted errors
Dominik Honnef [Wed, 28 Oct 2015 06:53:54 +0000 (07:53 +0100)]
cmd/go: use t.Fatalf for formatted errors

Change-Id: I657588ccc4a87e075d15acb9f4cd1e417bbd7960
Reviewed-on: https://go-review.googlesource.com/16421
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agoruntime/cgo: export _cgo_reginit on ppc64x
Michael Hudson-Doyle [Thu, 3 Sep 2015 20:31:03 +0000 (08:31 +1200)]
runtime/cgo: export _cgo_reginit on ppc64x

This is needed to make external linking work.

Change-Id: I4cf7edb4ea318849cab92a697952f8745eed40c4
Reviewed-on: https://go-review.googlesource.com/14237
Reviewed-by: Ian Lance Taylor <iant@golang.org>
9 years agocmd/go: implicitly include math in a shared library on arm
Michael Hudson-Doyle [Thu, 3 Sep 2015 00:33:06 +0000 (12:33 +1200)]
cmd/go: implicitly include math in a shared library on arm

In the same manner in which runtime/cgo is included on other architectures.

Change-Id: I90a5ad8585248b2566d763d33994a600508d89cb
Reviewed-on: https://go-review.googlesource.com/14221
Reviewed-by: Ian Lance Taylor <iant@golang.org>
9 years agocmd/compile, cmd/go: enable -buildmode=shared and related flags on linux/arm
Michael Hudson-Doyle [Wed, 28 Oct 2015 23:45:34 +0000 (12:45 +1300)]
cmd/compile, cmd/go: enable -buildmode=shared and related flags on linux/arm

Change-Id: I20840632771a250fb279df64d394135994482af8
Reviewed-on: https://go-review.googlesource.com/14186
Reviewed-by: David Crawshaw <crawshaw@golang.org>
9 years agoruntime: add missing word in comment
Ian Lance Taylor [Wed, 28 Oct 2015 16:57:04 +0000 (09:57 -0700)]
runtime: add missing word in comment

Change-Id: Iffe27445e35ec071cf0920a05c81b8a97a3ed712
Reviewed-on: https://go-review.googlesource.com/16431
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agoruntime: c-shared entrypoint for linux/arm64
David Crawshaw [Tue, 27 Oct 2015 23:47:54 +0000 (19:47 -0400)]
runtime: c-shared entrypoint for linux/arm64

Change-Id: I7dab124842f5209097a8d5a802fcbdde650654fa
Reviewed-on: https://go-review.googlesource.com/16395
Reviewed-by: Ian Lance Taylor <iant@golang.org>
9 years agocmd/go: allow -buildmode=pie for android/amd64.
Hyang-Ah Hana Kim [Tue, 20 Oct 2015 04:24:03 +0000 (00:24 -0400)]
cmd/go: allow -buildmode=pie for android/amd64.

For golang/go#10743 golang/go#10807

Change-Id: I35dedb52e2b47fe7ffc655d01f22ac05fe830bde
Reviewed-on: https://go-review.googlesource.com/16396
Reviewed-by: David Crawshaw <crawshaw@golang.org>
9 years agocmd/go: enable android/amd64 build.
Hyang-Ah Hana Kim [Fri, 16 Oct 2015 18:19:13 +0000 (14:19 -0400)]
cmd/go: enable android/amd64 build.

For golang/go#10743

Change-Id: Iec047821147a0e28edebf875fefe25993785702b
Reviewed-on: https://go-review.googlesource.com/15994
Reviewed-by: David Crawshaw <crawshaw@golang.org>
9 years agoruntime, cmd: TLS setup for android/amd64.
Hyang-Ah Hana Kim [Fri, 16 Oct 2015 18:04:29 +0000 (14:04 -0400)]
runtime, cmd: TLS setup for android/amd64.

Android linker does not handle TLS for us. We set up the TLS slot
for g, as darwin/386,amd64 handle instead. This is disgusting and
fragile. We will eventually fix this ugly hack by taking advantage
of the recent TLS IE model implementation. (Instead of referencing
an GOT entry, make the code sequence look into the TLS variable that
holds the offset.)

The TLS slot for g in android/amd64 assumes a fixed offset from %fs.
See runtime/cgo/gcc_android_amd64.c for details.

For golang/go#10743

Change-Id: I1a3fc207946c665515f79026a56ea19134ede2dd
Reviewed-on: https://go-review.googlesource.com/15991
Reviewed-by: David Crawshaw <crawshaw@golang.org>
9 years agocmd/internal/obj, cmd/link: support inital-exec TLS on arm64
Michael Hudson-Doyle [Wed, 28 Oct 2015 02:27:51 +0000 (15:27 +1300)]
cmd/internal/obj, cmd/link: support inital-exec TLS on arm64

Change-Id: Iaf9159a68fa395245bc20ccb4a2a377f89371a7e
Reviewed-on: https://go-review.googlesource.com/13996
Reviewed-by: David Crawshaw <crawshaw@golang.org>
9 years agocmd/internal/obj, cmd/link, runtime: native-ish support for tls on arm64
Michael Hudson-Doyle [Tue, 11 Aug 2015 02:10:03 +0000 (14:10 +1200)]
cmd/internal/obj, cmd/link, runtime: native-ish support for tls on arm64

Fixes #10560

Change-Id: Iedffd9c236c4fbb386c3afc52c5a1457f96ef122
Reviewed-on: https://go-review.googlesource.com/13991
Reviewed-by: David Crawshaw <crawshaw@golang.org>
9 years agonet/http/httputil: add hook for managing io.Copy buffers per request
Brad Fitzpatrick [Tue, 28 Apr 2015 01:10:20 +0000 (18:10 -0700)]
net/http/httputil: add hook for managing io.Copy buffers per request

Adds ReverseProxy.BufferPool for users with sensitive allocation
requirements. Permits avoiding 32 KB of io.Copy garbage per request.

Fixes #10277

Change-Id: I5dfd58fa70a363ead4be56405e507df90d871719
Reviewed-on: https://go-review.googlesource.com/9399
Reviewed-by: Keith Randall <khr@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agoruntime: don't use FP when calling nextSample in the Plan 9 sighandler
David du Colombier [Wed, 28 Oct 2015 05:44:26 +0000 (06:44 +0100)]
runtime: don't use FP when calling nextSample in the Plan 9 sighandler

In the Go signal handler on Plan 9, when a signal with
the _SigThrow flag is received, we call startpanic before
printing the stack trace.

The startpanic function calls systemstack which calls
startpanic_m. In the startpanic_m function, we call
allocmcache to allocate _g_.m.mcache. The problem is
that allocmcache calls nextSample, which does a floating
point operation to return a sampling point for heap profiling.

However, Plan 9 doesn't support floating point in the
signal handler.

This change adds a new function nextSampleNoFP, only
called when in the Plan 9 signal handler, which is
similar to nextSample, but avoids floating point.

