]> Cypherpunks repositories - gostls13.git/log
gostls13.git
7 years agocmd/go: restrict meta imports to valid schemes
Ian Lance Taylor [Thu, 15 Feb 2018 23:57:13 +0000 (15:57 -0800)]
cmd/go: restrict meta imports to valid schemes

Before this change, when using -insecure, we permitted any meta import
repo root as long as it contained "://". When not using -insecure, we
restrict meta import repo roots to be valid URLs. People may depend on
that somehow, so permit meta import repo roots to be invalid URLs, but
require them to have valid schemes per RFC 3986.

Fixes #23867

Change-Id: Iac666dfc75ac321bf8639dda5b0dba7c8840922d
Reviewed-on: https://go-review.googlesource.com/94603
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agonet/http: increase timeout length for TestOnlyWriteTimeout
Richard Miller [Fri, 16 Feb 2018 17:01:52 +0000 (17:01 +0000)]
net/http: increase timeout length for TestOnlyWriteTimeout

This test was sometimes timing out on the plan9/arm builder
(raspberry pi) when run in parallel with other network intensive
tests. It appears that tcp on the loopback interface could do
with some tuning for better performance on Plan 9, but until
that's done, increasing the timeout from 5 to 10 seconds allows
this test to pass.  This should have no effect on other platforms
where 5 seconds was already enough.

Change-Id: If310ee569cae8ca8f56346d84ce23803feb23a41
Reviewed-on: https://go-review.googlesource.com/94796
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agoruntime: don't ignore address hint for sysReserve in Plan 9
Richard Miller [Fri, 16 Feb 2018 15:20:04 +0000 (15:20 +0000)]
runtime: don't ignore address hint for sysReserve in Plan 9

On Plan 9, sysReserve was ignoring the address hint and allocating
memory wherever it is available.  This causes the new
TestArenaCollision test to fail on 32-bit Plan 9.  We now use the
address hint in the specific case where sysReserve is extending the
process address space at its end, and similarly we contract the
address space in the case where sysFree is releasing memory at
the end.

Fixes #23860

Change-Id: Ia5254779ba8f1698c999832720a88de400b5f91a
Reviewed-on: https://go-review.googlesource.com/94776
Reviewed-by: Austin Clements <austin@google.com>
Reviewed-by: David du Colombier <0intro@gmail.com>
7 years agocmd/compile: make math.Ceil/Floor/Round/Trunc intrinsics on arm64
Chad Rosier [Fri, 16 Feb 2018 14:22:32 +0000 (09:22 -0500)]
cmd/compile: make math.Ceil/Floor/Round/Trunc intrinsics on arm64

name       old time/op  new time/op  delta
Ceil        550ns ± 0%   486ns ± 7%  -11.64%  (p=0.000 n=13+18)
Floor       495ns ±19%   512ns ±12%     ~     (p=0.164 n=20+20)
Round       550ns ± 0%   487ns ± 8%  -11.49%  (p=0.000 n=12+19)
Trunc       563ns ± 7%   488ns ±13%  -13.44%  (p=0.000 n=15+2)

Change-Id: I53f234b160b3c026a277506e2cf977d150379464
Reviewed-on: https://go-review.googlesource.com/88295
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
7 years agoruntime: only run TestArenaCollision if the target can exec
Elias Naur [Fri, 16 Feb 2018 09:26:00 +0000 (10:26 +0100)]
runtime: only run TestArenaCollision if the target can exec

Replace the test for nacl with testenv.MustHaveExec to also skip
test on iOS.

Change-Id: I6822714f6d71533d1b18bbb7894f6ad339d8aea1
Reviewed-on: https://go-review.googlesource.com/94755
Run-TryBot: Elias Naur <elias.naur@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
7 years agocmd/compile: permit go:cgo_import_dynamic anywhere
Ian Lance Taylor [Wed, 14 Feb 2018 15:42:40 +0000 (07:42 -0800)]
cmd/compile: permit go:cgo_import_dynamic anywhere

It's used on Solaris to import symbols from shared libraries, e.g., in
golang.org/x/sys/unix and golang.org/x/net/internal/socket.
We could use a different directive but that would require build tags
in all the places that use it.

Updates #23672
Updates #23749

Change-Id: I47fcf72a6d2862e304204705979c2056c2f78ec5
Reviewed-on: https://go-review.googlesource.com/94018
Run-TryBot: Andrew Bonventre <andybons@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agodoc: add note about invalid flag errors to 1.10 release notes
Ian Lance Taylor [Thu, 15 Feb 2018 02:04:54 +0000 (18:04 -0800)]
doc: add note about invalid flag errors to 1.10 release notes

Updates #23672
Updates #23749

Change-Id: I85b6493bd1e4095398508b7ef395c35463dfbde1
Reviewed-on: https://go-review.googlesource.com/94158
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Andrew Bonventre <andybons@golang.org>
7 years agocmd/compile: arm64 intrinsics for math/bits.OnesCount
Balaram Makam [Tue, 30 Jan 2018 17:16:52 +0000 (12:16 -0500)]
cmd/compile: arm64 intrinsics for math/bits.OnesCount

This adds math/bits intrinsics for OnesCount on arm64.

name         old time/op  new time/op  delta
OnesCount    3.81ns ± 0%  1.60ns ± 0%  -57.96%  (p=0.000 n=7+8)
OnesCount8   1.60ns ± 0%  1.60ns ± 0%     ~     (all equal)
OnesCount16  2.41ns ± 0%  1.60ns ± 0%  -33.61%  (p=0.000 n=8+8)
OnesCount32  4.17ns ± 0%  1.60ns ± 0%  -61.58%  (p=0.000 n=8+8)
OnesCount64  3.80ns ± 0%  1.60ns ± 0%  -57.84%  (p=0.000 n=8+8)

Update #18616

Conflicts:
src/cmd/compile/internal/gc/asm_test.go

Change-Id: I63ac2f63acafdb1f60656ab8a56be0b326eec5cb
Reviewed-on: https://go-review.googlesource.com/90835
Run-TryBot: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
7 years agocmd/compile/internal/gc: use functype instead of OTFUNC
Matthew Dempsky [Fri, 7 Apr 2017 22:39:36 +0000 (15:39 -0700)]
cmd/compile/internal/gc: use functype instead of OTFUNC

Slightly simpler.

Change-Id: Ic3a96675c56cc8c2e336b932536c2247f8cbb96d
Reviewed-on: https://go-review.googlesource.com/39996
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
Reviewed-by: Robert Griesemer <gri@golang.org>
7 years agoruntime: replace _MaxMem with maxAlloc
Austin Clements [Tue, 2 Jan 2018 02:51:47 +0000 (21:51 -0500)]
runtime: replace _MaxMem with maxAlloc

Now that we have memLimit, also having _MaxMem is a bit confusing.

Replace it with maxAlloc, which better conveys what it limits. We also
define maxAlloc slightly differently: since it's now clear that it
limits allocation size, we can account for a subtle difference between
32-bit and 64-bit.

Change-Id: Iac39048018cc0dae7f0919e25185fee4b3eed529
Reviewed-on: https://go-review.googlesource.com/85890
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rick Hudson <rlh@golang.org>
7 years agoruntime: move comment about address space sizes to malloc.go
Austin Clements [Mon, 1 Jan 2018 22:53:59 +0000 (17:53 -0500)]
runtime: move comment about address space sizes to malloc.go

Currently there's a detailed comment in lfstack_64bit.go about address
space limitations on various architectures. Since that's now relevant
to malloc, move it to a more prominent place in the documentation for
memLimitBits.

Updates #10460.

Change-Id: If9708291cf3a288057b8b3ba0ba6a59e3602bbd6
Reviewed-on: https://go-review.googlesource.com/85889
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rick Hudson <rlh@golang.org>
7 years agoruntime: remove non-reserved heap logic
Austin Clements [Sun, 31 Dec 2017 00:35:46 +0000 (19:35 -0500)]
runtime: remove non-reserved heap logic

Currently large sysReserve calls on some OSes don't actually reserve
the memory, but just check that it can be reserved. This was important
when we called sysReserve to "reserve" many gigabytes for the heap up
front, but now that we map memory in small increments as we need it,
this complication is no longer necessary.

This has one curious side benefit: currently, on Linux, allocations
that are large enough to be rejected by mmap wind up freezing the
application for a long time before it panics. This happens because
sysReserve doesn't reserve the memory, so sysMap calls mmap_fixed,
which calls mmap, which fails because the mapping is too large.
However, mmap_fixed doesn't inspect *why* mmap fails, so it falls back
to probing every page in the desired region individually with mincore
before performing an (otherwise dangerous) MAP_FIXED mapping, which
will also fail. This takes a long time for a large region. Now this
logic is gone, so the mmap failure leads to an immediate panic.

Updates #10460.

Change-Id: I8efe88c611871cdb14f99fadd09db83e0161ca2e
Reviewed-on: https://go-review.googlesource.com/85888
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rick Hudson <rlh@golang.org>
7 years agoruntime: use sparse mappings for the heap
Austin Clements [Wed, 20 Dec 2017 06:05:23 +0000 (22:05 -0800)]
runtime: use sparse mappings for the heap

This replaces the contiguous heap arena mapping with a potentially
sparse mapping that can support heap mappings anywhere in the address
space.

This has several advantages over the current approach:

* There is no longer any limit on the size of the Go heap. (Currently
  it's limited to 512GB.) Hence, this fixes #10460.

* It eliminates many failures modes of heap initialization and
  growing. In particular it eliminates any possibility of panicking
  with an address space conflict. This can happen for many reasons and
  even causes a low but steady rate of TSAN test failures because of
  conflicts with the TSAN runtime. See #16936 and #11993.

* It eliminates the notion of "non-reserved" heap, which was added
  because creating huge address space reservations (particularly on
  64-bit) led to huge process VSIZE. This was at best confusing and at
  worst conflicted badly with ulimit -v. However, the non-reserved
  heap logic is complicated, can race with other mappings in non-pure
  Go binaries (e.g., #18976), and requires that the entire heap be
  either reserved or non-reserved. We currently maintain the latter
  property, but it's quite difficult to convince yourself of that, and
  hence difficult to keep correct. This logic is still present, but
  will be removed in the next CL.

* It fixes problems on 32-bit where skipping over parts of the address
  space leads to mapping huge (and never-to-be-used) metadata
  structures. See #19831.

This also completely rewrites and significantly simplifies
mheap.sysAlloc, which has been a source of many bugs. E.g., #21044,
 #20259, #18651, and #13143 (and maybe #23222).

This change also makes it possible to allocate individual objects
larger than 512GB. As a result, a few tests that expected huge
allocations to fail needed to be changed to make even larger
allocations. However, at the moment attempting to allocate a humongous
object may cause the program to freeze for several minutes on Linux as
we fall back to probing every page with addrspace_free. That logic
(and this failure mode) will be removed in the next CL.

Fixes #10460.
Fixes #22204 (since it rewrites the code involved).

