The new code is a bit less efficient,
but it does not involve altering the structure
of any linked lists.
This will make it easier to replace NodeLists
with Node slices.
We can return to a more efficient algorithm
when NodeLists have been replaced.
Martin Möhrmann [Sat, 27 Feb 2016 13:37:10 +0000 (14:37 +0100)]
fmt: refactor and unify float and complex formatting
Removes specialized functions for each verb and float/complex size
and replaces them with generic variants fmtFloat and
fmtComplex similar to other generic fmt functions.
Simplifies the complex formatting by relying on fmtFloat
to handle the verb and default precision selection.
Complex imaginary formatting does not need to clear the f.space flag
because the set f.plus flag will force a sign instead of a space.
Sets default precision for %b to -1 (same as %g and %G)
since precision for %b has no affect in strconv.AppendFloat.
Add more tests and group them a bit better.
Use local copies of +Inf,-Inf and NaN instead
of math package functions for testing.
Saves around 8kb in the go binary.
name old time/op new time/op delta
SprintfFloat-2 200ns ± 4% 196ns ± 4% -1.55% (p=0.007 n=20+20)
SprintfComplex-2 569ns ± 4% 570ns ± 3% ~ (p=0.804 n=20+20)
Change-Id: I36d35dab6f835fc2bd2c042ac97705868eb2446f
Reviewed-on: https://go-review.googlesource.com/20252 Reviewed-by: Rob Pike <r@golang.org>
Run-TryBot: Rob Pike <r@golang.org>
Martin Möhrmann [Thu, 3 Mar 2016 16:06:43 +0000 (17:06 +0100)]
fmt: optimize byte slice and array formatting for %v and %d
Instead of calling printArg in fmtBytes to format each byte call
the byte formatting functions directly since it is known each
element is of type byte.
Add more tests for byte slice and array formatting.
name old time/op new time/op delta
SprintfBytes-2 843ns ±16% 417ns ±11% -50.58% (p=0.000 n=20+20)
Change-Id: I5b907dbf52091e3de9710b09d67649c76f4c17e9
Reviewed-on: https://go-review.googlesource.com/20176
Run-TryBot: Rob Pike <r@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rob Pike <r@golang.org>
Matthew Dempsky [Sat, 5 Mar 2016 00:39:07 +0000 (16:39 -0800)]
cmd/compile: stop leaking 'any' type into user package
The new check corresponds to the (etype != TANY || Debug['A'] != 0)
that was lost in golang.org/cl/19936.
Fixes #14652.
Change-Id: Iec3788ff02529b3b0f0d4dd92ec9f3ef20aec849
Reviewed-on: https://go-review.googlesource.com/20271 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Lynn Boger [Tue, 16 Feb 2016 22:05:56 +0000 (16:05 -0600)]
crypto/cipher: xorBytes performance ppc64le/ppc64
Update supportsUnaligned in xor.go to be true for
GOARCH values ppc64le and ppc64. This allows the
xor of long buffers to be done on double words
(8 bytes) instead of a single byte at a time, which
significantly improves performance.
Fixes #14350
Change-Id: Iccc6b9d3df2e604a55f4c1e4890bdd3bb0d77ab0
Reviewed-on: https://go-review.googlesource.com/19519 Reviewed-by: Minux Ma <minux@golang.org>
Rhys Hiltner [Fri, 4 Mar 2016 06:37:14 +0000 (22:37 -0800)]
cmd/link/internal/ld: don't panic on short buildid
When the linker was written in C, command line arguments were passed
around as null-terminated byte arrays which encouraged checking
characters one at a time. In Go, that can easily lead to
out-of-bounds panics.
Use the more idiomatic strings.HasPrefix when checking cmd/link's -B
argument to avoid the panic, and replace the manual hex decode with
use of the encoding/hex package.
Austin Clements [Wed, 2 Mar 2016 02:30:26 +0000 (21:30 -0500)]
runtime: disable gcMarkRootCheck debugging check during STW
gcMarkRootCheck takes ~10ns per goroutine. This is just a debugging
check, so disable it (plus, if something is going to go wrong, it's
more likely to go wrong during concurrent mark).
We may be able to re-enable this later, or move it to after we've
started the world again. (But not for 1.6.x.)
