Skip to content

G
go
Switch branch/tag
  1. 14 Apr, 2016 24 commits
  3. 13 Apr, 2016 16 commits
    • Robert Griesemer's avatar
      cmd/compile: don't export unneeded OAS, OASWB nodes · 6e5027a3
      Robert Griesemer authored 
      Also:
- "rewrite" node Op in exporter for some nodes instead of importer
- more comments

Change-Id: I809e6754d14987b28f1da9379951ffa2e690c2a7
Reviewed-on: https://go-review.googlesource.com/22008Reviewed-by: default avatarMatthew Dempsky <mdempsky@google.com>
Run-TryBot: Robert Griesemer <gri@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
      6e5027a3
    • Lynn Boger's avatar
      syscall: fix epoll_event struct for ppc64le/ppc64 · 44f80f6d
      Lynn Boger authored 
      The existing epoll_event structure used by many of
the epoll_* syscalls was defined incorrectly
for use with ppc64le & ppc64 in the syscall
directory.  This resulted in the caller getting
incorrect information on return from these
syscalls.  This caused failures in fsnotify as
well as builds with upstream Docker.  The
structure is defined correctly in gccgo.

This adds a pad field that is expected for
these syscalls on ppc64le, ppc64.
Fixes #15135

Change-Id: If7e8ea9eb1d1ca5182c8dc0f935b334127341ffd
Reviewed-on: https://go-review.googlesource.com/21582Reviewed-by: default avatarIan Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
      44f80f6d
    • David Crawshaw's avatar
      cmd/compile, etc: use name for type pkgPath · f120936d
      David Crawshaw authored 
      By replacing the *string used to represent pkgPath with a
reflect.name everywhere, the embedded *string for package paths
inside the reflect.name can be replaced by an offset, nameOff.
This reduces the number of pointers in the type information.

This also moves all reflect.name types into the same section, making
it possible to use nameOff more widely in later CLs.

No significant binary size change for normal binaries, but:

linux/amd64 PIE:
	cmd/go: -440KB (3.7%)
	jujud:  -2.6MB (3.2%)

For #6853.

Change-Id: I3890b132a784a1090b1b72b32febfe0bea77eaee
Reviewed-on: https://go-review.googlesource.com/21395
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: default avatarIan Lance Taylor <iant@golang.org>
      f120936d
    • Brad Fitzpatrick's avatar
      runtime: rename os1_darwin.go to os_darwin.go · 73e2ad20
      Brad Fitzpatrick authored 
      Change-Id: If0e0bc5a85101db1e70faaab168fc2d12024eb93
Reviewed-on: https://go-review.googlesource.com/22005Reviewed-by: default avatarIan Lance Taylor <iant@golang.org>
      73e2ad20
    • Brad Fitzpatrick's avatar
      runtime: merge the darwin os*.go files together · d9712aa8
      Brad Fitzpatrick authored 
      Merge them together into os1_darwin.go. A future CL will rename it.

Change-Id: Ia4380d3296ebd5ce210908ce3582ff184566f692
Reviewed-on: https://go-review.googlesource.com/22004
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: default avatarIan Lance Taylor <iant@golang.org>
      d9712aa8
    • David Crawshaw's avatar
      cmd/link: handle long symbol names · 79048df2
      David Crawshaw authored 
      Fixes #15104.

Change-Id: I9ddfbbf39ef0a873b703ee3e04fbb7d1192f5f39
Reviewed-on: https://go-review.googlesource.com/21581
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: default avatarIan Lance Taylor <iant@golang.org>
      79048df2
    • Todd Neal's avatar
      cmd/compile: sort partitions by dom to speed up cse · 3ea7cfab
      Todd Neal authored 
      We do two O(n) scans of all values in an eqclass when computing
substitutions for CSE.

In unfortunate cases, like those found in #15112, we can have a large
eqclass composed of values found in blocks none of whom dominate the
other.  This leads to O(n^2) behavior. The elements are removed one at a
time, with O(n) scans each time.

This CL removes the linear scan by sorting the eqclass so that dominant
values will be sorted first.  As long as we also ensure we don't disturb
the sort order, then we no longer need to scan for the maximally
dominant value.

