gosweepone just switches to the system stack and calls sweepone.
sweepone doesn't need to run on the system stack, so this is pretty
pointless.
Historically, this was necessary because the sweeper was written in C
and hence needed to run on the system stack. gosweepone was the
function that Go code (specifically, bgsweep) used to call into the C
sweeper implementation. This probably became unnecessary in 2014 with
CL golang.org/cl/167540043, which ported the sweeper to Go.
This CL changes all callers of gosweepone to call sweepone and
eliminates gosweepone.
runtime: simplify free count calculation in (un)cacheSpan
For unclear reasons, cacheSpan and uncacheSpan compute the number of
elements in a span by dividing its size by the element size. This
number is simply available in the mspan structure, so just use it.
Austin Clements [Mon, 8 Oct 2018 17:51:10 +0000 (13:51 -0400)]
runtime: avoid tracking spans with no objects with mcentral
Lazy mcache flushing (golang.org/cl/134783) made it so that moving a
span from an mcache to an mcentral was sometimes responsible for
sweeping the span. However, it did a "preserving" sweep, which meant
it retained ownership, even if the sweeper swept all objects in the
span. As a result, we could put a completely unused span back in the
mcentral.
Fix this by first taking back ownership of the span into the mcentral
and moving it to the right mcentral list, and then doing a
non-preserving sweep. The non-preserving sweep will move the span to
the heap if it sweeps all objects.
Austin Clements [Mon, 8 Oct 2018 23:24:46 +0000 (19:24 -0400)]
runtime: fix non-in-use span check
Lazy mcache flushing (golang.org/cl/134783) introduced a second value
for sweepgen that indicates a span has been swept. I missed adding
this case to a sanity check in sweepone, so it can now panic if it
finds a non-in-use spans that's been swept *and* put in an mcache.
Fix this by adding the second sweepgen case to this check.
freeSpan currently takes a mysterious "acct int32" argument. This is
really just a boolean and actually just needs to match the "large"
argument to alloc in order to balance out accounting.
To make this clearer, replace acct with a "large bool" argument that
must match the call to mheap.alloc.
net/http/httputil: rewrite flushing code, disable on Server-Sent Events
* Rewrite the flushing code to not use a persistent goroutine, which
also simplifies testing.
* Define the meaning of a negative flush interval. Its meaning doesn't
change, but now it's locked in, and then we can use it to optimize
the performance of the non-buffered case to avoid use of an AfterFunc.
* Support (internal-only) special casing of FlushInterval values per
request/response.
* For now, treat Server-Sent Event responses as unbuffered. (or rather,
immediately flushed from the buffer per-write)
Marten Seemann [Fri, 17 Aug 2018 11:03:55 +0000 (18:03 +0700)]
crypto/tls: fix ServerHello SCT test
According to https://tools.ietf.org/html/rfc6962#section-3.3, the SCT
must be at least one byte long. The parsing code correctly checks for
this condition, but rarely the test does generate an empty SCT.
Ben Shi [Mon, 8 Oct 2018 02:20:03 +0000 (02:20 +0000)]
cmd/compile: combine similar code in amd64's assembly generator
BSFQ/BSRQ/BSFL/BSRL/SQRTSD have similar logic in amd64's assembly
generator. This CL combines them together while does not impact
generated amd64 code. The total size of
pkg/linux_amd64/cmd/compile/internal decreases about 1.8KB.
Change-Id: I5f3210c5178c20ac9108877c69f17234baf5b6b7
Reviewed-on: https://go-review.googlesource.com/c/140438
Run-TryBot: Ben Shi <powerman1st@163.com>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
Ben Shi [Sat, 6 Oct 2018 13:13:48 +0000 (13:13 +0000)]
cmd/compile: add indexed form for several 386 instructions
This CL implements indexed memory operands for the following instructions.
(ADD|SUB|MUL|AND|OR|XOR)Lload -> (ADD|SUB|MUL|AND|OR|XOR)Lloadidx4
(ADD|SUB|AND|OR|XOR)Lmodify -> (ADD|SUB|AND|OR|XOR)Lmodifyidx4
(ADD|AND|OR|XOR)Lconstmodify -> (ADD|AND|OR|XOR)Lconstmodifyidx4
1. The total size of pkg/linux_386/ decreases about 2.5KB, excluding
cmd/compile/ .
strconv: add example for QuoteRuneToGraphic and QuoteToGraphic functions
Change-Id: Ie5b2ef0087dbc7b8191de8c8b4190396631e3c7f
Reviewed-on: https://go-review.googlesource.com/c/137215
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org>
Robert Griesemer [Tue, 2 Oct 2018 22:55:38 +0000 (15:55 -0700)]
spec: distinguish between explicit and implicit conversions (clarification)
The spec used the term "conversion" somewhat indiscriminately for
explicit conversions that appear literally in the source, and implicit
conversions that are implied by the context of an expression.
Be clearer about it by defining the terms.
Also, state that integer to string conversions of the form string(x)
are never implicit. This clarifies situations where implicit conversions
might require an integer to change to a string (but don't and never have
done so). See line 3948.
Fixes #26313.
Change-Id: I8939466df6b5950933ae7c987662ef9f88827fda
Reviewed-on: https://go-review.googlesource.com/c/139099 Reviewed-by: Rob Pike <r@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org>
Alex Brainman [Sun, 7 Oct 2018 00:22:57 +0000 (11:22 +1100)]
net: skip TestUnixConnLocalWindows on windows/arm
Similarly to CL 138676, skip TestUnixConnLocalWindows on windows/arm.
Fixes #28061
Change-Id: I2270d2f9d268e85ea567be0c0c37c48e4d482282
Reviewed-on: https://go-review.googlesource.com/c/140397
Run-TryBot: Alex Brainman <alex.brainman@gmail.com> Reviewed-by: Ian Lance Taylor <iant@golang.org>
Ben Shi [Sat, 6 Oct 2018 03:35:17 +0000 (03:35 +0000)]
cmd/compile: combine similar code in amd64's assembly generator
This CL combines similar code in amd64's assembly generator. The
total size of pkg/linux_amd64/cmd/compile/ decreases about 4.5KB,
while the generated amd64 code is not affected.
Change-Id: I4cdbdd22bde8857aafdc29b47fa100a906fa1598
Reviewed-on: https://go-review.googlesource.com/c/140298
Run-TryBot: Ben Shi <powerman1st@163.com> Reviewed-by: Keith Randall <khr@golang.org>
Ben Shi [Fri, 5 Oct 2018 12:09:43 +0000 (12:09 +0000)]
cmd/compile: combine similar code in 386's assembly generator
The indexed MOVload and MOVstore have similar logic, and this CL
combine them together. The total size of pkg/linux_386/cmd/compile/
decreases about 4KB.
Change-Id: I06236a3542aaa3dfc113c49fe4c69d209e018dfe
Reviewed-on: https://go-review.googlesource.com/c/139958
Run-TryBot: Ben Shi <powerman1st@163.com>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org>
Reason for revert: It broke all the darwin builders; it's also not
obvious how the weird darwin versions (900, 1000) relate to the > 3.9
requisite, so I'm not sure how to decide about skipping in a robust
way. It's better to revert the check for now.
Fixes #28028
Change-Id: Ibbcb7bf7cd2136e0851ebd097a2bc4dec9f0ee18
Reviewed-on: https://go-review.googlesource.com/c/140217 Reviewed-by: Keith Randall <khr@golang.org>
Mikio Hara [Thu, 4 Oct 2018 23:28:20 +0000 (08:28 +0900)]
vendor: update golang_org/x/net/route from upstream
Updates the route package to git rev 146acd2 for:
- 146acd2 don't run NET_RT_IFLIST vs. NET_RT_IFLISTL test in 386 emulation (again)
Change-Id: I24de1eb31b2ca0e24cb9ab1648f7a71b5067cf97
Reviewed-on: https://go-review.googlesource.com/c/139937 Reviewed-by: Ian Lance Taylor <iant@golang.org>
This is similar to CL 136816 for x/sys/unix, changing the FreeBSD ABI to use 64-bit inodes in
Stat_t, Statfs_t, and Dirent types.
The changes are forward compatible, that is FreeBSD 10.x, 11.x continue to use their current sysnum numbers.
The affected types are converted to the new layout (with some overhead).
Thus the same statically linked binary should work using the native sysnums (without any conversion) on FreeBSD 12.
Breaking API changes in package syscall are:
Mknod takes a uint64 (C dev_t) instead of int.
Stat_t: Dev, Ino, Nlink, Rdev, Gen became uint64.
Atimespec, Mtimespec, Ctimespec, Birthtimespec renamed to Atim, Mtim, Ctim, Birthtim respectively.
Statfs_t: Mntonname and Mntfromname changed from [88]int8 to [1024]int8 arrays.