Change-Id: Iaa30437aa0f7c8c84d40afbab7567ad3bd5ea2de
Reviewed-on: https://go-review.googlesource.com/16307
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
9 years agocmd/link: remove unused field LSym.Hash
Michael Hudson-Doyle [Wed, 28 Oct 2015 03:35:22 +0000 (16:35 +1300)]
cmd/link: remove unused field LSym.Hash

Change-Id: Idc5c4a69919a8ed9d76d4a9cfd9827fb5c59dd11
Reviewed-on: https://go-review.googlesource.com/16389
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agotext/template: remove redundant indirect
Nodir Turakulov [Wed, 14 Oct 2015 04:34:47 +0000 (21:34 -0700)]
text/template: remove redundant indirect

Change-Id: I8cc9783fd044bed48347824dcf973c61c78275a5
Reviewed-on: https://go-review.googlesource.com/15833
Reviewed-by: Andrew Gerrand <adg@golang.org>
9 years agoruntime: invoke vsyscall helper via TCB when dynamic linking on linux/386
Michael Hudson-Doyle [Tue, 27 Oct 2015 03:40:30 +0000 (16:40 +1300)]
runtime: invoke vsyscall helper via TCB when dynamic linking on linux/386

The dynamic linker on linux/386 stores the address of the vsyscall helper at a
fixed offset from the %gs register on linux/386 for easy access from PIC code.

Change-Id: I635305cfecceef2289985d62e676e16810ed6b94
Reviewed-on: https://go-review.googlesource.com/16346
Reviewed-by: Ian Lance Taylor <iant@golang.org>
9 years agocmd/compile, cmd/go, cmd/link: enable -buildmode=shared and related flags on linux/386
Michael Hudson-Doyle [Tue, 27 Oct 2015 03:39:00 +0000 (16:39 +1300)]
cmd/compile, cmd/go, cmd/link: enable -buildmode=shared and related flags on linux/386

Change-Id: If3417135ca474468a480b08cf46334fda28f79b4
Reviewed-on: https://go-review.googlesource.com/16345
Reviewed-by: Ian Lance Taylor <iant@golang.org>
9 years agonet/http/httptest: close conns in StateNew on Server close
Brad Fitzpatrick [Tue, 27 Oct 2015 18:30:20 +0000 (18:30 +0000)]
net/http/httptest: close conns in StateNew on Server close

This part got dropped when we were debating between two solutions
in https://golang.org/cl/15151

Fixes #13032

Change-Id: I820b94f6c0c102ccf9342abf957328ea01f49a26
Reviewed-on: https://go-review.googlesource.com/16313
Reviewed-by: Austin Clements <austin@google.com>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agoreflect: Comment fix
kargakis [Tue, 27 Oct 2015 15:25:08 +0000 (16:25 +0100)]
reflect: Comment fix

Change-Id: I86cdd5c1d7b6f76d3474d180e75ea0c732241080
Reviewed-on: https://go-review.googlesource.com/16309
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
9 years agocmd/internal/obj/x86: remove REGTMP
Michael Hudson-Doyle [Tue, 27 Oct 2015 03:30:32 +0000 (16:30 +1300)]
cmd/internal/obj/x86: remove REGTMP

Nothing uses this.

Change-Id: Ibc13066940bd2ea5c74d955a67f9dc531bef2758
Reviewed-on: https://go-review.googlesource.com/16344
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
9 years agoruntime: eliminate some unnecessary uintptr conversions
Matthew Dempsky [Tue, 27 Oct 2015 00:53:22 +0000 (17:53 -0700)]
runtime: eliminate some unnecessary uintptr conversions

arena_{start,used,end} are already uintptr, so no need to convert them
to uintptr, much less to convert them to unsafe.Pointer and then to
uintptr.  No binary change to pkg/linux_amd64/runtime.a.

Change-Id: Ia4232ed2a724c44fde7eba403c5fe8e6dccaa879
Reviewed-on: https://go-review.googlesource.com/16339
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>

9 years agocmd/compile/internal/gc: introduce type for untyped constant kinds
Robert Griesemer [Mon, 26 Oct 2015 23:00:59 +0000 (16:00 -0700)]
cmd/compile/internal/gc: introduce type for untyped constant kinds

Change-Id: Ia34b6dd099d07d5e1d4bffe775a20fa92705fdb0
Reviewed-on: https://go-review.googlesource.com/16335
Run-TryBot: Robert Griesemer <gri@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
9 years agocmd/compile/internal/gc: introduce type for decl contexts/storage classes
Robert Griesemer [Mon, 26 Oct 2015 21:57:36 +0000 (14:57 -0700)]
cmd/compile/internal/gc: introduce type for decl contexts/storage classes

Change-Id: I956e27fa07f16060b8f41b986d991c36557f7c12
Reviewed-on: https://go-review.googlesource.com/16332
Reviewed-by: Keith Randall <khr@golang.org>
9 years agosyscall: define common notes on Plan 9
David du Colombier [Sat, 24 Oct 2015 11:28:00 +0000 (13:28 +0200)]
syscall: define common notes on Plan 9

There is no signal list on Plan 9, since notes
are strings. However, some programs expect
signals to be defined in the syscall package.
Hence, we define a list of the most common notes.

Updates #11975.

Change-Id: I852e14fd98777c9595a406e04125be1cbebed0fb
Reviewed-on: https://go-review.googlesource.com/16301
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: David du Colombier <0intro@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agoruntime: handle abort note on Plan 9
David du Colombier [Sat, 24 Oct 2015 16:14:51 +0000 (18:14 +0200)]
runtime: handle abort note on Plan 9

Implement an abort note on Plan 9, as an
equivalent of the SIGABRT signal on other
operating systems.

Updates #11975.

Change-Id: I010c9b10f2fbd2471aacd1d073368d975a2f0592
Reviewed-on: https://go-review.googlesource.com/16300
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: David du Colombier <0intro@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agoruntime: fix tiny allocator
Matthew Dempsky [Mon, 26 Oct 2015 19:38:47 +0000 (12:38 -0700)]
runtime: fix tiny allocator

When a new tiny block is allocated because we're allocating an object
that won't fit into the current block, mallocgc saves the new block if
it has more space leftover than the old block.  However, the logic for
this was subtly broken in golang.org/cl/2814, resulting in never
saving (or consequently reusing) a tiny block.

Change-Id: Ib5f6769451fb82877ddeefe75dfe79ed4a04fd40
Reviewed-on: https://go-review.googlesource.com/16330
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/internal/gc: escape binary export data so it contains no '$'
Robert Griesemer [Fri, 23 Oct 2015 23:01:09 +0000 (16:01 -0700)]
cmd/compile/internal/gc: escape binary export data so it contains no '$'

Necessary to ensure that subsequent tools can continue to find
then end of the export data section simply by searching for "$$".

Adjusted gcimporter used by go/types accordingly.

Also, fixed a bug in gcimporter related to reading export data
in debug format.

