Alex copied this from a temporary thing I had written for the runtime
package. In runtime, you can't really access other packages like
syscall, so everything has to be very manual. But in x/sys, we can do
things properly. So this reimplements the function in a more straight
forward way.
Change-Id: I1634904bb1e10f33252954ce02d4b17ae56592e5
Reviewed-on: https://go-review.googlesource.com/c/sys/+/298830
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
In anticipation of the next commit which adds win32 pipe APIs, add some
of the foundational NT APIs for that, which will be required for making
a robust Go pipe library. Also add a simple test case.
Change-Id: I898bd6c5265a8939a7f05a24c4d9b22941dc56b7
Reviewed-on: https://go-review.googlesource.com/c/sys/+/298171
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
It turns out that the proc thread update function doesn't actually
allocate new memory for its arguments and instead just copies the
pointer values into the preallocated memory. Since we were allocating
that memory as []byte, the garbage collector didn't scan it for pointers
to Go allocations and freed them. We _could_ fix this by requiring that
all users of this use runtime.KeepAlive for everything they pass to the
update function, but that seems harder than necessary. Instead, we can
just do the allocation as []unsafe.Pointer, which means the GC can
operate as intended and not free these from beneath our feet. In order
to ensure this remains true, we also add a test for this.
Updates golang/go#44662.
Change-Id: Iaa8b694a6682cc1876879632c7ba068e47b8666d
Reviewed-on: https://go-review.googlesource.com/c/sys/+/297331
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Trust: Bryan C. Mills <bcmills@google.com>
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
The prior function updated the TEB's LastStatus member, which is not
what we want to be doing here. It's also not consistent with Microsoft's
own Go code for their pipe library, which properly uses the "NoTeb"
variant like this commit.
For good measure, we add a simple test case to make sure these paths are
being exercised.
Change-Id: I4080898f704bdc93a6048001b06ffce516fb412d
Reviewed-on: https://go-review.googlesource.com/c/sys/+/298169
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
All references to "Errno" go through the syscalldot function, which
prepends "syscall." if the generation target is not in the syscall
package. Similarly so, we need a windowsdot function to prepend
"windows." to NTStatus.
Change-Id: I86affcf0155632ed5848ebabca97d175a041c6c3
Reviewed-on: https://go-review.googlesource.com/c/sys/+/298170
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
The same symbol name might be defined multiple times with different
function names or signatures. In that case, it's an error to redefine
the proc variable. So deduplicate these before printing.
Change-Id: I7a636cd44fb1ea6f51840ef3924d6316e266faa0
Reviewed-on: https://go-review.googlesource.com/c/sys/+/288332
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
The native NT API returns error values from a different namespace as the
usual Win32 one. This means it needs to be typed differently. This
commit adds broad support for using NTSTATUS values in a new type called
NTStatus.
First we add the type as a basic uint32. Then we add all of the
predefined constants from ntstatus.h, by augmenting mkerrors.bash to do
automatic extraction. There's a convenece way to convert an NT error to
a Win32 error, so we add the NTStatus.Errno() function. Since NTStatus
is an error type, we define an Error() function that returns a string by
asking ntdll.dll for its contents, in the exact same way that
syscall.Errno.Error() does, by calling FormatMessage. Since functions
need to actually use this, we add the rule that if a `//sys` declaration
returns an error value called "ntstatus", then the type underlying the
error interface is an NTStatus instead of an Errno. Finally we fix one
function that was returning an error interface of an Errno rather than
an NTStatus.
Change-Id: I06296b9563bbec526759d12a19f13ac6ad46dcc3
Reviewed-on: https://go-review.googlesource.com/c/sys/+/297330
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Trust: Alex Brainman <alex.brainman@gmail.com>
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
The third argument to GetQueuedCompletionStatus is a pointer to a
uintptr, not a uint32. Users of this functions have therefore been
corrupting their memory every time they used it. Either that memory
corruption was silent (dangerous), or their programs didn't work so they
chose a different API to use.
Updates golang/go#44538.
Change-Id: Ie1f66de11001cf9c8195afaa61f003a86f821a95
Reviewed-on: https://go-review.googlesource.com/c/sys/+/295174
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
This fixes the name of a struct to match Microsoft's documentation.
It also reverts a change I made in the last patchset of CL 285714. The
idea there was that a Go bool would suffice for a win32 BOOL, because of
little endian. But in fact, a value of 0xff000000 would be treated as
false by Go but true by C, which is a problem. So this changes to using
a vanilla int32 type, which matches the C "int" used in the Microsoft
headers.