Dirent: Fileno became uint64, Namlen uint16 and an additional field Off int64 (currently unused) was added.
The following commands were run to generate ztypes_* and zsyscall_* on FreeBSD-12.0-ALPHA6 systems (GOARCH=386 were run on the same amd64 host):
GOOS=freebsd GOARCH=amd64 ./mksyscall.pl -tags freebsd,amd64 syscall_bsd.go syscall_freebsd.go syscall_freebsd_amd64.go |gofmt >zsyscall_freebsd_amd64.go
GOOS=freebsd GOARCH=amd64 go tool cgo -godefs types_freebsd.go | GOOS=freebsd GOARCH=amd64 go run mkpost.go >ztypes_freebsd_amd64.go
GOOS=freebsd GOARCH=386 ./mksyscall.pl -l32 -tags freebsd,386 syscall_bsd.go syscall_freebsd.go syscall_freebsd_386.go |gofmt >zsyscall_freebsd_386.go
GOOS=freebsd GOARCH=386 go tool cgo -godefs types_freebsd.go | GOOS=freebsd GOARCH=386 go run mkpost.go >ztypes_freebsd_386.go
GOOS=freebsd GOARCH=arm ./mksyscall.pl -l32 -arm -tags freebsd,arm syscall_bsd.go syscall_freebsd.go syscall_freebsd_arm.go |gofmt >zsyscall_freebsd_arm.go
GOOS=freebsd GOARCH=arm go tool cgo -godefs -- -fsigned-char types_freebsd.go | GOOS=freebsd GOARCH=arm go run mkpost.go >ztypes_freebsd_arm.go
The Kevent struct was changed to use the FREEBSD_COMPAT11 version always (requiring the COMPAT_FREEBSD11 kernel option FreeBSD-12, this is the default).
The definitions of ifData were not updated, their functionality in has have been replaced by vendored golang.org/x/net/route.
freebsdVersion initialization was dropped from init() in favor of a sync.Once based wrapper - supportsABI().
Robert Griesemer [Fri, 5 Oct 2018 00:16:33 +0000 (17:16 -0700)]
go/types: remove work-around for issue #26390
This work-around is not needed anymore now that method
signatures are type-checked separately from their receiver
base types: no artificial cycles are introduced anymore
and so there is no need to artificially cut them.
Fixes #26854.
Change-Id: I2ef15ceeaa0b486f65f6cdc466d0cf06246c74d7
Reviewed-on: https://go-review.googlesource.com/c/139900 Reviewed-by: Alan Donovan <adonovan@google.com>
Robert Griesemer [Thu, 4 Oct 2018 23:43:34 +0000 (16:43 -0700)]
go/types: remove work-around for issue #26124
This work-around is not needed anymore now that method
signatures are type-checked separately from their receiver
base types: no artificial cycles are introduced anymore
and so there is no need to artificially cut them.
Updates #26124.
Change-Id: I9d50171f12dd8977116a5d3f63ac39a06b1cd492
Reviewed-on: https://go-review.googlesource.com/c/139899 Reviewed-by: Alan Donovan <adonovan@google.com>
Robert Griesemer [Thu, 4 Oct 2018 00:50:02 +0000 (17:50 -0700)]
go/types: prepare for delayed type-checking of methods to when they are used
Remove assumption that methods associated to concrete (non-interface)
types have a fully set up signature. Such methods are found through
LookupFieldOrMethod or lookupMethod, or indexed method access from
a Named type. Make sure that the method's signature is type-checked
before use in those cases.
(MethodSets also hold methods but the type checker is not using
them but for internal verification. API clients will be using it
after all methods have been type-checked.)
Some functions such as MissingMethod may now have to type-check a
method and for that they need a *Checker. Add helper functions as
necessary to provide the additional (receiver) parameter but permit
it to be nil if the respective functions are invoked through the API
(at which point we know that all methods have a proper signature and
thus we don't need the delayed type-check).
Since all package-level objects eventually are type-checked through
the top-level loop in Checker.packageObjects we are guaranteed that
all methods will be type-checked as well.
Updates #23203.
Updates #26854.