Change-Id: Iaea4ed05edd8a5bab28ebe5b19a4740f5e537d35
Reviewed-on: https://go-review.googlesource.com/16283
Reviewed-by: Chris Manghane <cmang@golang.org>
9 years agoruntime: partition data and BSS root marking
Austin Clements [Fri, 16 Oct 2015 20:52:26 +0000 (16:52 -0400)]
runtime: partition data and BSS root marking

Currently data and BSS root marking are each a single markroot job.
This makes them difficult to load balance, which can draw out mark
termination time if they are large.

Fix this by splitting both in to 256K chunks. While we're putting in
the infrastructure for dynamic roots, we also replace the fixed
sharding of the span roots with sharding in to fixed sizes. In
addition to helping balance root marking, this also paves the way to
parallelizing concurrent scan and to letting assists help with root
marking.

Updates #10345. This fixes the data and BSS aspects of that bug; it
does not partition scanning of large heap objects.

This has negligible effect on either the go1 benchmarks or the garbage
benchmark:

name              old time/op  new time/op  delta
XBenchGarbage-12  4.90ms ± 1%  4.91ms ± 2%   ~     (p=0.058 n=17+16)

name                      old time/op    new time/op    delta
BinaryTree17-12              3.11s ± 4%     3.12s ± 4%    ~     (p=0.512 n=20+20)
Fannkuch11-12                2.53s ± 2%     2.47s ± 2%  -2.28%  (p=0.000 n=20+18)
FmtFprintfEmpty-12          49.1ns ± 1%    50.0ns ± 4%  +1.68%  (p=0.008 n=18+20)
FmtFprintfString-12          170ns ± 0%     172ns ± 1%  +1.05%  (p=0.000 n=14+19)
FmtFprintfInt-12             174ns ± 1%     162ns ± 1%  -6.81%  (p=0.000 n=18+17)
FmtFprintfIntInt-12          284ns ± 1%     277ns ± 1%  -2.42%  (p=0.000 n=20+19)
FmtFprintfPrefixedInt-12     252ns ± 1%     244ns ± 1%  -2.84%  (p=0.000 n=18+20)
FmtFprintfFloat-12           317ns ± 0%     311ns ± 0%  -1.95%  (p=0.000 n=19+18)
FmtManyArgs-12              1.08µs ± 1%    1.11µs ± 1%  +3.43%  (p=0.000 n=18+19)
GobDecode-12                8.56ms ± 1%    8.61ms ± 1%  +0.50%  (p=0.020 n=20+20)
GobEncode-12                6.58ms ± 1%    6.57ms ± 1%    ~     (p=0.792 n=20+19)
Gzip-12                      317ms ± 3%     317ms ± 2%    ~     (p=0.840 n=19+19)
Gunzip-12                   41.6ms ± 0%    41.6ms ± 0%  +0.07%  (p=0.027 n=18+15)
HTTPClientServer-12         62.2µs ± 1%    62.3µs ± 1%    ~     (p=0.283 n=19+20)
JSONEncode-12               16.5ms ± 2%    16.5ms ± 1%    ~     (p=0.857 n=20+19)
JSONDecode-12               58.5ms ± 1%    61.3ms ± 1%  +4.67%  (p=0.000 n=18+17)
Mandelbrot200-12            3.84ms ± 0%    3.84ms ± 0%    ~     (p=0.259 n=17+17)
GoParse-12                  3.70ms ± 2%    3.74ms ± 2%  +0.96%  (p=0.009 n=19+20)
RegexpMatchEasy0_32-12       100ns ± 1%     100ns ± 0%  +0.31%  (p=0.040 n=19+15)
RegexpMatchEasy0_1K-12       340ns ± 1%     340ns ± 1%    ~     (p=0.411 n=17+19)
RegexpMatchEasy1_32-12      82.7ns ± 2%    82.3ns ± 1%    ~     (p=0.456 n=20+19)
RegexpMatchEasy1_1K-12       498ns ± 2%     495ns ± 0%    ~     (p=0.108 n=19+17)
RegexpMatchMedium_32-12      130ns ± 1%     130ns ± 2%    ~     (p=0.405 n=18+19)
RegexpMatchMedium_1K-12     39.4µs ± 2%    39.1µs ± 1%  -0.64%  (p=0.002 n=20+19)
RegexpMatchHard_32-12       2.03µs ± 2%    2.02µs ± 0%    ~     (p=0.561 n=20+17)
RegexpMatchHard_1K-12       61.1µs ± 2%    60.8µs ± 1%    ~     (p=0.615 n=19+18)
Revcomp-12                   532ms ± 2%     531ms ± 1%    ~     (p=0.470 n=19+19)
Template-12                 68.5ms ± 1%    69.1ms ± 1%  +0.87%  (p=0.000 n=17+17)
TimeParse-12                 344ns ± 2%     344ns ± 1%  +0.25%  (p=0.032 n=19+18)
TimeFormat-12                347ns ± 1%     362ns ± 1%  +4.27%  (p=0.000 n=17+19)
[Geo mean]                  62.3µs         62.3µs       -0.04%

name                      old speed      new speed      delta
GobDecode-12              89.6MB/s ± 1%  89.2MB/s ± 1%  -0.50%  (p=0.019 n=20+20)
GobEncode-12               117MB/s ± 1%   117MB/s ± 1%    ~     (p=0.797 n=20+19)
Gzip-12                   61.3MB/s ± 3%  61.2MB/s ± 2%    ~     (p=0.834 n=19+19)
Gunzip-12                  467MB/s ± 0%   466MB/s ± 0%  -0.07%  (p=0.027 n=18+15)
JSONEncode-12              117MB/s ± 2%   117MB/s ± 1%    ~     (p=0.851 n=20+19)
JSONDecode-12             33.2MB/s ± 1%  31.7MB/s ± 1%  -4.47%  (p=0.000 n=18+17)
GoParse-12                15.6MB/s ± 2%  15.5MB/s ± 2%  -0.95%  (p=0.008 n=19+20)
RegexpMatchEasy0_32-12     321MB/s ± 2%   320MB/s ± 1%  -0.57%  (p=0.002 n=17+17)
RegexpMatchEasy0_1K-12    3.01GB/s ± 1%  3.01GB/s ± 1%    ~     (p=0.132 n=17+18)
RegexpMatchEasy1_32-12     387MB/s ± 2%   389MB/s ± 1%    ~     (p=0.423 n=20+19)
RegexpMatchEasy1_1K-12    2.05GB/s ± 2%  2.06GB/s ± 0%    ~     (p=0.129 n=19+17)
RegexpMatchMedium_32-12   7.64MB/s ± 1%  7.66MB/s ± 1%    ~     (p=0.258 n=18+19)
RegexpMatchMedium_1K-12   26.0MB/s ± 2%  26.2MB/s ± 1%  +0.64%  (p=0.002 n=20+19)
RegexpMatchHard_32-12     15.7MB/s ± 2%  15.8MB/s ± 1%    ~     (p=0.510 n=20+17)
RegexpMatchHard_1K-12     16.8MB/s ± 2%  16.8MB/s ± 1%    ~     (p=0.603 n=19+18)
Revcomp-12                 477MB/s ± 2%   479MB/s ± 1%    ~     (p=0.470 n=19+19)
Template-12               28.3MB/s ± 1%  28.1MB/s ± 1%  -0.85%  (p=0.000 n=17+17)
[Geo mean]                 100MB/s        100MB/s       -0.26%

Change-Id: Ib0bfe0145675ce88c5a8791752f7486ac98805b4
Reviewed-on: https://go-review.googlesource.com/16043
Reviewed-by: Rick Hudson <rlh@golang.org>
9 years agoandroidtest.bash: set GOARM=7
David Crawshaw [Mon, 26 Oct 2015 15:15:09 +0000 (11:15 -0400)]
androidtest.bash: set GOARM=7

It's the only ARM version we have ever supported on android.
(Not setting it caused some builder timeouts.)

