1 # Copyright (C) all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
4 # base class for remote IPC calls and workqueues, requires Storable or Sereal
5 # - ipc_do and ipc_worker_* is for a single worker/producer and uses pipes
6 # - wq_io_do and wq_worker* is for a single producer and multiple workers,
7 # using SOCK_SEQPACKET for work distribution
8 # use ipc_do when you need work done on a certain process
9 # use wq_io_do when your work can be done on any idle worker
10 package PublicInbox::IPC;
13 use parent qw(Exporter);
15 use PublicInbox::DS qw(awaitpid);
16 use PublicInbox::Spawn;
17 use PublicInbox::OnDestroy;
18 use PublicInbox::WQWorker;
19 use Socket qw(AF_UNIX MSG_EOR SOCK_STREAM);
20 my $MY_MAX_ARG_STRLEN = 4096 * 33; # extra 4K for serialization
21 my $SEQPACKET = eval { Socket::SOCK_SEQPACKET() }; # portable enough?
22 our @EXPORT_OK = qw(ipc_freeze ipc_thaw);
24 # ->imports at BEGIN turns sereal_*_with_object into custom ops on 5.14+
25 # and eliminate method call overhead
28 require Sereal::Encoder;
29 require Sereal::Decoder;
30 Sereal::Encoder->import('sereal_encode_with_object');
31 Sereal::Decoder->import('sereal_decode_with_object');
32 ($enc, $dec) = (Sereal::Encoder->new, Sereal::Decoder->new);
36 if ($enc && $dec) { # should be custom ops
37 *ipc_freeze = sub ($) { sereal_encode_with_object $enc, $_[0] };
38 *ipc_thaw = sub ($) { sereal_decode_with_object $dec, $_[0], my $ret };
41 *ipc_freeze = \&Storable::freeze;
42 *ipc_thaw = \&Storable::thaw;
45 my $recv_cmd = PublicInbox::Spawn->can('recv_cmd4');
46 our $send_cmd = PublicInbox::Spawn->can('send_cmd4') // do {
47 require PublicInbox::CmdIPC4;
48 $recv_cmd //= PublicInbox::CmdIPC4->can('recv_cmd4');
49 PublicInbox::CmdIPC4->can('send_cmd4');
51 require PublicInbox::Syscall;
52 $recv_cmd //= PublicInbox::Syscall->can('recv_cmd4');
53 PublicInbox::Syscall->can('send_cmd4');
58 defined(my $len = <$r>) or return;
59 chop($len) eq "\n" or croak "no LF byte in $len";
60 defined(my $n = read($r, my $buf, $len)) or croak "read error: $!";
61 $n == $len or croak "short read: $n != $len";
67 my $buf = ipc_freeze($ref);
68 print $w length($buf), "\n", $buf or croak "print: $!";
71 sub ipc_return ($$$) {
72 my ($w, $ret, $exc) = @_;
73 _send_rec($w, $exc ? bless(\$exc, 'PublicInbox::IPC::Die') : $ret);
76 sub ipc_worker_loop ($$$) {
77 my ($self, $r_req, $w_res) = @_;
78 my ($rec, $wantarray, $sub, @args);
80 while ($rec = _get_rec($r_req)) {
81 ($wantarray, $sub, @args) = @$rec;
82 # no waiting if client doesn't care,
83 # this is the overwhelmingly likely case
84 if (!defined($wantarray)) {
85 eval { $self->$sub(@args) };
86 warn "$$ die: $@ (from nowait $sub)\n" if $@;
87 } elsif ($wantarray) {
88 my @ret = eval { $self->$sub(@args) };
89 ipc_return($w_res, \@ret, $@);
90 } else { # '' => wantscalar
91 my $ret = eval { $self->$sub(@args) };
92 ipc_return($w_res, \$ret, $@);
97 # starts a worker if Sereal or Storable is installed
98 sub ipc_worker_spawn {
99 my ($self, $ident, $oldset, $fields, @cb_args) = @_;
100 return if ($self->{-ipc_ppid} // -1) == $$; # idempotent
101 delete(@$self{qw(-ipc_req -ipc_res -ipc_ppid -ipc_pid)});
102 pipe(my ($r_req, $w_req)) or die "pipe: $!";
103 pipe(my ($r_res, $w_res)) or die "pipe: $!";
104 my $sigset = $oldset // PublicInbox::DS::block_signals();
105 $self->ipc_atfork_prepare;