Change-Id: I6e48f0016cefd498aa70b776e84a48215a9042c5
Reviewed-on: https://go-review.googlesource.com/c/139425 Reviewed-by: Alan Donovan <adonovan@google.com>
Robert Griesemer [Thu, 4 Oct 2018 19:59:09 +0000 (12:59 -0700)]
go/types: fix recvPtr helper (follow-up on https://golang.org/cl/139422)
The prior CL prepared go/types for the situation where methods might
not have a type-checked signature when being looked up. The respective
adjustments to recvPtr were not correct (but because so far method
signatures are type-checked in time, the bug didn't manifest itself).
Updates #23203.
Updates #26854.
Change-Id: I796691d11e6aac84396bdef802ad30715755fcc6
Reviewed-on: https://go-review.googlesource.com/c/139721 Reviewed-by: Alan Donovan <adonovan@google.com>
Austin Clements [Fri, 30 Jun 2017 16:10:01 +0000 (12:10 -0400)]
runtime: improve mheap.alloc doc and let compiler check system stack
The alloc_m documentation refers to concepts that don't exist (and
maybe never did?). alloc_m is also not the API entry point to span
allocation.
Hence, rewrite the documentation for alloc and alloc_m. While we're
here, document why alloc_m must run on the system stack and replace
alloc_m's hand-implemented system stack check with a go:systemstack
annotation.
Change-Id: I30e263d8e53c2774a6614e1b44df5464838cef09
Reviewed-on: https://go-review.googlesource.com/c/139459
Run-TryBot: Austin Clements <austin@google.com> Reviewed-by: Michael Knyszek <mknyszek@google.com> Reviewed-by: Keith Randall <khr@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Alex Brainman [Sat, 16 Jun 2018 06:23:52 +0000 (16:23 +1000)]
cmd/link: move .rel symbol from .rdata into .text
.rel symbol type is sym.SELFROSECT, and that makes .rel written
into .rdata section. But .rel stores code - jump table used for
external C functions. So we have to mark whole .rdata section
as executable (IMAGE_SCN_MEM_EXECUTE), because of .rel presence
in it.
Move .rel into .text section, and make .rdata section non executable.
I also had to move code that adjusted the size of .rel symbol
before calling textaddress, otherwise textaddress would not
calculate size of .text section correctly.
Fixes #25926
Change-Id: I4962f5de7b367410154c8709adfcd8472de9ac1a
Reviewed-on: https://go-review.googlesource.com/c/125455
Run-TryBot: Alex Brainman <alex.brainman@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org>
Keith Randall [Thu, 4 Oct 2018 23:32:21 +0000 (16:32 -0700)]
cmd/link: fix deferreturn location on wasm
On wasm, pcln tables are indexed by "resumption point ID" instead of
by pc offset. When finding a deferreturn call, we must find the
associated resumption point ID for the deferreturn call.
Update #27518
Fixes wasm bug introduced in CL 134637.
Jason Keene [Thu, 13 Sep 2018 02:35:58 +0000 (20:35 -0600)]
cmd/go: ensure git attributes are set
This change disables the export-subst and export-ignore attributes when
creating zip files for modules. This is done to prevent the ziphash for
a given repo/revision from differing based on variables such as git
version or size of repo. The full rational for this change is detailed
here:
Ben Shi [Tue, 11 Sep 2018 08:55:10 +0000 (08:55 +0000)]
cmd/internal/obj/arm64: simplify ADD and SUB
Currently "ADD $0x123456, Rs, Rd" will load pre-stored 0x123456
from the constant pool and use it for the addition. Total 12 bytes
are cost. And so does SUB.
This CL breaks it to "ADD 0x123000, Rs, Rd" + "ADD 0x000456, Rd, Rd".
Both "0x123000" and "0x000456" can be directly encoded into the
instruction binary code. So 4 bytes are saved.
1. The total size of pkg/android_arm64 decreases about 0.3KB.
2. The go1 benchmark show little regression (excluding noise).
Alberto Donizetti [Wed, 3 Oct 2018 19:11:31 +0000 (21:11 +0200)]
runtime: skip TestLldbPython when lldb is too old
The TestLldbPython test is known to fail with very old lldb releases
(3.8 and older). Skip the test when the lldb found on the system is
too old.
Fixes #22299
Change-Id: I8f78d6c0d995118f806dae87f3f04a9726473116
Reviewed-on: https://go-review.googlesource.com/c/139397 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Austin Clements [Thu, 4 Oct 2018 14:46:20 +0000 (10:46 -0400)]
runtime: clarify table of arena sizes
Currently the table of arena sizes mixes the number of entries in the
L1 with the size of the L2. While the size of the L2 is important,
this makes it hard to see what's actually going on because there's an
implicit factor of sys.PtrSize.