Change-Id: Id7cd306e916b3754e8dfe32bf11ec30cad3a13b4
Reviewed-on: https://go-review.googlesource.com/c/sys/+/285717
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
This function returns 0 and sets last error on failure. While this isn't
said explicitly on MSDN, there's no PID 0, and trivial reverse
engineering shows that this is the case. For example:
.text:00000001800035ED loc_1800035ED: ; CODE XREF: GetWindowThreadProcessId+23↑j
.text:00000001800035ED ; GetWindowThreadProcessId+3D↑j ...
.text:00000001800035ED mov ecx, 578h ; LastError
.text:00000001800035F2 call cs:__imp_RtlSetLastWin32Error
.text:00000001800035F9 nop dword ptr [rax+rax+00h]
.text:00000001800035FE
.text:00000001800035FE loc_1800035FE: ; CODE XREF: GetWindowThreadProcessId+65↑j
.text:00000001800035FE xor eax, eax
.text:0000000180003600 jmp short loc_1800035AA
This function was also just added by a user who is likely its only
consumer, so this error is not too late to fix.
Change-Id: I5dd24e78c006686bb8f1288ad0fe63cd67df56a6
Reviewed-on: https://go-review.googlesource.com/c/sys/+/285272
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Trust: Alex Brainman <alex.brainman@gmail.com>
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
One thing the newer notification API does not do is automatically
provide the service exit code in its notifier response. It requires
querying manually. Unfortunately, we weren't propagating this
information up from the lower level struct, so this commit copys that
information over.
Change-Id: I70c683007ce34ffab6196329acefc8443f921ebe
Reviewed-on: https://go-review.googlesource.com/c/sys/+/274577
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Trust: Alex Brainman <alex.brainman@gmail.com>
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
The "?" code assumed that the error value was always called "err", when
in reality it might be called something different (like "ret") or even
entirely absent. This commit makes the templating robust to that. At the
same time, we move a lot of the complexity out of the actual templates
and into helper functions, so that this remains easy to read.
Change-Id: I939d56413a24f0e3e1bbf13da5adf13e9401747a
Reviewed-on: https://go-review.googlesource.com/c/sys/+/275472
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Trust: Alex Brainman <alex.brainman@gmail.com>
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Currently, doing s.UpdateConfig(s.Config()) will destroy the service,
because s.Config() fails to populate the SidType member, but
UpdateConfig will set the SidType, so that expression effectively zeros
out the SidType. Fix this by having Config() fetch the SidType in the
same way that it fetches the other additional fields there, such as
DelayedStartUp.
Change-Id: Idb917ef1e942020499b411b7777b995c29f0e7d2
Reviewed-on: https://go-review.googlesource.com/c/sys/+/270897
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
The original IsWow64Process returns false on arm, always, and so
IsWow64Process2 was added to account for this scenario. This isn't
available on older versions of Windows, so we mark it as such using the
new '?' notation. Finally, we add a test to make sure this all works and
does the expected thing on different versions of Windows.
Change-Id: Ic0412578cfb3f4cf6c9dc92a0028abc579bf6c85
Reviewed-on: https://go-review.googlesource.com/c/sys/+/269077
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Some functions that x/sys/windows will eventually have won't be
available on older versions of Windows, and in general we can expect
Microsoft to continue to add functions to newer builds of Windows
10. Therefore, we introduces a new notation for marking functions that
might not exist, letting the caller handle the situation without a
panic.
Change-Id: Ia66bf4aab601357198872c5cd29b6ca7c3bc6969
Reviewed-on: https://go-review.googlesource.com/c/sys/+/269938
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Because we're expecting for future functions to be unavailable, we
should add an Unwrap() function to the DLLError struct, so that people
can test for this situation easily via:
if errors.Is(err, windows.ERROR_PROC_NOT_FOUND) { ... }
DLLError already was wrapping the underlying Errno error, but never got
the Go 1.13 helper method.
Update golang/go#42584
Change-Id: Ib916ddd55a2de29f988edaaf82f2ae0ce1b18e3b
Reviewed-on: https://go-review.googlesource.com/c/sys/+/269937
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
While most of our constant gathering can be automated, some win32 error
constants come from unusual corners of the WDK, such as setupapi.h. We
add those constants to a new file, so that they can be referenced as
constants by Go programs.
Change-Id: I9bce7fddf7aaa00204e6048b53f14e6ba18d5f13
Reviewed-on: https://go-review.googlesource.com/c/sys/+/268358
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Trust: Alex Brainman <alex.brainman@gmail.com>
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Simon Rozman <simon@rozman.si>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
CL 258038 improperly added a weird custom type to mkwinsyscall, rather
than doing the norm with wrapper functions. So, we revert the change to
mkwinsyscall and add the proper wrapper function to do the type
conversion.