For 1.6.x.
Fixes #14419.
name / 95%ile-time/markTerm old new delta
500kIdleGs-12 24.0ms ± 0% 18.9ms ± 6% -21.46% (p=0.000 n=15+20)
Austin Clements [Tue, 1 Mar 2016 20:09:24 +0000 (15:09 -0500)]
runtime: reset mark state before stopping the world
Currently we reset the mark state during STW sweep termination. This
involves looping over all of the goroutines. Each iteration of this
loop takes ~25ns, so at around 400k goroutines, we'll exceed our 10ms
pause goal.
However, it's safe to do this before we stop the world for sweep
termination because nothing is consuming this state yet. Hence, move
the reset to just before STW.
This isn't perfect: a long reset can still delay allocating goroutines
that block on GC starting. But it's certainly better to block some
things eventually than to block everything immediately.
For 1.6.x.
Fixes #14420.
name \ 95%ile-time/sweepTerm old new delta
500kIdleGs-12 11312µs ± 6% 18.9µs ± 6% -99.83% (p=0.000 n=16+20)
David Crawshaw [Sat, 27 Feb 2016 17:33:14 +0000 (08:33 -0900)]
cmd/link: deduplicate read-only string data
Many read-only strings in Go binaries are substrings of other read-only
strings. A common source is the text form of type information, which
will include both "struct { X int }" and "*struct { X int }" or
"*bytes.Reader" and "func(*bytes.Reader)" in the same binary.
Because this character data is referred to by separate string headers,
we can skip writing the smaller string and modify the pointer
relocation to point to the larger string. This CL does this
deduplication in the linker after the reachable set of strings has
been determined.
This removes 765KB from juju (1.4% without DWARF).
Link time goes at tip goes form 4.6s to 6.3s, but note that this CL
is part of a series that recently reduced link time from 9.6s.
For #6853.
Change-Id: Ib2087cf627c9f1e9a1181f9b4c8f81d1a3f42191
Reviewed-on: https://go-review.googlesource.com/19987 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Matthew Dempsky [Fri, 4 Mar 2016 15:55:39 +0000 (07:55 -0800)]
cmd/compile: small cleanup to syslook calls
Passing copy==1 to syslook is only necessary to support subsequent
calls to substArgTypes. typ2Itab and concatstring* don't have "any"
parameters, so no point in deep copying their function signatures at
every call site.
For a couple other syslook calls (makemap and conv[IET]2[IET]), move
them closer to their corresponding substArgTypes calls so it's easier
to see that all syslook(fn, 1) calls are necessary.
Change-Id: I4a0588ab2b8b5b8ce7a0a44b24c8cf8fda489af6
Reviewed-on: https://go-review.googlesource.com/20215
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org>
Robert Griesemer [Fri, 4 Mar 2016 01:39:55 +0000 (17:39 -0800)]
math/big: use correct precision in Float.Float32/64 for denormals
When a big.Float is converted to a denormal float32/64, the rounding
precision depends on the size of the denormal. Rounding may round up
and thus change the size (exponent) of the denormal. Recompute the
correct precision again for correct placement of the mantissa.
Fixes #14553.
Change-Id: Iedab5810a2d2a405cc5da28c6de7be34cb035b86
Reviewed-on: https://go-review.googlesource.com/20198 Reviewed-by: Alan Donovan <adonovan@google.com>
Michael Hudson-Doyle [Fri, 4 Mar 2016 02:21:49 +0000 (15:21 +1300)]
cmd/internal/obj/ppc64: eliminate ~700 conversions by tweaking signatures
This is mostly changing the opXXX helpers to take an int16 (matching Prog.As)
argument and return a uint32. The only bit that's not completely trivial is
passing -p.As to opirr to signal operating on a shifted constant, because AADD
+ ALAST overflows int16.
Matthew Dempsky [Thu, 3 Mar 2016 23:49:04 +0000 (15:49 -0800)]
cmd/compile: remove unnecessary conversions
Automated CL prepared by github.com/mdempsky/unconvert, except for
reverting changes to ssa/rewritegeneric.go (generated file) and
package big (vendored copy of math/big).