This slightly slows down compilebench and the x/benchmarks garbage
benchmark.

name       old time/op     new time/op     delta
Template       184ms ± 1%      185ms ± 1%    ~     (p=0.065 n=10+9)
Unicode       86.9ms ± 3%     86.3ms ± 1%    ~     (p=0.631 n=10+10)
GoTypes        599ms ± 0%      602ms ± 0%  +0.56%  (p=0.000 n=10+9)
Compiler       2.87s ± 1%      2.89s ± 1%  +0.51%  (p=0.002 n=9+10)
SSA            7.29s ± 1%      7.25s ± 1%    ~     (p=0.182 n=10+9)
Flate          118ms ± 2%      118ms ± 1%    ~     (p=0.113 n=9+9)
GoParser       147ms ± 1%      148ms ± 1%  +1.07%  (p=0.003 n=9+10)
Reflect        401ms ± 1%      404ms ± 1%  +0.71%  (p=0.003 n=10+9)
Tar            175ms ± 1%      175ms ± 1%    ~     (p=0.604 n=9+10)
XML            209ms ± 1%      210ms ± 1%    ~     (p=0.052 n=10+10)

(https://perf.golang.org/search?q=upload:20171231.4)

name                       old time/op  new time/op  delta
Garbage/benchmem-MB=64-12  2.23ms ± 1%  2.25ms ± 1%  +0.84%  (p=0.000 n=19+19)

(https://perf.golang.org/search?q=upload:20171231.3)

Relative to the start of the sparse heap changes (starting at and
including "runtime: fix various contiguous bitmap assumptions"),
overall slowdown is roughly 1% on GC-intensive benchmarks:

name        old time/op     new time/op     delta
Template        183ms ± 1%      185ms ± 1%  +1.32%  (p=0.000 n=9+9)
Unicode        84.9ms ± 2%     86.3ms ± 1%  +1.65%  (p=0.000 n=9+10)
GoTypes         595ms ± 1%      602ms ± 0%  +1.19%  (p=0.000 n=9+9)
Compiler        2.86s ± 0%      2.89s ± 1%  +0.91%  (p=0.000 n=9+10)
SSA             7.19s ± 0%      7.25s ± 1%  +0.75%  (p=0.000 n=8+9)
Flate           117ms ± 1%      118ms ± 1%  +1.10%  (p=0.000 n=10+9)
GoParser        146ms ± 2%      148ms ± 1%  +1.48%  (p=0.002 n=10+10)
Reflect         398ms ± 1%      404ms ± 1%  +1.51%  (p=0.000 n=10+9)
Tar             173ms ± 1%      175ms ± 1%  +1.17%  (p=0.000 n=10+10)
XML             208ms ± 1%      210ms ± 1%  +0.62%  (p=0.011 n=10+10)
[Geo mean]      369ms           373ms       +1.17%

(https://perf.golang.org/search?q=upload:20180101.2)

name                       old time/op  new time/op  delta
Garbage/benchmem-MB=64-12  2.22ms ± 1%  2.25ms ± 1%  +1.51%  (p=0.000 n=20+19)

(https://perf.golang.org/search?q=upload:20180101.3)

Change-Id: I5daf4cfec24b252e5a57001f0a6c03f22479d0f0
Reviewed-on: https://go-review.googlesource.com/85887
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rick Hudson <rlh@golang.org>
7 years agoruntime: eliminate most uses of mheap_.arena_*
Austin Clements [Tue, 19 Dec 2017 04:35:34 +0000 (20:35 -0800)]
runtime: eliminate most uses of mheap_.arena_*

This replaces all uses of the mheap_.arena_* fields outside of
mallocinit and sysAlloc. These fields fundamentally assume a
contiguous heap between two bounds, so eliminating these is necessary
for a sparse heap.

Many of these are replaced with checks for non-nil spans at the test
address (which in turn checks for a non-nil entry in the heap arena
array). Some of them are just for debugging and somewhat meaningless
with a sparse heap, so those we just delete.

Updates #10460.

Change-Id: I8345b95ffc610aed694f08f74633b3c63506a41f
Reviewed-on: https://go-review.googlesource.com/85886
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rick Hudson <rlh@golang.org>
7 years agoruntime: make span map sparse
Austin Clements [Wed, 13 Dec 2017 21:09:02 +0000 (16:09 -0500)]
runtime: make span map sparse

This splits the span map into separate chunks for every 64MB of the
heap. The span map chunks now live in the same indirect structure as
the bitmap.

Updates #10460.

This causes a slight improvement in compilebench and the x/benchmarks
garbage benchmark. I'm not sure why it improves performance.

name       old time/op     new time/op     delta
Template       185ms ± 1%      184ms ± 1%    ~            (p=0.315 n=9+10)
Unicode       86.9ms ± 1%     86.9ms ± 3%    ~            (p=0.356 n=9+10)
GoTypes        602ms ± 1%      599ms ± 0%  -0.59%         (p=0.002 n=9+10)
Compiler       2.89s ± 0%      2.87s ± 1%  -0.50%          (p=0.003 n=9+9)
SSA            7.25s ± 0%      7.29s ± 1%    ~            (p=0.400 n=9+10)
Flate          118ms ± 1%      118ms ± 2%    ~            (p=0.065 n=10+9)
GoParser       147ms ± 2%      147ms ± 1%    ~            (p=0.549 n=10+9)
Reflect        403ms ± 1%      401ms ± 1%  -0.47%         (p=0.035 n=9+10)
Tar            176ms ± 1%      175ms ± 1%  -0.59%         (p=0.013 n=10+9)
XML            211ms ± 1%      209ms ± 1%  -0.83%        (p=0.011 n=10+10)

(https://perf.golang.org/search?q=upload:20171231.1)

name                       old time/op  new time/op  delta
Garbage/benchmem-MB=64-12  2.24ms ± 1%  2.23ms ± 1%  -0.36%  (p=0.001 n=20+19)

(https://perf.golang.org/search?q=upload:20171231.2)

Change-Id: I2563f8704ab9812434947faf293c5327f9b0d07a
Reviewed-on: https://go-review.googlesource.com/85885
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rick Hudson <rlh@golang.org>
7 years agoruntime: abstract remaining mheap.spans access
Austin Clements [Wed, 13 Dec 2017 21:03:23 +0000 (16:03 -0500)]
runtime: abstract remaining mheap.spans access

This abstracts the remaining direct accesses to mheap.spans into new
mheap.setSpan and mheap.setSpans methods.

For #10460.

Change-Id: Id1db8bc5e34a77a9221032aa2e62d05322707364
Reviewed-on: https://go-review.googlesource.com/85884
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rick Hudson <rlh@golang.org>
7 years agoruntime: make the heap bitmap sparse
Austin Clements [Sat, 9 Dec 2017 03:57:53 +0000 (22:57 -0500)]
runtime: make the heap bitmap sparse

This splits the heap bitmap into separate chunks for every 64MB of the
heap and introduces an index mapping from virtual address to metadata.
It modifies the heapBits abstraction to use this two-level structure.
Finally, it modifies heapBitsSetType to unroll the bitmap into the
object itself and then copy it out if the bitmap would span
discontiguous bitmap chunks.

This is a step toward supporting general sparse heaps, which will
eliminate address space conflict failures as well as the limit on the
heap size.

It's also advantageous for 32-bit. 32-bit already supports
discontiguous heaps by always starting the arena at address 0.
However, as a result, with a contiguous bitmap, if the kernel chooses
a high address (near 2GB) for a heap mapping, the runtime is forced to
map up to 128MB of heap bitmap. Now the runtime can map sections of
the bitmap for just the parts of the address space used by the heap.

Updates #10460.

This slightly slows down the x/garbage and compilebench benchmarks.
However, I think the slowdown is acceptably small.

name        old time/op     new time/op     delta
Template        178ms ± 1%      180ms ± 1%  +0.78%    (p=0.029 n=10+10)
Unicode        85.7ms ± 2%     86.5ms ± 2%    ~       (p=0.089 n=10+10)
GoTypes         594ms ± 0%      599ms ± 1%  +0.70%    (p=0.000 n=9+9)
Compiler        2.86s ± 0%      2.87s ± 0%  +0.40%    (p=0.001 n=9+9)
SSA             7.23s ± 2%      7.29s ± 2%  +0.94%    (p=0.029 n=10+10)
Flate           116ms ± 1%      117ms ± 1%  +0.99%    (p=0.000 n=9+9)
GoParser        146ms ± 1%      146ms ± 0%    ~       (p=0.193 n=10+7)
Reflect         399ms ± 0%      403ms ± 1%  +0.89%    (p=0.001 n=10+10)
Tar             173ms ± 1%      174ms ± 1%  +0.91%    (p=0.013 n=10+9)
XML             208ms ± 1%      210ms ± 1%  +0.93%    (p=0.000 n=10+10)
[Geo mean]      368ms           371ms       +0.79%

name                       old time/op  new time/op  delta
Garbage/benchmem-MB=64-12  2.17ms ± 1%  2.21ms ± 1%  +2.15%  (p=0.000 n=20+20)

Change-Id: I037fd283221976f4f61249119d6b97b100bcbc66
Reviewed-on: https://go-review.googlesource.com/85883
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rick Hudson <rlh@golang.org>
7 years agoruntime: fix various contiguous bitmap assumptions
Austin Clements [Sat, 9 Dec 2017 03:24:59 +0000 (22:24 -0500)]
runtime: fix various contiguous bitmap assumptions

There are various places that assume the heap bitmap is contiguous and
scan it sequentially. We're about to split up the heap bitmap. This
commit modifies all of these except heapBitsSetType to use the
heapBits abstractions so they can transparently switch to a
discontiguous bitmap.

Updates #10460. This is a step toward supporting sparse heaps.

Change-Id: I2f3994a5785e4dccb66602fb3950bbd290d9392c
Reviewed-on: https://go-review.googlesource.com/85882
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rick Hudson <rlh@golang.org>
7 years agoruntime: lay out heap bitmap forward in memory
Austin Clements [Thu, 23 Jun 2016 20:25:50 +0000 (14:25 -0600)]
runtime: lay out heap bitmap forward in memory

Currently the heap bitamp is laid in reverse order in memory relative
to the heap itself. This was originally done out of "excessive
cleverness" so that computing a bitmap pointer could load only the
arena_start field and so that heaps could be more contiguous by
growing the arena and the bitmap out from a common center point.

However, this appears to have no actual performance benefit, it
complicates nearly every use of the bitmap, and it makes already
confusing code more confusing. Furthermore, it's still possible to use
a single field (the new bitmap_delta) for the bitmap pointer
computation by employing slightly different excessive cleverness.

Hence, this CL puts the bitmap into forward order.

This is a (very) updated version of CL 9404.

Change-Id: I743587cc626c4ecd81e660658bad85b54584108c
Reviewed-on: https://go-review.googlesource.com/85881
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rick Hudson <rlh@golang.org>
7 years agoruntime: use spanOf* more widely
Austin Clements [Mon, 4 Dec 2017 16:02:59 +0000 (11:02 -0500)]
runtime: use spanOf* more widely

The logic in the spanOf* functions is open-coded in a lot of places
right now. Replace these with calls to the spanOf* functions.

Change-Id: I3cc996aceb9a529b60fea7ec6fef22008c012978
Reviewed-on: https://go-review.googlesource.com/85880
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rick Hudson <rlh@golang.org>
7 years agoruntime: consolidate mheap.lookup* and spanOf*
Austin Clements [Mon, 4 Dec 2017 15:58:15 +0000 (10:58 -0500)]
runtime: consolidate mheap.lookup* and spanOf*

I think we'd forgotten about the mheap.lookup APIs when we introduced
spanOf*, but, at any rate, the spanOf* functions are used far more
widely at this point, so this CL eliminates the mheap.lookup*
functions in favor of spanOf*.

Change-Id: I15facd0856e238bb75d990e838a092b5bef5bdfc
Reviewed-on: https://go-review.googlesource.com/85879
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rick Hudson <rlh@golang.org>
7 years agoruntime: split object finding out of heapBitsForObject
Austin Clements [Tue, 12 Dec 2017 00:40:12 +0000 (19:40 -0500)]
runtime: split object finding out of heapBitsForObject

heapBitsForObject does two things: it finds the base of the object and
it creates the heapBits for the base of the object. There are several
places where we just care about the base of the object. Furthermore,
greyobject only needs the heapBits in the checkmark path and can
easily compute them only when needed. Once we eliminate passing the
heap bits to grayobject, almost all uses of heapBitsForObject don't
need the heap bits.

Hence, this splits heapBitsForObject into findObject and
heapBitsForAddr (the latter already exists), removes the hbits
argument to grayobject, and replaces all heapBitsForObject calls with
calls to findObject.

In addition to making things cleaner overall, heapBitsForAddr is going
to get more expensive shortly, so it's important that we don't do it
needlessly.

Note that there's an interesting performance pitfall here. I had
originally moved findObject to mheap.go, since it made more sense
there. However, that leads to a ~2% slow down and a whopping 11%
increase in L1 icache misses on both the x/garbage and compilebench
benchmarks. This suggests we may want to be more principled about
this, but, for now, let's just leave findObject in mbitmap.go.

(I tried to make findObject small enough to inline by splitting out
the error case, but, sadly, wasn't quite able to get it under the
inlining budget.)

