| Age | Commit message (Collapse) | Author |
|---|
|
Zeroing the upper part of X15 may make the CPU think it is
"dirty" and slow down SSE operations. For now, just not zeroing
the upper part, and construct a zero value on the fly if we need
a 256- or 512-bit zero value. Maybe VZEROUPPER works better than
explicitly zeroing X15, but we need to evaluate.
Long term, we probably want to move more things from SSE to AVX.
This essentially undoes CL 698237 and CL 698238, except keeping
using X15 for 128-bit zeroing for SIMD.
Change-Id: I1564e6332c4c57f9721397c92c7c734c5497534c
Reviewed-on: https://go-review.googlesource.com/c/go/+/728240
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: David Chase <drchase@google.com>
|
|
If simd experiment is not enabled, the compiler doesn't use the
AVX part of the register. So only zero it with the SSE instruction.
Change-Id: Ia3bdf34a9ed273128db2ee0f4f5db6f7cc76a975
Reviewed-on: https://go-review.googlesource.com/c/go/+/724720
Reviewed-by: Junyang Shao <shaojunyang@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: David Chase <drchase@google.com>
|
|
On AMD64, we reserve the X15 register as the zero register.
Currently we use an SSE instruction to zero it, and we only use
it in SSE contexts. When the machine supports AVX, the high bits
of the register is not necessarily zeroed.
Now that the compiler generates AVX code for SIMD, it would be
great to have a zero register in the AVX context. This CL zeroes
the whole X15 register if AVX is supported.
Change-Id: I4dc803362f2e007b1614b90de435fbb7814cebc7
Reviewed-on: https://go-review.googlesource.com/c/go/+/698237
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Junyang Shao <shaojunyang@google.com>
Reviewed-by: David Chase <drchase@google.com>
|
|
Factor out the code related to doing calls using the Windows stdcall
calling convention into a separate package. This will allow us to
reuse it in other low-level packages that can't depend on syscall.
Updates #51087.
Cq-Include-Trybots: luci.golang.try:gotip-windows-arm64,gotip-windows-amd64-longtest,gotip-solaris-amd64
Change-Id: I68640b07091183b50da6bef17406c10a397896e9
Reviewed-on: https://go-review.googlesource.com/c/go/+/689156
Reviewed-by: Michael Pratt <mpratt@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
|
|
The syscall_SyscallX functions currently discard the nargs parameter
when calling syscall_SyscallN. This precludes some optimizations
down the line. For example, on amd64, a syscall that takes 0 arguments
don't need to set any of the params passing registers (CX, DX, R8, and
R9).
This CL updates all syscall_SyscallX functions so they call
syscall_SyscallN with an argument slice of the right length.
While here, remove the hack in syscall_SyscallN to support less than 4
arguments, and update instead asmstdcall on amd64 to properly handle
this case.
Change-Id: I0328e14f34c2b000fde06cc6a579b09e8c32f2b9
Reviewed-on: https://go-review.googlesource.com/c/go/+/563315
TryBot-Result: Gopher Robot <gobot@golang.org>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Run-TryBot: Quim Muntal <quimmuntal@gmail.com>
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Keith Randall <khr@google.com>
Reviewed-by: Than McIntosh <thanm@google.com>
|
|
The Windows unhandled exception mechanism fails to call the callback
set in SetUnhandledExceptionFilter if the stack can't be correctly
unwound.
Some cgo glue code was not properly chaining the frame pointer, making
the stack unwind to fail in case of an exception inside a cgo call.
This CL fix that and adds a test case to avoid regressions.
Fixes #50951
Change-Id: Ic782b5257fe90b05e3def8dbf0bb8d4ed37a190b
Reviewed-on: https://go-review.googlesource.com/c/go/+/525475
Reviewed-by: Bryan Mills <bcmills@google.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
Run-TryBot: Quim Muntal <quimmuntal@gmail.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
|
|
Windows APIs are normally not arch-specific, so it's better to
implement them in Go instead of assembly.
It was previously implemented in assembly because it was the only way
to support calls without a valid g. This CL defines a new function,
stdcall_no_g, that can be used in such cases.
While here, I've also replaced the use of the deprecated syscall
NtWaitForSingleObject with WaitForSingleObject. The former may
give the illusion of being more accurate, as it takes a higher
resolution timeout, but it's not. Windows time resolution is 15.6ms,
and can be as high as 1ms when using a high resolution timer, which
WaitForSingleObject supports.
Change-Id: I903400220ade4d4ccc15685c8da47182430f8686
Reviewed-on: https://go-review.googlesource.com/c/go/+/526477
Reviewed-by: Bryan Mills <bcmills@google.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Run-TryBot: Quim Muntal <quimmuntal@gmail.com>
|
|
Windows APIs are normally not arch-specific, so it's better to
implement them in Go instead of assembly.
It was previously implemented in assembly because it was the only way
to support calls without a valid g. This CL defines a new function,
stdcall_no_g, that can be used in such cases.