Change-Id: I64dc4199f14077c7b6a2f334b12249d4a785eadd
Reviewed-on: https://go-review.googlesource.com/20089
Run-TryBot: Matthew Dempsky <mdempsky@google.com> Reviewed-by: Dave Cheney <dave@cheney.net>
Spencer Nelson [Wed, 2 Mar 2016 20:29:30 +0000 (15:29 -0500)]
cmd/vet: Use function signature to find format string index.
cmd/vet's printf checker currently uses a hardcoded map of function
names to expected positions of format strings. We can be a bit more
precise than this by looking up the signature of the function, which
helps when libraries implement functions like Errorf or Logf with
extra arguments like log levels or error codes.
Specifically, the format string param is assumed to be the last string
parameter of the called function.
Fixes #12294.
Change-Id: Icf10ebb819bba91fa1c4109301417042901e34c7
Reviewed-on: https://go-review.googlesource.com/20163
Run-TryBot: Rob Pike <r@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Rob Pike <r@golang.org>
Ian Lance Taylor [Thu, 3 Mar 2016 22:11:17 +0000 (14:11 -0800)]
cmd/compile: add setNodeSeq, and use it in bimport.go
Also rewrite bexport.go to use nodeSeqIterate.
The new setNodeSeq is a transitional generic function to set either a
NodeList or a slice to either a NodeList or a slice. This should permit
us to flip fields from *NodeList to []*Node, or Nodes, without changing
other code.
Ian Lance Taylor [Thu, 3 Mar 2016 19:30:17 +0000 (11:30 -0800)]
cmd/compile: add nodeSeqIterator interface
I tried to write a program to convert *NodeList to Node, but ran into
too many problem cases. I'm backing off and trying a more iterative
approach using interfaces.
This CL adds an interface for iteration over either a *NodeList or a
Nodes. I changed typechecklist to use it, to show how it works. After
NodeList is eliminated, we can change the typechecklist parameter type
to Nodes.
Passes toolstash -cmp.
Change-Id: I5c7593714b020d20868b99151b1e7cadbbdbc397
Reviewed-on: https://go-review.googlesource.com/20190
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
David Crawshaw [Thu, 3 Mar 2016 15:51:30 +0000 (10:51 -0500)]
cmd/link: combine all string data into one symbol
This CL introduces a mergestrings pass after the reachability
analysis to combine all reachable go.string."..." character data
symbols into a single symbol.
Shrinks juju by 1.2mb (1.5%).
Shrinks cmd/go by 0.5% when building without DWARF.
No noticable effect on linker speed.
Change-Id: I2ba3e60bf418f65766bda257f6ca9eea26d895b6
Reviewed-on: https://go-review.googlesource.com/20165
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org>
Keith Randall [Thu, 3 Mar 2016 04:54:41 +0000 (20:54 -0800)]
cmd/compile: fix fieldtrack experiment
The ODOTPTRs introduced in CL #19814 don't have field names,
just offsets. The fieldtrack experiment crashes when
examining them. Instead, just ignore them. We'll never track
these fields anyway.
It would be nice to have the runtime type struct build in the
compiler (like we do sudog, for example) so we could use its
fieldnames. Doesn't seem worth it just for this CL.
Change-Id: I5e75024f5a8333eb7439543b3f466ea40213a1b9
Reviewed-on: https://go-review.googlesource.com/20157 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Michael Munday [Thu, 3 Mar 2016 02:24:04 +0000 (21:24 -0500)]
cmd/link: add Minalign field to ld.Arch
The new Minalign field sets the minimum alignment for all symbols.
This is required for the upcoming s390x port which requires symbols
be 2-byte aligned for efficient relative addressing.
All preexisting architectures have Minalign set to 1 which means
that this commit should have no effect.
I tested values of 2, 4 and 8 on linux amd64 and the tests appear to
pass. Increasing Minalign to 16 appears to break the runtime. I
think this is due to assumptions made about the layout of module
data.
toolstash -cmp on linux amd64 shows no changes due to this commit.
Resolves #14604
Change-Id: I0fe042d52c4e4732eba5fabcd0c31102a2408764
Reviewed-on: https://go-review.googlesource.com/20149 Reviewed-by: Minux Ma <minux@golang.org>
Run-TryBot: Minux Ma <minux@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Mikio Hara [Wed, 2 Mar 2016 11:08:18 +0000 (20:08 +0900)]
net: deduplicate TCP socket code
This change consolidates functions and methods related to TCPAddr,
TCPConn and TCPListener for maintenance purpose, especially for
documentation. Also refactors Dial error code paths.
