1 |
=head1 NAME |
2 |
|
3 |
Async::Interrupt - allow C/XS libraries to interrupt perl asynchronously |
4 |
|
5 |
=head1 SYNOPSIS |
6 |
|
7 |
use Async::Interrupt; |
8 |
|
9 |
=head1 DESCRIPTION |
10 |
|
11 |
This module implements a single feature only of interest to advanced perl |
12 |
modules, namely asynchronous interruptions (think "UNIX signals", which |
13 |
are very similar). |
14 |
|
15 |
Sometimes, modules wish to run code asynchronously (in another thread), |
16 |
and then signal the perl interpreter on certain events. One common way is |
17 |
to write some data to a pipe and use an event handling toolkit to watch |
18 |
for I/O events. Another way is to send a signal. Those methods are slow, |
19 |
and in the case of a pipe, also not asynchronous - it won't interrupt a |
20 |
running perl interpreter. |
21 |
|
22 |
This module implements asynchronous notifications that enable you to |
23 |
signal running perl code form another thread, asynchronously, without |
24 |
issuing syscalls. |
25 |
|
26 |
It works by creating an C<Async::Interrupt> object for each such use. This |
27 |
object stores a perl and/or a C-level callback that is invoked when the |
28 |
C<Async::Interrupt> object gets signalled. It is executed at the next time |
29 |
the perl interpreter is running (i.e. it will interrupt a computation, but |
30 |
not an XS function or a syscall). |
31 |
|
32 |
You can signal the C<Async::Interrupt> object either by calling it's C<< |
33 |
->signal >> method, or, more commonly, by calling a C function. |
34 |
|
35 |
The C<< ->signal_func >> returns the address of the C function that is to |
36 |
be called (plus an argument to be used during the call). The signalling |
37 |
function also takes an integer argument in the range SIG_ATOMIC_MIN to |
38 |
SIG_ATOMIC_MAX (guaranteed to allow at least 0..127). |
39 |
|
40 |
Since this kind of interruption is fast, but can only interrupt a |
41 |
I<running> interpreter, there is optional support for also signalling a |
42 |
pipe - that means you can also wait for the pipe to become readable (e.g. |
43 |
via L<EV> or L<AnyEvent>). This, of course, incurs the overhead of a |
44 |
C<read> and C<write> syscall. |
45 |
|
46 |
=over 4 |
47 |
|
48 |
=cut |
49 |
|
50 |
package Async::Interrupt; |
51 |
|
52 |
no warnings; |
53 |
|
54 |
BEGIN { |
55 |
$VERSION = '0.02'; |
56 |
|
57 |
require XSLoader; |
58 |
XSLoader::load Async::Interrupt::, $VERSION; |
59 |
} |
60 |
|
61 |
our $DIED = sub { warn "$@" }; |
62 |
|
63 |
=item $async = new Async::Interrupt key => value... |
64 |
|
65 |
Creates a new Async::Interrupt object. You may only use async |
66 |
notifications on this object while it exists, so you need to keep a |
67 |
reference to it at all times while it is used. |
68 |
|
69 |
Optional constructor arguments include (normally you would specify at |
70 |
least one of C<cb> or C<c_cb>). |
71 |
|
72 |
=over 4 |
73 |
|
74 |
=item cb => $coderef->($value) |
75 |
|
76 |
Registers a perl callback to be invoked whenever the async interrupt is |
77 |
signalled. |
78 |
|
79 |
Note that, since this callback can be invoked at basically any time, it |
80 |
must not modify any well-known global variables such as C<$/> without |
81 |
restoring them again before returning. |
82 |
|
83 |
The exceptions are C<$!> and C<$@>, which are saved and restored by |
84 |
Async::Interrupt. |
85 |
|
86 |
If the callback should throw an exception, then it will be caught, |
87 |
and C<$Async::Interrupt::DIED> will be called with C<$@> containing |
88 |
the exception. The default will simply C<warn> about the message and |
89 |
continue. |
90 |
|
91 |
=item c_cb => [$c_func, $c_arg] |
92 |
|
93 |
Registers a C callback the be invoked whenever the async interrupt is |
94 |
signalled. |
95 |
|
96 |
The C callback must have the following prototype: |
97 |
|
98 |
void c_func (pTHX_ void *c_arg, int value); |
99 |
|
100 |
Both C<$c_func> and C<$c_arg> must be specified as integers/IVs, and |
101 |
C<$value> is the C<value> passed to some earlier call to either C<$signal> |
102 |
or the C<signal_func> function. |
103 |
|
104 |
Note that, because the callback can be invoked at almost any time, you |
105 |
have to be careful at saving and restoring global variables that Perl |
106 |
might use (the exception is C<errno>, which is saved and restored by |
107 |
Async::Interrupt). The callback itself runs as part of the perl context, |
108 |
so you can call any perl functions and modify any perl data structures (in |
109 |
which case the requirements set out for C<cb> apply as well). |