Change-Id: I98134e4ce6bf4b52e1384fe84bddeedb00e18c0b
Reviewed-on: https://go-review.googlesource.com/c/sys/+/268777
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
Trust: Alex Brainman <alex.brainman@gmail.com>
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
We already provide ByteSliceFromString and BytePtrFromString, and on
Windows we provide UTF16FromString, UTF16PtrFromString, UTF16ToString,
and UTF16PtrToString. So this commit fills in the remaining two
Byte-oriented functions: ByteSliceToString and BytePtrToString. Since
the existing two are available on windows, unix, and plan9, we add the
remaining two to the same places. This helps eliminate unsafe pointer
options in addition to triggering checkptr in Go 1.15, by eliminating
the use of prior idioms for these types of casts.
Change-Id: I85ae49f2756e142c2462fe8b2428216b6992e089
Reviewed-on: https://go-review.googlesource.com/c/sys/+/263757
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Trust: Brad Fitzpatrick <bradfitz@golang.org>
Trust: Jason A. Donenfeld <Jason@zx2c4.com>
This test's existence was predicated upon assumptions about the full
range of known data types and known data into those types. However,
we've learned from Microsoft that there are several undocumented secret
registry types that are in use by various parts of Windows, and we've
learned from inspection that many Microsoft uses of registry types don't
strictly adhere to the recommended value size. It's therefore foolhardy
to make any assumptions about what goes in and out of the registry, and
so this test is meaningless and error-prone.
Updates golang/go#35084
Change-Id: Ie545229afd8dc5bde90fffa0f735f7102cd4a6eb
Reviewed-on: https://go-review.googlesource.com/c/sys/+/203605
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
According to MSDN, "If the data has the REG_SZ, REG_MULTI_SZ or
REG_EXPAND_SZ type, this size includes any terminating null character or
characters unless the data was stored without them. [...] If the data
has the REG_SZ, REG_MULTI_SZ or REG_EXPAND_SZ type, the string may not
have been stored with the proper terminating null characters. Therefore,
even if the function returns ERROR_SUCCESS, the application should
ensure that the string is properly terminated before using it;
otherwise, it may overwrite a buffer."
It's therefore dangerous to pass it off unbounded as we do, and in fact
this led to crashes on real systems.
Change-Id: I2ab324e85f75dc3e4d6d62fec3b96937fec77510
Reviewed-on: https://go-review.googlesource.com/c/sys/+/202957
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
While the other BSDs use __sysctl as the name, Darwin now uses sysctl,
without the leading underscores, and considers __sysctl to be "private".
Using __sysctl leads to App Store rejections, and Go's syscall package
already uses the proper syscall. So this commit changes Darwin's syscall
to use it too here, while reverting a recent commit that removed it all
together on arm and arm64.
This reverts CL 194097.
Fixesgolang/go#35103
Updates golang/go#34133
Updates golang/go#35101
Change-Id: Ic72d5e7a435b99fe62c533b77b2c3790590f4c9e
Reviewed-on: https://go-review.googlesource.com/c/sys/+/202837
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: David Crawshaw <crawshaw@golang.org>
Reviewed-by: Tobias Klauser <tobias.klauser@gmail.com>
Chmod toggles the FILE_ATTRIBUTES_READONLY flag depending on the
permission bits. That's a bit odd but I guess some compromises were made
at some point and this is what was chosen to map to a Unix concept that
Windows doesn't really have in the same way. That's fine. However, the
logic used in Chmod was forgotten from Open, which then manifested
itself in various places, most recently, go modules' read-only behavior.
This corresonds with the syscall CL 202439.
Updates golang/go#35033
Change-Id: Id8e74c5205057a74a35eda213516780b79a2aed2
Reviewed-on: https://go-review.googlesource.com/c/sys/+/202440
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bryan C. Mills <bcmills@google.com>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
ShellExecute is an ancient API with an unusual return signature. It
pretends to return an HINSTANCE for backwards compatibility with Windows
3, but it's actualy a fake HINSTANCE. What's really happening here,
according to MSDN, is that it returns either an Windows error less
than or equal to 32, or it succeeds. So we adjust the return value
accordingly.
Prior to this commit, it was impossible to tell whether this command had
succeeded. For example, when using the "runas" verb, ShellExecute did
not correctly indicate whether or not permission was granted.
