David Crawshaw [Thu, 25 Aug 2016 18:31:50 +0000 (14:31 -0400)]
cmd/compile: qualify unexported fields of unnamed types
The compiler was canonicalizing unnamed types of the form
struct { i int }
across packages, even though an unexported field i should not be
accessible from other packages.
The fix requires both qualifying the field name in the string used by
the compiler to distinguish the type, and ensuring the struct's pkgpath
is set in the rtype version of the data when the type being written is
not part of the localpkg.
Fixes #16616
Change-Id: Ibab160b8b5936dfa47b17dbfd48964a65586785b
Reviewed-on: https://go-review.googlesource.com/27791
Run-TryBot: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org> Reviewed-by: Robert Griesemer <gri@golang.org>
Hiroshi Ioka [Fri, 26 Aug 2016 08:44:30 +0000 (17:44 +0900)]
os: fix build error on plan9
https://go-review.googlesource.com/#/c/27580 added the test.
However the test use syscall.ELOOP which is not defined on plan9.
Move test code from "os_test.go" to "os_windows_test.go" to prevent
build error.
Change-Id: Ie7f05bfb9ab229e06a8e82a4b3b8a7ca82d4663b
Reviewed-on: https://go-review.googlesource.com/27833
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: David du Colombier <0intro@gmail.com> Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Hiroshi Ioka [Thu, 18 Aug 2016 07:10:28 +0000 (16:10 +0900)]
os: prevent infinite symlink loop of Stat on Windows
The Windows version of Stat calls Readlink iteratively until
reaching a non-symlink file.
If the given file is a circular symlink, It never stops.
This CL defines the maximum number of symlink loop count.
If the loop count will exceed that number, Stat will return error.
Robert Griesemer [Thu, 25 Aug 2016 23:53:10 +0000 (16:53 -0700)]
cmd/compile: fail gracefully on export format skew
Import errors due to unexpected format are virtually
always due to version skew. Don't panic but report a
good error message (incl. hint that the imported package
needs to be reinstalled) if not in debugFormat mode.
Recognize export data format version and store it so
it can be used to automatically handle minor version
differences. We did this before, but not very well.
No export data format changes.
Manually tested with corrupted export data.
For #16881.
Change-Id: I53ba98ef747b1c81033a914bb61ee52991f35a90
Reviewed-on: https://go-review.googlesource.com/27814 Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Keith Randall [Tue, 23 Aug 2016 23:49:28 +0000 (16:49 -0700)]
cmd/compile: inline atomics from runtime/internal/atomic on amd64
Inline atomic reads and writes on amd64. There's no reason
to pay the overhead of a call for these.
To keep atomic loads from being reordered, we make them
return a <value,memory> tuple.
Change the meaning of resultInArg0 for tuple-generating ops
to mean the first part of the result tuple, not the second.
This means we can always put the store part of the tuple last,
matching how arguments are laid out. This requires reordering
the outputs of add32carry and sub32carry and their descendents
in various architectures.
benchmark old ns/op new ns/op delta
BenchmarkAtomicLoad64-8 2.09 0.26 -87.56%
BenchmarkAtomicStore64-8 7.54 5.72 -24.14%
Add missing function prototypes.
Fix function prototypes.
Use FP references instead of SP references.
Fix variable names.
Update comments.
Clean up whitespace. (Not for vet.)
Joe Tsai [Sat, 20 Aug 2016 08:46:32 +0000 (01:46 -0700)]
archive/tar: isolate regular and sparse file handling as methods
Factor out the regular file handling logic into handleRegularFile
from nextHeader. We will need to reuse this logic when fixing #15573
in a future CL.
Factor out the sparse file handling logic into handleSparseFile.
Currently this logic is split between nextHeader (for GNU sparse
files) and Next (for PAX sparse files). Instead, we move this
related code into a single method.
There is no overall logic change. Thus, no unit tests.
Sina Siadat [Fri, 17 Jun 2016 16:32:59 +0000 (21:02 +0430)]
net/http: send Content-Range if no byte range overlaps
RFC 7233, section 4.4 says:
>>>
For byte ranges, failing to overlap the current extent means that the
first-byte-pos of all of the byte-range-spec values were greater than the
current length of the selected representation. When this status code is
generated in response to a byte-range request, the sender SHOULD generate a
Content-Range header field specifying the current length of the selected
representation
<<<
Thus, we should send the Content-Range only if none of the ranges
overlap.
cmd/compile: when inlining ==, don’t take the address of the values
This CL reworks walkcompare for clarity and concision.
It also makes one significant functional change.
(The functional change is hard to separate cleanly
from the cleanup, so I just did them together.)
When inlining and unrolling an equality comparison
for a small struct or array, compare the elements like:
The result is the same, but taking the address
and working through the indirect
forces the backends to generate less efficient code.
This is only an improvement with the SSA backend.
However, every port but s390x now has a working
SSA backend, and switching to the SSA backend
by default everywhere is a priority for Go 1.8.
It thus seems reasonable to start to prioritize
SSA performance over the old backend.
Ian Lance Taylor [Thu, 7 Jul 2016 00:11:29 +0000 (17:11 -0700)]
path/filepath: don't return SkipDir at top
If the walker function called on a top-level file returns SkipDir,
then (before this change) Walk would return SkipDir, which the
documentation implies will not happen.
Fixes #16280.
Change-Id: I37d63bdcef7af4b56e342b624cf0d4b42e65c297
Reviewed-on: https://go-review.googlesource.com/24780
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Cherry Zhang [Mon, 22 Aug 2016 16:25:23 +0000 (12:25 -0400)]
cmd/compile: get MIPS64 SSA working
- implement *, /, %, shifts, Zero, Move.
- fix mistakes in comparison.
- fix floating point rounding.
- handle RetJmp in assembler (which was not handled, as a consequence
Duff's device was disabled in the old backend.)
Ian Lance Taylor [Wed, 24 Aug 2016 23:06:10 +0000 (16:06 -0700)]
runtime: remove unused type sigtabtt
The type sigtabtt was introduced by an automated tool in
https://golang.org/cl/167550043. It was the Go version of the C type
SigTab. However, when the C code using SigTab was converted to Go in
https://golang.org/cl/168500044 it was rewritten to use a different Go
type, sigTabT, rather than sigtabtt (the difference being that sigTabT
uses string where sigtabtt uses *int8 from the C type char*). So this is
just a dreg from the conversion that was never actually used.
Change-Id: I2ec6eb4b25613bf5e5ad1dbba1f4b5ff20f80f55
Reviewed-on: https://go-review.googlesource.com/27691
Run-TryBot: Ian Lance Taylor <iant@golang.org> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
cmd/compile: optimize integer "in range" expressions
Use unsigned comparisons to reduce from
two comparisons to one for integer "in range"
checks, such as a <= b && b < c.
We already do this for bounds checks.
Extend it to user code.
This is much easier to do in the front end than SSA.
A back end optimization would be more powerful,
but this is a good start.
This reduces the power of some of SSA prove
inferences (#16653), but those regressions appear
to be rare and not worth holding this CL for.
Dave Cheney [Wed, 24 Aug 2016 10:27:38 +0000 (20:27 +1000)]
cmd/link/internal/ld: move ld.Cpos to coutbuf.Offset
This change moves the ld.Cpos function to a method on coutbuf. This is
part of a larger change that makes ld.outbuf look more like a bio.Buf in
an effort to eventually replace the former with the latter.
Change-Id: I506f7131935a2aa903fa302a0fab0c5be50220fd
Reviewed-on: https://go-review.googlesource.com/27578
Run-TryBot: Dave Cheney <dave@cheney.net> Reviewed-by: David Crawshaw <crawshaw@golang.org> Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Dave Cheney [Wed, 24 Aug 2016 10:44:09 +0000 (20:44 +1000)]
cmd/internal/*: only call ctx.Bso.Flush when something has been written
In many places where ctx.Bso.Flush is used as the target for some debug
logging, ctx.Bso.Flush is called unconditionally. In the majority of
cases where debug logging is not enabled, this means Flush is called
many times when there is nothing to be flushed (it will be called anyway
when ctx.Bso is eventually closed), sometimes in a loop.
Avoid this by moving the ctx.Bso.Flush call into the same condition
block as the debug print. This pattern was previously applied
sporadically.
Change-Id: I0444cb235cc8b9bac51a59b2e44e59872db2be06
Reviewed-on: https://go-review.googlesource.com/27579
Run-TryBot: Dave Cheney <dave@cheney.net>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: David Crawshaw <crawshaw@golang.org> Reviewed-by: Matthew Dempsky <mdempsky@google.com>
cmd/asm, go/build: invoke cmd/asm only once per package
Prior to this CL, cmd/go invoked cmd/asm once
for every assembly file.
The exec and cmd/asm startup overhead dwarfed
the actual time spent assembling.
This CL adds support to cmd/asm to process
multiple input files and uses it in cmd/go.
This cuts 10% off the wall time for 'go build -a math'.
Robert Griesemer [Wed, 24 Aug 2016 18:33:55 +0000 (11:33 -0700)]
spec: complete list of special comma-ok forms
The enumerations didn't include the syntactic form where the lhs is
full variable declaration with type specification, as in:
var x, ok T = ...
Fixes #15782.
Change-Id: I0f7bafc37dc9dcf62cdb0894a0d157074ccd4b3e
Reviewed-on: https://go-review.googlesource.com/27670 Reviewed-by: Matthew Dempsky <mdempsky@google.com> Reviewed-by: Ian Lance Taylor <iant@golang.org> Reviewed-by: Rob Pike <r@golang.org>
Keith Randall [Sun, 3 Jul 2016 00:19:25 +0000 (17:19 -0700)]
cmd/objdump: implement objdump of .o files
Update goobj reader so it can provide all the information
necessary to disassemble .o (and .a) files.
Grab architecture of .o files from header.
.o files have relocations in them. This CL also contains a simple
mechanism to disassemble relocations and add relocation info as an extra
column in the output.
Fixes #13862
Change-Id: I608fd253ff1522ea47f18be650b38d528dae9054
Reviewed-on: https://go-review.googlesource.com/24818 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Ian Lance Taylor [Wed, 24 Aug 2016 00:47:11 +0000 (17:47 -0700)]
net: use runtime.Keepalive for *netFD values
The net package sets a finalizer on *netFD. I looked through all the
uses of *netFD in the package, looking for each case where a *netFD
was passed as an argument and the final reference to the argument was
not a function or method call. I added a call to runtime.KeepAlive after
each such final reference (there were only three).
The code is safe today without the KeepAlive calls because the compiler
keeps arguments alive for the duration of the function. However, that is
not a language requirement, so adding the KeepAlive calls ensures that
this code remains safe even if the compiler changes in the future.
Change-Id: I4e2bd7c5a946035dc509ccefb4828f72335a9ee3
Reviewed-on: https://go-review.googlesource.com/27650
Run-TryBot: Ian Lance Taylor <iant@golang.org> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Ian Lance Taylor [Tue, 23 Aug 2016 22:56:43 +0000 (15:56 -0700)]
os: use runtime.Keepalive for *Process values
The os package sets a finalizer on *Process. I looked through all the
uses of *Process in the package, looking for each case where a *Process
was passed as an argument and the final reference to the argument was
not a function or method call. I added a call to runtime.KeepAlive after
each such final reference (there were only three).
The code is safe today without the KeepAlive calls because the compiler
keeps arguments alive for the duration of the function. However, that is
not a language requirement, so adding the KeepAlive calls ensures that
this code remains safe even if the compiler changes in the future.
I also removed an existing unnecessry call to runtime.KeepAlive. The
syscall.Syscall function is handled specially by the compiler to keep
its arguments alive.
Change-Id: Ibd2ff20b31ed3de4f6a59dd1633c1b44001d91d9
Reviewed-on: https://go-review.googlesource.com/27637
Run-TryBot: Ian Lance Taylor <iant@golang.org> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Jess Frazelle [Fri, 20 May 2016 23:17:27 +0000 (16:17 -0700)]
cmd/go: updates go get to return exit status 0 for test file only pkgs
Updates the behavior of `go get` to return exit status 0 when a
requested package only contains test files.
Fixes #15093
Change-Id: I76b80517d58748090f5e8c6f41178361e2d7ca54
Reviewed-on: https://go-review.googlesource.com/23314 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Keith Randall [Tue, 23 Aug 2016 17:43:47 +0000 (10:43 -0700)]
cmd/compile: clean up ctz ops
Now that we have ops that can return 2 results, have BSF return a result
and flags. We can then get rid of the redundant comparison and use CMOV
instead of CMOVconst ops.
Get rid of a bunch of the ops we don't use. Ctz{8,16}, plus all the Clzs,
and CMOVNEs. I don't think we'll ever use them, and they would be easy
to add back if needed.
Change-Id: I8858a1d017903474ea7e4002fc76a6a86e7bd487
Reviewed-on: https://go-review.googlesource.com/27630
Run-TryBot: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: David Chase <drchase@google.com>
Robert Griesemer [Tue, 23 Aug 2016 23:02:19 +0000 (16:02 -0700)]
cmd/compile: export package for _ (blank) struct fields
Blank struct fields are regular unexported fields. Two
blank fields are different if they are from different
packages. In order to correctly differentiate them, the
compiler needs the package information. Add it to the
export data.
Fixes #15514.
Change-Id: I421aaca22b542fcd0d66b2d2db777249cad78df6
Reviewed-on: https://go-review.googlesource.com/27639 Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Ross Light [Tue, 16 Aug 2016 18:10:36 +0000 (11:10 -0700)]
reflect: document equality guarantee for Type
The only previous mention of this property was in the String() method.
Since this is the only way to uniquely identify a type and we can't
change this property without breaking the Go 1 guarantee, it seems
better to document this property than hiding it on a method.
Fixes #16348
Change-Id: I3d25f7d6e6007e3c15c2e13010869888d0181fc2
Reviewed-on: https://go-review.googlesource.com/27170 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Than McIntosh [Mon, 11 Jul 2016 18:26:18 +0000 (14:26 -0400)]
test: add test for gccgo issue #15722
Change-Id: I4faf9a55414e217f0c48528efb13ab8fdcd9bb16
Reviewed-on: https://go-review.googlesource.com/24845 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
Kevin Burke [Sun, 21 Aug 2016 05:05:47 +0000 (22:05 -0700)]
cmd/compile: precompute constant square roots
If a program wants to evaluate math.Sqrt for any constant value
(for example, math.Sqrt(3)), we can replace that expression with
its evaluation (1.7320508075688772) at compile time, instead of
generating a SQRT assembly command or equivalent.
Adds tests that math.Sqrt generates the correct values. I also
compiled a short program and verified that the Sqrt expression was
replaced by a constant value in the "after opt" step.
Adds a short doc to the top of generic.rules explaining what the file
does and how other files interact with it.
Gabriel Russell [Tue, 23 Aug 2016 12:31:26 +0000 (08:31 -0400)]
time: fix optional fractional seconds range err
The optional fractional seconds overrides any range error
from the second parsing. Instead don't look for optional fractional
seconds if a range error has occured.
the correlation "some error on reading after reading operation" looks a
bit confusing because the operation on the ReadFrom method of TCPConn is
actually "writing after reading." To clarify and avoid confusion, this
change sets "readfrom" to the Op field of outer-most OpError instead of
"read."
Michael Matloob [Tue, 23 Aug 2016 12:17:58 +0000 (08:17 -0400)]
debug/macho: fix comment on Section64
Change-Id: I7c809ec385b56ebb2ec784a1479d466df6ab4d1a
Reviewed-on: https://go-review.googlesource.com/27565 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Ian Lance Taylor [Tue, 23 Aug 2016 13:14:31 +0000 (06:14 -0700)]
doc: add note about CL 24180 to go1.8.txt
Change-Id: Ie2bef1c181a49d7a02ed8068894d2bd81fc5bafa
Reviewed-on: https://go-review.googlesource.com/27566 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Ian Lance Taylor [Wed, 6 Jul 2016 22:30:33 +0000 (15:30 -0700)]
runtime: add missing race and msan checks to reflect functions
Add missing race and msan checks to reflect.typedmmemove and
reflect.typedslicecopy. Missing these checks caused the race detector
to miss races and caused msan to issue false positive errors.