106 my $seed = rand(0xffffffff);
107 my $pid = fork // die "fork: $!";
110 eval { Net::SSLeay::randomize() };
111 eval { PublicInbox::DS->Reset };
112 delete @$self{qw(-wq_s1 -wq_s2 -wq_workers -wq_ppid)};
113 $w_req = $r_res = undef;
114 $w_res->autoflush(1);
115 $SIG{$_} = 'IGNORE' for (qw(TERM INT QUIT));
117 # ensure we properly exit even if warn() dies:
118 my $end = PublicInbox::OnDestroy->new($$, sub { exit(!!$@) });
121 local @$self{keys %$fields} = values(%$fields);
122 my $on_destroy = $self->ipc_atfork_child;
123 local @SIG{keys %SIG} = values %SIG;
124 PublicInbox::DS::sig_setmask($sigset);
125 ipc_worker_loop($self, $r_req, $w_res);
127 warn "worker $ident PID:$$ died: $@\n" if $@;
128 undef $end; # trigger exit
130 PublicInbox::DS::sig_setmask($sigset) unless $oldset;
131 $r_req = $w_res = undef;
132 $w_req->autoflush(1);
133 $self->{-ipc_req} = $w_req;
134 $self->{-ipc_res} = $r_res;
135 $self->{-ipc_ppid} = $$;
136 awaitpid($pid, \&ipc_worker_reap, $self, @cb_args);
137 $self->{-ipc_pid} = $pid;
140 sub ipc_worker_reap { # awaitpid callback
141 my ($pid, $self, $cb, @args) = @_;
142 delete $self->{-wq_workers}->{$pid};
143 return $cb->($pid, $self, @args) if $cb;
146 # TERM(15) is our default exit signal, PIPE(13) is likely w/ pager
147 warn "$self->{-wq_ident} PID:$pid died \$?=$?\n" if $s != 15 && $s != 13
150 # for base class, override in sub classes
151 sub ipc_atfork_prepare {}
153 sub wq_atexit_child {}
155 sub ipc_atfork_child {
157 my $io = delete($self->{-ipc_atfork_child_close}) or return;
162 # idempotent, can be called regardless of whether worker is active or not
163 sub ipc_worker_stop {
165 my ($pid, $ppid) = delete(@$self{qw(-ipc_pid -ipc_ppid)});
166 my ($w_req, $r_res) = delete(@$self{qw(-ipc_req -ipc_res)});
167 if (!$w_req && !$r_res) {
168 die "unexpected PID:$pid without IPC pipes" if $pid;
171 die 'no PID with IPC pipes' unless $pid;
172 $w_req = $r_res = undef;
173 awaitpid($pid) if $$ == $ppid; # for non-event loop
176 # use this if we have multiple readers reading curl or "pigz -dc"
177 # and writing to the same store
180 $f // die 'BUG: no filename given';
181 require PublicInbox::Lock;
182 $self->{-ipc_lock} //= bless { lock_path => $f }, 'PublicInbox::Lock'
185 sub _wait_return ($$) {
186 my ($r_res, $sub) = @_;
187 my $ret = _get_rec($r_res) // die "no response on $sub";
188 die $$ret if ref($ret) eq 'PublicInbox::IPC::Die';
189 wantarray ? @$ret : $$ret;
192 # call $self->$sub(@args), on a worker if ipc_worker_spawn was used
194 my ($self, $sub, @args) = @_;
195 if (my $w_req = $self->{-ipc_req}) { # run in worker
196 my $ipc_lock = $self->{-ipc_lock};
197 my $lock = $ipc_lock ? $ipc_lock->lock_for_scope : undef;
198 if (defined(wantarray)) {
199 my $r_res = $self->{-ipc_res} or die 'no ipc_res';
200 _send_rec($w_req, [ wantarray, $sub, @args ]);
201 _wait_return($r_res, $sub);
202 } else { # likely, fire-and-forget into pipe
203 _send_rec($w_req, [ undef , $sub, @args ]);
205 } else { # run locally
210 # needed when there's multiple IPC workers and the parent forking
211 # causes newer siblings to inherit older siblings sockets
212 sub ipc_sibling_atfork_child {
214 my ($pid, undef) = delete(@$self{qw(-ipc_pid -ipc_ppid)});
215 delete(@$self{qw(-ipc_req -ipc_res)});
216 $pid == $$ and die "BUG: $$ ipc_atfork_child called on itself";