110 |
|
111 |
=item pipe => [$fileno_or_fh_for_reading, $fileno_or_fh_for_writing] |
112 |
|
113 |
Specifies two file descriptors (or file handles) that should be signalled |
114 |
whenever the async interrupt is signalled. This means a single octet will |
115 |
be written to it, and before the callback is being invoked, it will be |
116 |
read again. Due to races, it is unlikely but possible that multiple octets |
117 |
are written. It is required that the file handles are both in nonblocking |
118 |
mode. |
119 |
|
120 |
(You can get a portable pipe and set non-blocking mode portably by using |
121 |
e.g. L<AnyEvent::Util> from the L<AnyEvent> distribution). |
122 |
|
123 |
The object will keep a reference to the file handles. |
124 |
|
125 |
This can be used to ensure that async notifications will interrupt event |
126 |
frameworks as well. |
127 |
|
128 |
=back |
129 |
|
130 |
=cut |
131 |
|
132 |
sub new { |
133 |
my ($class, %arg) = @_; |
134 |
|
135 |
bless \(_alloc $arg{cb}, @{$arg{c_cb}}[0,1], @{$arg{pipe}}[0,1]), $class |
136 |
} |
137 |
|
138 |
=item ($signal_func, $signal_arg) = $async->signal_func |
139 |
|
140 |
Returns the address of a function to call asynchronously. The function has |
141 |
the following prototype and needs to be passed the specified C<$c_arg>, |
142 |
which is a C<void *> cast to C<IV>: |
143 |
|
144 |
void (*signal_func) (void *signal_arg, int value) |
145 |
|
146 |
An example call would look like: |
147 |
|
148 |
signal_func (signal_arg, 0); |
149 |
|
150 |
The function is safe to call from within signal and thread contexts, at |
151 |
any time. The specified C<value> is passed to both C and Perl callback. |
152 |
|
153 |
C<$value> must be in the valid range for a C<sig_atomic_t> (0..127 is |
154 |
portable). |
155 |
|
156 |
If the function is called while the Async::Interrupt object is already |
157 |
signaled but before the callbacks are being executed, then the stored |
158 |
C<value> is either the old or the new one. Due to the asynchronous |
159 |
nature of the code, the C<value> can even be passed to two consecutive |
160 |
invocations of the callback. |
161 |
|
162 |
=item $async->signal ($value=0) |
163 |
|
164 |
This signals the given async object from Perl code. Semi-obviously, this |
165 |
will instantly trigger the callback invocation. |
166 |
|
167 |
C<$value> must be in the valid range for a C<sig_atomic_t> (0..127 is |
168 |
portable). |
169 |
|
170 |
=item $async->block |
171 |
|
172 |
=item $async->unblock |
173 |
|
174 |
Sometimes you need a "critical section" of code that will not be |
175 |
interrupted by an Async::Interrupt. This can be implemented by calling C<< |
176 |
$async->block >> before the critical section, and C<< $async->unblock >> |
177 |
afterwards. |
178 |
|
179 |
Note that there must be exactly one call of C<unblock> for every previous |
180 |
call to C<block> (i.e. calls can nest). |
181 |
|
182 |
Since ensuring this in the presence of exceptions and threads is |
183 |
usually more difficult than you imagine, I recommend using C<< |
184 |
$async->scoped_block >> instead. |
185 |
|
186 |
=item $async->scope_block |
187 |
|
188 |
This call C<< $async->block >> and installs a handler that is called when |
189 |
the current scope is exited (via an exception, by canceling the Coro |
190 |
thread, by calling last/goto etc.). |
191 |
|
192 |
This is the recommended (and fastest) way to implement critical sections. |
193 |
|
194 |
=cut |
195 |
|
196 |
1; |
197 |
|
198 |
=back |
199 |
|
200 |
=head1 EXAMPLE |
201 |
|
202 |
There really should be a complete C/XS example. Bug me about it. |
203 |
|
204 |
=head1 IMPLEMENTATION DETAILS AND LIMITATIONS |
205 |
|
206 |
This module works by "hijacking" SIGKILL, which is guaranteed to be always |
207 |
available in perl, but also cannot be caught, so is always available. |
208 |
|
209 |
Basically, this module fakes the receive of a SIGKILL signal and |
210 |
then catches it. This makes normal signal handling slower (probably |
211 |
unmeasurably), but has the advantage of not requiring a special runops nor |
212 |
slowing down normal perl execution a bit. |
213 |
|
214 |
It assumes that C<sig_atomic_t> and C<int> are both exception-safe to |
215 |
modify (C<sig_atomic_> is used by this module, and perl itself uses |
216 |
C<int>, so we can assume that this is quite portable, at least w.r.t. |
217 |
signals). |
218 |
|
219 |
=head1 AUTHOR |
220 |
|
221 |
Marc Lehmann <schmorp@schmorp.de> |
222 |
http://home.schmorp.de/ |
223 |
|
224 |
=cut |
225 |
|