This changes the L2 column to say both the number of entries and the
size that results in. This should hopefully make the relations between
the columns of the table clearer, since they can now be plugged
directly into the given formula.
Robert Griesemer [Wed, 3 Oct 2018 21:40:09 +0000 (14:40 -0700)]
go/types: determine hasPtrRecv property from source rather than type
LookupFieldOrMethod needs to know if a method receiver is a pointer
type. Until now this was computed from the the method signature's
receiver, which required the method signature to be type-checked.
Furthermore, it required the receiver to be set before the method
signature was fully type-checked in some cases (see issue #6638).
This CL remembers this property during object resolution, when we
know it from the source.
With this CL, method signatures don't need to be type-checked before
they can be looked up; this is a first step towards separating
type checking of types and type-checking of associated methods.
Updates #23203.
Updates #26854.
Change-Id: Ie3eb7976e8fe8176ea1b284fa7471a4b7000f80b
Reviewed-on: https://go-review.googlesource.com/c/139422
Run-TryBot: Robert Griesemer <gri@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Alan Donovan <adonovan@google.com>
Matthew Dempsky [Wed, 3 Oct 2018 22:47:16 +0000 (15:47 -0700)]
cmd/compile: handle TPTR32 like TPTR64 in smallintconst
In preparation for followup CL merging TPTR32 and TPTR64, move TPTR32
from the small-types fast path to the generic 64-bit fallback code so
that it's in the same case clause as TPTR64.
This should be safe, but theoretically it could change semantics
because TPTR32 used to always be assumed to be "small", whereas now it
will only be considered small for values less than 1<<31.
This change is done in a separate CL so that it's more easily
identified by git bisection in case it does introduce regressions.
Change-Id: I6c7bb253d4e4d95c530a6e05a1147905674b55ca
Reviewed-on: https://go-review.googlesource.com/c/139517
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Robert Griesemer <gri@golang.org>
Kir Kolyshkin [Thu, 4 Oct 2018 01:41:18 +0000 (18:41 -0700)]
crypto/x509: fix getting user home dir on darwin
As pointed out in https://github.com/golang/go/issues/26463,
HOME (or equivalent) environment variable (rather than the
value obtained by parsing /etc/passwd or the like) should be
used to obtain user's home directory.
Since commit fa1a49aa556d8 there's a method to obtain
user's home directory -- use it here.
Michael Fraenkel [Thu, 4 Oct 2018 00:42:05 +0000 (20:42 -0400)]
expvar: add Map.Delete
Fixes #13491
Change-Id: Ic0525d8ee90f47d0d23c1485919aee13d2400494
Reviewed-on: https://go-review.googlesource.com/c/139537
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org>
Richard Musiol [Tue, 25 Sep 2018 11:45:08 +0000 (13:45 +0200)]
syscall: use asynchronous operations on js/wasm
This commit makes syscall on js/wasm use the asynchronous variants
of functions in Node.js' fs module. This enables concurrency
and allows the API of the fs module to be implemented with an
alternative backend that only supports asynchronous operations.
Brad Fitzpatrick [Wed, 3 Oct 2018 22:18:07 +0000 (22:18 +0000)]
crypto/rand: warn to stderr if blocked 60+ sec on first Reader.Read call
Fixes #22614
Change-Id: I220afbaaeab4dec6d59eeeef12107234a77f1587
Reviewed-on: https://go-review.googlesource.com/c/139419 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Brad Fitzpatrick [Wed, 3 Oct 2018 21:21:55 +0000 (21:21 +0000)]
os: add UserHomeDir
Fixes #26463
Change-Id: Iaef1c7456ffaeadeead6027a37d09c44a3d05bd5
Reviewed-on: https://go-review.googlesource.com/c/139418 Reviewed-by: Ian Lance Taylor <iant@golang.org>
Martin Möhrmann [Sun, 26 Aug 2018 12:22:39 +0000 (14:22 +0200)]
strings: correctly handle invalid utf8 sequences in Map
When an invalid UTF-8 byte sequence is decoded in a range loop over a string
a utf8.RuneError rune is returned. This is not distinguishable from decoding
the valid '\uFFFD' sequence representing utf8.RuneError from a string without
further checks within the range loop.
The previous Map code did not do any extra checks and would thereby not map
invalid UTF-8 byte sequences correctly when those were mapping to utf8.RuneError.
