Sophie

Sophie

distrib > Mandriva > 2010.0 > i586 > media > contrib-release > by-pkgid > cd14cddf3b3ceaf1193157472227757a > files > 46

parrot-doc-1.6.0-1mdv2010.0.i586.rpm

# Copyright (C) 2001-2006, Parrot Foundation.
# $Id: events.pod 39944 2009-07-08 02:48:40Z coke $

=head1 NAME

docs/dev/events.pod - Design Notes for Events

=head1 VERSION

This document describes the current state, which might not be the final
implementation.

=head1 Overview

Parrot has to deal with asynchronous events (from timers, signals, async IO,
notifications, and so on). This document describes the current implementation.

=head1 Prelims

As there is currently no good test if a threading library is included at link
time, its assumed, that platforms having B<PARROT_HAS_HEADER_PTHREAD> link
against B<libpthread>.

=head1 DESCRIPTION

On construction of the first interpreter (the one with no
B<parent_interpreter>) two threads are started: The B<event_thread>, which
manages the static global B<event_queue> and the B<io_thread> which is
responsible for signal and IO related events.

=head2 Events

Events can be either timed (they are due after some elapsed time) or untimed.
For the former there is one API call: B<Parrot_new_timer_event>

=head2 The B<event_thread>

The B<event_thread> holds the B<event_queue> mutex first. When there is no
event entry in the B<event_queue>, the B<event_thread> waits on the event
condition until an event arrives. When there is an event with a timed entry, a
timed wait is performed. (Waiting on the condition releases the mutex, so that
other threads can insert events into the B<event_queue>.)

When an event arrives (or the timeout was reached) the B<event_thread> pops off
all events and places the queue entries into the interpreter's B<task_queue>.
This also enables event checking in the interpreter's run-core.

When the popped off entry is a timed event and has a repeat interval, the entry
is duplicated and reinserted with the interval added to the current time.

=head2 B<Signals>

All signals that should be handled inside Parrot are blocked in all threads and
only enabled in the B<io_thread>. The signal handler functions just sets an
atomic flag, that this signal arrived and returns. This finally interrupts the
select(2) loop in the B<io_thread>.

=head2 The B<io_thread>

The B<io_thread> sleeps in a select(2) loop, which is interrupted when either a
signal arrives or when one of the file descriptors has a ready condition.
Additionally the file descriptor set contains the reader end of an internal
pipe, which is used by other threads to communicate with the B<io_thread>.

Signal events like SIGINT are broadcasted to all running interpreters, which
then throw an appropriate exception.

=head2 The interpreter event checking code

We cannot interrupt the interpreter at arbitrary points and run some different
code (e.g. a PASM subroutine handling timer events). So when an event is put
into the interpreter's B<task_queue> the opcode dispatch table for the
interpreter is changed.

Plain function cores get a function table with all entries filled with the
B<check_events__> opcode. This opcode pops off and finally handles the event.
The same scheme works for the CGOTO core, where the address table is replaced.
The switched core does an explicit check if events are to be handled.

Prederefed and especially the CGP core don't have an opcode dispatch table that
is checked during running the opcodes. When an event is scheduled, the event
handler replaces backward branches in the opcode image with the
B<check_events__> opcode.

The JIT core doesn't handle events yet.

After all events are popped off and handled, the opcode dispatch table is
restored to its original, and the B<check_events__> reexecutes the same
instruction again, which is now the real one and thus normal execution flow
continues.

This scheme has zero overhead in the absence of scheduled events for all cores
except switched and JIT.

=head1 Missing

=over 4

=item Synchronous event API

Sync events could be placed directly into the interpreter's task queue.

=item Async IO

That depends probably on the underlying OS, i.e. if it does async IO or we have
to do it.

=item Event priorities

=item A lot more

=back

=head1 Author

Leopold Toetsch C<lt@toetsch.at>

=cut

# vim: expandtab shiftwidth=2 tw=70: