1 # Copyright (C) 2020-2021 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 package PublicInbox::IPC;
8 use Carp qw(confess croak);
9 use PublicInbox::DS qw(dwaitpid);
10 use PublicInbox::Spawn;
12 use Socket qw(AF_UNIX MSG_EOR);
13 my $SEQPACKET = eval { Socket::SOCK_SEQPACKET() }; # portable enough?
14 use constant PIPE_BUF => $^O eq 'linux' ? 4096 : POSIX::_POSIX_PIPE_BUF();
16 # ->imports at BEGIN turns sereal_*_with_object into custom ops on 5.14+
17 # and eliminate method call overhead
20 require Sereal::Encoder;
21 require Sereal::Decoder;
22 Sereal::Encoder->import('sereal_encode_with_object');
23 Sereal::Decoder->import('sereal_decode_with_object');
24 ($enc, $dec) = (Sereal::Encoder->new, Sereal::Decoder->new);
28 if ($enc && $dec) { # should be custom ops
29 *freeze = sub ($) { sereal_encode_with_object $enc, $_[0] };
30 *thaw = sub ($) { sereal_decode_with_object $dec, $_[0], my $ret };
32 eval { # some distros have Storable as a separate package from Perl
34 Storable->import(qw(freeze thaw));
36 } // warn("Storable (part of Perl) missing: $@\n");
39 my $recv_cmd = PublicInbox::Spawn->can('recv_cmd4');
40 my $send_cmd = PublicInbox::Spawn->can('send_cmd4') // do {
41 require PublicInbox::CmdIPC4;
42 $recv_cmd //= PublicInbox::CmdIPC4->can('recv_cmd4');
43 PublicInbox::CmdIPC4->can('send_cmd4');
45 require PublicInbox::CmdIPC1;
46 $recv_cmd = PublicInbox::CmdIPC1->can('recv_cmd1');
47 PublicInbox::CmdIPC1->can('send_cmd1');
52 defined(my $len = <$r>) or return;
53 chop($len) eq "\n" or croak "no LF byte in $len";
54 defined(my $n = read($r, my $buf, $len)) or croak "read error: $!";
55 $n == $len or croak "short read: $n != $len";
61 my $buf = freeze($ref);
62 length($buf) . "\n" . $buf;
67 print $w _pack_rec($ref) or croak "print: $!";
70 sub ipc_return ($$$) {
71 my ($w, $ret, $exc) = @_;
72 _send_rec($w, $exc ? bless(\$exc, 'PublicInbox::IPC::Die') : $ret);
75 sub ipc_worker_loop ($$$) {
76 my ($self, $r_req, $w_res) = @_;
77 my ($rec, $wantarray, $sub, @args);
79 while ($rec = _get_rec($r_req)) {
80 ($wantarray, $sub, @args) = @$rec;
81 # no waiting if client doesn't care,
82 # this is the overwhelmingly likely case
83 if (!defined($wantarray)) {
84 eval { $self->$sub(@args) };
85 warn "$$ die: $@ (from nowait $sub)\n" if $@;
86 } elsif ($wantarray) {
87 my @ret = eval { $self->$sub(@args) };
88 ipc_return($w_res, \@ret, $@);
89 } else { # '' => wantscalar
90 my $ret = eval { $self->$sub(@args) };
91 ipc_return($w_res, \$ret, $@);
96 # starts a worker if Sereal or Storable is installed
97 sub ipc_worker_spawn {
98 my ($self, $ident, $oldset) = @_;
99 return unless $enc; # no Sereal or Storable
100 return if ($self->{-ipc_ppid} // -1) == $$; # 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();
106 $self->ipc_atfork_parent;
107 defined(my $pid = fork) or die "fork: $!";
109 eval { PublicInbox::DS->Reset };
110 $self->{-ipc_parent_pid} = $parent;
111 $w_req = $r_res = undef;
112 $w_res->autoflush(1);
113 $SIG{$_} = 'IGNORE' for (qw(TERM INT QUIT));
115 PublicInbox::DS::sig_setmask($oldset);