Fix this by adding the extra checks necessary to distinguish the decoding
of invalid utf8 byte sequences from decoding the sequence for utf8.RuneError
when the mapping of a rune is utf8.RuneError.
This fix does not result in a measureable performance regression:
name old time/op new time/op delta
ByteByteMap 1.05µs ± 3% 1.03µs ± 3% ~ (p=0.118 n=10+10)
Map/identity/ASCII 169ns ± 2% 170ns ± 1% ~ (p=0.501 n=9+10)
Map/identity/Greek 298ns ± 1% 303ns ± 4% ~ (p=0.338 n=10+10)
Map/change/ASCII 323ns ± 3% 325ns ± 4% ~ (p=0.679 n=8+10)
Map/change/Greek 628ns ± 5% 635ns ± 1% ~ (p=0.460 n=10+9)
MapNoChanges 120ns ± 4% 119ns ± 1% ~ (p=0.496 n=10+9)
This commit adds AIX operating system to internal/poll package for ppc64
architecture.
Updates: #25893
Change-Id: I9b1da9255012de58f16547c1b18f8840485da170
Reviewed-on: https://go-review.googlesource.com/c/138717
Run-TryBot: Ian Lance Taylor <iant@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org>
This commit adds AIX operating system to runtime package for ppc64
architecture.
Only new files and minor changes are in this commit. Others
modifications in files like asm_ppc64.s will be in separate commits.
Updates: #25893
Change-Id: I9c5e073f5f3debb43b004ad1167694a5afd31cfd
Reviewed-on: https://go-review.googlesource.com/c/138716
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org>
Austin Clements [Fri, 17 Aug 2018 19:42:19 +0000 (15:42 -0400)]
runtime: fix race between unminit and Windows profile loop
Currently, the Windows profile loop isn't robust against racing with
unminit. For example,
T1 is running profileloop1, T2 is another thread
T1: thread := atomic.Loaduintptr(&T2.thread)
T2: calls unminit, which does CloseHandle(T2.thread)
T1: attempts to suspends T2
In this case the SuspendThread will fail, but currently we ignore this
failure and forge ahead, which will cause further failures and
probably bad profile data.
Handle this race by defending against SuspendThread failing. If
SuspendThread succeeds, then we know the thread is no longer going
anywhere.
Austin Clements [Mon, 27 Aug 2018 01:30:19 +0000 (21:30 -0400)]
runtime: skip debug call injection tests under a debugger
The debug call injection tests will freeze when run under a debugger
because they depend on catching SIGTRAP, which is usually swallowed by
a debugger.
Keith Randall [Tue, 11 Sep 2018 22:14:28 +0000 (15:14 -0700)]
runtime: on a signal, set traceback address to a deferreturn call
When a function triggers a signal (like a segfault which translates to
a nil pointer exception) during execution, a sigpanic handler is just
below it on the stack. The function itself did not stop at a
safepoint, so we have to figure out what safepoint we should use to
scan its stack frame.
Previously we used the site of the most recent defer to get the live
variables at the signal site. That answer is not quite correct, as
explained in #27518. Instead, use the site of a deferreturn call.
It has all the right variables marked as live (no args, all the return
values, except those that escape to the heap, in which case the
corresponding PAUTOHEAP variables will be live instead).
This CL requires stack objects, so that all the local variables
and args referenced by the deferred closures keep the right variables alive.
Keith Randall [Fri, 7 Sep 2018 21:55:09 +0000 (14:55 -0700)]
cmd/compile,runtime: remove ambiguously live logic
The previous CL introduced stack objects. This CL removes the old
ambiguously live liveness analysis. After this CL we're relying
on stack objects exclusively.
Update a bunch of liveness tests to reflect the new world.
Keith Randall [Sun, 2 Sep 2018 03:16:39 +0000 (20:16 -0700)]
cmd/compile,runtime: implement stack objects
Rework how the compiler+runtime handles stack-allocated variables
whose address is taken.
Direct references to such variables work as before. References through
pointers, however, use a new mechanism. The new mechanism is more
precise than the old "ambiguously live" mechanism. It computes liveness
at runtime based on the actual references among objects on the stack.
Each function records all of its address-taken objects in a FUNCDATA.
These are called "stack objects". The runtime then uses that
information while scanning a stack to find all of the stack objects on
a stack. It then does a mark phase on the stack objects, using all the
pointers found on the stack (and ancillary structures, like defer
records) as the root set. Only stack objects which are found to be
live during this mark phase will be scanned and thus retain any heap
objects they point to.
