http: clarify intent for persistence

[public-inbox.git] / lib / PublicInbox / HTTP.pm

diff --git a/lib/PublicInbox/HTTP.pm b/lib/PublicInbox/HTTP.pm
index 6c4c20d7cf801663af01beb71d3676bc941c6479..393451221112478d832c82142d37e7c3faf57757 100644 (file)
--- a/lib/PublicInbox/HTTP.pm
+++ b/lib/PublicInbox/HTTP.pm
@@ -13,7 +13,7 @@ use warnings;
 use base qw(Danga::Socket);
 use fields qw(httpd env rbuf input_left remote_addr remote_port);
 use Fcntl qw(:seek);
-use HTTP::Parser::XS qw(parse_http_request); # supports pure Perl fallback
+use Plack::HTTPParser qw(parse_http_request); # XS or pure Perl
 use HTTP::Status qw(status_message);
 use HTTP::Date qw(time2str);
 use IO::File;
@@ -24,6 +24,12 @@ use constant {
        CHUNK_MAX_HDR => 256,
 };
 
+# Use the same configuration parameter as git since this is primarily
+# a slow-client sponge for git-http-backend
+# TODO: support per-respository http.maxRequestBuffer somehow...
+our $MAX_REQUEST_BUFFER = $ENV{GIT_HTTP_MAX_REQUEST_BUFFER} ||
+                       (10 * 1024 * 1024);
+
 my $null_io = IO::File->new('/dev/null', '<');
 my $http_date;
 my $prev = 0;
@@ -70,7 +76,11 @@ sub rbuf_process {
 
        # We do not support Trailers in chunked requests, for now
        # (they are rarely-used and git (as of 2.7.2) does not use them)
-       return quit($self, 400) if $r == -1 || $env{HTTP_TRAILER};
+       if ($r == -1 || $env{HTTP_TRAILER} ||
+                       # this length-check is necessary for PURE_PERL=1:
+                       ($r == -2 && length($self->{rbuf}) > 0x4000)) {
+               return quit($self, 400);
+       }
        return $self->watch_read(1) if $r < 0; # incomplete
        $self->{rbuf} = substr($self->{rbuf}, $r);
 
@@ -141,7 +151,7 @@ sub response_header_write {
        my $proto = $env->{SERVER_PROTOCOL} or return; # HTTP/0.9 :P
        my $status = $res->[0];
        my $h = "$proto $status " . status_message($status) . "\r\n";
-       my ($len, $chunked);
+       my $term;
        my $headers = $res->[1];
 
        for (my $i = 0; $i < @$headers; $i += 2) {
@@ -149,33 +159,32 @@ sub response_header_write {
                my $v = $headers->[$i + 1];
                next if $k =~ /\A(?:Connection|Date)\z/i;
 
-               $len = $v if $k =~ /\AContent-Length\z/i;
-               if ($k =~ /\ATransfer-Encoding\z/i && $v =~ /\bchunked\b/i) {
-                       $chunked = 1;
+               if ($k =~ /\AContent-Length\z/ ||
+                   ($k =~ /\ATransfer-Encoding\z/i && $v =~ /\bchunked\b/i)) {
+                       $term = 1;
                }
-
                $h .= "$k: $v\r\n";
        }
 
        my $conn = $env->{HTTP_CONNECTION} || '';
-       my $alive = (defined($len) || $chunked) &&
-                       ($proto eq 'HTTP/1.1' && $conn !~ /\bclose\b/i) ||
-                       ($conn =~ /\bkeep-alive\b/i);
+       my $alive = $term &&
+                       (($proto eq 'HTTP/1.1' && $conn !~ /\bclose\b/i) ||
+                        ($conn =~ /\bkeep-alive\b/i));
 
        $h .= 'Connection: ' . ($alive ? 'keep-alive' : 'close');
        $h .= "\r\nDate: " . http_date() . "\r\n\r\n";
 
-       if (($len || $chunked) && $env->{REQUEST_METHOD} ne 'HEAD') {
+       if ($term && $env->{REQUEST_METHOD} ne 'HEAD') {
                more($self, $h);
        } else {
                $self->write($h);
        }
-       ($alive, $chunked);
+       $alive;
 }
 
 sub response_write {
        my ($self, $env, $res) = @_;
-       my ($alive, $chunked) = response_header_write($self, $env, $res);
+       my $alive = response_header_write($self, $env, $res);
        my $write = sub { $self->write($_[0]) };
        my $close = sub {
                if ($alive) {
@@ -228,6 +237,10 @@ sub input_prepare {
        my $input = $null_io;
        my $len = $env->{CONTENT_LENGTH};
        if ($len) {
+               if ($len > $MAX_REQUEST_BUFFER) {
+                       quit($self, 413);
+                       return;
+               }
                $input = IO::File->new_tmpfile;
        } elsif (env_chunked($env)) {
                $len = CHUNK_START;
@@ -247,7 +260,7 @@ sub env_chunked { ($_[0]->{HTTP_TRANSFER_ENCODING} || '') =~ /\bchunked\b/i }
 
 sub write_err {
        my ($self) = @_;
-       my $err = $self->{env}->{'psgi.errors'};
+       my $err = $self->{httpd}->{env}->{'psgi.errors'};
        my $msg = $! || '(zero write)';
        $err->print("error buffering to input: $msg\n");
        quit($self, 500);
@@ -260,7 +273,7 @@ sub recv_err {
                $self->{input_left} = $len;
                return;
        }
-       my $err = $self->{env}->{'psgi.errors'};
+       my $err = $self->{httpd}->{env}->{'psgi.errors'};
        $err->print("error reading for input: $! ($len bytes remaining)\n");
        quit($self, 500);
 }
@@ -302,6 +315,9 @@ sub event_read_input_chunked { # unlikely...
                if ($len == CHUNK_START) {
                        if ($$rbuf =~ s/\A([a-f0-9]+).*?\r\n//i) {
                                $len = hex $1;
+                               if (($len + -s $input) > $MAX_REQUEST_BUFFER) {
+                                       return quit($self, 413);
+                               }
                        } elsif (length($$rbuf) > CHUNK_MAX_HDR) {
                                return quit($self, 400);
                        }
@@ -354,6 +370,13 @@ sub quit {
 sub event_hup { $_[0]->close }
 sub event_err { $_[0]->close }
 
+sub write ($$) : method {
+       my PublicInbox::HTTP $self = $_[0];
+       return 1 if (defined($_[1]) && ref($_[1]) eq '' && $_[1] eq '');
+
+       $self->SUPER::write($_[1]);
+}
+
 # for graceful shutdown in PublicInbox::Daemon:
 sub busy () {
        my ($self) = @_;