Change-Id: I26a223b918c6c462b06ac256bdacf9ddb78752bc
Reviewed-on: https://go-review.googlesource.com/c/go/+/526476
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Cherry Mui <cherryyz@google.com>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Run-TryBot: Quim Muntal <quimmuntal@gmail.com>
Reviewed-by: Bryan Mills <bcmills@google.com>
|
|
This CL fixes the links to Microsoft documentation in the Go source
code. Some links were broken and some others were outdated.
Change-Id: I4c3bcd3aa3c07a31be1b7f94c25339dcc2e771e8
Reviewed-on: https://go-review.googlesource.com/c/go/+/527556
Reviewed-by: Heschi Kreinick <heschi@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Run-TryBot: Quim Muntal <quimmuntal@gmail.com>
Auto-Submit: Quim Muntal <quimmuntal@gmail.com>
|
|
Having to pass a dummy pointer to the libcall.args field is a bit
annoying. This change allows nil to be passed instead.
windows/arm and windows/arm64 already support nil libcall.args.
Change-Id: I07a2bdb7d1f76b13d125397ff5177337c43536a3
Reviewed-on: https://go-review.googlesource.com/c/go/+/526016
Reviewed-by: Ian Lance Taylor <iant@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Bryan Mills <bcmills@google.com>
|
|
This reapplies CL 191759, which was reverted in CL 192622.
Wine fixed the compatibility issue more than 3 years
ago, in version 5.10 (see [1]). We no longer have to keep the compatibility hack on our side.
Updates #34021
[1]: https://github.com/wine-mirror/wine/commit/1ae10889647c1c84c36660749508a42e99e64a5e
Change-Id: I3b77701d01fdf58fbf350321fc0a957c0f247d32
Reviewed-on: https://go-review.googlesource.com/c/go/+/526358
Reviewed-by: Daniel Martí <mvdan@mvdan.cc>
Reviewed-by: Cherry Mui <cherryyz@google.com>
Reviewed-by: Bryan Mills <bcmills@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Quim Muntal <quimmuntal@gmail.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
|
|
The Go runtime allocates the TLS slot in the TEB TLS slots instead of
using the TEB arbitrary pointer. See CL 431775 for more context.
The problem is that the TEB TLS slots array only has capacity for 64
indices, allocating more requires some complex logic that we don't
support yet.
Although the Go runtime only allocates one index, a Go DLL can be
loaded in a process with more than 64 TLS slots allocated,
in which case it abort.
This CL avoids aborting by falling back to the older behavior, that
is to use the TEB arbitrary pointer.
Fixes #59213
Change-Id: I39c73286fe2da95aa9c5ec5657ee0979ecbec533
Reviewed-on: https://go-review.googlesource.com/c/go/+/486816
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
Run-TryBot: Quim Muntal <quimmuntal@gmail.com>
Reviewed-by: Bryan Mills <bcmills@google.com>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
|
|
For #59670.
Change-Id: I0efa743edc08e48dc8d906803ba45e9f641369db
Reviewed-on: https://go-review.googlesource.com/c/go/+/486977
Reviewed-by: Cherry Mui <cherryyz@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Auto-Submit: Austin Clements <austin@google.com>
Run-TryBot: Austin Clements <austin@google.com>
|
|
This reverts commit CL 486381.
Submitted out of order and breaks bootstrap.
Change-Id: Ia472111cb966e884a48f8ee3893b3bf4b4f4f875
Reviewed-on: https://go-review.googlesource.com/c/go/+/486915
Reviewed-by: David Chase <drchase@google.com>
TryBot-Bypass: Austin Clements <austin@google.com>
|
|
For #59670.
Change-Id: I4476d6f92663e8a825d063d6e6a7fc9a2ac99d4d
Reviewed-on: https://go-review.googlesource.com/c/go/+/486381
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Cherry Mui <cherryyz@google.com>
|
|
Most of these are one-off mistakes. Only one file was all spaces.
Change-Id: I277c3ce4a4811aa4248c90676f66bc775ae8d062
Reviewed-on: https://go-review.googlesource.com/c/go/+/478976
Run-TryBot: Michael Pratt <mpratt@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Cherry Mui <cherryyz@google.com>
|
|
This CL moves the usleep2HighRes from assembly to good old Go.
This is safe because since CL 288793 usleep is always called with
a g, else one wold have to call usleep_no_g. This condition was
not enforced when high resolution timers were first implemented
on Windows (CL 248699), so the implementation was done in assembly.
Other than removing a bunch of obscure assembly code, this CL makes
high resolution timers work on windows arm/arm64 by free, as the
system calls are the same in all windows platforms.