A subsequent CL will remove all the "ambiguously live" logic from
the compiler, so that the stack object tracing will be required.
For this CL, the stack tracing is all redundant with the current
ambiguously live logic.
Matthew Dempsky [Wed, 3 Oct 2018 18:06:55 +0000 (11:06 -0700)]
cmd/compile/internal/gc: remove binary package export format
This CL removes all unused code from bimport.go and bexport.go.
In the interest of keeping this CL strictly delete-only and easier to
review, the task of consolidating the vestigial code elsewhere is left
to future CLs.
Change-Id: Ib757cc27e3fe814cbf534776d026e4d4cddfc6db
Reviewed-on: https://go-review.googlesource.com/c/139338
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Robert Griesemer <gri@golang.org>
Matthew Dempsky [Wed, 3 Oct 2018 17:56:23 +0000 (10:56 -0700)]
cmd/compile/internal/gc: disable binary package export format
The new indexed package export format appears stable, and no reports
of needing to revert back to binary package export.
This CL disables the binary package export format by mechanically
replacing 'flagiexport' with 'true', and then superficial code
cleanups to keep the resulting code idiomatic. The resulting dead code
is removed in a followup CL.
Change-Id: Ic30d85f78778a31d279a56b9ab14e80836d50135
Reviewed-on: https://go-review.googlesource.com/c/139337
Run-TryBot: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Robert Griesemer <gri@golang.org>
Florian Uekermann [Fri, 1 Dec 2017 19:01:55 +0000 (20:01 +0100)]
time: return first error in unsuccessful calls to LoadLocation
Unsuccessful calls to LoadLocation previously returned the first
error encountered while traversing the default list of sources, but
ignored errors from sources specified by ZONEINFO. Whether errors
indicating missing zones or sources were ignored in this process
differed between kinds of sources.
With this change, unsuccessful calls to LoadLocation always return
the first error, not counting errors indicating missing zones or
sources.
David Url [Thu, 23 Aug 2018 15:28:59 +0000 (17:28 +0200)]
net/http: log call site which causes multiple header writes
If an illegal header write is detected, find the first caller outside of
net/http using runtime.CallersFrames and include the call site in the log
message.
When deadline has already passed,
current context is canceled before return cancel function.
So is unnecessary to call cancel with remove from parent again
in return cancel function.
Cherry Zhang [Wed, 3 Oct 2018 02:04:45 +0000 (22:04 -0400)]
cmd/compile: fix type of OffPtr in some optimization rules
In some optimization rules the type of generated OffPtr was
incorrectly set to the type of the pointee, instead of the
pointer. When the OffPtr value is spilled, this may generate
a spill of the wrong type, e.g. a floating point spill of an
integer (pointer) value. On Wasm, this leads to invalid
bytecode.
Fixes #27961.
Change-Id: I5d464847eb900ed90794105c0013a1a7330756cc
Reviewed-on: https://go-review.googlesource.com/c/139257
Run-TryBot: Cherry Zhang <cherryyz@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Keith Randall <khr@golang.org> Reviewed-by: Richard Musiol <neelance@gmail.com>
Anton Gyllenberg [Tue, 2 Oct 2018 22:47:28 +0000 (22:47 +0000)]
cmd/go: prevent infinite loop in QueryPackage()
p = path.Dir(p) converges to either "." or "/". The current
implementation of modload.QueryPackage() has a loop that
terminates only on ".", not "/". This leads to the go command
hanging in an infinite loop if the user manages to supply
a file path starting with "/" as package path.
An example of the issue is if the user (incorrectly) attempts
to use an absolute directory path in an import statement within
a module (import "/home/bob/myproj") and then runs go list.
This commit adds AIX operating system to cmd/dist package for ppc64
architecture.
The stack guard is increased because of syscalls made inside the runtime
which need a larger stack.
Disable cmd/vet/all tests until aix/ppc64 is fully available.
Change-Id: I7e3caf86724249ae564a152d90c1cbd4de288814
Reviewed-on: https://go-review.googlesource.com/c/138715
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Ian Lance Taylor <iant@golang.org>
Eduard Urbach [Tue, 24 Apr 2018 14:28:19 +0000 (14:28 +0000)]
net/http: explain more how Pusher.Push works
This will clarify that the resources are not completely pushed yet when `Push` returns and that it starts a separate goroutine. This might be implementation dependant but as I believe there is currently only one implementation it should be added to the documentation of the interface which most people will look up first.
net/http: make Transport send WebSocket upgrade requests over HTTP/1
WebSockets requires HTTP/1 in practice (no spec or implementations
work over HTTP/2), so if we get an HTTP request that looks like it's
trying to initiate WebSockets, use HTTP/1, like browsers do.
