async: scope: add completion sender

Add a mechanism to await the completion of an async::scope,
equivalent to a 'join' on the spawned tasks it contains.

Signed-off-by: Patrick Williams <patrick@stw

async: scope: add completion sender

Add a mechanism to await the completion of an async::scope,
equivalent to a 'join' on the spawned tasks it contains.

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I669680441704cdb59aa982d97bed610178c498b9

show more ...

async: add scope

In order to be able to create multiple subtasks as a collection
and in order to make sure we keep good track of any tasks that
are created by the context, add a `scope` as a contain

async: add scope

In order to be able to create multiple subtasks as a collection
and in order to make sure we keep good track of any tasks that
are created by the context, add a `scope` as a container of subtasks.
This is partially modelled after the C++ P2519 proposal.

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I6e99b2fa2829d80c320491991f7e038f122963a9

show more ...

async: add sleep_for sender

Sometimes it is useful to do the equivalent of
`std::this_thread::sleep_for` in a co-routine context. Add a
sender-based implementation to async.

Signed-off-by: Patrick

async: add sleep_for sender

Sometimes it is useful to do the equivalent of
`std::this_thread::sleep_for` in a co-routine context. Add a
sender-based implementation to async.

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I9991eb40b7a1b12e61511f1200bc99fdcdbccf0a

show more ...

event: add oneshot timer support

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I9785fa52e38202821258ad162e6f9168c71e53f7

async: context: support stopping the context

Up until this point, the async::context would run forever and have
no way of exiting the process, which isn't very useful. Add some
support for stop con

async: context: support stopping the context

Up until this point, the async::context would run forever and have
no way of exiting the process, which isn't very useful. Add some
support for stop conditions so that we can ask the context to stop
and cleanly exit.

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: Ia5967162fb103d4b4b0490d8cbbef12bdb361bac

show more ...

event: add a simple wrapper around sd-event

In order to facilitate stop-conditions on the async::context,
we need to be able to escape the wait on the dbus fd. sd-event
will allow us to do this and

event: add a simple wrapper around sd-event

In order to facilitate stop-conditions on the async::context,
we need to be able to escape the wait on the dbus fd. sd-event
will allow us to do this and also build other co-routine primitives
in the future (such as `co_await async::delay(1s)` backed by sd-event
timers). Define a simple wrapper around sd-event here which can
be leveraged by the async framework.

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I945416cbce3c98d98135c42cb7ca249a57cb9a60

show more ...

async: add match

Define a class which will register for a dbus signal match
and generate Senders that await new signals.

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: Ib1d77708d74e

async: add match

Define a class which will register for a dbus signal match
and generate Senders that await new signals.

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: Ib1d77708d74e6063bcaa08fc76fb98667ee7a2cb

show more ...

async: add client-object proxy

Add a class which acts as a client-side proxy to a object. The
proxy holds the object address (service, path, interface) and
creates Senders for standard dbus operati

async: add client-object proxy

Add a class which acts as a client-side proxy to a object. The
proxy holds the object address (service, path, interface) and
creates Senders for standard dbus operations: method-call,
get-property, set-property, get-all-properties. These Senders
can also be used in co-routine contexts (via co_await).

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I4b2e50a8b55d10975984ac4244bb075db8746fbf

show more ...

async: add context support

In order to handle async operations, we need some kind of
'run-loop' around the sd_bus_wait/sd_bus_process process.
Define a 'context' to hold the bus and manage the run-l

async: add context support

In order to handle async operations, we need some kind of
'run-loop' around the sd_bus_wait/sd_bus_process process.
Define a 'context' to hold the bus and manage the run-loop.

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I16099dd2e15809776827dd85327d950af3f8f5b2

show more ...

async: add coroutine task support

Add sdbusplus::async::task<...> template which works with the
std::executors proposal (P2300) and test-cases.

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
C

async: add coroutine task support

Add sdbusplus::async::task<...> template which works with the
std::executors proposal (P2300) and test-cases.

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: Ie63791daeab90ae1cc3862bb30878531b1775fa7

show more ...

std::execution: import from p2300 prototype

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I47725915b544af2b71ce0f423f70f453f8bcbc39

Allow clients to disable coroutine support

This commit allows clients to opt out of the boost coroutines in
asio::connection. This was added later after the initial development of
the asio bindings

Allow clients to disable coroutine support

This commit allows clients to opt out of the boost coroutines in
asio::connection. This was added later after the initial development of
the asio bindings, and imposes unfortunate handler requirements upon
clients, namely, that boost coroutines will not compile when exceptions
have been disabled via -fno-exceptions. For clients wishing to disable
exceptions, this is unfortunate, and prevents binary size optimizations
like:
https://gerrit.openbmc-project.xyz/c/openbmc/bmcweb/+/46841

Therefore, this patchset allows clients to define
SDBUSPLUS_DISABLE_BOOST_COROUTINES which prevents
sdbusplus::asio::connection from pulling in boost/asio/spawn.hpp, and
disallows the use of yield_method_call (which requires a yield context).