Change-Id: Ie60554d6465798bacb9a225c4ead7e8dd62bce14
Reviewed-on: https://go-review.googlesource.com/c/sys/+/199521
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Nenad Kozul <knenad@gmail.com>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
This allows us to modify this file and fix it more fluidly. Users can
invoke it from go generate via:
go run golang.org/x/sys/windows/mkwinsyscall
This was taken from Go repo commit 6b85fa80.
Updates golang/go#34388
Change-Id: I8dc39eed96b2499ccbde53554b3e16e6c1f6aa98
Reviewed-on: https://go-review.googlesource.com/c/sys/+/198637
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
From MSDN on CreateServiceW:
The returned handle is only valid for the process that called
CreateService. It can be closed by calling the CloseServiceHandle
function.
This isn't an actual kernel object to be closed with CloseHandle.
Change-Id: Iee225a666576d57a6c4864abef20206d54cbbce2
Reviewed-on: https://go-review.googlesource.com/c/sys/+/198298
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Reviewed-by: Jason A. Donenfeld <Jason@zx2c4.com>
There's no point in adding a function call to retrieve a constant, or
worse, a syscall to retrieve a constant. These are fixed and baked so
deep into NT they'll never change. So let's benefit from the obvious
optimization and make these constants. Go easily inlines the function
calls as well. We also take the opportunity to sunset
OpenCurrentProcessToken and restore its original behavior, since users
should be invoking this deliberately with the correct access mask.
Change-Id: I92f7de56c0fcf5b69b59f5a79d2828c7ddf3c8f6
Reviewed-on: https://go-review.googlesource.com/c/sys/+/196800
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
This function shouldn't return an error. Like other String() functions
everywhere in Golang, this should instead return empty or a token value
during an error, so that it can be passed to %v and similar. Also, allow
for SID strings of maximum size.
Change-Id: Ib6d8407f8ad0bdabcb22c31b8f387594f2ea7672
Reviewed-on: https://go-review.googlesource.com/c/sys/+/196799
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
GetCurrentProcess and GetCurrentThread return -1 and -2 respectively. We
could arguably hard code those values, but MSDN cautions not to; I'm
sure this advice is old now, given that the psuedo handles for tokens
(not processes/threads) are now implemented with inline functions in the
headers for Windows 8, but anyway, we'll follow Microsoft's advice.
However, regardless of that, these functions never ever return an error.
MSDN doesn't indicate that they do, reverse engineering the functions
doesn't indicate that they do, and checking against 0 is just plain
wrong, considering 0!=INVALID_HANDLE_VALUE; however
INVALID_HANDLE_VALUE==-1, so that's not correct either. In fact,
checking any value and returning any error does not make sense.
Incidently having to check code for the pseudo handle is more verbose
too.
In order to make this function do the correct thing and meet the spec,
remove the error value from the return.
Change-Id: If03c9dab001be3bf5a04999aef20dbfcf8a4f405
Reviewed-on: https://go-review.googlesource.com/c/sys/+/196798
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
The SECURITY_ATTRIBUTES struct always takes a SECURITY_DESCRIPTOR
pointer. Now that we've defined SECURITY_DESCRIPTOR, make
SECURITY_ATTRIBUTES properly specify the type. This eliminates the need
for terrible uintptr(unsafe.Pointer(...)) casts everywhere.
Change-Id: Ibbc85524cfe33589d43f963e10aa19d7f47686f2
Reviewed-on: https://go-review.googlesource.com/c/sys/+/196797
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
This adds the basic foundation for dealing with security descriptors and
access control lists. The basic creators and getters are included in
this patch. These are some of the most fundamental security objects on NT,
and any work with the security API is fairly limited without it. These
are "core" NT structures.
Change-Id: I9a6399cb6ee41a825de30d5364ab69102d5f6d57
Reviewed-on: https://go-review.googlesource.com/c/sys/+/195498
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jason A. Donenfeld <Jason@zx2c4.com>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Windows type PBOOL is a pointer to a 4 byte value, where 0 means false
and not-0 means true. That means we should use uint32 here, not bool,
since Go bools can be 1 byte. This commit was re-generated using
mksyscall_windows.go from CL 196122.
Updates: golang/go#34364
Change-Id: I8e83b9a09c0b58d14ac9a7dee316553940ac6ee3
Reviewed-on: https://go-review.googlesource.com/c/sys/+/196123
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Jason A. Donenfeld <Jason@zx2c4.com>
There are a few functions to control the behavior of shutdown and
logout, both for what the current process does during shutdown, and also
whether or not the current process is running in an interactive session.