220 my ($self, $s2, $len, $full_stream) = @_;
221 my @fds = $recv_cmd->($s2, my $buf, $len // $MY_MAX_ARG_STRLEN);
222 return if scalar(@fds) && !defined($fds[0]);
223 my $n = length($buf) or return 0;
226 if (open(my $cmdfh, '+<&=', $fd)) {
227 $self->{$nfd++} = $cmdfh;
228 $cmdfh->autoflush(1);
230 die "$$ open(+<&=$fd) (FD:$nfd): $!";
233 while ($full_stream && $n < $len) {
234 my $r = sysread($s2, $buf, $len - $n, $n) // croak "read: $!";
235 croak "read EOF after $n/$len bytes" if $r == 0;
238 # Sereal dies on truncated data, Storable returns undef
239 my $args = ipc_thaw($buf) // die "thaw error on buffer of size: $n";
241 my $sub = shift @$args;
242 eval { $self->$sub(@$args) };
243 warn "$$ $0 wq_worker: $sub: $@" if $@;
244 delete @$self{0..($nfd-1)};
248 sub wq_worker_loop ($$) {
249 my ($self, $bcast2) = @_;
250 my $wqw = PublicInbox::WQWorker->new($self, $self->{-wq_s2});
251 PublicInbox::WQWorker->new($self, $bcast2) if $bcast2;
252 PublicInbox::DS->SetPostLoopCallback(sub { $wqw->{sock} });
253 PublicInbox::DS::event_loop();
254 PublicInbox::DS->Reset;
257 sub do_sock_stream { # via wq_io_do, for big requests
258 my ($self, $len) = @_;
259 recv_and_run($self, my $s2 = delete $self->{0}, $len, 1);
263 my ($self, $sub, @args) = @_;
264 if (my $wkr = $self->{-wq_workers}) {
265 my $buf = ipc_freeze([$sub, @args]);
266 for my $bcast1 (values %$wkr) {
267 my $sock = $bcast1 // $self->{-wq_s1} // next;
268 send($sock, $buf, MSG_EOR) // croak "send: $!";
269 # XXX shouldn't have to deal with EMSGSIZE here...
272 eval { $self->$sub(@args) };
273 warn "wq_broadcast: $@" if $@;
277 sub stream_in_full ($$$) {
278 my ($s1, $fds, $buf) = @_;
279 socketpair(my $r, my $w, AF_UNIX, SOCK_STREAM, 0) or
280 croak "socketpair: $!";
281 my $n = $send_cmd->($s1, [ fileno($r) ],
282 ipc_freeze(['do_sock_stream', length($buf)]),
283 MSG_EOR) // croak "sendmsg: $!";
285 $n = $send_cmd->($w, $fds, $buf, 0) // croak "sendmsg: $!";
286 while ($n < length($buf)) {
287 my $x = syswrite($w, $buf, length($buf) - $n, $n) //
288 croak "syswrite: $!";
289 croak "syswrite wrote 0 bytes" if $x == 0;
294 sub wq_io_do { # always async
295 my ($self, $sub, $ios, @args) = @_;
296 if (my $s1 = $self->{-wq_s1}) { # run in worker
297 my $fds = [ map { fileno($_) } @$ios ];
298 my $buf = ipc_freeze([$sub, @args]);
299 if (length($buf) > $MY_MAX_ARG_STRLEN) {
300 stream_in_full($s1, $fds, $buf);
302 my $n = $send_cmd->($s1, $fds, $buf, MSG_EOR);
303 return if defined($n); # likely
305 croak "sendmsg: $! (check RLIMIT_NOFILE)";
306 $!{EMSGSIZE} ? stream_in_full($s1, $fds, $buf) :
307 croak("sendmsg: $!");
310 @$self{0..$#$ios} = @$ios;
311 eval { $self->$sub(@args) };
312 warn "wq_io_do: $@" if $@;
313 delete @$self{0..$#$ios}; # don't close
318 my ($self, $wantarray, $sub, @args) = @_;
320 my @ret = eval { $self->$sub(@args) };
321 ipc_return($self->{0}, \@ret, $@);
322 } else { # '' => wantscalar
323 my $ret = eval { $self->$sub(@args) };
324 ipc_return($self->{0}, \$ret, $@);
329 my ($self, $sub, @args) = @_;
330 if (defined(wantarray)) {
331 pipe(my ($r, $w)) or die "pipe: $!";
332 wq_io_do($self, 'wq_sync_run', [ $w ], wantarray, $sub, @args);
334 _wait_return($r, $sub);
336 wq_io_do($self, $sub, [], @args);
340 sub prepare_nonblock {
341 ($_[0]->{-wq_s1} // die 'BUG: no {-wq_s1}')->blocking(0);
342 require PublicInbox::WQBlocked;