Change-Id: I41ecf78026fd7e11e85258a411ae074a77e8c7fc
Reviewed-on: https://go-review.googlesource.com/c/go/+/471142
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
Reviewed-by: Bryan Mills <bcmills@google.com>
Auto-Submit: Michael Pratt <mpratt@google.com>
Run-TryBot: Quim Muntal <quimmuntal@gmail.com>
|
|
This CL removes some NOFRAME flags on Windows assembly files
for several reasons:
- windows/386 does not use a frame pointer
- Leaf frameless functions already skip the frame pointer
- Some non-leaf functions do not contain enough dragons to justify
not using the frame pointer
Updates #58378
Change-Id: I31e71bf7f769e1957a4adba91778da5af66ce1e4
Reviewed-on: https://go-review.googlesource.com/c/go/+/466835
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Keith Randall <khr@google.com>
Run-TryBot: Quim Muntal <quimmuntal@gmail.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
|
|
runtime·cbctxts has been unused since CL 258938, but it was left over.
Change-Id: I374ad26e668a36994e41f5d17593b33090bdc644
Reviewed-on: https://go-review.googlesource.com/c/go/+/463119
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Bryan Mills <bcmills@google.com>
Run-TryBot: Quim Muntal <quimmuntal@gmail.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
|
|
This CL marks non-leaf nosplit assembly functions as NOFRAME to avoid
relying on the implicit amd64 NOFRAME heuristic, where NOSPLIT functions
without stack were also marked as NOFRAME.
Updates #57302
Updates #40044
Change-Id: Ia4d26f8420dcf2b54528969ffbf40a73f1315d61
Reviewed-on: https://go-review.googlesource.com/c/go/+/459395
Reviewed-by: Cherry Mui <cherryyz@google.com>
Run-TryBot: Quim Muntal <quimmuntal@gmail.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
|
|
This CL removes badsignal2 function, as it is unused on Windows.
badsignal2 was originally intended to abort the process when
an exception was raised on a non-Go thread, following the same approach
as Linux and others.
Since it was added, back on https://golang.org/cl/5797068, it has caused
several issues on Windows, see #8224 and #50877. That's because we can't
know wether the signal is bad or not, as our trap might not be at the
end of the exception handler chain.
To fix those issues, https://golang.org/cl/104200046 and CL 442896
stopped calling badsignal2, and CL 458135 removed one last incorrect
call on amd64 and 386.
Change-Id: I5bd31ee2672118ae0f1a2c8b46a1bb0f4893a011
Reviewed-on: https://go-review.googlesource.com/c/go/+/463116
Reviewed-by: Bryan Mills <bcmills@google.com>
Run-TryBot: Quim Muntal <quimmuntal@gmail.com>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
|
|
This CL factors out part of the Windows sigtramp implementation, which
was duplicated in all four architectures. The new common code is
implemented in Go rather than in assembly, which will make Windows
error handling easier to reason and maintain.
While here, implement the control flow guard workaround on
windows/386, which almost comes for free.
Change-Id: I0bf38c28c54793225126e161bd95527a62de05e0
Reviewed-on: https://go-review.googlesource.com/c/go/+/458135
Run-TryBot: Quim Muntal <quimmuntal@gmail.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Bryan Mills <bcmills@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
|
|
This CL redesign how we get the TLS pointer on windows/amd64.
We were previously reading it from the [TEB] arbitrary data slot,
located at 0x28(GS), which can only hold 1 TLS pointer.
With this CL, we will read the TLS pointer from the TEB TLS slot array,
located at 0x1480(GS). The TLS slot array can hold multiple
TLS pointers, up to 64, so multiple Go runtimes running on the
same thread can coexists with different TLS.
Each new TLS slot has to be allocated via [TlsAlloc],
which returns the slot index. This index can then be used to get the
slot offset from GS with the following formula: 0x1480 + index*8
The slot index is fixed per Go runtime, so we can store it
in runtime.tls_g and use it latter on to read/update the TLS pointer.
Loading the TLS pointer requires the following asm instructions:
MOVQ runtime.tls_g, AX
MOVQ AX(GS), AX
Notice that this approach is also implemented on windows/arm64.
[TEB]: https://en.wikipedia.org/wiki/Win32_Thread_Information_Block
[TlsAlloc]: https://learn.microsoft.com/en-us/windows/win32/api/processthreadsapi/nf-processthreadsapi-tlsalloc
Updates #22192
Change-Id: Idea7119fd76a3cd083979a4d57ed64b552fa101b
Reviewed-on: https://go-review.googlesource.com/c/go/+/431775
Reviewed-by: Cherry Mui <cherryyz@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Quim Muntal <quimmuntal@gmail.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
|
|
The stack pointer must lie within system stack limits
when Control Flow Guard (CFG) is enabled on Windows.
This CL updates runtime.sigtramp to honor this restriction by
porting some code from the windows/arm64 version, which
already supports CFG.
Fixes #53560
Change-Id: I7f88f9ae788b2bac38aac898b2567f1bea62f8f3
Reviewed-on: https://go-review.googlesource.com/c/go/+/437559
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
Run-TryBot: Michael Pratt <mpratt@google.com>
|
|
On 64-bit, this is more efficient, and on ARM64, this prevents the time
from moving backwards due to the weaker memory model. On ARM32 due to
the weaker memory model, we issue a memory barrier.
Updates #48072.