Change-Id: I7bcb92f383ade565d22a9f2494e4c66fd513fb10
Reviewed-on: https://go-review.googlesource.com/85878
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rick Hudson <rlh@golang.org>
7 years agoruntime: replace mlookup and findObject with heapBitsForObject
Austin Clements [Mon, 4 Dec 2017 15:43:11 +0000 (10:43 -0500)]
runtime: replace mlookup and findObject with heapBitsForObject

These functions all serve essentially the same purpose. mlookup is
used in only one place and findObject in only three. Use
heapBitsForObject instead, which is the most optimized implementation.

(This may seem slightly silly because none of these uses care about
the heap bits, but we're about to split up the functionality of
heapBitsForObject anyway. At that point, findObject will rise from the
ashes.)

Change-Id: I906468c972be095dd23cf2404a7d4434e802f250
Reviewed-on: https://go-review.googlesource.com/85877
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rick Hudson <rlh@golang.org>
7 years agoruntime: validate lfnode addresses
Austin Clements [Mon, 4 Dec 2017 00:28:28 +0000 (19:28 -0500)]
runtime: validate lfnode addresses

Change-Id: Ic8c506289caaf6218494e5150d10002e0232feaa
Reviewed-on: https://go-review.googlesource.com/85876
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rick Hudson <rlh@golang.org>
7 years agoruntime: expand/update lfstack address space assumptions
Austin Clements [Sun, 3 Dec 2017 23:08:57 +0000 (18:08 -0500)]
runtime: expand/update lfstack address space assumptions

I was spelunking Linux's address space code and found that some of the
information about maximum virtual addresses in lfstack's comments was
out of date. This expands and updates the comment.

Change-Id: I9f54b23e6b266b3c5cc20259a849231fb751f6e7
Reviewed-on: https://go-review.googlesource.com/85875
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Rick Hudson <rlh@golang.org>
7 years agocmd/compile: improve absorb shifts optimization for arm64
Chad Rosier [Thu, 15 Feb 2018 19:49:03 +0000 (14:49 -0500)]
cmd/compile: improve absorb shifts optimization for arm64

Current absorb shifts optimization can generate dead Value nodes which increase
use count of other live nodes. It will impact other optimizations (such as
combined loads) which are enabled based on specific use count. This patch fixes
the issue by decreasing the use count of nodes referenced by dead Value nodes
generated by absorb shifts optimization.

Performance impacts on go1 benchmarks (data collected on A57@2GHzx8):

name                     old time/op    new time/op    delta
BinaryTree17-8              6.28s ± 2%     6.24s ± 1%     ~     (p=0.065 n=10+9)
Fannkuch11-8                6.32s ± 0%     6.33s ± 0%   +0.17%  (p=0.000 n=10+10)
FmtFprintfEmpty-8          98.9ns ± 0%    99.2ns ± 0%   +0.34%  (p=0.000 n=9+7)
FmtFprintfString-8          183ns ± 1%     182ns ± 1%   -1.01%  (p=0.005 n=9+10)
FmtFprintfInt-8             199ns ± 1%     202ns ± 1%   +1.41%  (p=0.000 n=10+9)
FmtFprintfIntInt-8          272ns ± 1%     276ns ± 3%   +1.36%  (p=0.015 n=10+10)
FmtFprintfPrefixedInt-8     367ns ± 1%     369ns ± 1%   +0.68%  (p=0.042 n=10+10)
FmtFprintfFloat-8           491ns ± 1%     493ns ± 1%     ~     (p=0.064 n=10+10)
FmtManyArgs-8              1.31µs ± 1%    1.32µs ± 1%   +0.39%  (p=0.042 n=8+9)
GobDecode-8                17.0ms ± 2%    16.2ms ± 2%   -4.74%  (p=0.000 n=10+10)
GobEncode-8                13.7ms ± 2%    13.4ms ± 1%   -2.40%  (p=0.000 n=10+9)
Gzip-8                      844ms ± 0%     737ms ± 0%  -12.70%  (p=0.000 n=10+10)
Gunzip-8                   84.4ms ± 1%    83.9ms ± 0%   -0.55%  (p=0.000 n=10+8)
HTTPClientServer-8          122µs ± 1%     124µs ± 1%   +1.75%  (p=0.000 n=10+9)
JSONEncode-8               34.9ms ± 1%    32.4ms ± 0%   -7.11%  (p=0.000 n=10+9)
JSONDecode-8                150ms ± 0%     146ms ± 1%   -2.84%  (p=0.000 n=7+10)
Mandelbrot200-8            10.0ms ± 0%    10.0ms ± 0%     ~     (p=0.529 n=10+10)
GoParse-8                  8.18ms ± 1%    8.03ms ± 0%   -1.93%  (p=0.000 n=10+10)
RegexpMatchEasy0_32-8       209ns ± 0%     209ns ± 0%     ~     (p=0.248 n=10+9)
RegexpMatchEasy0_1K-8       789ns ± 1%     790ns ± 0%     ~     (p=0.361 n=10+10)
RegexpMatchEasy1_32-8       202ns ± 0%     202ns ± 1%     ~     (p=0.137 n=8+10)
RegexpMatchEasy1_1K-8      1.12µs ± 2%    1.12µs ± 1%     ~     (p=0.810 n=10+10)
RegexpMatchMedium_32-8      298ns ± 0%     298ns ± 0%     ~     (p=0.443 n=10+9)
RegexpMatchMedium_1K-8     83.0µs ± 5%    78.6µs ± 0%   -5.37%  (p=0.000 n=10+10)
RegexpMatchHard_32-8       4.32µs ± 0%    4.26µs ± 0%   -1.47%  (p=0.000 n=10+10)
RegexpMatchHard_1K-8        132µs ± 4%     126µs ± 0%   -4.41%  (p=0.000 n=10+9)
Revcomp-8                   1.11s ± 0%     1.11s ± 0%   +0.14%  (p=0.017 n=10+9)
Template-8                  155ms ± 1%     155ms ± 1%     ~     (p=0.796 n=10+10)
TimeParse-8                 774ns ± 1%     785ns ± 1%   +1.41%  (p=0.001 n=10+10)
TimeFormat-8                788ns ± 1%     806ns ± 1%   +2.24%  (p=0.000 n=10+9)

name                     old speed      new speed      delta
GobDecode-8              45.2MB/s ± 2%  47.5MB/s ± 2%   +4.96%  (p=0.000 n=10+10)
GobEncode-8              56.0MB/s ± 2%  57.4MB/s ± 1%   +2.44%  (p=0.000 n=10+9)
Gzip-8                   23.0MB/s ± 0%  26.3MB/s ± 0%  +14.55%  (p=0.000 n=10+10)
Gunzip-8                  230MB/s ± 1%   231MB/s ± 0%   +0.55%  (p=0.000 n=10+8)
JSONEncode-8             55.6MB/s ± 1%  59.9MB/s ± 0%   +7.65%  (p=0.000 n=10+9)
JSONDecode-8             12.9MB/s ± 0%  13.3MB/s ± 1%   +2.94%  (p=0.000 n=7+10)
GoParse-8                7.08MB/s ± 1%  7.22MB/s ± 0%   +1.95%  (p=0.000 n=10+10)
RegexpMatchEasy0_32-8     153MB/s ± 0%   153MB/s ± 0%   -0.16%  (p=0.023 n=10+10)
RegexpMatchEasy0_1K-8    1.30GB/s ± 1%  1.30GB/s ± 0%     ~     (p=0.393 n=10+10)
RegexpMatchEasy1_32-8     158MB/s ± 0%   158MB/s ± 0%     ~     (p=0.684 n=10+10)
RegexpMatchEasy1_1K-8     915MB/s ± 2%   918MB/s ± 1%     ~     (p=0.796 n=10+10)
RegexpMatchMedium_32-8   3.35MB/s ± 0%  3.35MB/s ± 0%     ~     (p=1.000 n=10+9)
RegexpMatchMedium_1K-8   12.3MB/s ± 5%  13.0MB/s ± 0%   +5.56%  (p=0.000 n=10+10)
RegexpMatchHard_32-8     7.40MB/s ± 0%  7.51MB/s ± 0%   +1.50%  (p=0.000 n=10+10)
RegexpMatchHard_1K-8     7.75MB/s ± 4%  8.10MB/s ± 0%   +4.52%  (p=0.000 n=10+8)
Revcomp-8                 229MB/s ± 0%   228MB/s ± 0%   -0.14%  (p=0.017 n=10+9)
Template-8               12.5MB/s ± 1%  12.5MB/s ± 1%     ~     (p=0.780 n=10+10)

Change-Id: I103389f168eac79f6af44e8fef93acc2a7a4ac96
Reviewed-on: https://go-review.googlesource.com/88415
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
7 years agocompiler: honor //line directives in DWARF variable file/line attrs
Than McIntosh [Tue, 6 Feb 2018 14:36:13 +0000 (09:36 -0500)]
compiler: honor //line directives in DWARF variable file/line attrs

During DWARF debug generation, the DW_AT_decl_line / DW_AT_decl_file
attributes for variable DIEs were being computed without taking into
account the possibility of "//line" directives. Fix things up to use
the correct src.Pos methods to pick up this info.

Fixes #23704.

Change-Id: I88c21a0e0a9602392be229252d856a6d665868e2
Reviewed-on: https://go-review.googlesource.com/92255
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Heschi Kreinick <heschi@google.com>
7 years agointernal/trace: link user span start and end events
Hana Kim [Tue, 13 Feb 2018 22:18:08 +0000 (17:18 -0500)]
internal/trace: link user span start and end events

Also add testdata for version 1.11 including UserTaskSpan test trace.

Change-Id: I673fb29bb3aee96a14fadc0ab860d4f5832143f5
Reviewed-on: https://go-review.googlesource.com/93795
Reviewed-by: Heschi Kreinick <heschi@google.com>
7 years agocmd/cgo: delete double space in comment
zaq1tomo [Thu, 15 Feb 2018 16:12:24 +0000 (16:12 +0000)]
cmd/cgo: delete double space in comment

delete double space from comment

Change-Id: I71af5c1149941575016f79a91269f128b1fc16af
GitHub-Last-Rev: aba8874bd362d05d6c29c8647049369dfcd796f5
GitHub-Pull-Request: golang/go#23851
Reviewed-on: https://go-review.googlesource.com/94415
Reviewed-by: Ian Lance Taylor <iant@golang.org>
7 years agodebug/dwarf: formStrp uses a 64-bit value for 64-bit DWARF
Ian Lance Taylor [Sat, 16 Dec 2017 02:02:57 +0000 (18:02 -0800)]
debug/dwarf: formStrp uses a 64-bit value for 64-bit DWARF

No test as the only system I know that uses 64-bit DWARF is AIX.

Change-Id: I24e225253075be188845656b6778993c2d24ebf5
Reviewed-on: https://go-review.googlesource.com/84379
Run-TryBot: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
7 years agoruntime/trace: implement annotation API
Hana Kim [Tue, 5 Dec 2017 23:02:10 +0000 (18:02 -0500)]
runtime/trace: implement annotation API

This implements the annotation API proposed in golang.org/cl/63274.

traceString is updated to protect the string map with trace.stringsLock
because the assumption that traceString is called by a single goroutine
(either at the beginning of tracing and at the end of tracing when
dumping all the symbols and function names) is no longer true.

traceString is used by the annotation apis (NewContext, StartSpan, Log)
to register frequently appearing strings (task and span names, and log
keys) after this change.

NewContext -> one or two records (EvString, EvUserTaskCreate)
end function -> one record (EvUserTaskEnd)
StartSpan -> one or two records (EvString, EvUserSpan)
span end function -> one or two records (EvString, EvUserSpan)
Log -> one or two records (EvString, EvUserLog)

EvUserLog record is of the typical record format written by traceEvent
except that it is followed by bytes that represents the value string.

In addition to runtime/trace change, this change includes
corresponding changes in internal/trace to parse the new record types.

Future work to improve efficiency:
  More efficient unique task id generation instead of atomic. (per-P
  counter).
  Instead of a centralized trace.stringsLock, consider using per-P
  string cache or something more efficient.