For the code in issue #15112:

Before:
real    1m26.094s
user    1m30.776s
sys     0m1.125s

Aefter:
real    0m52.099s
user    0m56.829s
sys     0m1.092s

Updates #15112

Change-Id: Ic4f8680ed172e716232436d31963209c146ef850
Reviewed-on: https://go-review.googlesource.com/21981Reviewed-by: default avatarJosh Bleecher Snyder <josharian@gmail.com>
Run-TryBot: Todd Neal <todd@tneal.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
      3ea7cfab
    • Austin Clements's avatar
      runtime/internal/atomic: rename Storep1 to StorepNoWB · 4721ea6a
      Austin Clements authored 
      Make it clear that the point of this function stores a pointer
*without* a write barrier.

sed -i -e 's/Storep1/StorepNoWB/' $(git grep -l Storep1)

Updates #15270.

Change-Id: Ifad7e17815e51a738070655fe3b178afdadaecf6
Reviewed-on: https://go-review.googlesource.com/21994Reviewed-by: default avatarBrad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: default avatarMichael Matloob <matloob@golang.org>
      4721ea6a
    • Dmitry Vyukov's avatar
      cmd/pprof/internal/profile: always subtract 1 from PCs · da6205b6
      Dmitry Vyukov authored 
      Go runtime never emits PCs that are not a return address
(except for cpu profiler).

Change-Id: I08d9dc5c7c71e23f34f2f0c16f8baeeb4f64fcd6
Reviewed-on: https://go-review.googlesource.com/21735Reviewed-by: default avatarIan Lance Taylor <iant@golang.org>
      da6205b6
    • Robert Griesemer's avatar
      cmd/compile, go/importer: minor cleanups · eb79f21c
      Robert Griesemer authored 
      Change-Id: Ic7a1fb0dbbf108052c970a4a830269a5673df7df
Reviewed-on: https://go-review.googlesource.com/21963Reviewed-by: default avatarMatthew Dempsky <mdempsky@google.com>
      eb79f21c
    • Robert Griesemer's avatar
      cmd/compile: move more compiler specifics into compiler specific export section · 7d0d1222
      Robert Griesemer authored 
      Instead of indicating with each function signature if it has an inlineable
body, collect all functions in order and export function bodies with function
index in platform-specific section.

Moves this compiler specific information out of the platform-independent
export data section, and removes an int value for all functions w/o body.
Also simplifies the code a bit.

Change-Id: I8b2d7299dbe81f2706be49ecfb9d9f7da85fd854
Reviewed-on: https://go-review.googlesource.com/21939Reviewed-by: default avatarMatthew Dempsky <mdempsky@google.com>
      7d0d1222
    • Austin Clements's avatar
      runtime/internal/atomic: remove write barrier from Storep1 on s390x · d8e8fc29
      Austin Clements authored 
      atomic.Storep1 is not supposed to invoke a write barrier (that's what
atomicstorep is for), but currently does on s390x. This causes a panic
in runtime.mapzero when it tries to use atomic.Storep1 to store what's
actually a scalar.

Fix this by eliminating the write barrier from atomic.Storep1 on
s390x. Also add some documentation to atomicstorep to explain the
difference between these.

Fixes #15270.

Change-Id: I291846732d82f090a218df3ef6351180aff54e81
Reviewed-on: https://go-review.googlesource.com/21993Reviewed-by: default avatarBrad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Austin Clements <austin@google.com>
Reviewed-by: default avatarMichael Munday <munday@ca.ibm.com>
      d8e8fc29
    • David Chase's avatar
      cmd/compile: move spills to loop exits when easy. · 6b85a45e
      David Chase authored 
      For call-free inner loops.

Revised statistics:
  85 inner loop spills sunk
 341 inner loop spills remaining
1162 inner loop spills that were candidates for sinking
     ended up completely register allocated
 119 inner loop spills could have been sunk were used in
     "shuffling" at the bottom of the loop.
   1 inner loop spill not sunk because the register assigned
     changed between def and exit,

 Understanding how to make an inner loop definition not be
 a candidate for from-memory shuffling (to force the shuffle
 code to choose some other value) should pick up some of the
 119 other spills disqualified for this reason.

 Modified the stats printing based on feedback from Austin.

Change-Id: If3fb9b5d5a028f42ccc36c4e3d9e0da39db5ca60
Reviewed-on: https://go-review.googlesource.com/21037Reviewed-by: default avatarKeith Randall <khr@golang.org>
Run-TryBot: David Chase <drchase@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
      6b85a45e
    • Lynn Boger's avatar
      runtime: improve memmove performance ppc64,ppc64le · c4807d4c
      Lynn Boger authored 
      This change improves the performance of memmove
on ppc64 & ppc64le mainly for moves >=32 bytes.
In addition, the test to detect backward moves
 was enhanced to avoid backward moves if source
and dest were in different types of storage, since
backward moves might not always be efficient.

Fixes #14507

The following shows some of the improvements from the test
in the runtime package:

BenchmarkMemmove32                   4229.56      4717.13      1.12x
BenchmarkMemmove64                   6156.03      7810.42      1.27x
BenchmarkMemmove128                  7521.69      12468.54     1.66x
BenchmarkMemmove256                  6729.90      18260.33     2.71x
BenchmarkMemmove512                  8521.59      18033.81     2.12x
BenchmarkMemmove1024                 9760.92      25762.61     2.64x
BenchmarkMemmove2048                 10241.00     29584.94     2.89x
BenchmarkMemmove4096                 10399.37     31882.31     3.07x

BenchmarkMemmoveUnalignedDst16       1943.69      2258.33      1.16x
BenchmarkMemmoveUnalignedDst32       3885.08      3965.81      1.02x
BenchmarkMemmoveUnalignedDst64       5121.63      6965.54      1.36x
BenchmarkMemmoveUnalignedDst128      7212.34      11372.68     1.58x
BenchmarkMemmoveUnalignedDst256      6564.52      16913.59     2.58x
BenchmarkMemmoveUnalignedDst512      8364.35      17782.57     2.13x
BenchmarkMemmoveUnalignedDst1024     9539.87      24914.72     2.61x
BenchmarkMemmoveUnalignedDst2048     9199.23      21235.11     2.31x
BenchmarkMemmoveUnalignedDst4096     10077.39     25231.99     2.50x

BenchmarkMemmoveUnalignedSrc32       3249.83      3742.52      1.15x
BenchmarkMemmoveUnalignedSrc64       5562.35      6627.96      1.19x
BenchmarkMemmoveUnalignedSrc128      6023.98      10200.84     1.69x
BenchmarkMemmoveUnalignedSrc256      6921.83      15258.43     2.20x
BenchmarkMemmoveUnalignedSrc512      8593.13      16541.97     1.93x
BenchmarkMemmoveUnalignedSrc1024     9730.95      22927.84     2.36x
BenchmarkMemmoveUnalignedSrc2048     9793.28      21537.73     2.20x
BenchmarkMemmoveUnalignedSrc4096     10132.96     26295.06     2.60x

Change-Id: I73af59970d4c97c728deabb9708b31ec7e01bdf2
Reviewed-on: https://go-review.googlesource.com/21990Reviewed-by: default avatarBill O'Farrell <billotosyr@gmail.com>
Reviewed-by: default avatarBrad Fitzpatrick <bradfitz@golang.org>
      c4807d4c
    • David Crawshaw's avatar
      cmd/link: use a switch for name prefix switching · 66afbf10
      David Crawshaw authored 
      Minor cleanup.

Change-Id: I7574f58a7e55c2bb798ebe9c7c98d36b8c258fb8
Reviewed-on: https://go-review.googlesource.com/21982
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: default avatarBrad Fitzpatrick <bradfitz@golang.org>
      66afbf10
    • David Crawshaw's avatar
      cmd/compile, etc: store method tables as offsets · 7d469179
      David Crawshaw authored 
      This CL introduces the typeOff type and a lookup method of the same
name that can turn a typeOff offset into an *rtype.

In a typical Go binary (built with buildmode=exe, pie, c-archive, or
c-shared), there is one moduledata and all typeOff values are offsets
relative to firstmoduledata.types. This makes computing the pointer
cheap in typical programs.

With buildmode=shared (and one day, buildmode=plugin) there are
multiple modules whose relative offset is determined at runtime.
We identify a type in the general case by the pair of the original
*rtype that references it and its typeOff value. We determine
the module from the original pointer, and then use the typeOff from
there to compute the final *rtype.

To ensure there is only one *rtype representing each type, the
runtime initializes a typemap for each module, using any identical
type from an earlier module when resolving that offset. This means
that types computed from an offset match the type mapped by the
pointer dynamic relocations.

A series of followup CLs will replace other *rtype values with typeOff
(and name/*string with nameOff).

For types created at runtime by reflect, type offsets are treated as
global IDs and reference into a reflect offset map kept by the runtime.

darwin/amd64:
	cmd/go:  -57KB (0.6%)
	jujud:  -557KB (0.8%)

linux/amd64 PIE:
	cmd/go: -361KB (3.0%)
	jujud:  -3.5MB (4.2%)

For #6853.

Change-Id: Icf096fd884a0a0cb9f280f46f7a26c70a9006c96
Reviewed-on: https://go-review.googlesource.com/21285Reviewed-by: default avatarIan Lance Taylor <iant@golang.org>
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
      7d469179
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7