345 sub wq_nonblock_do { # always async
346 my ($self, $sub, @args) = @_;
347 my $buf = ipc_freeze([$sub, @args]);
348 if ($self->{wqb}) { # saturated once, assume saturated forever
349 $self->{wqb}->flush_send($buf);
351 $send_cmd->($self->{-wq_s1}, [], $buf, MSG_EOR) //
352 ($!{EAGAIN} ? PublicInbox::WQBlocked->new($self, $buf)
353 : croak("sendmsg: $!"));
357 sub _wq_worker_start {
358 my ($self, $oldset, $fields, $one, @cb_args) = @_;
359 my ($bcast1, $bcast2);
360 $one or socketpair($bcast1, $bcast2, AF_UNIX, $SEQPACKET, 0) or
361 die "socketpair: $!";
362 my $seed = rand(0xffffffff);
363 my $pid = fork // die "fork: $!";
366 eval { Net::SSLeay::randomize() };
368 eval { PublicInbox::DS->Reset };
369 delete @$self{qw(-wq_s1 -wq_ppid)};
370 $self->{-wq_worker_nr} =
371 keys %{delete($self->{-wq_workers}) // {}};
372 $SIG{$_} = 'DEFAULT' for (qw(TTOU TTIN TERM QUIT INT CHLD));
373 local $0 = $one ? $self->{-wq_ident} :
374 "$self->{-wq_ident} $self->{-wq_worker_nr}";
375 # ensure we properly exit even if warn() dies:
376 my $end = PublicInbox::OnDestroy->new($$, sub { exit(!!$@) });
379 local @$self{keys %$fields} = values(%$fields);
380 my $on_destroy = $self->ipc_atfork_child;
381 local @SIG{keys %SIG} = values %SIG;
382 PublicInbox::DS::sig_setmask($oldset);
383 wq_worker_loop($self, $bcast2);
385 warn "worker $self->{-wq_ident} PID:$$ died: $@" if $@;
386 undef $end; # trigger exit
388 $self->{-wq_workers}->{$pid} = $bcast1;
389 awaitpid($pid, \&ipc_worker_reap, $self, @cb_args);
393 # starts workqueue workers if Sereal or Storable is installed
394 sub wq_workers_start {
395 my ($self, $ident, $nr_workers, $oldset, $fields, @cb_args) = @_;
396 ($send_cmd && $recv_cmd && defined($SEQPACKET)) or return;
397 return if $self->{-wq_s1}; # idempotent
398 $self->{-wq_s1} = $self->{-wq_s2} = undef;
399 socketpair($self->{-wq_s1}, $self->{-wq_s2}, AF_UNIX, $SEQPACKET, 0) or
400 die "socketpair: $!";
401 $self->ipc_atfork_prepare;
402 $nr_workers //= $self->{-wq_nr_workers}; # was set earlier
403 my $sigset = $oldset // PublicInbox::DS::block_signals();
404 $self->{-wq_workers} = {};
405 $self->{-wq_ident} = $ident;
406 my $one = $nr_workers == 1;
407 $self->{-wq_nr_workers} = $nr_workers;
408 for (1..$nr_workers) {
409 _wq_worker_start($self, $sigset, $fields, $one, @cb_args);
411 PublicInbox::DS::sig_setmask($sigset) unless $oldset;
412 $self->{-wq_ppid} = $$;
417 if (my $wqb = delete $self->{wqb}) {
420 delete @$self{qw(-wq_s1 -wq_s2)} or return;
421 return if ($self->{-wq_ppid} // -1) != $$;
422 awaitpid($_) for keys %{$self->{-wq_workers}};
426 my ($self, $sig) = @_;
427 kill($sig // 'TERM', keys %{$self->{-wq_workers}});
432 my $ppid = $self->{-wq_ppid};
433 wq_kill($self) if $ppid && $ppid == $$;
435 ipc_worker_stop($self);
438 sub detect_nproc () {
439 # _SC_NPROCESSORS_ONLN = 84 on both Linux glibc and musl
440 return POSIX::sysconf(84) if $^O eq 'linux';
441 return POSIX::sysconf(58) if $^O eq 'freebsd';
444 # getconf(1) is POSIX, but *NPROCESSORS* vars are not
445 for (qw(_NPROCESSORS_ONLN NPROCESSORS_ONLN)) {
446 `getconf $_ 2>/dev/null` =~ /^(\d+)$/ and return $1;
448 for my $nproc (qw(nproc gnproc)) { # GNU coreutils nproc
449 `$nproc 2>/dev/null` =~ /^(\d+)$/ and return $1;
452 # should we bother with `sysctl hw.ncpu`? Those only give
453 # us total processor count, not online processor count.