Change-Id: If4695716c3039d8af14e14808af217f5c99fc93a
Reviewed-on: https://go-review.googlesource.com/c/go/+/361057
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Austin Clements <austin@google.com>
|
|
This CL adds a new syscall.SyscallN API.
The proposal discussion also suggests the API should not only for
Windows but other platforms. However, the existing API set already
contain differences between platforms, hence the CL only implements
the Windows platform.
Moreover, although the API offers variadic parameters, the permitted
parameters remains up to a limit, which is selected as 42, and arguably
large enough.
Fixes #46552
Change-Id: I66b49988a304d9fc178c7cd5de46d0b75e167a4f
Reviewed-on: https://go-review.googlesource.com/c/go/+/336550
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Matthew Dempsky <mdempsky@google.com>
Trust: Matthew Dempsky <mdempsky@google.com>
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
|
|
assembly functions
There are a few assembly functions in the runtime that are marked
as ABIInternal, solely because funcPC can get the right address.
The functions themselves do not actually follow ABIInternal (or
irrelevant). Now we have internal/abi.FuncPCABI0, use that, and
un-mark the functions.
Also un-mark assembly functions that are only called in assembly.
For them, it only matters if the caller and callee are consistent.
Change-Id: I240e126ac13cb362f61ff8482057ee9f53c24097
Reviewed-on: https://go-review.googlesource.com/c/go/+/321950
Trust: Cherry Mui <cherryyz@google.com>
Run-TryBot: Cherry Mui <cherryyz@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
|
|
Currently if a signal lands on a non-Go thread that's handled by the Go
handler, Go will emit a message. However, unlike everywhere else in the
runtime, Go will not abort the process after, and the signal handler
will try to continue executing.
This leads to cascading failures and possibly even memory corruption.
For #45638.
Change-Id: I546f4e82f339d555bed295528d819ac883b92bc6
Reviewed-on: https://go-review.googlesource.com/c/go/+/316809
Trust: Michael Knyszek <mknyszek@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
|
|
This is a step toward separating whether time.now is implemented in
assembly from whether we are using faketime.
Change-Id: I8bf059b44a103b034835e3d3b799319cc05e9552
Reviewed-on: https://go-review.googlesource.com/c/go/+/314273
Trust: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
|
|
Change-Id: I1a583d3da9cca4ac51f3fec9b508b7638b452d60
Reviewed-on: https://go-review.googlesource.com/c/go/+/314270
Trust: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
|
|
Change-Id: I2770195cb53220948081a6265f891ef064e4f763
Reviewed-on: https://go-review.googlesource.com/c/go/+/313629
Trust: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
|
|
Currently we call runtimeQPC as ABIInternal because it shaves off 24
bytes by not having an extra wrapper, and at the time we were exceeding
the nosplit stack limit in some cases.
However, this code was written before we had the regabiargs GOEXPERIMENT
flag, and wasn't properly flagged. Naturally, with regabiargs enabled,
it leads to garbage being returned, because it needs to store
runtimeQPC's result to the stack.
We didn't notice this because today runtimeQPC is only used in Wine, not
on any native Windows platform.
Back when I wrote this code, it appeared to be necessary on even native
Windows, but it turns out that's not true anymore. Turn it back into a
native call through a wrapper.
For #40724.
Change-Id: Ia2e5901965ef46c5f299daccef49952026854fe6
Reviewed-on: https://go-review.googlesource.com/c/go/+/312429
Trust: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
|
|
There are several assembly functions that transition from the Windows
ABI to the Go ABI. These all need to save all registers that are
callee-save in the Windows ABI and caller-save in the Go ABI and
prepare the register state for Go. However, they all do this slightly
differently and most of them don't save the necessary XMM registers
for this transition (which could corrupt them in the C caller).
Furthermore, now that we have a carefully specified Go ABI, it's clear
that none of these actually get all of the details 100% right.
So, unify this code into two macros in a shared header in
runtime/cgo/abi_amd64.h that handle all necessary registers and setup
and use these macros everywhere on Windows that handles transitions
from C to Go.
Change-Id: I62f41345a507aad1ca383814ac8b7e2a9ffb821e
Reviewed-on: https://go-review.googlesource.com/c/go/+/309769
Trust: Austin Clements <austin@google.com>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
|
|
than 16 args
Fixes #45524
Change-Id: Id867f45ea98689b73d5b1b141c19317bc7608b05
GitHub-Last-Rev: e9b09fb557dda291fb6cf27c185063c26832a15b
GitHub-Pull-Request: golang/go#45531
Reviewed-on: https://go-review.googlesource.com/c/go/+/309390
Reviewed-by: El Mostafa Idrassi <el.mostafa.idrassi@gmail.com>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Trust: Alex Brainman <alex.brainman@gmail.com>
Trust: Alberto Donizetti <alb.donizetti@gmail.com>
Run-TryBot: Alex Brainman <alex.brainman@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
|
|
This function is no longer used.
Eliminating this actually fixes several problems:
- It made assumptions about what registers memclrNoHeapPointers would
preserve. Besides being an abstraction violation and lurking
maintenance issue, this actively became a problem for regabi because
the call to memclrNoHeapPointers now happens through an ABI wrapper,
which is generated by the compiler and hence we can't easily control
what registers it clobbers.
- The amd64 implementation (at least), does not interact with the host
ABI correctly. Notably, it doesn't save many of the registers that
are callee-save in the host ABI but caller-save in the Go ABI.
- It interacts strangely with the NOSPLIT checker because it allocates
an entire M and G on its stack. It worked around this on arm64, and
happened to do things the NOSPLIT checker couldn't track on 386 and
amd64, and happened to be *4 bytes* below the limit on arm (so any
addition to the m or g structs would cause a NOSPLIT failure). See
CL 309031 for a more complete explanation.
Fixes #45530.
Updates #40724.
Change-Id: Ic70d4d7e1c17f1d796575b3377b8529449e93576
Reviewed-on: https://go-review.googlesource.com/c/go/+/309634
Trust: Austin Clements <austin@google.com>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
|
|
This replaces the externalthreadhandler-based implementation of
profileloop with one that uses newm to start a new thread. This is a
step toward eliminating externalthreadhandler.
For #45530.
Change-Id: Id8e5540423fe2d2004024b649afec6998f77b092
Reviewed-on: https://go-review.googlesource.com/c/go/+/309633
Trust: Austin Clements <austin@google.com>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
|
|
This replaces the externalthreadhandler-based implementation of
ctrlhandler with one based on compileCallback. This is a step toward
eliminating externalthreadhandler.
For #45530.
Change-Id: I2de2f2f37777af292db67ccf8057b7566aab81f8
Reviewed-on: https://go-review.googlesource.com/c/go/+/309632
Trust: Austin Clements <austin@google.com>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Michael Pratt <mpratt@google.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
|
|
Fixes #45498
Change-Id: I89365f3517bc84376f0f580c64a57f38aaba0cbb
Reviewed-on: https://go-review.googlesource.com/c/go/+/308997
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Trust: Alex Brainman <alex.brainman@gmail.com>
|
|
The covers three kinds of uses:
1. Calls of closures from assembly. These are always ABIInternal calls
without wrappers. I went through every indirect call in the runtime
and I think mcall is the only case of assembly calling a Go closure in
a way that's affected by ABIInternal. systemstack also calls a
closure, but it takes no arguments.
2. Calls of Go functions that expect raw ABIInternal pointers. I also
only found one of these: callbackasm1 -> cgocallback on Windows. These
are trickier to find, though.
3. Finally, I found one case on NetBSD where new OS threads were
directly calling the Go runtime entry-point from assembly via a PC,
rather than going through a wrapper. This meant new threads may not
have special registers set up. In this case, a change on all other
OSes had already forced new thread entry to go through an ABI wrapper,
so I just caught NetBSD up with that change.
With this change, I'm able to run a "hello world" with
GOEXPERIMENT=regabi,regabiargs.
For #40724.
Change-Id: I2a6d0e530c4fd4edf13484d923891c6160d683aa
Reviewed-on: https://go-review.googlesource.com/c/go/+/305669
Trust: Austin Clements <austin@google.com>
Run-TryBot: Austin Clements <austin@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
|
|
On Windows, when calling into needm in cgocallback on a new thread that
is unknown to the Go runtime, we currently call through an ABI wrapper.
The ABI wrapper tries to restore the G register from TLS.
On other platforms, TLS is set up just enough that the wrapper will
simply load a nil g from TLS, but on Windows TLS isn't set up at all, so
there's nowhere for the wrapper to load from.
So, bypass the wrapper in the call to needm. needm takes no arguments
and returns no results so there are no special ABI considerations,
except that we must clear X15 which is used as a zero register in Go
code (a function normally performed by the ABI wrapper). needm is also
otherwise already special and carefully crafted to avoid doing anything
that would require a valid G or M, at least until it is able to create
one.
While we're here, this change simplifies setg so that it doesn't set up
TLS on Windows and instead provides an OS-specific osSetupTLS to do
that.
The result of this is that setg(nil) no longer clears the TLS space
pointer on Windows. There's exactly one place this is used (dropm) where
it doesn't matter anymore, and an empty TLS means that setg's wrapper
will crash on the return path. Another result is that the G slot in the
TLS will be properly cleared, however, which isn't true today.
For #40724.
Change-Id: I65c3d924a3b16abe667b06fd91d467d6d5da31d7
Reviewed-on: https://go-review.googlesource.com/c/go/+/303070
Trust: Michael Knyszek <mknyszek@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
|
|
In the runtime there are Windows-specific assembly routines that are
address-taken via funcPC and are not intended to be called through a
wrapper. Mark them as ABIInternal so that we don't grab the wrapper,
because that will break in all sorts of contexts.
For #40724.
For #44065.