116 my $on_destroy = $self->ipc_atfork_child;
117 eval { ipc_worker_loop($self, $r_req, $w_res) };
118 die "worker $ident PID:$$ died: $@\n" if $@;
121 PublicInbox::DS::sig_setmask($sigset) unless $oldset;
122 $r_req = $w_res = undef;
123 $w_req->autoflush(1);
124 $self->{-ipc_req} = $w_req;
125 $self->{-ipc_res} = $r_res;
126 $self->{-ipc_ppid} = $$;
127 $self->{-ipc_pid} = $pid;
130 sub ipc_worker_reap { # dwaitpid callback
131 my ($self, $pid) = @_;
132 warn "PID:$pid died with \$?=$?\n" if $?;
135 # for base class, override in sub classes
136 sub ipc_atfork_parent {}
137 sub ipc_atfork_child {}
139 # should only be called inside the worker process
140 sub ipc_worker_exit {
141 my (undef, $code) = @_;
145 # idempotent, can be called regardless of whether worker is active or not
146 sub ipc_worker_stop {
148 my ($pid, $ppid) = delete(@$self{qw(-ipc_pid -ipc_ppid)});
149 my ($w_req, $r_res) = delete(@$self{qw(-ipc_req -ipc_res)});
150 if (!$w_req && !$r_res) {
151 die "unexpected PID:$pid without IPC pipes" if $pid;
154 die 'no PID with IPC pipes' unless $pid;
155 _send_rec($w_req, [ undef, 'ipc_worker_exit', 0 ]);
156 $w_req = $r_res = undef;
158 # allow any sibling to send ipc_worker_exit, but siblings can't wait
159 return if $$ != $ppid;
160 dwaitpid($pid, \&ipc_worker_reap, $self);
163 # use this if we have multiple readers reading curl or "pigz -dc"
164 # and writing to the same store
167 require PublicInbox::Lock;
168 $self->{-ipc_lock} //= bless { lock_path => $f }, 'PublicInbox::Lock'
171 sub ipc_async_wait ($$) {
172 my ($self, $max) = @_; # max == -1 to wait for all
173 my $aif = $self->{-async_inflight} or return;
174 my $r_res = $self->{-ipc_res} or die 'BUG: no ipc_res';
175 while (my ($sub, $bytes, $cb, $cb_arg) = splice(@$aif, 0, 4)) {
176 my $ret = _get_rec($r_res) //
177 die "no response on $sub (req.size=$bytes)";
178 $self->{-async_inflight_bytes} -= $bytes;
180 eval { $cb->($cb_arg, $ret) };
181 warn "E: $sub callback error: $@\n" if $@;
182 return if --$max == 0;
186 # call $self->$sub(@args), on a worker if ipc_worker_spawn was used
188 my ($self, $sub, @args) = @_;
189 if (my $w_req = $self->{-ipc_req}) { # run in worker
190 my $ipc_lock = $self->{-ipc_lock};
191 my $lock = $ipc_lock ? $ipc_lock->lock_for_scope : undef;
192 if (defined(wantarray)) {
193 my $r_res = $self->{-ipc_res} or die 'BUG: no ipc_res';
194 ipc_async_wait($self, -1);
195 _send_rec($w_req, [ wantarray, $sub, @args ]);
196 my $ret = _get_rec($r_res) // die "no response on $sub";
197 die $$ret if ref($ret) eq 'PublicInbox::IPC::Die';
198 wantarray ? @$ret : $$ret;
199 } else { # likely, fire-and-forget into pipe
200 _send_rec($w_req, [ undef , $sub, @args ]);
202 } else { # run locally
208 my ($self, $sub, $sub_args, $cb, $cb_arg) = @_;
209 if (my $w_req = $self->{-ipc_req}) { # run in worker
210 my $rec = _pack_rec([ 1, $sub, @$sub_args ]);
211 my $cur_bytes = \($self->{-async_inflight_bytes} //= 0);
212 while (($$cur_bytes + length($rec)) > PIPE_BUF) {
213 ipc_async_wait($self, 1);