Change-Id: I26061434252ff2a236bb31d95787a1c582d24b3f
Reviewed-on: https://go-review.googlesource.com/16295
Reviewed-by: Hyang-Ah Hana Kim <hyangah@gmail.com>
9 years agoruntime: use a 64kb system stack on arm
David Crawshaw [Tue, 15 Sep 2015 16:22:46 +0000 (12:22 -0400)]
runtime: use a 64kb system stack on arm

I went looking for an arm system whose stacks are by default smaller
than 64KB. In fact the smallest common linux target I could find was
Android, which like iOS uses 1MB stacks.

Fixes #11873

Change-Id: Ieeb66ad095b3da18d47ba21360ea75152a4107c6
Reviewed-on: https://go-review.googlesource.com/14602
Reviewed-by: Michael Hudson-Doyle <michael.hudson@canonical.com>
Reviewed-by: Minux Ma <minux@golang.org>
9 years agocmd/compile/internal/gc: make embedded unexported structs RO
Marcel van Lohuizen [Mon, 31 Aug 2015 20:24:07 +0000 (22:24 +0200)]
cmd/compile/internal/gc: make embedded unexported structs RO

gc will need to be rebuild.

Package that assume f.PkgPath != nil means a field is unexported and
must be ignored must be revised to check for
f.PkgPath != nil && !f.Anonymous,
so that they do try to walk into the embedded fields to look for
exported fields contained within.

Closes #12367, fixes #7363, fixes #11007, and fixes #7247.

Change-Id: I16402ee21ccfede80f277f84b3995cf26e97433d
Reviewed-on: https://go-review.googlesource.com/14085
Reviewed-by: Russ Cox <rsc@golang.org>
9 years agoencoding/xml: check for exported fields in embedded structs
Marcel van Lohuizen [Fri, 28 Aug 2015 08:21:45 +0000 (10:21 +0200)]
encoding/xml: check for exported fields in embedded structs

Addresses issue #12367.

Must be checked in before CL 14010.

Change-Id: I4523a1de112ed02371504e27882659bce8028a9f
Reviewed-on: https://go-review.googlesource.com/14012
Reviewed-by: Russ Cox <rsc@golang.org>
9 years agoencoding/json: check for exported fields in embedded structs
Marcel van Lohuizen [Fri, 28 Aug 2015 08:17:05 +0000 (10:17 +0200)]
encoding/json: check for exported fields in embedded structs

Addresses issue #12367.

Must be checked in before CL 14010.

Change-Id: I7233c3a62d4f55d0ac7e8a87df5fc4ee7beb7207
Reviewed-on: https://go-review.googlesource.com/14011
Reviewed-by: Russ Cox <rsc@golang.org>
9 years agounicode/utf8: added benchmarks
Marcel van Lohuizen [Wed, 9 Sep 2015 08:43:06 +0000 (10:43 +0200)]
unicode/utf8: added benchmarks

Cover some functions that weren't benched before and add InString
variants if the underlying implementation is different.

Note: compare (Valid|RuneCount)InString* to their (Valid|RuneCount)*
counterparts. It shows, somewhat unexpectedly, that ranging over
a string is *much* slower than using calls to DecodeRune.

Results:
In order to avoid a discrepancy in measuring the performance
of core we could leave the names of the string-based measurements
unchanged and suffix the added alternatives with Bytes.

Compared to old:
BenchmarkRuneCountTenASCIIChars-8        44.3          12.4          -72.01%
BenchmarkRuneCountTenJapaneseChars-8     167           67.1          -59.82%
BenchmarkEncodeASCIIRune-8               3.37          3.44          +2.08%
BenchmarkEncodeJapaneseRune-8            7.19          7.24          +0.70%
BenchmarkDecodeASCIIRune-8               5.41          5.53          +2.22%
BenchmarkDecodeJapaneseRune-8            8.17          8.41          +2.94%

All benchmarks:
BenchmarkRuneCountTenASCIIChars-8            100000000         12.4 ns/op
BenchmarkRuneCountTenJapaneseChars-8         20000000         67.1 ns/op
BenchmarkRuneCountInStringTenASCIIChars-8    30000000         44.5 ns/op
BenchmarkRuneCountInStringTenJapaneseChars-8 10000000        165 ns/op
BenchmarkValidTenASCIIChars-8                100000000         12.5 ns/op
BenchmarkValidTenJapaneseChars-8             20000000         71.1 ns/op
BenchmarkValidStringTenASCIIChars-8          30000000         50.0 ns/op
BenchmarkValidStringTenJapaneseChars-8       10000000        161 ns/op
BenchmarkEncodeASCIIRune-8                   500000000          3.44 ns/op
BenchmarkEncodeJapaneseRune-8                200000000          7.24 ns/op
BenchmarkDecodeASCIIRune-8                   300000000          5.53 ns/op
BenchmarkDecodeJapaneseRune-8                200000000          8.41 ns/op
BenchmarkFullASCIIRune-8                     500000000          3.91 ns/op
BenchmarkFullJapaneseRune-8                  300000000          4.22 ns/op

Change-Id: I674d2ee4917b975a37717bbfa1082cc84dcd275e
Reviewed-on: https://go-review.googlesource.com/14431
Reviewed-by: Russ Cox <rsc@golang.org>
9 years agoreflect: adjust access to unexported embedded structs
Marcel van Lohuizen [Fri, 28 Aug 2015 07:33:51 +0000 (09:33 +0200)]
reflect: adjust access to unexported embedded structs

This CL changes reflect to allow access to exported fields and
methods in unexported embedded structs for gccgo and after gc
has been adjusted to disallow access to embedded unexported structs.

Adresses #12367, #7363, #11007, and #7247.

Change-Id: If80536eab35abcd25300d8ddc2d27d5c42d7e78e
Reviewed-on: https://go-review.googlesource.com/14010
Reviewed-by: Russ Cox <rsc@golang.org>
9 years agoruntime: copy sqrt normalization bugfix from math
Caleb Spare [Fri, 23 Oct 2015 21:49:56 +0000 (14:49 -0700)]
runtime: copy sqrt normalization bugfix from math

This copies the change from CL 16158 (applied as
22d4c8bf13d5edf4670dbdaf0854d653d9c2b81a).