Change-Id: I12a728786786f423e5b229f8622e4a80ec27a31c
Reviewed-on: https://go-review.googlesource.com/c/go/+/302109
Trust: Michael Knyszek <mknyszek@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
|
|
This changes makes it so that nanotimeQPC calls nanotime1 without an ABI
wrapper by specifying the ABIInternal version directly. The reason why
this is necessary is because ABI wrappers typically require additional
stack space, and nanotimeQPC is used deep within nosplit contexts,
and with the ABI wrappers now enabled, this exhausts the stack guard
space held for nosplit functions. Rather than increase the stack guard,
we choose to do this.
For #40724.
Change-Id: Ia9173ca903335a9d6f380f57f4a45e49b58da6bb
Reviewed-on: https://go-review.googlesource.com/c/go/+/303069
Trust: Michael Knyszek <mknyszek@google.com>
Run-TryBot: Michael Knyszek <mknyszek@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
|
|
No change to actual runtime, but helps reduce the laundry list
of functions.
mcall, morestack, and asmcgocall are not actually top-of-frame,
so those need more attention in follow-up CLs.
mstart moved to assembly so that it can be marked TOPFRAME.
Since TOPFRAME also tells DWARF consumers not to unwind
this way, this change should also improve debuggers a
marginal amount.
This CL is part of a stack adding windows/arm64
support (#36439), intended to land in the Go 1.17 cycle.
This CL is, however, not windows/arm64-specific.
It is cleanup meant to make the port (and future ports) easier.
Change-Id: If1e0d46ca973de5e46b62948d076f675f285b5d9
Reviewed-on: https://go-review.googlesource.com/c/go/+/288802
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
|
|
Both asmcgocall and systemstack need to save the calling Go code's
context for use by traceback, but they do it differently.
Systemstack's appraoch is better, because it doesn't require a
special case in traceback.
So make them both use that.
While we are here, the fake mstart caller in systemstack is
no longer needed and can be removed.
(traceback knows to stop in systemstack because of the writes to SP.)
Also remove the fake mstarts in sys_windows_*.s.
And while we are there, fix the control flow guard code in sys_windows_arm.s.
The current code is using pointers to a stack frame that technically is gone
once we hit the RET instruction. Clearly it's working OK, but better not to depend
on data below SP being preserved, even for just a few instructions.
Store the value we need in other registers instead.
(This code is only used for pushing a sigpanic call, which does not
actually return to the site of the fault and therefore doesn't need to
preserve any of the registers.)
This CL is part of a stack adding windows/arm64
support (#36439), intended to land in the Go 1.17 cycle.
This CL is, however, not windows/arm64-specific.
It is cleanup meant to make the port (and future ports) easier.
Change-Id: Id1e3ef5e54f7ad786e4b87043f2626eba7c3bbd9
Reviewed-on: https://go-review.googlesource.com/c/go/+/288799
Trust: Russ Cox <rsc@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
|
|
This is dead code and need not be ported to each architecture.
This CL is part of a stack adding windows/arm64
support (#36439), intended to land in the Go 1.17 cycle.
This CL is, however, not windows/arm64-specific.
It is cleanup meant to make the port (and future ports) easier.
Change-Id: I2d0072b377f73e49d7158ea304670c26f5486c59
Reviewed-on: https://go-review.googlesource.com/c/go/+/288794
Trust: Russ Cox <rsc@golang.org>
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jason A. Donenfeld <Jason@zx2c4.com>
|
|
During a cgocallback, the runtime calls needm to get an m.
The calls made during needm cannot themselves assume that
there is an m or a g (which is attached to the m).
In the old days of making direct system calls, the only thing
you had to do for such functions was mark them //go:nosplit,
to avoid the use of g in the stack split prologue.
But now, on operating systems that make system calls through
shared libraries and use code that saves state in the g or m
before doing so, it's not safe to assume g exists. In fact, it is
not even safe to call getg(), because it might fault deferencing
the TLS storage to find the g pointer (that storage may not be
initialized yet, at least on Windows, and perhaps on other systems
in the future).
The specific routines that are problematic are usleep and osyield,
which are called during lock contention in lockextra, called
from needm.
All this is rather subtle and hidden, so in addition to fixing the
problem on Windows, this CL makes the fact of not running on
a g much clearer by introducing variants usleep_no_g and
osyield_no_g whose names should make clear that there is no g.
And then we can remove the various sketchy getg() == nil checks
in the existing routines.
As part of this cleanup, this CL also deletes onosstack on Windows.
onosstack is from back when the runtime was implemented in C.
It predates systemstack but does essentially the same thing.
Instead of having two different copies of this code, we can use
systemstack consistently. This way we need not port onosstack
to each architecture.
This CL is part of a stack adding windows/arm64
support (#36439), intended to land in the Go 1.17 cycle.
This CL is, however, not windows/arm64-specific.
It is cleanup meant to make the port (and future ports) easier.
