We are using `epoll()` in edge-triggerred mode, which requires
us to hit `EAGAIN` on `write()` before we are guaranteed
to receive an `EPOLLOUT` event.
`_SocketStreamConsumer` looks at `_NativeSocket.writeAvailable`
after issuing a `write` to decide whether it should continue
writing or pause and wait until event handler sends us `writeEvent`.
However we did not previously set `writeAvailable` to `true` until
the first `writeEvent` arrived, which lead to a hang on some
Linux servers: we would write a small amount of bytes into a pipe
and then wait for `writeEvent`, which would never come, as `epoll()`
is only guaranteed to wake up with `EPOLLOUT` if we hit `EAGAIN`
on `write()`.
This CL also changes `_NativeSocket.nativeWrite` implementation
to write bytes into the file descriptor until it gets EAGAIN.
This reduces the number of roundtrips between Dart and
native.
Unfortunately this CL does not come with a regression test
because it is relying on concrete `epoll()` behaviour and it this
behaviour does not reproduce for me.
TEST=Tested manually on the affected servers.
FIXED=b/286346121
Fixed: 286346121
Change-Id: I37fef9aa12b1da724b035aa9577b414a8057217e
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/308101
Reviewed-by: Martin Kustermann <kustermann@google.com>
Commit-Queue: Slava Egorov <vegorov@google.com>
This adds [sendMessage] and [receiveMessage] methods to [RawSocket] class. They are only supported on Linux at the moment as they require connection opened as unix domain socket connection.
This introduces [SocketControlMessage] class that represents a message sent via socket and also introduces [ResourceHandle] class that wraps [RandomAccessFile], [Socket]/[RawSocket]/[RawDatagramSocket], [Stdin]/[Stdout] for
marshalling/unmarshalling purposes.
Underlying OS implementation supports various kinds of control messages that can be passed via sockets, this CL only adds support for sending/receiving opened file description handles.
When receiving a message recipient can attempt to extract handles out of it via [tryExtractHandles]. It returns [null] if message has no handles to extract.
This is continuation of the work started on https://dart-review.googlesource.com/c/sdk/+/205067.
Bug: https://github.com/dart-lang/sdk/issues/46328
TEST=unix_socket_test
Change-Id: Ic9125b51dc80b677452e454366bae4118c298081
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/212036
Commit-Queue: Alexander Aprelev <aam@google.com>
Reviewed-by: Lasse R.H. Nielsen <lrn@google.com>
Reviewed-by: Zach Anderson <zra@google.com>
This relands https://dart-review.googlesource.com/c/sdk/+/205633
but without renaming TARGET_OS_IPHONE to DART_TARGET_OS_IPHONE.
It also changes uses of TARGET_OS_IOS to
DART_TARGET_OS_MACOS_IOS to be consistent with the rest of the
VM.
TargetConditionals.h for XCode 13 defines several
TARGET_OS_* preprocessor symbols that confuse the
Dart build. There is probably a more targeted fix
for this, but renaming the symbols that Dart uses
will also prevent this problem if more symbols
are added to the platform headers in the future.
See: https://github.com/dart-lang/sdk/issues/46499
TEST=It builds.
Change-Id: Ie775c19dd23cfdf5f65e5ebc6ee4ec3a561676fa
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/205860
Commit-Queue: Zach Anderson <zra@google.com>
Reviewed-by: Alexander Aprelev <aam@google.com>
TargetConditionals.h for XCode 13 defines several
TARGET_OS_* preprocessor symbols that confuse the
Dart build. There is probably a more targeted fix
for this, but renaming the symbols that Dart uses
will also prevent this problem if more symbols
are added to the platform headers in the future.
See: https://github.com/dart-lang/sdk/issues/46499
TEST=It builds.
Change-Id: I3b33a03b4a9a14b76d55fe12f8cdefec4b3c3664
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/205633
Commit-Queue: Zach Anderson <zra@google.com>
Reviewed-by: Siva Annamalai <asiva@google.com>
Support Unix domain sockets communication on Linux, MacOS and Android.
Changes:
1. Add a field for InternetAddressType named unix.
2. Constructor of InternetAddress gains one more optional field: type. InternetAddress(String address, {InternetAddressType type});
3. Add another constructor to InternetAddress which taks raw address/path for ip/unix addresses as an argument. InternetAddress.fromRawAddress(Uint8List rawAddress, {InternetAddressType type});
The operation for unix domain sockets communication is basically the same as normal sockets except an InternetAddress with type unix should be passed.