Updates #13013

Change-Id: Id7d02e63d92806f06a4e064a91b2fb6574fe385f
Reviewed-on: https://go-review.googlesource.com/16291
Reviewed-by: Minux Ma <minux@golang.org>
Run-TryBot: Minux Ma <minux@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agocmd/cgo: clarify C.GoStringN and C.GoBytes docs
Shenghou Ma [Fri, 16 Oct 2015 21:48:17 +0000 (17:48 -0400)]
cmd/cgo: clarify C.GoStringN and C.GoBytes docs

Fixes #12427.

Change-Id: I46725620c1e163f3b60ffcd85e5388fa646f074d
Reviewed-on: https://go-review.googlesource.com/15997
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
9 years agonet/url: fix mentions of URL.EscapedPath method
Shenghou Ma [Thu, 8 Oct 2015 00:07:56 +0000 (20:07 -0400)]
net/url: fix mentions of URL.EscapedPath method

Fixes #12862.

Change-Id: I6921ae31bd5515f344fd97d08eafc317228b98a0
Reviewed-on: https://go-review.googlesource.com/15590
Reviewed-by: Andrew Gerrand <adg@golang.org>
9 years agocmd/compile/internal/gc: remove "goto ret" in walkexpr
Håvard Haugen [Sun, 20 Sep 2015 21:28:34 +0000 (23:28 +0200)]
cmd/compile/internal/gc: remove "goto ret" in walkexpr

Passes go build -a -toolexec 'toolstash -cmp' std cmd.

Change-Id: Ib3d2c50601546495e7f1ab153d2978b1e3774101
Reviewed-on: https://go-review.googlesource.com/14800
Run-TryBot: Robert Griesemer <gri@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
9 years agocmd/go, cmd/link: -buildmode=pie for linux/amd64
David Crawshaw [Mon, 19 Oct 2015 20:31:20 +0000 (16:31 -0400)]
cmd/go, cmd/link: -buildmode=pie for linux/amd64

Depends on external linking right now. I have no immediate use for
this, but wanted to check how hard it is to support as android/amd64
is coming and it will require PIE.

Change-Id: I65c6b19159f40db4c79cf312cd0368c2b2527bfd
Reviewed-on: https://go-review.googlesource.com/16072
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agomath: fix normalization bug in pure-Go sqrt
Caleb Spare [Thu, 22 Oct 2015 03:35:22 +0000 (20:35 -0700)]
math: fix normalization bug in pure-Go sqrt

Fixes #13013

Change-Id: I6cf500eacdce76e303fc1cd92dd1c80eef0986bc
Reviewed-on: https://go-review.googlesource.com/16158
Reviewed-by: Andrew Gerrand <adg@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agoruntime: replace is{plan9,solaris,windows} with GOOS tests
Matthew Dempsky [Tue, 9 Jun 2015 22:24:38 +0000 (15:24 -0700)]
runtime: replace is{plan9,solaris,windows} with GOOS tests

Change-Id: I27589395f547c5837dc7536a0ab5bc7cc23a4ff6
Reviewed-on: https://go-review.googlesource.com/10872
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agomath/big: fix SetMantExp comment
David Crawshaw [Fri, 23 Oct 2015 16:21:02 +0000 (12:21 -0400)]
math/big: fix SetMantExp comment

Change-Id: If30cf9c94b58e18564db46c15c6f5cc14ec1a6fa
Reviewed-on: https://go-review.googlesource.com/16271
Reviewed-by: Robert Griesemer <gri@golang.org>
9 years agocmd/compile/internal/gc: update old c-style comments
Jeremy Jackins [Thu, 22 Oct 2015 00:51:12 +0000 (09:51 +0900)]
cmd/compile/internal/gc: update old c-style comments

Update old c-style comments to look like Go comments. Also replace some
lingering references to old .c files that don't exist anymore.

Change-Id: I72b2407a40fc76c23e9048643e0622fd70b4cf90
Reviewed-on: https://go-review.googlesource.com/16190
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
9 years agoruntime: account for cpu affinity in windows NumCPU
Alex Brainman [Thu, 17 Sep 2015 00:43:18 +0000 (10:43 +1000)]
runtime: account for cpu affinity in windows NumCPU

Fixes #11671

Change-Id: Ide1f8d92637dad2a2faed391329f9b6001789b76
Reviewed-on: https://go-review.googlesource.com/14742
Reviewed-by: Russ Cox <rsc@golang.org>
Run-TryBot: Alex Brainman <alex.brainman@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agodatabase/sql: use RWMutex for driver registration
Nathan VanBenschoten [Fri, 23 Oct 2015 01:50:06 +0000 (21:50 -0400)]
database/sql: use RWMutex for driver registration

Change-Id: I0f494c9f17cb6bb0cf5e7214cf033fdbd48f27f7
Reviewed-on: https://go-review.googlesource.com/16240
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

9 years agoregexp: fix slice bounds out of range panics
Didier Spezia [Fri, 28 Aug 2015 17:36:35 +0000 (17:36 +0000)]
regexp: fix slice bounds out of range panics

Regular expressions involving a (x){0} term are
simplified by removing this term from the
expression, just before the expression is compiled.

The number of subexpressions is evaluated before
the simplification. The number of capture instructions
in the compiled expressions is not necessarily in line
with the number of subexpressions.

When the ReplaceAll(String) methods are used, a number
of capture slots (nmatch) is evaluated as 2*(s+1)
(s being the number of subexpressions).

In some case, it can be higher than the number of capture
instructions evaluated at compile time, resulting in a
panic when the internal slices of regexp.machine
are resized to this value.

Fixed by capping the number of capture slots to the number
of capture instructions.

I must say I do not really see the benefits of setting
nmatch lower than re.prog.NumCap using this 2*(s+1) formula,
so perhaps this can be further simplified.

Fixes #11178
Fixes #11176

Change-Id: I21415e8ef2dd5f2721218e9a679f7f6bfb76ae9b
Reviewed-on: https://go-review.googlesource.com/14013
Reviewed-by: Russ Cox <rsc@golang.org>
9 years agostrings: Add examples for HasPrefix and HasSuffix
Gaurish Sharma [Sun, 12 Jul 2015 15:24:00 +0000 (20:54 +0530)]
strings: Add examples for HasPrefix and HasSuffix

These methods didn't had any examples, so added them. Examples makes things more clear
diff --git a/src/strings/example_test.go b/src/strings/example_test.go
index 7243e16..b7763bb 100644
--- a/src/strings/example_test.go
+++ b/src/strings/example_test.go
@@ -223,3 +223,19 @@ func ExampleTrimPrefix() {
  fmt.Print("Hello" + s)
  // Output: Hello, world!
 }
+
+func ExampleHasPrefix() {
+ fmt.Println(strings.HasPrefix("hello", "hell"))
+ fmt.Println(strings.HasPrefix("hello", "heaven"))
+ // Output:
+ // true
+ // false
+}
+
+func ExampleHasSuffix() {
+ fmt.Println(strings.HasSuffix("hello", "llo"))
+ fmt.Println(strings.HasSuffix("hello", "hell"))
+ // Output:
+ // true
+ // false
+}

Change-Id: I5d451c669bd05e19a2afc33ed2ec59b280c2c2d9
Reviewed-on: https://go-review.googlesource.com/12065
Reviewed-by: Russ Cox <rsc@golang.org>
9 years agocmd/compile/internal/gc: remove special handling for -0.0
Robert Griesemer [Thu, 22 Oct 2015 21:14:41 +0000 (14:14 -0700)]
cmd/compile/internal/gc: remove special handling for -0.0

Per the latest spec change, Go doesn't have -0 constants.