R=go1.11

Change-Id: Iec9276c6c51e5be441ccd52dec270f1e3b153970
Reviewed-on: https://go-review.googlesource.com/71690
Reviewed-by: Austin Clements <austin@google.com>
7 years agoruntime/trace: user annotation API
Hana Kim [Thu, 9 Nov 2017 16:39:10 +0000 (11:39 -0500)]
runtime/trace: user annotation API

This CL presents the proposed user annotation API skeleton.
This CL bumps up the trace version to 1.11.

Design doc https://goo.gl/iqJfJ3

Implementation CLs are followed.

The API introduces three basic building blocks. Log, Span, and Task.

Log is for basic logging. When called, the message will be recorded
to the trace along with timestamp, goroutine id, and stack info.

   trace.Log(ctx, messageType message)

Span can be thought as an extension of log to record interesting
time interval during a goroutine's execution. A span is local to a
goroutine by definition.

   trace.WithSpan(ctx, "doVeryExpensiveOp", func(ctx context) {
      /* do something very expensive */
   })

Task is higher-level concept that aids tracing of complex operations
that encompass multiple goroutines or are asynchronous.
For example, an RPC request, a HTTP request, a file write, or a
batch job can be traced with a Task.

Note we chose to design the API around context.Context so it allows
easier integration with other tracing tools, often designed around
context.Context as well. Log and WithSpan APIs recognize the task
information embedded in the context and record it in the trace as
well. That allows the Go execution tracer to associate and group
the spans and log messages based on the task information.

In order to create a Task,

   ctx, end := trace.NewContext(ctx, "myTask")
   defer end()

The Go execution tracer measures the time between the task created
and the task ended for the task latency.

More discussion history in golang.org/cl/59572.

Update #16619

R=go1.11

Change-Id: I59a937048294dafd23a75cf1723c6db461b193cd
Reviewed-on: https://go-review.googlesource.com/63274
Reviewed-by: Austin Clements <austin@google.com>
7 years agocmd/compile: convert untyped bool for OIF and OFOR
Michael Fraenkel [Thu, 15 Feb 2018 17:28:48 +0000 (12:28 -0500)]
cmd/compile: convert untyped bool for OIF and OFOR

Updates #23834.

Change-Id: I92aca9108590a0c7de774f4fad7ded97105e3cb8
Reviewed-on: https://go-review.googlesource.com/94475
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
7 years agocmd/asm, cmd/internal/obj/ppc64: add Immediate Shifted opcodes for ppc64x
Carlos Eduardo Seo [Wed, 3 Jan 2018 19:55:40 +0000 (17:55 -0200)]
cmd/asm, cmd/internal/obj/ppc64: add Immediate Shifted opcodes for ppc64x

This change adds ADD/AND/OR/XOR Immediate Shifted instructions for
ppc64x so they are usable in Go asm code. These instructions were
originally present in asm9.go, but they were only usable in that
file (as -AADD, -AANDCC, -AOR, -AXOR). These old mnemonics are now
removed.

Updates #23845

Change-Id: Ifa2fac685e8bc628cb241dd446adfc3068181826
Reviewed-on: https://go-review.googlesource.com/94115
Reviewed-by: Lynn Boger <laboger@linux.vnet.ibm.com>
7 years agoall: drop support for Windows Vista or below (Windows XP)
Mikio Hara [Tue, 13 Feb 2018 20:33:15 +0000 (05:33 +0900)]
all: drop support for Windows Vista or below (Windows XP)

Per the notice in the Go 1.10 release notes, this change drops the
support for Windows Vista or below (including Windows XP) and
simplifies the code for the sake of maintenance.

There is one exception to the above. The code related to DLL and
system calls still remains in the runtime package. The remaining code
will be refined and used for supporting upcoming Windows versions in
future.

Updates #17245
Fixes #23072

Change-Id: I9e2821721f25ef9b83dfbf85be2b7ee5d9023aa5
Reviewed-on: https://go-review.googlesource.com/94255
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agonet, internal/poll, net/internal/socktest: set SOCK_{CLOEXEC,NONBLOCK} atomically...
Tobias Klauser [Thu, 15 Feb 2018 11:20:27 +0000 (12:20 +0100)]
net, internal/poll, net/internal/socktest: set SOCK_{CLOEXEC,NONBLOCK} atomically on NetBSD

NetBSD supports the SOCK_CLOEXEC and SOCK_NONBLOCK flags to the socket
syscall since version 6.0. The same version also introduced the paccept
syscall which can be used to implement syscall.Accept4.

Follows CL 40895

Change-Id: I9e4e1829b0382744c7799f4e58929a53b4e193f7
Reviewed-on: https://go-review.googlesource.com/94295
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Benny Siegert <bsiegert@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agocmd/compile/internal/syntax: don't assume (operator) ~ means operator ^
Robert Griesemer [Thu, 15 Feb 2018 04:54:28 +0000 (20:54 -0800)]
cmd/compile/internal/syntax: don't assume (operator) ~ means operator ^

The scanner assumed that ~ really meant ^, which may be helpful when
coming from C. But ~ is not a valid Go token, and pretending that it
should be ^ can lead to confusing error messages. Better to be upfront
about it and complain about the invalid character in the first place.

This was code "inherited" from the original yacc parser which was
derived from a C compiler. It's 10 years later and we can probably
assume that people are less confused about C and Go.

Fixes #23587.

Change-Id: I8d8f9b55b0dff009b75c1530d729bf9092c5aea6
Reviewed-on: https://go-review.googlesource.com/94160
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
7 years agocmd/vendor/github.com/google/pprof: refresh from upstream
Daniel Martí [Mon, 12 Feb 2018 16:34:48 +0000 (16:34 +0000)]
cmd/vendor/github.com/google/pprof: refresh from upstream

Updating to commit 0e0e5b7254e076a62326ab7305ba49e8515f0c91
from github.com/google/pprof

Recent modifications to the vendored pprof, such as skipping
TestWebInterface to avoid starting a web browser, have all been fixed
upstream.

Change-Id: I72e11108c438e1573bf2f9216e76d157378e8d45
Reviewed-on: https://go-review.googlesource.com/93375
Run-TryBot: Daniel Martí <mvdan@mvdan.cc>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agoruntime: move ELF structure definitions into own files
Tobias Klauser [Wed, 14 Feb 2018 13:26:32 +0000 (14:26 +0100)]
runtime: move ELF structure definitions into own files

Move the ELF32 and ELF64 structure definitions into their own files so
they can be reused when vDSO support is added for other architectures.

Change-Id: Id0171b4e5cea4add8635743c881e3bf3469597af
Reviewed-on: https://go-review.googlesource.com/93995
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
7 years agoruntime: speed up stack copying a little
Josh Bleecher Snyder [Sun, 1 Oct 2017 06:06:49 +0000 (23:06 -0700)]
runtime: speed up stack copying a little

Remove a branch and a stack spill.

name                old time/op  new time/op  delta
StackCopy-8         79.2ms ± 1%  79.1ms ± 2%    ~     (p=0.063 n=96+95)
StackCopyNoCache-8   121ms ± 1%   120ms ± 2%  -0.46%  (p=0.000 n=97+88)

Change-Id: Ifcbbb05d773178fad84cb11a9a6768ace69fcf24
Reviewed-on: https://go-review.googlesource.com/94029
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
7 years agoruntime: simplify amd64 memmove of 3/4 bytes
Josh Bleecher Snyder [Mon, 13 Nov 2017 06:11:33 +0000 (22:11 -0800)]
runtime: simplify amd64 memmove of 3/4 bytes

Change-Id: I132d3627ae301b68bf87eacb5bf41fd1ba2dcd91
Reviewed-on: https://go-review.googlesource.com/94025
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agoruntime: fix minor doc typos in amd64 memmove
Josh Bleecher Snyder [Mon, 13 Nov 2017 06:11:11 +0000 (22:11 -0800)]
runtime: fix minor doc typos in amd64 memmove

Change-Id: Ic1ce2f93d6a225699e9ce5307d62cdda8f97630d
Reviewed-on: https://go-review.googlesource.com/94024
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agoruntime: short-circuit typedslicecopy when dstp == srcp
Josh Bleecher Snyder [Mon, 13 Nov 2017 06:11:51 +0000 (22:11 -0800)]
runtime: short-circuit typedslicecopy when dstp == srcp

If copying from a slice to itself, skip the write barriers
and actual memory copies.

This happens in practice in code like this snippet from
the trim pass in the compiler, when k ends up being 0:

copy(s.Values[k:], s.Values[:m])

Change-Id: Ie6924acfd56151f874d87f1d7f1f74320b4c4f10
Reviewed-on: https://go-review.googlesource.com/94023
Run-TryBot: Josh Bleecher Snyder <josharian@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
7 years agoruntime: use new instead of newobject to create hmap in makemap
Martin Möhrmann [Sat, 3 Feb 2018 15:29:54 +0000 (16:29 +0100)]
runtime: use new instead of newobject to create hmap in makemap

The runtime.hmap type is known at compile time.
Using new(hmap) avoids loading the hmap type from the maptype
supplied as an argument to makemap which is only known at runtime.

This change makes makemap consistent with makemap_small
by using new(hmap) instead of newobject in both functions.

Change-Id: Ia47acfda527e8a71d15a1a7a4c2b54fb923515eb
Reviewed-on: https://go-review.googlesource.com/91775
Run-TryBot: Martin Möhrmann <moehrmann@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
7 years agoruntime: improve test file naming
Martin Möhrmann [Mon, 29 Jan 2018 21:57:54 +0000 (22:57 +0100)]
runtime: improve test file naming

The runtime builtin functions that are tested in append_test.go
are defined in slice.go. Renaming the test file to slice_test.go
makes this relation explicit with a common file name prefix.

Change-Id: I2f89ec23a6077fe6b80d2161efc760df828c8cd4
Reviewed-on: https://go-review.googlesource.com/90655
Run-TryBot: Martin Möhrmann <moehrmann@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
7 years agocmd/go: add options to security whitelist
Ian Lance Taylor [Wed, 14 Feb 2018 01:43:16 +0000 (17:43 -0800)]
cmd/go: add options to security whitelist

Also permit passing flags to pkg-config, as we used to.

Also change the error message to refer to https://golang.org/s/invalidflag.

Fixes #23749

Change-Id: I3fbeb4c346610e6fd55e8720e720b0a40e352ab5
Reviewed-on: https://go-review.googlesource.com/93836
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agodoc: remove draft notice from Go 1.10 release notes
Ian Lance Taylor [Thu, 15 Feb 2018 02:08:19 +0000 (18:08 -0800)]
doc: remove draft notice from Go 1.10 release notes

Change-Id: I063b556b59ecb8a8f4a3f121c32982b90eee2d74
Reviewed-on: https://go-review.googlesource.com/94159
Reviewed-by: Andrew Bonventre <andybons@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agotest: add new test that gccgo failed to compile
Ian Lance Taylor [Fri, 2 Feb 2018 05:38:49 +0000 (21:38 -0800)]
test: add new test that gccgo failed to compile

Updates #22305

Change-Id: I0e6bbd880599fc1b70d0378b746d162d4a846c65
Reviewed-on: https://go-review.googlesource.com/91556
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agocmd/compile/internal/syntax: fix syntax.Parse doc string, improved tests
Robert Griesemer [Thu, 15 Feb 2018 00:59:36 +0000 (16:59 -0800)]
cmd/compile/internal/syntax: fix syntax.Parse doc string, improved tests

1) Fix the doc string for syntax.Parse: The returned AST is
always nil if there was an error and an error handler is missing.

2) Adjust the syntax Print and Dump tests such that they print and
dump the AST even in the presence of errors.

Change-Id: If658eabdcc83f578d815070bc65d1a5f6cfaddfc
Reviewed-on: https://go-review.googlesource.com/94157
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
7 years agocmd/compile/internal/syntax: more tolerant handling of missing function invocation...
Robert Griesemer [Thu, 15 Feb 2018 00:57:28 +0000 (16:57 -0800)]
cmd/compile/internal/syntax: more tolerant handling of missing function invocation in go/defer

Assume that an expression that is not a function call in a defer/go
statement is indeed a function that is just missing its invocation.
Report the error but continue with a sane syntax tree.

Fixes #23586.