Change-Id: I3352de1fd0a3c26267c6e209063e6e86abd26187
Reviewed-on: https://go-review.googlesource.com/c/go/+/288793
Trust: Russ Cox <rsc@golang.org>
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Reviewed-by: Jason A. Donenfeld <Jason@zx2c4.com>
|
|
This change adjusts usleep2HighRes so it does not crash when TLS is
not configured. When g is not available, usleep2HighRes just calls
usleep2 instead.
Updates #8687
Change-Id: Idbb80f7b71d1da350a6a7df7c49154eb1ffe29a8
Reviewed-on: https://go-review.googlesource.com/c/go/+/271907
Run-TryBot: Alex Brainman <alex.brainman@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Jason A. Donenfeld <Jason@zx2c4.com>
Reviewed-by: Simon Rozman <simon@rozman.si>
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Trust: Alex Brainman <alex.brainman@gmail.com>
|
|
This redesigns the way calls work from C to exported Go functions. It
removes several steps from the call path, makes cmd/cgo no longer
sensitive to the Go calling convention, and eliminates the use of
reflectcall from cgo.
In order to avoid generating a large amount of FFI glue between the C
and Go ABIs, the cgo tool has long depended on generating a C function
that marshals the arguments into a struct, and then the actual ABI
switch happens in functions with fixed signatures that simply take a
pointer to this struct. In a way, this CL simply pushes this idea
further.
Currently, the cgo tool generates this argument struct in the exact
layout of the Go stack frame and depends on reflectcall to unpack it
into the appropriate Go call (even though it's actually
reflectcall'ing a function generated by cgo).
In this CL, we decouple this struct from the Go stack layout. Instead,
cgo generates a Go function that takes the struct, unpacks it, and
calls the exported function. Since this generated function has a
generic signature (like the rest of the call path), we don't need
reflectcall and can instead depend on the Go compiler itself to
implement the call to the exported Go function.
One complication is that syscall.NewCallback on Windows, which
converts a Go function into a C function pointer, depends on
cgocallback's current dynamic calling approach since the signatures of
the callbacks aren't known statically. For this specific case, we
continue to depend on reflectcall. Really, the current approach makes
some overly simplistic assumptions about translating the C ABI to the
Go ABI. Now we're at least in a much better position to do a proper
ABI translation.
For comparison, the current cgo call path looks like:
GoF (generated C function) ->
crosscall2 (in cgo/asm_*.s) ->
_cgoexp_GoF (generated Go function) ->
cgocallback (in asm_*.s) ->
cgocallback_gofunc (in asm_*.s) ->
cgocallbackg (in cgocall.go) ->
cgocallbackg1 (in cgocall.go) ->
reflectcall (in asm_*.s) ->
_cgoexpwrap_GoF (generated Go function) ->
p.GoF
Now the call path looks like:
GoF (generated C function) ->
crosscall2 (in cgo/asm_*.s) ->
cgocallback (in asm_*.s) ->
cgocallbackg (in cgocall.go) ->
cgocallbackg1 (in cgocall.go) ->
_cgoexp_GoF (generated Go function) ->
p.GoF
Notably:
1. We combine _cgoexp_GoF and _cgoexpwrap_GoF and move the combined
operation to the end of the sequence. This combined function also
handles reflectcall's previous role.
2. We combined cgocallback and cgocallback_gofunc since the only
purpose of having both was to convert a raw PC into a Go function
value. We instead construct the Go function value in cgocallbackg1.
3. cgocallbackg1 no longer reaches backwards through the stack to get
the arguments to cgocallback_gofunc. Instead, we just pass the
arguments down.
4. Currently, we need an explicit msanwrite to mark the results struct
as written because reflectcall doesn't do this. Now, the results are
written by regular Go assignments, so the Go compiler generates the
necessary MSAN annotations. This also means we no longer need to track
the size of the arguments frame.
Updates #40724, since now we don't need to teach cgo about the
register ABI or change how it uses reflectcall.
Change-Id: I7840489a2597962aeb670e0c1798a16a7359c94f
Reviewed-on: https://go-review.googlesource.com/c/go/+/258938
Trust: Austin Clements <austin@google.com>
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
|
|
@jstarks suggested that recent versions of Windows provide access to high resolution timers. See
https://github.com/golang/go/issues/8687#issuecomment-656259353
for details.
I tried to run this C program on my Windows 10 computer
```
#include <stdio.h>
#include <Windows.h>
#pragma comment(lib, "Winmm.lib")
// Apparently this is already defined when I use msvc cl.