Change-Id: Ic2bcdc3bf507d121ed204f30f6744bb8764202c0
Reviewed-on: https://go-review.googlesource.com/16232
Reviewed-by: Chris Manghane <cmang@golang.org>
9 years agocmd/compile/internal/gc: compact binary export format
Robert Griesemer [Fri, 14 Aug 2015 02:05:37 +0000 (19:05 -0700)]
cmd/compile/internal/gc: compact binary export format

The binary import/export format is significantly more
compact than the existing textual format. It should
also be faster to read and write (to be measured).

Use -newexport to enable, for instance:
export GO_GCFLAGS=-newexport; make.bash

The compiler can import packages using both the old
and the new format ("mixed mode").

Missing: export info for inlined functions bodies
(performance issue, does not affect correctness).

Disabled by default until we have inlined function
bodies and confirmation of no regression and equality
of binaries.

For #6110.
For #1909.

This change depends on:

   https://go-review.googlesource.com/16220
   https://go-review.googlesource.com/16222

(already submitted) for all.bash to work.

Some initial export data sizes for std lib packages. This data
is without exported functions with inlineable function bodies.

Package                                       old      new    new/old

archive/tar.................................13875.....3883    28%
archive/zip.................................19464.....5046    26%
bufio....................................... 7733.....2222    29%
bytes.......................................10342.....3347    32%
cmd/addr2line.................................242.......26    11%
cmd/api.....................................39305....10368    26%
cmd/asm/internal/arch.......................27732.....7939    29%
cmd/asm/internal/asm........................35264....10295    29%
cmd/asm/internal/flags........................629......178    28%
cmd/asm/internal/lex........................39248....11128    28%
cmd/asm.......................................306.......26     8%
cmd/cgo.....................................40197....10570    26%
cmd/compile/internal/amd64...................1106......214    19%
cmd/compile/internal/arm....................27891.....7710    28%
cmd/compile/internal/arm64....................891......153    17%
cmd/compile/internal/big....................21637.....8336    39%
cmd/compile/internal/gc....................109845....29727    27%
cmd/compile/internal/mips64...................972......168    17%
cmd/compile/internal/ppc64....................972......168    17%
cmd/compile/internal/x86.....................1104......195    18%
cmd/compile...................................329.......26     8%
cmd/cover...................................12986.....3749    29%
cmd/dist......................................477.......67    14%
cmd/doc.....................................23043.....6793    29%
cmd/expdump...................................167.......26    16%
cmd/fix......................................1190......208    17%
cmd/go......................................26399.....5629    21%
cmd/gofmt.....................................499.......26     5%
cmd/internal/gcprog..........................1342......490    37%
cmd/internal/goobj...........................2690......980    36%
cmd/internal/obj/arm........................32740....10057    31%
cmd/internal/obj/arm64......................46542....15364    33%
cmd/internal/obj/mips.......................42140....13731    33%
cmd/internal/obj/ppc64......................42140....13731    33%
cmd/internal/obj/x86........................52732....19015    36%
cmd/internal/obj............................36729....11690    32%
cmd/internal/objfile........................36365....10287    28%
cmd/link/internal/amd64.....................45893....12220    27%
cmd/link/internal/arm.........................307.......96    31%
cmd/link/internal/arm64.......................345.......98    28%
cmd/link/internal/ld.......................109300....46326    42%
cmd/link/internal/ppc64.......................344.......99    29%
cmd/link/internal/x86.........................334......107    32%
cmd/link......................................314.......26     8%
cmd/newlink..................................8110.....2544    31%
cmd/nm........................................210.......26    12%
cmd/objdump...................................244.......26    11%
cmd/pack....................................14248.....4066    29%
cmd/pprof/internal/commands..................5239.....1285    25%
cmd/pprof/internal/driver...................37967.....8860    23%
cmd/pprof/internal/fetch....................30962.....7337    24%
cmd/pprof/internal/plugin...................47734.....7719    16%
cmd/pprof/internal/profile..................22286.....6922    31%
cmd/pprof/internal/report...................31187.....7838    25%
cmd/pprof/internal/svg.......................4315......965    22%
cmd/pprof/internal/symbolizer...............30051.....7397    25%
cmd/pprof/internal/symbolz..................28545.....6949    24%
cmd/pprof/internal/tempfile.................12550.....3356    27%
cmd/pprof.....................................563.......26     5%
cmd/trace....................................1455......636    44%
cmd/vendor/golang.org/x/arch/arm/armasm....168035....64737    39%
cmd/vendor/golang.org/x/arch/x86/x86asm.....26871.....8578    32%
cmd/vet.....................................38980.....9913    25%
cmd/vet/whitelist.............................102.......49    48%
cmd/yacc.....................................2518......926    37%
compress/bzip2...............................6326......129     2%
compress/flate...............................7069.....2541    36%
compress/gzip...............................20143.....5069    25%
compress/lzw..................................828......295    36%
compress/zlib...............................10676.....2692    25%
container/heap................................523......181    35%
container/list...............................3517......740    21%
container/ring................................881......229    26%
crypto/aes....................................550......187    34%
crypto/cipher................................1966......825    42%
crypto.......................................1836......646    35%
crypto/des....................................632......235    37%
crypto/dsa..................................18718.....5035    27%
crypto/ecdsa................................23131.....6097    26%
crypto/elliptic.............................20790.....5740    28%
crypto/hmac...................................455......186    41%
crypto/md5...................................1375......171    12%
crypto/rand.................................18132.....4748    26%
crypto/rc4....................................561......240    43%
crypto/rsa..................................22094.....6380    29%
crypto/sha1..................................1416......172    12%
crypto/sha256.................................551......238    43%
crypto/sha512.................................839......378    45%
crypto/subtle................................1153......250    22%
crypto/tls..................................58203....17984    31%
crypto/x509/pkix............................29447.....8161    28%
database/sql/driver..........................3318.....1096    33%
database/sql................................11258.....3942    35%
debug/dwarf.................................18416.....7006    38%
debug/elf...................................57530....21014    37%
debug/gosym..................................4992.....2058    41%
debug/macho.................................23037.....6538    28%
debug/pe....................................21063.....6619    31%
debug/plan9obj...............................2467......802    33%
encoding/ascii85.............................1523......360    24%
encoding/asn1................................1718......527    31%
encoding/base32..............................2642......686    26%
encoding/base64..............................3077......800    26%
encoding/binary..............................4727.....1040    22%
encoding/csv................................12223.....2850    23%
encoding......................................383......217    57%
encoding/gob................................37563....10113    27%
encoding/hex.................................1327......390    29%
encoding/json...............................30897.....7804    25%
encoding/pem..................................595......200    34%
encoding/xml................................37798.....9336    25%
errors........................................274.......36    13%
expvar.......................................3155.....1021    32%
flag........................................19860.....2849    14%
fmt..........................................3137.....1263    40%
go/ast......................................44729....13422    30%
go/build....................................16336.....4657    29%
go/constant..................................3703......846    23%
go/doc.......................................9877.....2807    28%
go/format....................................5472.....1575    29%
go/importer..................................4980.....1301    26%
go/internal/gccgoimporter....................5587.....1525    27%
go/internal/gcimporter.......................8979.....2186    24%
go/parser...................................20692.....5304    26%
go/printer...................................7015.....2029    29%
go/scanner...................................9719.....2824    29%
go/token.....................................7933.....2465    31%
go/types....................................64569....19978    31%
hash/adler32.................................1176......176    15%
hash/crc32...................................1663......360    22%
hash/crc64...................................1587......306    19%
hash/fnv.....................................3964......260     7%
hash..........................................591......278    47%
html..........................................217.......74    34%
html/template...............................69623....12588    18%
image/color/palette...........................315.......98    31%
image/color..................................5565.....1036    19%
image/draw...................................6917.....1028    15%
image/gif....................................8894.....1654    19%
image/internal/imageutil.....................9112.....1476    16%
image/jpeg...................................6647.....1026    15%
image/png....................................6906.....1069    15%
image.......................................28992.....6139    21%
index/suffixarray...........................17106.....4773    28%
internal/singleflight........................1614......506    31%
internal/testenv............................12212.....3152    26%
internal/trace...............................2762.....1323    48%
io/ioutil...................................13502.....3682    27%
io...........................................6765.....2482    37%
log.........................................11620.....3317    29%
log/syslog..................................13516.....3821    28%
math/big....................................21819.....8320    38%
math/cmplx...................................2816......438    16%
math/rand....................................2317......929    40%
math.........................................7511.....2444    33%
mime/multipart..............................12679.....3360    27%
mime/quotedprintable.........................5458.....1235    23%
mime.........................................6076.....1628    27%
net/http/cgi................................59796....17173    29%
net/http/cookiejar..........................14781.....3739    25%
net/http/fcgi...............................57861....16426    28%
net/http/httptest...........................84100....24365    29%
net/http/httputil...........................67763....18869    28%
net/http/internal............................6907......637     9%
net/http/pprof..............................57945....16316    28%
net/http....................................95391....30210    32%
net/internal/socktest........................4555.....1453    32%
net/mail....................................14481.....3608    25%
net/rpc/jsonrpc.............................33335......988     3%
net/rpc.....................................79950....23106    29%
net/smtp....................................57790....16468    28%
net/textproto...............................11356.....3248    29%
net/url......................................3123.....1009    32%
os/exec.....................................20738.....5769    28%
os/signal.....................................437......167    38%
os..........................................24875.....6668    27%
path/filepath...............................11340.....2826    25%
path..........................................778......285    37%
reflect.....................................15469.....5198    34%
regexp......................................13627.....4661    34%
regexp/syntax................................5539.....2249    41%
runtime/debug................................9275.....2322    25%
runtime/pprof................................1355......477    35%
runtime/race...................................39.......17    44%
runtime/trace.................................228.......92    40%
runtime.....................................13498.....1821    13%
sort.........................................2848......842    30%
strconv......................................2947.....1252    42%
strings......................................7983.....2456    31%
sync/atomic..................................2666.....1149    43%
sync.........................................2568......845    33%
syscall.....................................81252....38398    47%
testing/iotest...............................2444......302    12%
testing/quick...............................18890.....5076    27%
testing.....................................16502.....4800    29%
text/scanner.................................6849.....2052    30%
text/tabwriter...............................6607.....1863    28%
text/template/parse.........................22978.....6183    27%
text/template...............................64153....11518    18%
time........................................12103.....3546    29%
unicode......................................9706.....3320    34%
unicode/utf16................................1055......148    14%
unicode/utf8.................................1118......513    46%
vendor/golang.org/x/net/http2/hpack..........8905.....2636    30%