Change-Id: I6a1135bbdd7f4e4fc745ccf8f95dec5272b6839b
Bug: https://github.com/dart-lang/sdk/issues/21403
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/125932
Commit-Queue: Zichang Guo <zichangguo@google.com>
Reviewed-by: Siva Annamalai <asiva@google.com>
Prior to this change, connection errors such as "Connection refused"
etc, are caught either by some error condition being detected by
the eventhandler thread and that error condition being propagated
to the Dart thread, or by the first read or write after a
"successful" connection failing with an error.
However, posix documentation for non-blocking sockets recommends
either checking getsockopt(SO_ERROR), or calling connect() a
second time following the first write event on a socket after
the first connect call returns EINPROGRESS.
On Fuchsia, this check is mandatory because errors like
"Connection refused" are neither signaled on the socket's
underlying OS handle, nor indicated by error returns from
read() and write().
This change adds the check, which is optional on other platforms,
but mandatory on Fuchsia, to the Dart socket connection loop
after the first write signal is asserted.
Change-Id: I8aeea4665913f79e8138bdd830676ae43066a959
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/104263
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Zach Anderson <zra@google.com>
Change some static_libraries to source_sets to make ODR violations link-time errors.
This is needed to enable (stop suppressing) -fvisibility=hidden in Fuchsia product builds.
Change-Id: I699cec8d4b516beab9cebf9db0a522a7ff99e004
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/99822
Reviewed-by: Zach Anderson <zra@google.com>
Commit-Queue: Ryan Macnak <rmacnak@google.com>
This allows developers to have more fine grained control over socket
options supported by their platforms, particularly when there is not a
nice way to encapsulate differences between IPv4 and IPv6 options (as
with IP_MULTICAST_IF and IPV6_MULTICAST_IF). It also begins the work
of exposing socket level and option values, although keeping it for now
only to a minimum necessary to assist with setting the multicast
interface for datagram sockets.
This CL also marks `multicastInterface` as deprecated.
Bug: https://github.com/dart-lang/sdk/issues/17057
Change-Id: I39b3bf3d32d39de1c777acea4425d6eb2226355d
Reviewed-on: https://dart-review.googlesource.com/c/89164
Commit-Queue: Zach Anderson <zra@google.com>
Reviewed-by: Zach Anderson <zra@google.com>
Makes public some include paths needed for the dart_runner to depend on
dart_io.
Also updates some dart:io implementations to give and OSError
instead of crashing when something is unimplemented.
Change-Id: I862fc7cc43f56e74de791ecc021b88238b54a8e5
Reviewed-on: https://dart-review.googlesource.com/c/84920
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Zach Anderson <zra@google.com>
Read events should only be re-enabled when all available bytes have
been read, or there is an error.
related: FL-109
Change-Id: Iefd4ff3d7f4b79c08ee3d2bca7cdfef70de66ce9
Reviewed-on: https://dart-review.googlesource.com/75200
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Ben Konyi <bkonyi@google.com>
This updates the include path by which the runtime include fuchsia's
netconfig header to its current location and adds an explicit GN build
dependency on the target defining the interface so we can configure
include paths in a nicer way.
Change-Id: I2bddc452e5ded5de44c7a7087ac5457f0f07609b
Reviewed-on: https://dart-review.googlesource.com/11761
Reviewed-by: Zach Anderson <zra@google.com>
This CL rewrites the EventHandler on Fuchsia to use
the low-level ports system call API instead of the
epoll() emulation in musl. This will allow the
Magenta team to remove the epoll() emulation and
the deprecated system call API it is based on.
The new ports API doesn't provide epoll()-like
edge-triggering that clients of the EventHandler
expect. Therefore, this CL emulates it by adding
a level of indirection through a new "IOHandle"
object, which the socket code now uses instead
of file descriptors. This is similar to what we do
on Windows. See the comment at the top of
eventhandler_fuchsia.cc for more details.
Fuchsia issue US-251
R=asiva@google.com
Review-Url: https://codereview.chromium.org/2910853002 .
Revert "Pulled a significant portion of Socket implementation into BaseSocket in order to prepare for the synchronous socket impementations which will also use these common methods."
This reverts commit 0422920142.
Revert "Updated incorrect call to Write in process_win.cc"
This reverts commit 9e1435148b.
Revert "Fixed incorrect call to Write from process_win.cc"
This reverts commit 11dcab2bed3c9ec8ec76bc200206ac202a4a5681.
Fixed incorrect call to Write from process_win.cc
BUG=
Review-Url: https://codereview.chromium.org/2791163004 .
Vyacheslav Egorov