The below code is a port of the MSDN example code to Go using one of the
added new functions:
https://docs.microsoft.com/en-us/windows/win32/shutdown/how-to-shut-down-the-system
func shutdownLikeMSDNDoes() error {
seShutdownName, err := windows.UTF16PtrFromString("SeShutdownPrivilege")
if err != nil {
return err
}
var shutdownPriv windows.Tokenprivileges
err = windows.LookupPrivilegeValue(nil, seShutdownName, &shutdownPriv.Privileges[0].Luid)
if err != nil {
return err
}
shutdownPriv.Privileges[0].Attributes = windows.SE_PRIVILEGE_ENABLED
shutdownPriv.PrivilegeCount = 1
process, err := windows.GetCurrentProcess()
if err != nil {
return err
}
var token windows.Token
err = windows.OpenProcessToken(process, windows.TOKEN_ADJUST_PRIVILEGES | windows.TOKEN_QUERY, &token)
if err != nil {
return err
}
defer token.Close()
err = windows.AdjustTokenPrivileges(token, false, &shutdownPriv, 0, nil, nil)
if err != nil {
return err
}
err = windows.ExitWindowsEx(windows.EWX_SHUTDOWN | windows.EWX_FORCE,
windows.SHTDN_REASON_MAJOR_OPERATINGSYSTEM | windows.SHTDN_REASON_MINOR_UPGRADE | windows.SHTDN_REASON_FLAG_PLANNED)
if err != nil {
return err
}
return nil
}
Note, though, that this function doesn't set the token privs back to how
they were before, which isn't good. A more robust method than the MSDN
one above would be to duplicate&impersonate.
Fixes: golang/go#34271
Change-Id: Ibe55ddd35b709d9ab793cb9af47c39901c5e5c69
Reviewed-on: https://go-review.googlesource.com/c/sys/+/195497
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Bruce Downs <bruceadowns@gmail.com>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
A usual thing to ask is, "Is my current token in group X?" The right way
of doing such a thing is:
processToken, err := windows.OpenCurrentProcessToken()
if err != nil {
return false, err
}
defer processToken.Close()
var checkableToken windows.Token
err = windows.DuplicateTokenEx(token, windows.TOKEN_QUERY | windows.TOKEN_IMPERSONATE, nil, windows.SecurityIdentification, windows.TokenImpersonation, &checkableToken)
if err != nil {
return false, err
}
defer checkableToken.Close()
isMember, err := checkableToken.IsMember(someSID)
return isMember && err == nil, nil
This is the same flow that's used by, for example, shell32's internal
_LUAIsTokenAdmin function.
However, this all fails unless the original token is opened with
duplicate access. So this commit adjusts OpenCurrentProcessToken to do
the right thing.
Change-Id: I18efdfde43097ea9d10758018b0df132fba819f5
Reviewed-on: https://go-review.googlesource.com/c/sys/+/192337
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Simon Rozman <simon@rozman.si>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Indeed Go has mutexes of its own, but these are considerably
different from the native Windows ones, that can work across processes
and be put in various namespaces. They're an essential part of Windows
systems programming and important for interfacing with various external
interfaces.
Change-Id: I03987800ed1c134442321678c2c7d7aa359ecb36
Reviewed-on: https://go-review.googlesource.com/c/sys/+/192497
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
The Go runtime disables priority boosting because it interferes with the
scheduler timer threads in some cases. But some apps might want to
re-enable it for specific reasons, so this commit adds the function to
do so.
Change-Id: Ida68b9f2b188560b46c322197461dd06453329e2
Reviewed-on: https://go-review.googlesource.com/c/sys/+/191839
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
While RtlGetVersion was added so that users can get the Windows version
that isn't affected by manifesting, RtlGetVersion is still stubbed out
by the application compatibility layer (aclayers.dll and apphelp.dll)
for certain processes, such as msiexec.exe, rendering these functions
useless for actually determining the underlying operating system. This
matters in the case of msiexec.exe using a custom action DLL to install
a kernel driver, which of course is version specific. This is also
useful, it turns out, for the C runtime library, in which Microsoft uses
this function too. It's existed as a stable interface since Windows XP,
has Wine support, and is used in a decent amount of software.
Change-Id: If391e43bc6d798eff6803d5a7aa6a179f2b31d88
Reviewed-on: https://go-review.googlesource.com/c/sys/+/188119
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Other functions, like GetVersion(), will lie about the OS version
depending on various win32 and manifest compatibility shims in place.
Calling RtlGetVersion is the proper way to retrieve the true OS version.