All packages                              3518505  1017774    29%

Change-Id: Id657334f276383ff1e6fa91472d3d1db5a03349c
Reviewed-on: https://go-review.googlesource.com/13937
Run-TryBot: Robert Griesemer <gri@golang.org>
Reviewed-by: Chris Manghane <cmang@golang.org>
9 years agocmd/link/internal/ld: export data may be marked with $$ or $$B
Robert Griesemer [Thu, 22 Oct 2015 17:41:15 +0000 (10:41 -0700)]
cmd/link/internal/ld: export data may be marked with $$ or $$B

Preparation for dealing with binary export format. Accept $$B
as marker for export data. For now, skip that data if found.

Change-Id: I464ba22aaedcf349725379d91070fc900d93b7a2
Reviewed-on: https://go-review.googlesource.com/16222
Reviewed-by: Chris Manghane <cmang@golang.org>
9 years agocmd/internal/goobj: permit 0 bytes in export data (prep. for new export format)
Robert Griesemer [Thu, 22 Oct 2015 17:08:45 +0000 (10:08 -0700)]
cmd/internal/goobj: permit 0 bytes in export data (prep. for new export format)

Change-Id: Ida9a82a5420a41ef656388866a1cf8fadf12ccc6
Reviewed-on: https://go-review.googlesource.com/16220
Reviewed-by: Chris Manghane <cmang@golang.org>
9 years agocmd/compile/internal/gc: fixed comments (cosmetic change)
Robert Griesemer [Thu, 22 Oct 2015 17:04:23 +0000 (10:04 -0700)]
cmd/compile/internal/gc: fixed comments (cosmetic change)

Change-Id: I9cbbfcf55ce3cccb6f14b9ffe6777567dcdad9c9
Reviewed-on: https://go-review.googlesource.com/16179
Reviewed-by: Chris Manghane <cmang@golang.org>
9 years agoruntime: add pcvalue cache to improve stack scan speed
Austin Clements [Thu, 13 Aug 2015 03:43:43 +0000 (23:43 -0400)]
runtime: add pcvalue cache to improve stack scan speed

The cost of scanning large stacks is currently dominated by the time
spent looking up and decoding the pcvalue table. However, large stacks
are usually large not because they contain calls to many different
functions, but because they contain many calls to the same, small set
of recursive functions. Hence, walking large stacks tends to make the
same pcvalue queries many times.

Based on this observation, this commit adds a small, very simple, and
fast cache in front of pcvalue lookup. We thread this cache down from
operations that make many pcvalue calls, such as gentraceback, stack
scanning, and stack adjusting.

This simple cache works well because it has minimal overhead when it's
not effective. I also tried a hashed direct-map cache, CLOCK-based
replacement, round-robin replacement, and round-robin with lookups
disabled until there had been at least 16 probes, but none of these
approaches had obvious wins over the random replacement policy in this
commit.