//#define CREATE_WAITABLE_TIMER_HIGH_RESOLUTION = 0x00000002;
int usleep(HANDLE timer, LONGLONG d) {
LARGE_INTEGER liDueTime;
DWORD ret;
LARGE_INTEGER StartingTime, EndingTime, ElapsedMicroseconds;
LARGE_INTEGER Frequency;
QueryPerformanceFrequency(&Frequency);
QueryPerformanceCounter(&StartingTime);
liDueTime.QuadPart = d;
liDueTime.QuadPart = liDueTime.QuadPart * 10; // us into 100 of ns units
liDueTime.QuadPart = -liDueTime.QuadPart; // negative for relative dure time
if (!SetWaitableTimer(timer, &liDueTime, 0, NULL, NULL, 0)) {
printf("SetWaitableTimer failed: errno=%d\n", GetLastError());
return 1;
}
ret = WaitForSingleObject(timer, INFINITE);
if (ret != WAIT_OBJECT_0) {
printf("WaitForSingleObject failed: ret=%d errno=%d\n", ret, GetLastError());
return 1;
}
QueryPerformanceCounter(&EndingTime);
ElapsedMicroseconds.QuadPart = EndingTime.QuadPart - StartingTime.QuadPart;
ElapsedMicroseconds.QuadPart *= 1000000;
ElapsedMicroseconds.QuadPart /= Frequency.QuadPart;
printf("delay is %lld us - slept for %lld us\n", d, ElapsedMicroseconds.QuadPart);
return 0;
}
int testTimer(DWORD createFlag)
{
HANDLE timer;
timer = CreateWaitableTimerEx(NULL, NULL, createFlag, TIMER_ALL_ACCESS);
if (timer == NULL) {
printf("CreateWaitableTimerEx failed: errno=%d\n", GetLastError());
return 1;
}
usleep(timer, 1000LL);
usleep(timer, 100LL);
usleep(timer, 10LL);
usleep(timer, 1LL);
CloseHandle(timer);
return 0;
}
int main()
{
printf("\n1. CREATE_WAITABLE_TIMER_HIGH_RESOLUTION is off - timeBeginPeriod is off\n");
testTimer(0);
printf("\n2. CREATE_WAITABLE_TIMER_HIGH_RESOLUTION is on - timeBeginPeriod is off\n");
testTimer(CREATE_WAITABLE_TIMER_HIGH_RESOLUTION);
timeBeginPeriod(1);
printf("\n3. CREATE_WAITABLE_TIMER_HIGH_RESOLUTION is off - timeBeginPeriod is on\n");
testTimer(0);
printf("\n4. CREATE_WAITABLE_TIMER_HIGH_RESOLUTION is on - timeBeginPeriod is on\n");
testTimer(CREATE_WAITABLE_TIMER_HIGH_RESOLUTION);
}
```
and I see this output
```
1. CREATE_WAITABLE_TIMER_HIGH_RESOLUTION is off - timeBeginPeriod is off
delay is 1000 us - slept for 4045 us
delay is 100 us - slept for 3915 us
delay is 10 us - slept for 3291 us
delay is 1 us - slept for 2234 us
2. CREATE_WAITABLE_TIMER_HIGH_RESOLUTION is on - timeBeginPeriod is off
delay is 1000 us - slept for 1076 us
delay is 100 us - slept for 569 us
delay is 10 us - slept for 585 us
delay is 1 us - slept for 17 us
3. CREATE_WAITABLE_TIMER_HIGH_RESOLUTION is off - timeBeginPeriod is on
delay is 1000 us - slept for 742 us
delay is 100 us - slept for 893 us
delay is 10 us - slept for 414 us
delay is 1 us - slept for 920 us
4. CREATE_WAITABLE_TIMER_HIGH_RESOLUTION is on - timeBeginPeriod is on
delay is 1000 us - slept for 1466 us
delay is 100 us - slept for 559 us
delay is 10 us - slept for 535 us
delay is 1 us - slept for 5 us
```
That shows, that indeed using CREATE_WAITABLE_TIMER_HIGH_RESOLUTION
will provide sleeps as low as about 500 microseconds, while our
current approach provides about 1 millisecond sleep.
New approach also does not require for timeBeginPeriod to be on,
so this change solves long standing problem with go programs draining
laptop battery, because it calls timeBeginPeriod.
This change will only run on systems where
CREATE_WAITABLE_TIMER_HIGH_RESOLUTION flag is available. If not
available, the runtime will fallback to original code that uses
timeBeginPeriod.
This is how this change affects benchmark reported in issue #14790
name old time/op new time/op delta
ChanToSyscallPing 1.05ms ± 2% 0.68ms ±11% -35.43% (p=0.000 n=10+10)
The benchmark was run with GOMAXPROCS set to 1.
Fixes #8687
Updates #14790
Change-Id: I5b97ba58289c088c17c05292e12e45285c467eae
Reviewed-on: https://go-review.googlesource.com/c/go/+/248699
Run-TryBot: Alex Brainman <alex.brainman@gmail.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Alex Brainman <alex.brainman@gmail.com>
Reviewed-by: Austin Clements <austin@google.com>
|
|
RELNOTE=yes
Fixes #37273
Change-Id: Iedb7eab185dfeccb1b26902ef36411d2c53ea3e0
GitHub-Last-Rev: bbe30ba45d4a1bd53757b5824ad28024d5e2b179
GitHub-Pull-Request: golang/go#37380
Reviewed-on: https://go-review.googlesource.com/c/go/+/220578
Run-TryBot: Alex Brainman <alex.brainman@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
|