Change-Id: I2bd6d097dd763df51617cd825dc0ad300abf6212
Reviewed-on: https://go-review.googlesource.com/c/sys/+/182718
Run-TryBot: Jason A. Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matt Layher <mdlayher@gmail.com>
Reviewed-by: Jason A. Donenfeld <Jason@zx2c4.com>
This adds the recommended API for determining well-known folder paths,
such as where to place application configuration data. The MSDN
documentation mentions an optimization for the "current user" by passing
NULL as the token, so we provide both variants.
Updates golang/go#32248
Change-Id: I4a2d5d833543e6a0ba8f318944dd6493a0ec31d3
Reviewed-on: https://go-review.googlesource.com/c/sys/+/181637
Reviewed-by: Jason Donenfeld <Jason@zx2c4.com>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Run-TryBot: Jason Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
The SCM can be locked by NT. While traditionally any process could lock
the SCM using "LockServiceDatabase", Microsoft removed this
functionality because it created so many bugs, and that function now
does nothing. However, the system itself, via the "NT Service Control
Manager", is still allowed to lock the SCM.
For example, at boot time on Windows 8.1, the SCM is locked after a
service is started until that service reports itself in a running state.
This poses a bit of a problem: it's useful to install device drivers
from inside services as part of their initialization, and mark the
service as having started only after the device has installed. But
device installation might potentially load new drivers, and drivers
themselves exist as a special type of service. This means that if a
driver is installed before marking the service as started, the entire
SCM will deadlock, and the OS will be partially unresponsive for a
minute or two.
Fortunately Microsoft supplies an API for exactly this purpose. The
solution is to mark the service as started before installing device
drivers, only under the circumstance that the SCM is locked. So, this
commit adds the proper API for determining this. It can be used like
this:
if m, err := mgr.Connect(); err == nil {
if lockStatus, err := m.LockStatus(); err == nil && lockStatus.IsLocked {
log.Printf("SCM locked for %v by %s, marking service as started", lockStatus.Age, lockStatus.Owner)
changes <- svc.Status{State: svc.Running}
}
m.Disconnect()
}
deviceDriver.Install()
This creates messages like the following, indicating that this API
works:
SCM locked for 1s by .\NT Service Control Manager, marking service as started
Change-Id: Ic2f5b387e23efc3a287b2ab96ff84b357b712e36
Reviewed-on: https://go-review.googlesource.com/c/sys/+/180977
Run-TryBot: Jason Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
This corrects the Windows int type to be the more correct int32 Go
analog, as well as not using GetLastError() for the error value of the
GUID string parsing function.
Change-Id: I9716f991ef649f7d299295e3f4e75d3986ec3a74
Reviewed-on: https://go-review.googlesource.com/c/sys/+/181397
Run-TryBot: Jason Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Many service attributes are settable using the ordinary CreateService
function, but ones added later in Windows need to be set using the
ChangeServiceConfig2 function. One of these is the Description field,
which is nicely plumbed behind the scenes, so that users of the API can
set it, and the mgr package will just figure out what to do with it.
Another one that works exactly the same way is SidType. Support for its
constants was added in 30999d6 ("windows: add missing service
constants"), but it wasn't actually built into the mgr package's
configuration struct, creating inconstancies in interface.
This commit rectifies that by adding proper support to mgr's config
struct.
Change-Id: I4f148f2d2477a03a65e8a571a8401392f6fe9f28
Reviewed-on: https://go-review.googlesource.com/c/sys/+/180978
Run-TryBot: Jason Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Matt Layher <mdlayher@gmail.com>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Virtually every project that works with x/sys/window's GUIDs winds up
rolling their own version of this in one way or another. So let's add
the correct win32 wrappers for it, so that these are always generated,
parsed, and converted in the uniform correct way.
Change-Id: I35f4b4ab5fc681d3e16fc5bbaf2cb20031eb3f12
Reviewed-on: https://go-review.googlesource.com/c/sys/+/180938
Run-TryBot: Jason Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Rather than having to write grotesque things like:
groups, _ := processToken.GetTokenGroups()
for _, g := range (*[(1 << 28) - 1]windows.SIDAndAttributes)(unsafe.Pointer(&groups.Groups[0]))[:groups.GroupCount] {
// ...
}
Users can now write clean things like this:
groups, _ := processToken.GetTokenGroups()
for _, g := range groups.AllGroups() {
// ...
}
Change-Id: Ief06de6899c497175628ff51b9d6ae55a90d14f1
Reviewed-on: https://go-review.googlesource.com/c/sys/+/178857
Run-TryBot: Jason Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
I realize that at this time, x/sys/windows is most certainly not aiming
to be some sort of UI library or anything remotely close to that.