This nearly doubles the overall performance of the deep stack test
program from issue #10898:

name        old time/op  new time/op  delta
Issue10898   16.5s ±12%    9.2s ±12%  -44.37%  (p=0.008 n=5+5)

It's a very slight win on the garbage benchmark:

name              old time/op  new time/op  delta
XBenchGarbage-12  4.92ms ± 1%  4.89ms ± 1%  -0.75%  (p=0.000 n=18+19)

It's a wash (but doesn't harm performance) on the go1 benchmarks,
which don't have particularly deep stacks:

name                      old time/op    new time/op    delta
BinaryTree17-12              3.11s ± 2%     3.20s ± 3%  +2.83%  (p=0.000 n=17+20)
Fannkuch11-12                2.51s ± 1%     2.51s ± 1%  -0.22%  (p=0.034 n=19+18)
FmtFprintfEmpty-12          50.8ns ± 3%    50.6ns ± 2%    ~     (p=0.793 n=20+20)
FmtFprintfString-12          174ns ± 0%     174ns ± 1%  +0.17%  (p=0.048 n=15+20)
FmtFprintfInt-12             177ns ± 0%     165ns ± 1%  -6.99%  (p=0.000 n=17+19)
FmtFprintfIntInt-12          283ns ± 1%     284ns ± 0%  +0.22%  (p=0.000 n=18+15)
FmtFprintfPrefixedInt-12     243ns ± 1%     244ns ± 1%  +0.40%  (p=0.000 n=20+19)
FmtFprintfFloat-12           318ns ± 0%     319ns ± 0%  +0.27%  (p=0.001 n=19+20)
FmtManyArgs-12              1.12µs ± 0%    1.14µs ± 0%  +1.74%  (p=0.000 n=19+20)
GobDecode-12                8.69ms ± 0%    8.73ms ± 1%  +0.46%  (p=0.000 n=18+18)
GobEncode-12                6.64ms ± 1%    6.61ms ± 1%  -0.46%  (p=0.000 n=20+20)
Gzip-12                      323ms ± 2%     319ms ± 1%  -1.11%  (p=0.000 n=20+20)
Gunzip-12                   42.8ms ± 0%    42.9ms ± 0%    ~     (p=0.158 n=18+20)
HTTPClientServer-12         63.3µs ± 1%    63.1µs ± 1%  -0.35%  (p=0.011 n=20+20)
JSONEncode-12               16.9ms ± 1%    17.3ms ± 1%  +2.84%  (p=0.000 n=19+20)
JSONDecode-12               59.7ms ± 0%    58.5ms ± 0%  -2.05%  (p=0.000 n=19+17)
Mandelbrot200-12            3.92ms ± 0%    3.91ms ± 0%  -0.16%  (p=0.003 n=19+19)
GoParse-12                  3.79ms ± 2%    3.75ms ± 2%  -0.91%  (p=0.005 n=20+20)
RegexpMatchEasy0_32-12       102ns ± 1%     101ns ± 1%  -0.80%  (p=0.001 n=14+20)
RegexpMatchEasy0_1K-12       337ns ± 1%     346ns ± 1%  +2.90%  (p=0.000 n=20+19)
RegexpMatchEasy1_32-12      84.4ns ± 2%    84.3ns ± 2%    ~     (p=0.743 n=20+20)
RegexpMatchEasy1_1K-12       502ns ± 1%     505ns ± 0%  +0.64%  (p=0.000 n=20+20)
RegexpMatchMedium_32-12      133ns ± 1%     132ns ± 1%  -0.85%  (p=0.000 n=20+19)
RegexpMatchMedium_1K-12     40.1µs ± 1%    39.8µs ± 1%  -0.77%  (p=0.000 n=18+18)
RegexpMatchHard_32-12       2.08µs ± 1%    2.07µs ± 1%  -0.55%  (p=0.001 n=18+19)
RegexpMatchHard_1K-12       62.4µs ± 1%    62.0µs ± 1%  -0.74%  (p=0.000 n=19+19)
Revcomp-12                   545ms ± 2%     545ms ± 3%    ~     (p=0.771 n=19+20)
Template-12                 73.7ms ± 1%    72.0ms ± 0%  -2.33%  (p=0.000 n=20+18)
TimeParse-12                 358ns ± 1%     351ns ± 1%  -2.07%  (p=0.000 n=20+20)
TimeFormat-12                369ns ± 1%     356ns ± 0%  -3.53%  (p=0.000 n=20+18)
[Geo mean]                  63.5µs         63.2µs       -0.41%

name                      old speed      new speed      delta
GobDecode-12              88.3MB/s ± 0%  87.9MB/s ± 0%  -0.43%  (p=0.000 n=18+17)
GobEncode-12               116MB/s ± 1%   116MB/s ± 1%  +0.47%  (p=0.000 n=20+20)
Gzip-12                   60.2MB/s ± 2%  60.8MB/s ± 1%  +1.13%  (p=0.000 n=20+20)
Gunzip-12                  453MB/s ± 0%   453MB/s ± 0%    ~     (p=0.160 n=18+20)
JSONEncode-12              115MB/s ± 1%   112MB/s ± 1%  -2.76%  (p=0.000 n=19+20)
JSONDecode-12             32.5MB/s ± 0%  33.2MB/s ± 0%  +2.09%  (p=0.000 n=19+17)
GoParse-12                15.3MB/s ± 2%  15.4MB/s ± 2%  +0.92%  (p=0.004 n=20+20)
RegexpMatchEasy0_32-12     311MB/s ± 1%   314MB/s ± 1%  +0.78%  (p=0.000 n=15+19)
RegexpMatchEasy0_1K-12    3.04GB/s ± 1%  2.95GB/s ± 1%  -2.90%  (p=0.000 n=19+19)
RegexpMatchEasy1_32-12     379MB/s ± 2%   380MB/s ± 2%    ~     (p=0.779 n=20+20)
RegexpMatchEasy1_1K-12    2.04GB/s ± 1%  2.02GB/s ± 0%  -0.62%  (p=0.000 n=20+20)
RegexpMatchMedium_32-12   7.46MB/s ± 1%  7.53MB/s ± 1%  +0.86%  (p=0.000 n=20+19)
RegexpMatchMedium_1K-12   25.5MB/s ± 1%  25.7MB/s ± 1%  +0.78%  (p=0.000 n=18+18)
RegexpMatchHard_32-12     15.4MB/s ± 1%  15.5MB/s ± 1%  +0.62%  (p=0.000 n=19+19)
RegexpMatchHard_1K-12     16.4MB/s ± 1%  16.5MB/s ± 1%  +0.82%  (p=0.000 n=20+19)
Revcomp-12                 466MB/s ± 2%   466MB/s ± 3%    ~     (p=0.765 n=19+20)
Template-12               26.3MB/s ± 1%  27.0MB/s ± 0%  +2.38%  (p=0.000 n=20+18)
[Geo mean]                97.8MB/s       98.0MB/s       +0.23%

Change-Id: I281044ae0b24990ba46487cacbc1069493274bc4
Reviewed-on: https://go-review.googlesource.com/13614
Reviewed-by: Keith Randall <khr@golang.org>