Change-Id: Ib45ebac57c83b3e39ae4a1b137ffa291dec5b50d
Reviewed-on: https://go-review.googlesource.com/94156
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
7 years agocmd/compile: fix typechecking of untyped boolean expressions
Matthew Dempsky [Wed, 14 Feb 2018 22:08:17 +0000 (14:08 -0800)]
cmd/compile: fix typechecking of untyped boolean expressions

Previously, if we typechecked a statement like

    var x bool = p1.f == p2.f && p1.g == p2.g

we would correctly update the '&&' node's type from 'untyped bool' to
'bool', but the '==' nodes would stay 'untyped bool'. This is
inconsistent, and caused consistency checks during walk to fail.

This CL doesn't pass toolstash because it seems to slightly affect the
register allocator's heuristics. (Presumably 'untyped bool's were
previously making it all the way through SSA?)

Fixes #23414.

Change-Id: Ia85f8cfc69b5ba35dfeb157f4edf57612ecc3285
Reviewed-on: https://go-review.googlesource.com/94022
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
7 years agocmd/compile: allow converting defined string types to []rune
Kunpei Sakai [Sat, 13 Jan 2018 19:05:21 +0000 (04:05 +0900)]
cmd/compile: allow converting defined string types to []rune

Fixes #23298

Change-Id: I107c6f3a80db83f063c0daf262c6e7f7492e4d4c
Reviewed-on: https://go-review.googlesource.com/87695
Run-TryBot: Kunpei Sakai <namusyaka@gmail.com>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
7 years agocmd/compile/internal/syntax: follow Go naming conventions for error methods
Robert Griesemer [Wed, 14 Feb 2018 23:22:44 +0000 (15:22 -0800)]
cmd/compile/internal/syntax: follow Go naming conventions for error methods

Also, remove parser.error method (in favor of parser.errorAt) as it's only
used twice.

This is a purely cosmetic change.

Change-Id: Idb3b8b50f1c2e4d10de2ffb1c1184ceba8f7de8a
Reviewed-on: https://go-review.googlesource.com/94030
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
7 years agocmd/compile: fix constant folding of right shifts on s390x
Keith Randall [Wed, 14 Feb 2018 22:21:31 +0000 (14:21 -0800)]
cmd/compile: fix constant folding of right shifts on s390x

Repeat previous fix on amd64 for s390x.
Sub-word right shifts should sign extend before shifting.

Update #23812

Change-Id: I2d770190c7d8a22310b0dbd9facb3fb05afa362a
Reviewed-on: https://go-review.googlesource.com/94028
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agocmd/compile: fix printing of untyped types in -W output
Matthew Dempsky [Wed, 14 Feb 2018 21:51:51 +0000 (13:51 -0800)]
cmd/compile: fix printing of untyped types in -W output

It's always useful to distinguish "bool" and "string" from "untyped
bool" and "untyped string", so change typefmt to do this
unconditionally.

Also, while here, replace a bare 0 with its named constant FErr.

Fixes #23833.

Change-Id: I3fcb8d7204686937439caaaf8b3973fc236d0387
Reviewed-on: https://go-review.googlesource.com/94021
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agocmd/compile: fix go:uintptrescapes tag for unnamed parameters
Cherry Zhang [Fri, 8 Dec 2017 03:01:42 +0000 (22:01 -0500)]
cmd/compile: fix go:uintptrescapes tag for unnamed parameters

The tag was overwritten by the code for special handling unnamed
parameters.

Fixes #23045.

Change-Id: Ie2e1db3e902a07a2bbbc2a3424cea300f0a42cc3
Reviewed-on: https://go-review.googlesource.com/82775
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
7 years agoencoding/json: make error capture logic in recover more type safe
Joe Tsai [Wed, 6 Dec 2017 06:53:48 +0000 (22:53 -0800)]
encoding/json: make error capture logic in recover more type safe

Rather than only ignoring runtime.Error panics, which are a very
narrow set of possible panic values, switch it such that the json
package only captures panic values that have been properly wrapped
in a jsonError struct. This ensures that only intentional panics
originating from the json package are captured.

Fixes #23012

Change-Id: I5e85200259edd2abb1b0512ce6cc288849151a6d
Reviewed-on: https://go-review.googlesource.com/94019
Run-TryBot: Joe Tsai <thebrokentoaster@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agogo/types: make gotype continue after syntax errors
Robert Griesemer [Wed, 14 Feb 2018 00:37:51 +0000 (16:37 -0800)]
go/types: make gotype continue after syntax errors

This avoids odd behavior where sometimes a lot of useful
errors are not reported simply because of a small syntax
error.

Tested manually with non-existing files. (We cannot easily
add an automatic test because this is a stand-alone binary
in this directory that must be built manually.)

Fixes #23593.

Change-Id: Iff90f95413bed7d1023fa0a5c9eb0414144428a9
Reviewed-on: https://go-review.googlesource.com/93815
Reviewed-by: Alan Donovan <adonovan@google.com>
7 years agocmd/compile: constant fold !true and !false
Keith Randall [Mon, 22 Jan 2018 18:06:10 +0000 (10:06 -0800)]
cmd/compile: constant fold !true and !false

Constant fold Not of boolean constants.

Noticed while working on #23504.

Change-Id: I965705154ee7348a1a159fad4e029b922d3171b3
Reviewed-on: https://go-review.googlesource.com/88956
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Martin Möhrmann <moehrmann@google.com>
7 years agotest: add test case for incorrect gccgo compilation error
Ian Lance Taylor [Thu, 25 Jan 2018 18:33:00 +0000 (10:33 -0800)]
test: add test case for incorrect gccgo compilation error

Updates #23489

Change-Id: Ie846ccfe4c4d9295857f5da6863ac8f2ac0f2f6a
Reviewed-on: https://go-review.googlesource.com/89935
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agotest: add a test that gccgo miscompiled
Ian Lance Taylor [Fri, 2 Feb 2018 19:16:56 +0000 (11:16 -0800)]
test: add a test that gccgo miscompiled

Updates #20923

Change-Id: Ia1210ea3dec39e5db2521aeafca24d6e731f0c93
Reviewed-on: https://go-review.googlesource.com/91657
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agotest: add test for rounding to positive zero
Ian Lance Taylor [Mon, 5 Feb 2018 03:15:18 +0000 (19:15 -0800)]
test: add test for rounding to positive zero

Per the language spec clarification in https://golang.org/cl/14727.

Updates #12576
Updates #12621

Change-Id: I1e459c3c11a571bd29582761faacaa9ca3178ba6
Reviewed-on: https://go-review.googlesource.com/91895
Run-TryBot: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Robert Griesemer <gri@golang.org>
Reviewed-by: Robert Griesemer <gri@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

7 years agocmd/compile/internal/ssa: don't spill register offsets on amd64
Ilya Tocar [Mon, 4 Dec 2017 20:24:16 +0000 (14:24 -0600)]
cmd/compile/internal/ssa: don't spill register offsets on amd64

Transform (ADDQconst SP) into (LEA SP), because lea is rematerializeable,
so this avoids register spill. We can't mark ADDQconst as rematerializeable,
because it clobbers flags. This makes go binary ~2kb smaller.

For reference here is generated code for function from bug report.
Before:
        CALL    "".g(SB)
        MOVBLZX (SP), AX
        LEAQ    8(SP), DI
        TESTB   AX, AX
        JEQ     15
        MOVQ    "".p(SP), SI
        DUFFCOPY        $196
        MOVQ    $0, (SP)
        PCDATA  $0, $1
        CALL    "".h(SB)
        RET
        MOVQ    DI, ""..autotmp_2-8(SP) // extra spill
        PCDATA  $0, $2
        CALL    "".g(SB)
        MOVQ    ""..autotmp_2-8(SP), DI // extra register fill
        MOVQ    "".p(SP), SI
        DUFFCOPY        $196
        MOVQ    $1, (SP)
        PCDATA  $0, $1
        CALL    "".h(SB)
        JMP     14
        END

After:
        CALL    "".g(SB)
        MOVBLZX (SP), AX
        TESTB   AX, AX
        JEQ     15
        LEAQ    8(SP), DI
        MOVQ    "".p(SP), SI
        DUFFCOPY        $196
        MOVQ    $0, (SP)
        PCDATA  $0, $1
        CALL    "".h(SB)
        RET
        PCDATA  $0, $0  // no spill
        CALL    "".g(SB)
        LEAQ    8(SP), DI // rematerialized instead
        MOVQ    "".p(SP), SI
        DUFFCOPY        $196
        MOVQ    $1, (SP)
        PCDATA  $0, $1
        CALL    "".h(SB)
        JMP     14
        END

Fixes #22947

Change-Id: I8f33b860dc6c8828373477171b172ca2ce30074f
Reviewed-on: https://go-review.googlesource.com/81815
Run-TryBot: Ilya Tocar <ilya.tocar@intel.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
7 years agocmd/compile/internal/amd64: update popcnt code generation
Ilya Tocar [Thu, 7 Dec 2017 16:56:45 +0000 (10:56 -0600)]
cmd/compile/internal/amd64: update popcnt code generation

Popcnt has false dependency on output register and generates
MOVQ $0, reg to break it. But recently we switched MOVQ $0, reg
encoding from xor reg, reg  to actual mov $0, reg. This CL updates
code generation for popcnt to use actual XOR.

Change-Id: I4c1fc11e85758b53ba2679165fa55614ec54b27d
Reviewed-on: https://go-review.googlesource.com/82516
Run-TryBot: Ilya Tocar <ilya.tocar@intel.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
7 years agocmd/compile/internal: pass LocalSlot values, not pointers
Heschi Kreinick [Tue, 30 Jan 2018 00:21:25 +0000 (19:21 -0500)]
cmd/compile/internal: pass LocalSlot values, not pointers

Because getStackOffset is a function pointer, the compiler assumes that
its arguments escape. Pass a value instead to avoid heap allocations.

Change-Id: Ib94e5941847f134cd00e873040a4d7fcf15ced26
Reviewed-on: https://go-review.googlesource.com/92397
Run-TryBot: Heschi Kreinick <heschi@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Chase <drchase@google.com>
7 years agocmd/compile/internal: reuse memory for valueToProgAfter
Heschi Kreinick [Mon, 29 Jan 2018 22:01:41 +0000 (17:01 -0500)]
cmd/compile/internal: reuse memory for valueToProgAfter

Not a big improvement, but does help edge cases like the SSA package.
Change-Id: I40e531110b97efd5f45955be477fd0f4faa8d545
Reviewed-on: https://go-review.googlesource.com/92396
Run-TryBot: Heschi Kreinick <heschi@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Chase <drchase@google.com>
7 years agocmd/compile/internal/ssa: use math/bits for register sets
Heschi Kreinick [Mon, 29 Jan 2018 21:09:11 +0000 (16:09 -0500)]
cmd/compile/internal/ssa: use math/bits for register sets

Using bits.TrailingZeroes instead of iterating over each bit is a small
but easy win for the common case of only one or two registers being set.

I copied in the implementation for use with pre-1.9 bootstraps.

Change-Id: Ieaa768554d7d5239a5617fbf34f1ee0b32ce1de5
Reviewed-on: https://go-review.googlesource.com/92395
Run-TryBot: Heschi Kreinick <heschi@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Chase <drchase@google.com>
7 years agocmd/compile/internal/ssa: reduce location list memory use
Heschi Kreinick [Mon, 22 Jan 2018 22:06:26 +0000 (17:06 -0500)]
cmd/compile/internal/ssa: reduce location list memory use

Put everything that showed up in the allocation profile into the cache,
and reuse it across functions.