However, MessageBox is sort of the one universal Windows API that's used
even for console programs. It's *the* classic debugging helper. It's
even pretty customary for console programs to display their usage
message in a message box rather than stdout. I realize this might be
somewhat disturbing to consider if you're in the blissful Unix utopia,
but on Windows, this is just sort of how things roll. Easy access to
MessageBox() is important.
Change-Id: I16183b69e8a27a5ddaf73d1d4e106bb7b201a6f0
Reviewed-on: https://go-review.googlesource.com/c/sys/+/178899
Run-TryBot: Jason Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
This is the way to do things like execute a process elevated with UAC
and interact with that whole system. It turns out to be quite important
for writing Windows software.
Change-Id: I5e05dc9b89ea308d42ac86ba563fd01922fc940c
Reviewed-on: https://go-review.googlesource.com/c/sys/+/178898
Run-TryBot: Jason Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
These allow actual inspection of SIDs. For example, it might be
desirable to iterate through the group SIDs in order to find one set by
SERVICE_CONFIG_SERVICE_SID_INFO:
for _, g := range groups {
if g.Attributes != windows.SE_GROUP_ENABLED|windows.SE_GROUP_ENABLED_BY_DEFAULT|windows.SE_GROUP_OWNER {
continue
}
if !g.Sid.IsValid() {
continue
}
if g.Sid.IdentifierAuthority() != windows.SECURITY_NT_AUTHORITY {
continue
}
if g.Sid.SubAuthorityCount() < 6 || g.Sid.SubAuthority(0) != 80 {
continue
}
sid = g.Sid
break
}
Another usage of the APIs added would be to find if a user is in the
administrator group with either an elevated or unelevated token:
isAdmin := false
for _, g := range groups {
if g.Attributes&(windows.SE_GROUP_ENABLED|windows.SE_GROUP_USE_FOR_DENY_ONLY) == 0 {
continue
}
if !g.Sid.IsWellKnown(windows.WinBuiltinAdministratorsSid) {
continue
}
isAdmin = true
break
}
Change-Id: I8f8dc8d37b71ec58fd51e21ea1f1b3aada6d66b0
Reviewed-on: https://go-review.googlesource.com/c/sys/+/177841
Run-TryBot: Jason Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
This is what everybody winds up doing with this object, so we make it
somewhat nicer than copying and pasting this everywhere or using type
aliases.
Change-Id: I3e12395cadfe212a7d01ce86478de9486383729a
Reviewed-on: https://go-review.googlesource.com/c/sys/+/178577
Run-TryBot: Jason Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
The GetCurrent*Token variety of functions are actually implemented as
inline functions in the header files of the SDK. Attempting to call out
to these as library functions is an error. This commit also adds a test
to ensure that these work as expected.
Change-Id: I105f1ca1a8936114fe61bc22188200c31f240a23
Reviewed-on: https://go-review.googlesource.com/c/sys/+/177840
Run-TryBot: Jason Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
This adds a number of useful comments for ChangeServiceConfig2. It looks
like the MingW headers have these in two different places, awkwardly,
and whoever imported these constants missed the second half of them,
whereas the real Microsoft header files have them all together in one
place.
Change-Id: I723c4bcd86e5a1b905cb738c5c2b49805a354af4
Reviewed-on: https://go-review.googlesource.com/c/sys/+/177842
Run-TryBot: Jason Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Byte slices aren't necessarily aligned, which means casting them to
integer types and dereferencing may result in an unaligned load. This
is mostly fine on Intel but isn't necessarily fine on other platforms.
Any good compiler will generate optimal code for the platform using the
pattern of this commit.
Change-Id: I6dd8debad1cb850b8562ee96ae0f366d1f822a6f
Reviewed-on: https://go-review.googlesource.com/c/sys/+/176857
Run-TryBot: Jason Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
The svc package exposes svc.SessionChange, but it's impossible to do
anything with them without these structures, and without being able to
enumerate them prior to events, the events themselves aren't useful, so
we add the enumeration functions as well.
Change-Id: I14c932dfe97c6712fd4868c1b3a0e3a61a6a562c
Reviewed-on: https://go-review.googlesource.com/c/sys/+/176623
Run-TryBot: Jason Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
This lays the groundwork for service notification and tracking by adding
the required API functions. Users can make notifiers directly using it,
or later if we're feeling ambitious, we can see if we can come up with a
generalized solution in x/windows/svc.
Change-Id: I80503cc27970fbb23bf17cd8bc50eaa7787aa6bd
Reviewed-on: https://go-review.googlesource.com/c/sys/+/176624
Run-TryBot: Jason Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
These are extremely useful functions and core to the Windows security
API. They are so useful, in fact, that most of these were taken right
out of the Go repo's internal/syscall/windows package.