This is part of a series of commits to make WebSockets work over
httputil.ReverseProxy. See #26937.
Austin Clements [Thu, 16 Aug 2018 16:32:46 +0000 (12:32 -0400)]
runtime: eliminate work.markrootdone and second root marking pass
Before STW and concurrent GC were unified, there could be either one
or two root marking passes per GC cycle. There were several tasks we
had to make sure happened once and only once (whether that was at the
beginning of concurrent mark for concurrent GC or during mark
termination for STW GC). We kept track of this in work.markrootdone.
Now that STW and concurrent GC both use the concurrent marking code
and we've eliminated all work done by the second root marking pass, we
only ever need a single root marking pass. Hence, we can eliminate
work.markrootdone and all of the code that's conditional on it.
Austin Clements [Thu, 23 Aug 2018 17:14:19 +0000 (13:14 -0400)]
runtime: flush mcaches lazily
Currently, all mcaches are flushed during STW mark termination as a
root marking job. This is currently necessary because all spans must
be out of these caches before sweeping begins to avoid races with
allocation and to ensure the spans are in the state expected by
sweeping. We do it as a root marking job because mcache flushing is
somewhat expensive and O(GOMAXPROCS) and this parallelizes the work
across the Ps. However, it's also the last remaining root marking job
performed during mark termination.
This CL moves mcache flushing out of mark termination and performs it
lazily. We keep track of the last sweepgen at which each mcache was
flushed and as each P is woken from STW, it observes that its mcache
is out-of-date and flushes it.
The introduces a complication for spans cached in stale mcaches. These
may now be observed by background or proportional sweeping or when
attempting to add a finalizer, but aren't in a stable state. For
example, they are likely to be on the wrong mcentral list. To fix
this, this CL extends the sweepgen protocol to also capture whether a
span is cached and, if so, whether or not its cache is stale. This
protocol blocks asynchronous sweeping from touching cached spans and
makes it the responsibility of mcache flushing to sweep the flushed
spans.
This eliminates the last mark termination root marking job, which
means we can now eliminate that entire infrastructure.
Updates #26903. This implements lazy mcache flushing.
Austin Clements [Thu, 16 Aug 2018 16:25:38 +0000 (12:25 -0400)]
runtime: eliminate blocking GC work drains
Now work.helperDrainBlock is always false, so we can remove it and
code paths that only ran when it was true. That means we no longer use
the gcDrainBlock mode of gcDrain, so we can eliminate that. That means
we no longer use gcWork.get, so we can eliminate that. That means we
no longer use getfull, so we can eliminate that.
Updates #26903. This is a follow-up to unifying STW GC and concurrent GC.
Austin Clements [Thu, 16 Aug 2018 16:17:32 +0000 (12:17 -0400)]
runtime: clean up remaining mark work check
Now that STW GC marking is unified with concurrent marking, there
should never be mark work remaining in mark termination. Hence, we can
make that check unconditional.
Updates #26903. This is a follow-up to unifying STW GC and concurrent GC.
Austin Clements [Mon, 13 Aug 2018 20:30:54 +0000 (16:30 -0400)]
runtime: implement STW GC in terms of concurrent GC
Currently, STW GC works very differently from concurrent GC. The
largest differences in that in concurrent GC, all marking work is done
by background mark workers during the mark phase, while in STW GC, all
marking work is done by gchelper during the mark termination phase.
This is a consequence of the evolution of Go's GC from a STW GC by
incrementally moving work from STW mark termination into concurrent
mark. However, at this point, the STW code paths exist only as a
debugging mode. Having separate code paths for this increases the
maintenance burden and complexity of the garbage collector. At the
same time, these code paths aren't tested nearly as well, making it
far more likely that they will bit-rot.
This CL reverses the relationship between STW GC, by re-implementing
STW GC in terms of concurrent GC.
This builds on the new scheduled support for disabling user goroutine
scheduling. During sweep termination, it disables user scheduling, so
when the GC starts the world again for concurrent mark, it's really
only "concurrent" with itself.
There are several code paths that were specific to STW GC that are now
vestigial. We'll remove these in the follow-up CLs.