The followup changes will update comments and examples.
Joe Tsai [Wed, 2 Mar 2016 20:21:53 +0000 (12:21 -0800)]
compress/bzip2: refactor unit tests
Over the years as more bugs were discovered with the bzip2 library,
new Tests were appended the unit tests and the tests became gnarly.
Clean up the tests to be more consistent with modern Go style in
addition to coalescing common tests into a general version that
iterates over a list of input/output pairs. This has the advantage that
the input, output, and test code are all in the same area, rather than
being sprawled around the test file.
There is no loss of test coverage.
Change-Id: I377ed89378f0b89763d4a56ffc37b22d9c2a369e
Reviewed-on: https://go-review.googlesource.com/20133
Run-TryBot: Joe Tsai <joetsai@digital-static.net>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Matthew Dempsky [Thu, 3 Mar 2016 02:03:50 +0000 (18:03 -0800)]
cmd/compile: eliminate arch-specific typedefs
Arch backends already provide us Widthint and Widthptr, which is ample
information to figure out how to define the universal "int", "uint",
and "uintptr" types. No need for providing a generic typedef
mechanism beyond that.
Change-Id: I35c0c17a67c80605a9208b93d77d6960b2cbb17d
Reviewed-on: https://go-review.googlesource.com/20153
Run-TryBot: Matthew Dempsky <mdempsky@google.com> Reviewed-by: Dave Cheney <dave@cheney.net> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Alex Brainman [Wed, 2 Mar 2016 05:58:40 +0000 (16:58 +1100)]
misc/cgo/test: rewrite windows version of mysleep
Latest version of gcc (tdm-1) 5.1.0 refuses to compile our code
on windows/386 (see issue for details). Rewrite the code.
Fixes #14328
Change-Id: I70f4f063282bd2958cd2175f3974369dd49dd8dc
Reviewed-on: https://go-review.googlesource.com/20008 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Robert Griesemer [Fri, 23 Oct 2015 01:56:45 +0000 (18:56 -0700)]
cmd/compile/internal/gc: export & import function bodies, but don't hook up
Function bodies are not yet hooked up because the node structure is not
100% correct. This commit establishes that we can correctly write bodies
out and read them in again.
- export and import all exported inlined function bodies:
(export GO_GCFLAGS="-newexport"; sh all.bash) working
- inlined functions are not yet hooked up (just dropped on the floor)
- improved tracing output and error messages
- make mkbuiltin.go work for both textual and binary export data
so we can run tests with the new format
Change-Id: I70dc4de419df1b604389c3747041d6dba8730b0b
Reviewed-on: https://go-review.googlesource.com/16284 Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Change-Id: I43de9eee2d96f9c46f7b0fbe099a0571164324f5
Reviewed-on: https://go-review.googlesource.com/20107 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Tamir Duberstein [Fri, 26 Feb 2016 19:17:29 +0000 (14:17 -0500)]
crypto/tls: check errors from (*Conn).writeRecord
This promotes a connection hang during TLS handshake to a proper error.
This doesn't fully address #14539 because the error reported in that
case is a write-on-socket-not-connected error, which implies that an
earlier error during connection setup is not being checked, but it is
an improvement over the current behaviour.
Updates #14539.
Change-Id: I0571a752d32d5303db48149ab448226868b19495
Reviewed-on: https://go-review.googlesource.com/19990 Reviewed-by: Adam Langley <agl@golang.org>
David Crawshaw [Wed, 2 Mar 2016 15:01:01 +0000 (10:01 -0500)]
cmd/link: C is gone, remove check for Go calling C
It looks like the compiler still uses the Cfunc flag for functions
marked as //go:systemstack, but if I'm reading this right, that
doesn't apply here and the linker no longer needs Cfunc.