Elias Naur [Mon, 27 Jun 2016 19:38:04 +0000 (21:38 +0200)]
time: load time zones from the system tzdata file on Android
Android timezones are in a packed format, different from the separate
files of a regular Unix system. This CL contain the necessary code to
parse the packed tzdata file and extract time zones from it. It also
adds a basic test to ensure the new parser works.
Fixes #13581
Change-Id: Idebe73726c3d4c2de89dd6ae1d7d19f975207500
Reviewed-on: https://go-review.googlesource.com/24494 Reviewed-by: Hyang-Ah Hana Kim <hyangah@gmail.com> Reviewed-by: David Crawshaw <crawshaw@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
cmd/compile: use a map to track const switch cases
This is simpler than the sorting technique.
It also allows us to simplify or eliminate
some of the sorting decisions.
Most important, sorting will not work when case clauses
represent ranges of integers: There is no correct
sort order that allows overlap detection by comparing
neighbors. Using a map allows of a cheap, simple
approach to ranges, namely to insert every int
in the map. The equivalent approach for sorting
means juggling temporary Nodes for every int,
which is a lot more expensive.
Dhaivat Pandit [Mon, 22 Aug 2016 17:29:02 +0000 (10:29 -0700)]
net/http/httptest: updated example to use Result()
example for httptest.Recorder was inspecting Recoder directly.
Using Result() to convert Recorder into a http.Response yields a much
better user experience.
Ian Lance Taylor [Tue, 23 Aug 2016 04:49:32 +0000 (21:49 -0700)]
cmd/link: fix handling of cgo_dynamic_interpreter
CL 27473 accidentally changed `!Debug['I']` to `*flagInterpreter != ""`.
Since the old `Debug['I']` was set when the new *flagInterpreter was
set, this inverted the sense of the condition. The effect was to always
ignore the cgo_dynamic_interpreter setting from runtime/cgo. This worked
OK when the default interpreter was the correct one, but failed when it
was not, as is currently the case on, at least, PPC64 and ARM.
This CL restores the old behavior by using a separate variable to track
whether the -I flag was used, just as we used to.
Change-Id: Icf9b65fa41349ed2e4de477fec0a557ef1eb8189
Reviewed-on: https://go-review.googlesource.com/27562
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Brad Fitzpatrick [Mon, 22 Aug 2016 22:22:44 +0000 (22:22 +0000)]
io: fix comment in test
Updates #16795
Change-Id: I0bcc34bb5a92a2c480aebfb0eb6ba57bcc7f7cfd
Reviewed-on: https://go-review.googlesource.com/27551 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Eric Lagergren [Sat, 20 Aug 2016 22:46:06 +0000 (15:46 -0700)]
encoding/xml: do not ignore error return from copyValue
The error return from copyValue was ignored causing some XML attribute
parsing to swallow an error.
Additionally, type MyMarshalerAttrTest had no UnmarshalXMLAttr method
causing marshalTests not to be symmetrical and the test suite to fail
for test case 101.
Dmitry Vyukov [Mon, 8 Aug 2016 14:02:09 +0000 (16:02 +0200)]
runtime: speed up StartTrace with lots of blocked goroutines
In StartTrace we emit EvGoCreate for all existing goroutines.
This includes stack unwind to obtain current stack.
Real Go programs can contain hundreds of thousands of blocked goroutines.
For such programs StartTrace can take up to a second (few ms per goroutine).
Obtain current stack ID once and use it for all EvGoCreate events.
This speeds up StartTrace with 10K blocked goroutines from 20ms to 4 ms
(win for StartTrace called from net/http/pprof hander will be bigger
as stack is deeper).
Change-Id: I9e5ff9468331a840f8fdcdd56c5018c2cfde61fc
Reviewed-on: https://go-review.googlesource.com/25573
Run-TryBot: Dmitry Vyukov <dvyukov@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Hyang-Ah Hana Kim <hyangah@gmail.com>