Tested:
Compiled with
https://gerrit.openbmc-project.xyz/c/openbmc/bmcweb/+/46841 and observed
a good build. See the aforementioned patchset for why this is useful,
but TL;DR, it reduced the binary size of bmcweb by ~30% for no lost
user-facing functionality.

Signed-off-by: Ed Tanous <edtanous@google.com>
Change-Id: I5339c3ff288b333fb0466fb2e6ac00f5b65964b4

show more ...

clang-format: update with latest

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I19356ecda2940a32661f54c0b54c74acc891f9c4

asio/object_server: Use nullptr

Signed-off-by: Brad Bishop <bradleyb@fuzziesquirrel.com>
Change-Id: Ia38e1b7509c65fe065e1bbd1fa5f04da6355ed46

Fix instances of extra-semi

Signed-off-by: Brad Bishop <bradleyb@fuzziesquirrel.com>
Change-Id: I233bd736375e8fce3646232dd6d101989d941f4d

clang-format: update whitespace penalty

clang-format-15 seems to end up with a slightly different whitespace
break than clang-format-14 did. Tweak the PenaltyBreakAssignment
to favor not breaking a

clang-format: update whitespace penalty

clang-format-15 seems to end up with a slightly different whitespace
break than clang-format-14 did. Tweak the PenaltyBreakAssignment
to favor not breaking as often, which allows the two to be more
consistent.

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I49826e2011bce317b6b82dc985ddd35fa6eef2cb

show more ...

asio: fix instances of unused-exception-parameter

Signed-off-by: Brad Bishop <bradleyb@fuzziesquirrel.com>
Change-Id: Ibbcea5805e4916820f797988b9b95a92e6fe2823

Fix inconsistent-missing-destructor-override

Signed-off-by: Brad Bishop <bradleyb@fuzziesquirrel.com>
Change-Id: I1fcfcaf4b41319d7a0b11fb937df06bec41d7620

bus.hpp: fix missing-noreturn

Signed-off-by: Brad Bishop <bradleyb@fuzziesquirrel.com>
Change-Id: I734ec4f690c5a162e55ab1b69ac6b6139d942781

Fix instances of zero-as-null-pointer-constant

Signed-off-by: Brad Bishop <bradleyb@fuzziesquirrel.com>
Change-Id: I87e5e39202520bd7b6066fc6444362f32659049f

match: more string_view improvements

- It has been suggested that `string_view` is better than `const
string_view&`[1]. Switch all calls to that type.

- In some cases the optimizations usi

match: more string_view improvements

- It has been suggested that `string_view` is better than `const
string_view&`[1]. Switch all calls to that type.

- In some cases the optimizations using string_view allowed
non-zstrings to be passed to sdbus calls without compiler warning.
Revert back to sacrificing the slight performance of constructing
the string for safety.

[1]: https://quuxplusone.github.io/blog/2021/11/09/pass-string-view-by-value/

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I21a542fc9264b0f998e1c234064b6e6414327b82

show more ...

match: constexpr, noexcept, string_view-ify

Optimize the bus::match interfaces:

Add constexpr and noexcept markers on all functions which can be
done so, which potentially makes the resulti

match: constexpr, noexcept, string_view-ify

Optimize the bus::match interfaces:

Add constexpr and noexcept markers on all functions which can be
done so, which potentially makes the resulting binaries smaller.

Switch parameters and return values to string_view whenever
possible, which also makes resulting binaries smaller.

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I0a863b8a16349123d42011e08ab0d0464ea11548

show more ...

.gitignore: remove vim and update python

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: Idfffe485440ced20dd1a772e0cf2a4334da06f7a

Combine owned strings

There are two lists in object_server for storing strings that need their
lifetime extended for both methods and signal names. Neither list is
used independently, and exists en

Combine owned strings

There are two lists in object_server for storing strings that need their
lifetime extended for both methods and signal names. Neither list is
used independently, and exists entirely just to store strings to give
the vtables something to point to, so there's no sense in keeping two
lists.

As some point, these could arguably become static sets, so that strings
memory footprints are deduplicated between interfaces in the same
daemon, but that is a more invasive change.

Signed-off-by: Ed Tanous <edtanous@google.com>
Change-Id: Ie676a4ffb524387abf12a0ed20f50d109c85bc58

show more ...

Use built-in sd-bus methods for member name check

sd-bus has built-in methods for verifying member names that do a better
job checking than the method that was created here. Use them.

Signed-off-b

Use built-in sd-bus methods for member name check

sd-bus has built-in methods for verifying member names that do a better
job checking than the method that was created here. Use them.

Signed-off-by: Ed Tanous <edtanous@google.com>
Change-Id: I5b5389abf7f3033f3f4de344118bf9049d81be66

show more ...