After this CL, the overhead of enabling location lists is getting
pretty close to the desired 5%:

compilecmp -all -beforeflags -dwarflocationlists=0 -afterflags -dwarflocationlists=1 -n 30 4ebad42292b6a4090faf37753dd768d2965e38c4 4ebad42292b6a4090faf37753dd768d2965e38c4
compilecmp  -dwarflocationlists=0 4ebad42292b6a4090faf37753dd768d2965e38c4  -dwarflocationlists=1 4ebad42292b6a4090faf37753dd768d2965e38c4
benchstat -geomean  /tmp/869550129 /tmp/143495132
completed   30 of   30, estimated time remaining 0s (eta 3:24PM)
name        old time/op       new time/op       delta
Template          199ms ± 4%        209ms ± 6%   +5.17%  (p=0.000 n=29+30)
Unicode          99.2ms ± 8%      100.5ms ± 6%     ~     (p=0.112 n=30+30)
GoTypes           642ms ± 3%        684ms ± 3%   +6.54%  (p=0.000 n=29+30)
SSA               8.00s ± 1%        8.71s ± 1%   +8.78%  (p=0.000 n=29+29)
Flate             129ms ± 7%        134ms ± 5%   +3.77%  (p=0.000 n=30+30)
GoParser          157ms ± 4%        164ms ± 5%   +4.35%  (p=0.000 n=29+30)
Reflect           428ms ± 3%        450ms ± 4%   +5.09%  (p=0.000 n=30+30)
Tar               195ms ± 5%        204ms ± 8%   +4.78%  (p=0.000 n=30+30)
XML               228ms ± 4%        241ms ± 4%   +5.62%  (p=0.000 n=30+29)
StdCmd            15.4s ± 1%        16.7s ± 1%   +8.29%  (p=0.000 n=29+29)
[Geo mean]        476ms             502ms        +5.35%

name        old user-time/op  new user-time/op  delta
Template          294ms ±18%        304ms ±15%     ~     (p=0.242 n=29+29)
Unicode           182ms ±27%        172ms ±28%     ~     (p=0.104 n=30+30)
GoTypes           957ms ±15%       1016ms ±12%   +6.16%  (p=0.000 n=30+30)
SSA               13.3s ± 5%        14.3s ± 3%   +7.32%  (p=0.000 n=30+28)
Flate             188ms ±17%        193ms ±17%     ~     (p=0.288 n=28+29)
GoParser          232ms ±16%        238ms ±13%     ~     (p=0.065 n=30+29)
Reflect           585ms ±13%        620ms ±10%   +5.88%  (p=0.000 n=30+30)
Tar               298ms ±21%        332ms ±23%  +11.32%  (p=0.000 n=30+30)
XML               329ms ±17%        343ms ±12%   +4.18%  (p=0.032 n=30+30)
[Geo mean]        492ms             513ms        +4.13%

name        old alloc/op      new alloc/op      delta
Template         38.3MB ± 0%       40.3MB ± 0%   +5.29%  (p=0.000 n=30+30)
Unicode          29.3MB ± 0%       29.6MB ± 0%   +1.28%  (p=0.000 n=30+29)
GoTypes           110MB ± 0%        118MB ± 0%   +6.97%  (p=0.000 n=29+30)
SSA              1.48GB ± 0%       1.61GB ± 0%   +9.06%  (p=0.000 n=30+30)
Flate            24.8MB ± 0%       26.0MB ± 0%   +4.99%  (p=0.000 n=29+30)
GoParser         30.9MB ± 0%       32.2MB ± 0%   +4.20%  (p=0.000 n=30+30)
Reflect          76.8MB ± 0%       80.6MB ± 0%   +4.97%  (p=0.000 n=30+30)
Tar              39.6MB ± 0%       41.7MB ± 0%   +5.22%  (p=0.000 n=29+30)
XML              42.0MB ± 0%       45.4MB ± 0%   +8.22%  (p=0.000 n=29+30)
[Geo mean]       63.9MB            67.5MB        +5.56%

name        old allocs/op     new allocs/op     delta
Template           383k ± 0%         405k ± 0%   +5.69%  (p=0.000 n=30+30)
Unicode            343k ± 0%         346k ± 0%   +0.98%  (p=0.000 n=30+27)
GoTypes           1.15M ± 0%        1.22M ± 0%   +6.17%  (p=0.000 n=29+29)
SSA               12.2M ± 0%        13.2M ± 0%   +8.15%  (p=0.000 n=30+30)
Flate              234k ± 0%         249k ± 0%   +6.44%  (p=0.000 n=30+30)
GoParser           315k ± 0%         332k ± 0%   +5.31%  (p=0.000 n=30+28)
Reflect            972k ± 0%        1010k ± 0%   +3.89%  (p=0.000 n=30+30)
Tar                394k ± 0%         415k ± 0%   +5.35%  (p=0.000 n=28+30)
XML                404k ± 0%         429k ± 0%   +6.31%  (p=0.000 n=29+29)
[Geo mean]         651k              686k        +5.35%

Change-Id: Ia005a8d6b33ce9f8091322f004376a3d6e5c1a94
Reviewed-on: https://go-review.googlesource.com/89357
Run-TryBot: Heschi Kreinick <heschi@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Chase <drchase@google.com>
7 years agocmd/compile: reimplement location list generation
Heschi Kreinick [Thu, 26 Oct 2017 19:40:17 +0000 (15:40 -0400)]
cmd/compile: reimplement location list generation

Completely redesign and reimplement location list generation to be more
efficient, and hopefully not too hard to understand.

RegKills are gone. Instead of using the regalloc's liveness
calculations, redo them using the Ops' clobber information. Besides
saving a lot of Values, this avoids adding RegKills to blocks that would
be empty otherwise, which was messing up optimizations. This does mean
that it's much harder to tell whether the generation process is buggy
(there's nothing to cross-check it with), and there may be disagreements
with GC liveness. But the performance gain is significant, and it's nice
not to be messing with earlier compiler phases.

The intermediate representations are gone. Instead of producing
ssa.BlockDebugs, then dwarf.LocationLists, and then finally real
location lists, go directly from the SSA to a (mostly) real location
list. Because the SSA analysis happens before assembly, it stores
encoded block/value IDs where PCs would normally go. It would be easier
to do the SSA analysis after assembly, but I didn't want to retain the
SSA just for that.

Generation proceeds in two phases: first, it traverses the function in
CFG order, storing the state of the block at the beginning and end. End
states are used to produce the start states of the successor blocks. In
the second phase, it traverses in program text order and produces the
location lists. The processing in the second phase is redundant, but
much cheaper than storing the intermediate representation. It might be
possible to combine the two phases somewhat to take advantage of cases
where the CFG matches the block layout, but I haven't tried.

Location lists are finalized by adding a base address selection entry,
translating each encoded block/value ID to a real PC, and adding the
terminating zero entry. This probably won't work on OSX, where dsymutil
will choke on the base address selection. I tried emitting CU-relative
relocations for each address, and it was *very* bad for performance --
it uses more memory storing all the relocations than it does for the
actual location list bytes. I think I'm going to end up synthesizing the
relocations in the linker only on OSX, but TBD.

TestNexting needs updating: with more optimizations working, the
debugger doesn't stop on the continue (line 88) any more, and the test's
duplicate suppression kicks in. Also, dx and dy live a little longer
now, but they have the correct values.

Change-Id: Ie772dfe23a4e389ca573624fac4d05401ae32307
Reviewed-on: https://go-review.googlesource.com/89356
Run-TryBot: Heschi Kreinick <heschi@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Chase <drchase@google.com>
7 years agocmd/compile/internal: decouple scope tracking from location lists
Heschi Kreinick [Tue, 23 Jan 2018 19:10:08 +0000 (14:10 -0500)]
cmd/compile/internal: decouple scope tracking from location lists

We're trying to enable location lists by default, and it's easier to do
that if we don't have to worry about scope tracking at the same time.
We can evaluate their performance impact separately.

However, that does mean that "err" is ambiguous in the test case, so
rename it to err2 for now.

Change-Id: I24f119016185c52b7d9affc74207f6a5b450fb6f
Reviewed-on: https://go-review.googlesource.com/89355
Run-TryBot: Heschi Kreinick <heschi@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Chase <drchase@google.com>
7 years agoruntime: use private futexes on Linux
Ian Lance Taylor [Mon, 27 Nov 2017 23:40:28 +0000 (15:40 -0800)]
runtime: use private futexes on Linux

By default futexes are permitted in shared memory regions, which
requires the kernel to translate the memory address. Since our futexes
are never in shared memory, set FUTEX_PRIVATE_FLAG, which makes futex
operations slightly more efficient.

Change-Id: I2a82365ed27d5cd8d53c5382ebaca1a720a80952
Reviewed-on: https://go-review.googlesource.com/80144
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: David Crawshaw <crawshaw@golang.org>
7 years agocmd/asm: add PRFM instruction on ARM64
fanzha02 [Thu, 30 Nov 2017 08:30:53 +0000 (08:30 +0000)]
cmd/asm: add PRFM instruction on ARM64

The current assembler cannot handle PRFM(immediate) instruciton.
The fix creates a prfopfield struct that contains the eight
prefetch operations and the value to use in instruction. And add
the test cases.

Fixes #22932

Change-Id: I621d611bd930ef3c42306a4372447c46d53b2ccf
Reviewed-on: https://go-review.googlesource.com/81675
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
7 years agocmd/internal/obj/mips: support NEG, avoid crash with illegal instruction
Cherry Zhang [Thu, 25 Jan 2018 16:46:06 +0000 (11:46 -0500)]
cmd/internal/obj/mips: support NEG, avoid crash with illegal instruction

Add support of NEG{V,W} pseudo-instructions, which are translated
to a SUB instruction from R0 with proper width.

Also turn illegal instruction to UNDEF, to avoid crashing in
asmout when it tries to read the operands.

Fixes #23548.

Change-Id: I047b27559ccd9594c3dcf62ab039b636098f30a3
Reviewed-on: https://go-review.googlesource.com/89896
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Chase <drchase@google.com>
7 years agonacl*.bash: pass flags to make.bash
Cherry Zhang [Sat, 10 Feb 2018 20:32:28 +0000 (15:32 -0500)]
nacl*.bash: pass flags to make.bash

Just like all.bash passes flags to make.bash, I think it makes
sense that naclmake.bash and nacltest.bash do so as well. For
example, on a slow machine I can do "./nacltest.bash -v" to see
the build progress.

Change-Id: Id766dd590e6b83e8b5345822580dc1b05eac8ea3
Reviewed-on: https://go-review.googlesource.com/93117
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
7 years agocmd/compile: CALLudiv on nacl/arm doesn't clobber R12
Cherry Zhang [Fri, 9 Feb 2018 17:47:33 +0000 (12:47 -0500)]
cmd/compile: CALLudiv on nacl/arm doesn't clobber R12

On nacl/arm, R12 is clobbered by the RET instruction in function
that has a frame. runtime.udiv doesn't have a frame, so it does
not clobber R12.

Change-Id: I0de448749f615908f6659e92d201ba3eb2f8266d
Reviewed-on: https://go-review.googlesource.com/93116
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
7 years agoruntime/internal/atomic: add early nil check on ARM
Cherry Zhang [Mon, 12 Feb 2018 22:00:01 +0000 (17:00 -0500)]
runtime/internal/atomic: add early nil check on ARM

If nil, fault before taking the lock or calling into the kernel.

Change-Id: I013d78a5f9233c2a9197660025f679940655d384
Reviewed-on: https://go-review.googlesource.com/93636
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
7 years agoruntime/internal/atomic: unify sys_*_arm.s on non-linux
Cherry Zhang [Mon, 12 Feb 2018 19:27:46 +0000 (14:27 -0500)]
runtime/internal/atomic: unify sys_*_arm.s on non-linux

Updates #23778.

Change-Id: I80e57a15b6e3bbc2e25ea186399ff0e360fc5c21
Reviewed-on: https://go-review.googlesource.com/93635
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
7 years agocmd/compile: replace range loop over list of nodes with orderexprlistinplace
Martin Möhrmann [Wed, 17 Jan 2018 20:00:26 +0000 (21:00 +0100)]
cmd/compile: replace range loop over list of nodes with orderexprlistinplace

Replace explicit range loop that applies orderexprinplace on a
list of nodes with existing helper function orderexprlistinplace.

Passes toolstash -cmp.