Joe Tsai [Wed, 2 Mar 2016 11:22:00 +0000 (03:22 -0800)]
compress/bzip2: prevent zero-length Huffman codes
Unlike RFC 1951 (DEFLATE), bzip2 does not use zero-length Huffman codes
to indicate that the symbol is missing. Instead, bzip2 uses a sparse
bitmap to indicate which symbols are present. Thus, it is undefined what
happens when a length of zero is used. Thus, fix the parsing logic so that
the length cannot ever go below 1-bit similar to how the C logic does things.
David Crawshaw [Mon, 29 Feb 2016 22:07:50 +0000 (13:07 -0900)]
cmd/link: write DWARF PC table without seeking
This per-symbol table was written with the strategy:
1. record offset and write fake header
2. write body
3. seek back to fake header
4. write real header
This CL collects the per-symbol body into a []byte, then writes the
real header followed by the body to the output file. This saves two
seeks per-symbol and overwriting the fake header.
Small performance improvement (3.5%) in best-of-ten links of godoc:
tip: real 0m1.132s user 0m1.256s
this: real 0m1.090s user 0m1.210s
I'm not sure if the performance measured here alone justifies it,
but I think this is an easier to read style of code.
Marvin Stenger [Tue, 1 Mar 2016 20:18:56 +0000 (21:18 +0100)]
cmd/vet: polish output of shadow test
This commit modifies the style of a error message in case of -shadow.
Previously such a message would look like:
foo.go:42: declaration of err shadows declaration at shadow.go:13:
Changes of the commit include highlighting the variable name and
removing the ": "(space intended) at the end of the line:
foo.go:42: declaration of "err" shadows declaration at shadow.go:13
Fixes #14585.
Change-Id: Ia6a6bf396668dcba9a24f025a08d8826db31f434
Reviewed-on: https://go-review.googlesource.com/20093 Reviewed-by: Rob Pike <r@golang.org>
Brad Fitzpatrick [Tue, 1 Mar 2016 23:21:55 +0000 (23:21 +0000)]
all: single space after period.
The tree's pretty inconsistent about single space vs double space
after a period in documentation. Make it consistently a single space,
per earlier decisions. This means contributors won't be confused by
misleading precedence.
This CL doesn't use go/doc to parse. It only addresses // comments.
It was generated with:
$ perl -i -npe 's,^(\s*// .+[a-z]\.) +([A-Z]),$1 $2,' $(git grep -l -E '^\s*//(.+\.) +([A-Z])')
$ go test go/doc -update
Change-Id: Iccdb99c37c797ef1f804a94b22ba5ee4b500c4f7
Reviewed-on: https://go-review.googlesource.com/20022 Reviewed-by: Rob Pike <r@golang.org> Reviewed-by: Dave Day <djd@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Ian Lance Taylor [Tue, 1 Mar 2016 22:52:18 +0000 (14:52 -0800)]
cmd/compile: fix OADDSTR buffer size calculation
The size calculation has been wrong since this code was first committed
in https://golang.org/cl/3120. The effect was that the compiler always
allocated a temporary buffer on the stack for a non-escaping string
concatenation. This turns out to make no practical difference, as the
compiler always allocates a buffer of the same size (32 bytes) and the
runtime only uses the temporary buffer if the concatenated strings
fit (check is in rawstringtmp in runtime/string.go).
The effect of this change is to avoid generating a temporary buffer on
the stack that will not be used.
Change-Id: Id632bfe3d6c113c9934c018a2dd4bcbf1784a63d
Reviewed-on: https://go-review.googlesource.com/20112
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
This appears to make no difference in performance of real
code, meaning the compiler itself. The earlier version of
this has been stripped down to help make the cost of this
only-aesthetic-on-Intel phase be as cheap as possible (we
probably want information about inner loops for improving
register allocation, but because register allocation follows
close behind this pass, conceivably the information could be
reused -- so we might do this anyway just to normalize
output).
For a ./make.bash that takes 200 user seconds, about .75
second is reported in likelyadjust (summing nanoseconds
reported with -d=ssa/likelyadjust/time ).
Upstream predictions are respected.
Includes test, limited to build on amd64 only.
Did several iterations on the debugging output to allow
some rough checks on behavior.
Debug=1 logging notes agree/disagree with earlier passes,
allowing analysis like the following:
Run on make.bash:
GO_GCFLAGS=-d=ssa/likelyadjust/debug \
./make.bash >& lkly5.log