215 my $ipc_lock = $self->{-ipc_lock};
216 my $lock = $ipc_lock ? $ipc_lock->lock_for_scope : undef;
217 print $w_req $rec or croak "print: $!";
218 $$cur_bytes += length($rec);
219 push @{$self->{-async_inflight}},
220 $sub, length($rec), $cb, $cb_arg;
222 my $ret = [ eval { $self->$sub(@$sub_args) } ];
224 $ret = ( bless(\$exc, 'PublicInbox::IPC::Die') );
226 eval { $cb->($cb_arg, $ret) };
227 warn "E: $sub callback error: $@\n" if $@;
231 # needed when there's multiple IPC workers and the parent forking
232 # causes newer siblings to inherit older siblings sockets
233 sub ipc_sibling_atfork_child {
235 my ($pid, undef) = delete(@$self{qw(-ipc_pid -ipc_ppid)});
236 delete(@$self{qw(-ipc_req -ipc_res)});
237 $pid == $$ and die "BUG: $$ ipc_atfork_child called on itself";
240 sub wq_worker_loop ($$) {
241 my ($self, $s2) = @_;
243 my $len = $self->{wq_req_len} // (4096 * 33);
244 my ($rec, $sub, @args);
246 my (@fds) = $recv_cmd->($s2, $buf, $len) or return; # EOF
248 my @m = @{$self->{wq_open_modes} // [qw( +<&= >&= >&= )]};
250 my $mode = shift(@m);
251 if (open(my $fh, $mode, $fd)) {
254 die "$$ open($mode$fd) (FD:$i): $!";
257 # Sereal dies, Storable returns undef
259 die "thaw error on buffer of size:".length($buf);
260 ($sub, @args) = @$rec;
261 eval { $self->$sub(@args) };
262 warn "$$ wq_worker: $@" if $@;
263 delete @$self{0, 1, 2};
267 sub wq_do { # always async
268 my ($self, $sub, $in, $out, $err, @args) = @_;
269 if (my $s1 = $self->{-wq_seq}) { # run in worker
270 $_ = fileno($_) for ($in, $out, $err);
271 $send_cmd->($s1, $in, $out, $err,
272 freeze([$sub, @args]), MSG_EOR);
274 @$self{0, 1, 2} = ($in, $out, $err);
275 eval { $self->$sub(@args) };
276 warn "wq_do: $@" if $@;
277 delete @$self{0, 1, 2};
281 # starts workqueue workers if Sereal or Storable is installed
282 sub wq_workers_start {
283 my ($self, $ident, $nr_workers, $oldset) = @_;
284 ($enc && $send_cmd && $recv_cmd && defined($SEQPACKET)) or return;
285 return if $self->{-wq_seq}; # idempotent
287 socketpair($s1, $s2, AF_UNIX, $SEQPACKET, 0) or die "socketpair: $!";
288 my $sigset = $oldset // PublicInbox::DS::block_signals();
289 $self->ipc_atfork_parent;
291 $self->{-wq_workers} = {};
292 for my $i (0..($nr_workers - 1)) {
293 defined(my $pid = fork) or die "fork: $!";
295 eval { PublicInbox::DS->Reset };
297 $SIG{$_} = 'IGNORE' for (qw(TERM INT QUIT));
298 local $0 = $ident."[$i]";
299 PublicInbox::DS::sig_setmask($oldset);
300 my $on_destroy = $self->ipc_atfork_child;
301 eval { wq_worker_loop($self, $s2) };
302 die "worker $ident PID:$$ died: $@\n" if $@;
305 $self->{-wq_workers}->{$pid} = $i;
308 PublicInbox::DS::sig_setmask($sigset) unless $oldset;
310 $self->{-wq_seq} = $s1;
311 $self->{-wq_ppid} = $$;
316 delete $self->{-wq_seq} or return;
317 my $ppid = delete $self->{-wq_ppid} // die 'BUG: no wq_ppid';
318 my $workers = delete $self->{-wq_workers} // die 'BUG: no wq_workers';
319 return if $ppid != $$; # can't reap siblings or parents
320 for my $pid (keys %$workers) {
321 dwaitpid($pid, \&ipc_worker_reap, $self);