Change-Id: Ic8098ed08cf67f319de3faa83b00a5b73bbde95d
Reviewed-on: https://go-review.googlesource.com/88815
Reviewed-by: Daniel Martí <mvdan@mvdan.cc>
Run-TryBot: Daniel Martí <mvdan@mvdan.cc>
TryBot-Result: Gobot Gobot <gobot@golang.org>

7 years agogithub: add a Pull Request template
Andrew Bonventre [Wed, 14 Feb 2018 05:27:10 +0000 (00:27 -0500)]
github: add a Pull Request template

Change-Id: I02938b2435e3a98efea7ee5545a6f8f5f6f794b4
Reviewed-on: https://go-review.googlesource.com/93915
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agocmd/compile: generate tbz/tbnz when comparing against zero on arm64
Chad Rosier [Wed, 7 Feb 2018 20:37:33 +0000 (15:37 -0500)]
cmd/compile: generate tbz/tbnz when comparing against zero on arm64

The tbz/tbnz checks the sign bit to determine if the value is >= 0 or < 0.

go1 benchmark results:
name                   old speed      new speed      delta
JSONEncode             94.4MB/s ± 1%  95.7MB/s ± 0%  +1.36%  (p=0.000 n=10+9)
JSONDecode             19.7MB/s ± 1%  19.9MB/s ± 1%  +1.08%  (p=0.000 n=9+10)
Gzip                   45.5MB/s ± 0%  46.0MB/s ± 0%  +1.06%  (p=0.000 n=10+10)
Revcomp                 376MB/s ± 0%   379MB/s ± 0%  +0.69%  (p=0.000 n=10+10)
RegexpMatchHard_1K     12.6MB/s ± 0%  12.7MB/s ± 0%  +0.57%  (p=0.000 n=10+8)
RegexpMatchMedium_32   3.21MB/s ± 0%  3.22MB/s ± 0%  +0.31%  (p=0.000 n=9+10)
RegexpMatchEasy1_1K    1.27GB/s ± 0%  1.27GB/s ± 0%  +0.23%  (p=0.000 n=9+9)
RegexpMatchHard_32     11.4MB/s ± 0%  11.4MB/s ± 1%  +0.19%  (p=0.036 n=10+8)
RegexpMatchEasy0_1K    1.77GB/s ± 0%  1.77GB/s ± 0%  +0.13%  (p=0.000 n=9+10)
RegexpMatchMedium_1K   19.3MB/s ± 0%  19.3MB/s ± 0%  +0.04%  (p=0.008 n=10+8)
RegexpMatchEasy0_32     131MB/s ± 0%   131MB/s ± 0%    ~     (p=0.211 n=10+10)
GobDecode              57.5MB/s ± 1%  57.6MB/s ± 2%    ~     (p=0.469 n=10+10)
GobEncode              58.6MB/s ± 1%  58.5MB/s ± 2%    ~     (p=0.781 n=10+10)
GoParse                9.40MB/s ± 0%  9.39MB/s ± 0%  -0.19%  (p=0.005 n=10+9)
RegexpMatchEasy1_32     133MB/s ± 0%   133MB/s ± 0%  -0.48%  (p=0.000 n=10+10)
Template               20.9MB/s ± 0%  20.6MB/s ± 0%  -1.54%  (p=0.000 n=8+10)

Change-Id: I411efe44db35c3962445618d5a47c12e31b3925b
Reviewed-on: https://go-review.googlesource.com/92715
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
7 years agosyscall, os: use pipe2 syscall on NetBSD instead of pipe
Tobias Klauser [Wed, 14 Feb 2018 14:56:39 +0000 (14:56 +0000)]
syscall, os: use pipe2 syscall on NetBSD instead of pipe

The pipe2 syscall is part of NetBSD since version 6.0 and thus exists in
all officially supported versions (6.0 through 6.1 and 7.0+).

Follows CL 38426

Change-Id: I7b62b507300c3dfbcc6ae56408a7d7088ddccc77
Reviewed-on: https://go-review.googlesource.com/94035
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Benny Siegert <bsiegert@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

7 years agocmd/go: put "go help" list in the right order, take 2
Nate Wilkinson [Mon, 12 Feb 2018 17:33:27 +0000 (10:33 -0700)]
cmd/go: put "go help" list in the right order, take 2

The previous fix had "bug" and "build" in the wrong order.

Fixes #23791

Change-Id: I4897428516b159966c13c1054574c4f6fbf0fbac
Reviewed-on: https://go-review.googlesource.com/94017
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agosync: enable profiling of RWMutex
Lorenz Bauer [Wed, 10 Jan 2018 10:50:19 +0000 (10:50 +0000)]
sync: enable profiling of RWMutex

Include reader / writer interactions of RWMutex in the mutex profile.
Writer contention is already included in the profile, since a plain Mutex
is used to control exclusion.

Fixes #18496

Change-Id: Ib0dc1ffa0fd5e6d964a6f7764d7f09556eb63f00
Reviewed-on: https://go-review.googlesource.com/87095
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Peter Weinberger <pjw@google.com>
7 years agocrypto/cipher: add NewGCMWithNonceAndTagSize for custom tag sizes.
Conrado Gouvea [Fri, 14 Jul 2017 01:14:37 +0000 (22:14 -0300)]
crypto/cipher: add NewGCMWithNonceAndTagSize for custom tag sizes.

GCM allows using tag sizes smaller than the block size. This adds a
NewGCMWithNonceAndTagSize function which allows specifying the tag
size.

Fixes #19594

Change-Id: Ib2008c6f13ad6d916638b1523c0ded8a80eaf42d
Reviewed-on: https://go-review.googlesource.com/48510
Reviewed-by: Filippo Valsorda <hi@filippo.io>
Run-TryBot: Filippo Valsorda <hi@filippo.io>
TryBot-Result: Gobot Gobot <gobot@golang.org>

7 years agocrypto/rsa: improve error message for keys too short for PSS
Filippo Valsorda [Thu, 8 Feb 2018 15:19:24 +0000 (16:19 +0100)]
crypto/rsa: improve error message for keys too short for PSS

Fixes #23736

Change-Id: I850d91a512394c4292927d51c475064bfa4e3053
Reviewed-on: https://go-review.googlesource.com/92815
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agoreflect: add embedded field test
Ian Lance Taylor [Thu, 1 Feb 2018 01:12:29 +0000 (17:12 -0800)]
reflect: add embedded field test

Gccgo failed this test.

Updates #23620

Change-Id: I3979a6d3b87d2d014850accf9cb7f356349e6195
Reviewed-on: https://go-review.googlesource.com/91138
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
7 years agoruntime: remove extraneous stackPreempt setting
David Crawshaw [Sun, 24 Sep 2017 14:48:18 +0000 (10:48 -0400)]
runtime: remove extraneous stackPreempt setting

The stackguard is set to stackPreempt earlier in reentersyscall, and
as it comes with throwsplit = true there's no way for the stackguard
to be set to anything else by the end of reentersyscall.

Change-Id: I4e942005b22ac784c52398c74093ac887fc8ec24
Reviewed-on: https://go-review.googlesource.com/65673
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
7 years agocmd/compile/internal/ssa: optimize arm64 with FNMULS/FNMULD
Ben Shi [Wed, 7 Feb 2018 12:24:41 +0000 (12:24 +0000)]
cmd/compile/internal/ssa: optimize arm64 with FNMULS/FNMULD

FNMULS&FNMULD are efficient arm64 instructions, which can be used
to improve FP performance. This CL use them to optimize pairs of neg-mul
operations.

Here are benchmark test results on Raspberry Pi 3 with ArchLinux.

1. A special test case gets about 15% improvement.
(https://github.com/benshi001/ugo1/blob/master/fpmul_test.go)
FPMul-4                     485µs ± 0%     410µs ± 0%  -15.49%  (p=0.000 n=26+23)

2. There is little regression in the go1 benchmark (excluding noise).
name                     old time/op    new time/op    delta
BinaryTree17-4              42.0s ± 3%     42.1s ± 2%    ~     (p=0.542 n=39+40)
Fannkuch11-4                33.3s ± 3%     32.9s ± 1%    ~     (p=0.200 n=40+32)
FmtFprintfEmpty-4           534ns ± 0%     534ns ± 0%    ~     (all equal)
FmtFprintfString-4         1.09µs ± 1%    1.09µs ± 0%    ~     (p=0.950 n=32+32)
FmtFprintfInt-4            1.14µs ± 0%    1.14µs ± 1%    ~     (p=0.571 n=32+31)
FmtFprintfIntInt-4         1.79µs ± 3%    1.76µs ± 0%  -1.42%  (p=0.004 n=40+34)
FmtFprintfPrefixedInt-4    2.17µs ± 0%    2.17µs ± 0%    ~     (p=0.073 n=31+34)
FmtFprintfFloat-4          3.33µs ± 3%    3.28µs ± 0%  -1.46%  (p=0.001 n=40+34)
FmtManyArgs-4              7.28µs ± 6%    7.19µs ± 0%    ~     (p=0.641 n=40+33)
GobDecode-4                96.5ms ± 4%    96.5ms ± 9%    ~     (p=0.214 n=40+40)
GobEncode-4                79.5ms ± 0%    80.7ms ± 4%  +1.51%  (p=0.000 n=34+40)
Gzip-4                      4.53s ± 4%     4.56s ± 4%  +0.60%  (p=0.000 n=40+40)
Gunzip-4                    451ms ± 3%     442ms ± 0%  -1.93%  (p=0.000 n=40+32)
HTTPClientServer-4          530µs ± 1%     535µs ± 1%  +0.88%  (p=0.000 n=39+39)
JSONEncode-4                214ms ± 4%     211ms ± 0%    ~     (p=0.059 n=40+31)
JSONDecode-4                865ms ± 5%     864ms ± 4%  -0.06%  (p=0.003 n=40+40)
Mandelbrot200-4            52.0ms ± 3%    52.1ms ± 3%    ~     (p=0.556 n=40+40)
GoParse-4                  43.1ms ± 8%    42.1ms ± 0%    ~     (p=0.083 n=40+33)
RegexpMatchEasy0_32-4      1.02µs ± 3%    1.02µs ± 4%  +0.06%  (p=0.020 n=40+40)
RegexpMatchEasy0_1K-4      3.90µs ± 0%    3.96µs ± 3%  +1.58%  (p=0.000 n=31+40)
RegexpMatchEasy1_32-4       967ns ± 4%     981ns ± 3%  +1.40%  (p=0.000 n=40+40)
RegexpMatchEasy1_1K-4      6.41µs ± 4%    6.43µs ± 3%    ~     (p=0.386 n=40+40)
RegexpMatchMedium_32-4     1.76µs ± 3%    1.78µs ± 3%  +1.08%  (p=0.000 n=40+40)
RegexpMatchMedium_1K-4      561µs ± 0%     562µs ± 0%  +0.09%  (p=0.003 n=34+31)
RegexpMatchHard_32-4       31.5µs ± 2%    31.1µs ± 4%  -1.17%  (p=0.000 n=30+40)
RegexpMatchHard_1K-4        960µs ± 3%     950µs ± 4%  -1.02%  (p=0.016 n=40+40)
Revcomp-4                   7.79s ± 7%     7.79s ± 4%    ~     (p=0.859 n=40+40)
Template-4                  889ms ± 6%     872ms ± 3%  -1.86%  (p=0.025 n=40+31)
TimeParse-4                4.80µs ± 0%    4.89µs ± 3%  +1.71%  (p=0.001 n=31+40)
TimeFormat-4               4.70µs ± 1%    4.78µs ± 3%  +1.57%  (p=0.000 n=33+40)
[Geo mean]                  710µs          709µs       -0.13%

name                     old speed      new speed      delta
GobDecode-4              7.96MB/s ± 4%  7.96MB/s ± 9%    ~     (p=0.174 n=40+40)
GobEncode-4              9.65MB/s ± 0%  9.51MB/s ± 4%  -1.45%  (p=0.000 n=34+40)
Gzip-4                   4.29MB/s ± 4%  4.26MB/s ± 4%  -0.59%  (p=0.000 n=40+40)
Gunzip-4                 43.0MB/s ± 3%  43.9MB/s ± 0%  +1.90%  (p=0.000 n=40+32)
JSONEncode-4             9.09MB/s ± 4%  9.22MB/s ± 0%    ~     (p=0.429 n=40+31)
JSONDecode-4             2.25MB/s ± 5%  2.25MB/s ± 4%    ~     (p=0.278 n=40+40)
GoParse-4                1.35MB/s ± 7%  1.37MB/s ± 0%    ~     (p=0.071 n=40+25)
RegexpMatchEasy0_32-4    31.5MB/s ± 3%  31.5MB/s ± 4%  -0.08%  (p=0.018 n=40+40)
RegexpMatchEasy0_1K-4     263MB/s ± 0%   259MB/s ± 3%  -1.51%  (p=0.000 n=31+40)
RegexpMatchEasy1_32-4    33.1MB/s ± 4%  32.6MB/s ± 3%  -1.38%  (p=0.000 n=40+40)
RegexpMatchEasy1_1K-4     160MB/s ± 4%   159MB/s ± 3%    ~     (p=0.364 n=40+40)
RegexpMatchMedium_32-4    565kB/s ± 3%   562kB/s ± 2%    ~     (p=0.208 n=40+40)
RegexpMatchMedium_1K-4   1.82MB/s ± 0%  1.82MB/s ± 0%  -0.27%  (p=0.000 n=34+31)
RegexpMatchHard_32-4     1.02MB/s ± 3%  1.03MB/s ± 4%  +1.04%  (p=0.000 n=32+40)
RegexpMatchHard_1K-4     1.07MB/s ± 4%  1.08MB/s ± 4%  +0.94%  (p=0.003 n=40+40)
Revcomp-4                32.6MB/s ± 7%  32.6MB/s ± 4%    ~     (p=0.965 n=40+40)
Template-4               2.18MB/s ± 6%  2.22MB/s ± 3%  +1.83%  (p=0.020 n=40+31)
[Geo mean]               7.77MB/s       7.78MB/s       +0.16%