Change-Id: I13e34b830dd60f59fcae8085ae2be189d9cc9282
Reviewed-on: https://go-review.googlesource.com/c/sys/+/176625
Reviewed-by: Matt Layher <mdlayher@gmail.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Run-TryBot: Matt Layher <mdlayher@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
The two functions have the same levels of compatibility, but this latter
one gives us access to the process ID, which is important for things
like WFP whitelisting. The change required is fairly trivial too.
Change-Id: Ifb6b3ee3e897202b9cffa1388c53c25cbcfede61
Reviewed-on: https://go-review.googlesource.com/c/sys/+/173666
Run-TryBot: Jason Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
While the types_windows file previously had a small handful of types,
this forced application code to have an awkward mixture of artisanal
error constants and factory-ready ones. This commit adds the missing
ones and separates them into a new file, since they are quite numerous.
These also preserve the order of winerr.h, which should make it somewhat
easier to import new ones in the future.
Fixesgolang/go#31360
Change-Id: If2abc507a8884ec1641f0b17fe0c612a950d3644
Reviewed-on: https://go-review.googlesource.com/c/sys/+/170918
Reviewed-by: Jason Donenfeld <Jason@zx2c4.com>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Jason Donenfeld <Jason@zx2c4.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
We already have all the structs and constants for async I/O, and the
various functions of x/sys/windows take the right parameters. But we're
missing the final step of any ordinary async I/O routine: getting the
result of overlapped I/O. Without this, the rest of the plumbing
supported by this module isn't actually so useful. So add this small
oversight.
Change-Id: I0ce1a71bce06bc81a83f3b0ca10ad9c4b67af726
Reviewed-on: https://go-review.googlesource.com/c/sys/+/168521
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
The previous selection was a bit haphazard, and defining random
constants in an application and combining those with existing constants
in this library looks kind of weird. So instead let's add the actual set
of flags, instead of the strange prior assortment.
Change-Id: I6ca266cd80aa7a43e93e969e1dbb09c8c7bf12c7
Reviewed-on: https://go-review.googlesource.com/c/sys/+/168520
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
The security API is already quite extensive, but for some strange
reason, this essential and useful function was left out of the initial
port. So, we add it here, along with the relevant constants and a test
case.
Change-Id: I99568703565addf15603480f11b0edafdfc1718f
Reviewed-on: https://go-review.googlesource.com/c/sys/+/167378
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
Run-TryBot: Alex Brainman <alex.brainman@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
The %WINDIR% variable is an odd choice and not even entirely reliable.
Since Windows 2000, there has been a specific function for determining
this information, so let's use it. It's also a useful function in its
own right for folks who want to launch system tools in a somewhat safe
way, like netsh.exe.
Updates golang/go#14959
Updates golang/go#30642
Change-Id: Ic24baf37d14f2daced0c1db2771b5a673d2c8852
Reviewed-on: https://go-review.googlesource.com/c/sys/+/165759
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Alex Brainman <alex.brainman@gmail.com>
These DLLs, well advapi32.dll in particular, are vulnerable to classic
DLL directory injection attacks. The rest of x/sys/windows moved over to
the safe system loader, but apparently the svc package was forgotten.
This tidies up that oversight.
Change-Id: I330fa752cf2d49ccc5cf1bd60fb4bd612bd2b6b0
Reviewed-on: https://go-review.googlesource.com/c/sys/+/165758
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
ARM64 does this exact same thing, and in fact this is required on
Android 9, where Bionic now only allows lstat on 32-bit platforms
Folks are running into this when using Android on Chromebooks,
which are keeping AMD64 Android alive. So, we change the AMD64
behavior to model that of ARM64.
Change-Id: Id933cc22b42ab365c12980915557bdbf88661a4d
Reviewed-on: https://go-review.googlesource.com/c/153319
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
This is yet another syscall that needs to be changed around for
Android's seccomp. We already do this in MIPS64 and ARM64, but
apparently forgot about AMD64.
Change-Id: I42c8a4653a55d4714dacfaf7add14012f428c992
Reviewed-on: https://go-review.googlesource.com/109835
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
By simply rearranging the logic, we avoid the overhead of a superfluous
call to getsockopt. For, if p is already non empty, there's no point
in having to check if we need to attach dummy payload. This has
performance benefits when using send/recvmsg for high speed
communications.
Change-Id: If47ab5322464aad8e6d2603a369e95327223211f
Reviewed-on: https://go-review.googlesource.com/108337
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Jason A. Donenfeld