3. There is little change in the compilecmp benchmark (excluding noise).
name        old time/op       new time/op       delta
Template          2.37s ± 3%        2.35s ± 4%    ~     (p=0.529 n=10+10)
Unicode           1.38s ± 8%        1.36s ± 5%    ~     (p=0.247 n=10+10)
GoTypes           8.10s ± 2%        8.10s ± 2%    ~     (p=0.971 n=10+10)
Compiler          40.5s ± 4%        40.8s ± 1%    ~     (p=0.529 n=10+10)
SSA                115s ± 2%         115s ± 3%    ~     (p=0.684 n=10+10)
Flate             1.45s ± 5%        1.46s ± 3%    ~     (p=0.796 n=10+10)
GoParser          1.86s ± 4%        1.84s ± 2%    ~     (p=0.095 n=9+10)
Reflect           5.11s ± 2%        5.13s ± 2%    ~     (p=0.315 n=10+10)
Tar               2.22s ± 3%        2.23s ± 1%    ~     (p=0.299 n=9+7)
XML               2.72s ± 3%        2.72s ± 3%    ~     (p=0.912 n=10+10)
[Geo mean]        5.03s             5.02s       -0.21%

name        old user-time/op  new user-time/op  delta
Template          2.92s ± 2%        2.89s ± 1%    ~     (p=0.247 n=10+10)
Unicode           1.71s ± 5%        1.69s ± 4%    ~     (p=0.393 n=10+10)
GoTypes           9.78s ± 2%        9.76s ± 2%    ~     (p=0.631 n=10+10)
Compiler          49.1s ± 2%        49.1s ± 1%    ~     (p=0.796 n=10+10)
SSA                144s ± 1%         144s ± 2%    ~     (p=0.796 n=10+10)
Flate             1.74s ± 2%        1.73s ± 3%    ~     (p=0.842 n=10+9)
GoParser          2.23s ± 3%        2.25s ± 2%    ~     (p=0.143 n=10+10)
Reflect           5.93s ± 3%        5.98s ± 2%    ~     (p=0.211 n=10+9)
Tar               2.65s ± 2%        2.69s ± 3%  +1.51%  (p=0.010 n=9+10)
XML               3.25s ± 2%        3.21s ± 1%  -1.24%  (p=0.035 n=10+9)
[Geo mean]        6.07s             6.07s       -0.08%

name        old text-bytes    new text-bytes    delta
HelloSize         641kB ± 0%        641kB ± 0%    ~     (all equal)

name        old data-bytes    new data-bytes    delta
HelloSize        9.46kB ± 0%       9.46kB ± 0%    ~     (all equal)

name        old bss-bytes     new bss-bytes     delta
HelloSize         125kB ± 0%        125kB ± 0%    ~     (all equal)

name        old exe-bytes     new exe-bytes     delta
HelloSize        1.24MB ± 0%       1.24MB ± 0%    ~     (all equal)

Change-Id: Id095d998c380eef929755124084df02446a6b7c1
Reviewed-on: https://go-review.googlesource.com/92555
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
7 years agomisc/cgo/testcshared: increase sleep in TestUnexportedSymbols
Ian Lance Taylor [Wed, 14 Feb 2018 05:15:05 +0000 (21:15 -0800)]
misc/cgo/testcshared: increase sleep in TestUnexportedSymbols

Increase the sleep and wait for up to 2 seconds for the dup2.
Apparently it can sometimes take a long time.

Fixes #23784

Change-Id: I929530b057bbcd842b28a7640c39dd68d719ff7d
Reviewed-on: https://go-review.googlesource.com/93895
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agopath/filepath: fix escaped chars in Glob on non-Windows
Daniel Martí [Thu, 11 Jan 2018 20:23:22 +0000 (20:23 +0000)]
path/filepath: fix escaped chars in Glob on non-Windows

Backslashes are ignored in Match and Glob on Windows, since those
collide with the separator character. However, they should still work in
both functions on other operating systems.

hasMeta did not reflect this logic - it always treated a backslash as a
non-special character. Do that only on Windows.

Assuming this is what the TODO was referring to, remove it. There are no
other characters that scanChunk treats especially.

Fixes #23418.

Change-Id: Ie0bd795812e0ed9d8c8c1bbc3137f29d960cba84
Reviewed-on: https://go-review.googlesource.com/87455
Run-TryBot: Daniel Martí <mvdan@mvdan.cc>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
7 years agopath: remove filename mentions from pattern godocs
Daniel Martí [Tue, 16 Jan 2018 16:55:06 +0000 (16:55 +0000)]
path: remove filename mentions from pattern godocs

path.Match works purely with strings, not file paths. That's what sets
it apart from filepath.Match. For example, only filepath.Match will
change its behavior towards backslashes on Windows, to accomodate for
the file path separator on that system.

As such, path.Match should make no mention of file names. Nor should
path.ErrBadPattern mention globbing at all - the package has no notion
of globbing, and the error concerns only patterns.

For a similar reason, remove the mention of globbing from
filepath.ErrBadPattern. The error isn't reserved to just globbing, as it
can be returned from filepath.Match. And, as before, it only concerns
the patterns themselves.

Change-Id: I58a83ffa3e2549625d8e546ef916652525504bd1
Reviewed-on: https://go-review.googlesource.com/87857
Reviewed-by: Rob Pike <r@golang.org>
7 years agomath/big: fix %s verbs in Float tests error messages
Alberto Donizetti [Wed, 10 Jan 2018 16:31:25 +0000 (17:31 +0100)]
math/big: fix %s verbs in Float tests error messages

Fatalf calls in two Float tests use the %s verb with Floats values,
which is not allowed and results in failure messages that look like
this:

    float_test.go:1385: i = 0, prec = 1, ToZero:
                     %!s(*big.Float=1) [0]
                /    %!s(*big.Float=1) [0]
                =    %!s(*big.Float=0.0625)
                want %!s(*big.Float=1)

Switch to %v.

Change-Id: Ifdc80bf19c91ca1b190f6551a6d0a51b42ed5919
Reviewed-on: https://go-review.googlesource.com/87199
Run-TryBot: Alberto Donizetti <alb.donizetti@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agocmd/compile: change type of clear argument of ordercopyexpr to bool
Martin Möhrmann [Wed, 17 Jan 2018 19:26:23 +0000 (20:26 +0100)]
cmd/compile: change type of clear argument of ordercopyexpr to bool

ordercopyexpr is only called with 0 or 1 as value for the clear
argument. The clear variable in ordercopyexpr is only used in the
call to ordertemp which has a clear argument of type bool.

Change the clear argument of ordercopyexpr from int to bool and change
calls to ordercopyexpr to use false instead of 0 and true instead of 1.

Passes toolstash -cmp.

Change-Id: Ic264aafd3b0c8b99f6ef028ffaa2e30f23f9125a
Reviewed-on: https://go-review.googlesource.com/88115
Reviewed-by: Daniel Martí <mvdan@mvdan.cc>
7 years agointernal/cpu: make arm64 capability bits naming less verbose
Martin Möhrmann [Sun, 4 Feb 2018 20:23:25 +0000 (21:23 +0100)]
internal/cpu: make arm64 capability bits naming less verbose

This makes the constant names less verbose and aligns them more
with the Linux kernel which uses HWCAP_XXX for the constant names.

Change-Id: Ia7d079b59b57978adc045945951eaa1d99b41fac
Reviewed-on: https://go-review.googlesource.com/91738
Run-TryBot: Martin Möhrmann <moehrmann@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
7 years agoruntime: add symbol for AT_FDCWD on Linux amd64 and mips64x
Tobias Klauser [Tue, 13 Feb 2018 16:07:54 +0000 (17:07 +0100)]
runtime: add symbol for AT_FDCWD on Linux amd64 and mips64x

Also order the syscall number list by numerically for mips64x.

Follow-up for CL 92895.

Change-Id: I5f01f8c626132a06160997fce8a2aef0c486bb1c
Reviewed-on: https://go-review.googlesource.com/93616
Run-TryBot: Tobias Klauser <tobias.klauser@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

7 years agodoc/articles/wiki: highlight the use of _ warning
Agniva De Sarker [Thu, 11 Jan 2018 16:49:01 +0000 (22:19 +0530)]
doc/articles/wiki: highlight the use of _ warning

This moves the paragraph mentioning the use of _ higher up
to emphasize the warning and thereby reducing chances of getting
stuck.

Fixes #22617

Change-Id: I64352a3e966a22d86fc9d381332bade49d74714a
Reviewed-on: https://go-review.googlesource.com/87375
Reviewed-by: Ian Lance Taylor <iant@golang.org>
7 years agoencoding/hex: fix potential incorrect Dumper output when Close is called multiple...
Tim Cooper [Fri, 26 Jan 2018 20:36:15 +0000 (16:36 -0400)]
encoding/hex: fix potential incorrect Dumper output when Close is called multiple times

Fixes #23574

Change-Id: I69573de47daa6fd53cc99a78c0c4b867460242e3
Reviewed-on: https://go-review.googlesource.com/90275
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>

7 years agocmd/compile: fix constant folding of right shifts
Keith Randall [Tue, 13 Feb 2018 20:33:55 +0000 (12:33 -0800)]
cmd/compile: fix constant folding of right shifts

The sub-word shifts need to sign-extend before shifting, to avoid
bringing in data from higher in the argument.

Fixes #23812

Change-Id: I0a95a0b49c48f3b40b85765bb4a9bb492be0cd73
Reviewed-on: https://go-review.googlesource.com/93716
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
7 years agogo/build: add go1.11 build tag
Mikio Hara [Tue, 13 Feb 2018 22:20:26 +0000 (07:20 +0900)]
go/build: add go1.11 build tag

As usual, adding go1.11 early in the cycle so that we can start
regression testing of the master toolchain.

Change-Id: Ie96eca7223722d60d7acc6b3b996b76740c36419
Reviewed-on: https://go-review.googlesource.com/93775
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Mikio Hara <mikioh.mikioh@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>

7 years agoruntime/trace: fix TestTraceSymbolize when GOMAXPROCS=1
David du Colombier [Tue, 13 Feb 2018 21:11:20 +0000 (22:11 +0100)]
runtime/trace: fix TestTraceSymbolize when GOMAXPROCS=1

CL 92916 added the GOMAXPROCS test in TestTraceSymbolize.

This test only succeeds when the value of GOMAXPROCS changes.

Since the test calls runtime.GOMAXPROCS(1), it will fails
on machines where GOMAXPROCS=1.

This change fixes the test by calling runtime.GOMAXPROCS(oldGoMaxProcs+1).

Fixes #23816.

Change-Id: I1183dbbd7db6077cbd7fa0754032ff32793b2195
Reviewed-on: https://go-review.googlesource.com/93735
Run-TryBot: David du Colombier <0intro@gmail.com>
Reviewed-by: Hyang-Ah Hana Kim <hyangah@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>