File

net.server_epoll: Call :shutdown() on TLS sockets when supported Comment from Matthew: This fixes a potential issue where the Prosody process gets blocked on sockets waiting for them to close. Unlike non-TLS sockets, closing a TLS socket sends layer 7 data, and this can cause problems for sockets which are in the process of being cleaned up. This depends on LuaSec changes which are not yet upstream. From Martijn's original email: So first my analysis of luasec. in ssl.c the socket is put into blocking mode right before calling SSL_shutdown() inside meth_destroy(). My best guess to why this is is because meth_destroy is linked to the __close and __gc methods, which can't exactly be called multiple times and luasec does want to make sure that a tls session is shutdown as clean as possible. I can't say I disagree with this reasoning and don't want to change this behaviour. My solution to this without changing the current behaviour is to introduce a shutdown() method. I am aware that this overlaps in a conflicting way with tcp's shutdown method, but it stays close to the OpenSSL name. This method calls SSL_shutdown() in the current (non)blocking mode of the underlying socket and returns a boolean whether or not the shutdown is completed (matching SSL_shutdown()'s 0 or 1 return values), and returns the familiar ssl_ioerror() strings on error with a false for completion. This error can then be used to determine if we have wantread/wantwrite to finalize things. Once meth_shutdown() has been called once a shutdown flag will be set, which indicates to meth_destroy() that the SSL_shutdown() has been handled by the application and it shouldn't be needed to set the socket to blocking mode. I've left the SSL_shutdown() call in the LSEC_STATE_CONNECTED to prevent TOCTOU if the application reaches a timeout for the shutdown code, which might allow SSL_shutdown() to clean up anyway at the last possible moment. Another thing I've changed to luasec is the call to socket_setblocking() right before calling close(2) in socket_destroy() in usocket.c. According to the latest POSIX[0]: Note that the requirement for close() on a socket to block for up to the current linger interval is not conditional on the O_NONBLOCK setting. Which I read to mean that removing O_NONBLOCK on the socket before close doesn't impact the behaviour and only causes noise in system call tracers. I didn't touch the windows bits of this, since I don't do windows. For the prosody side of things I've made the TLS shutdown bits resemble interface:onwritable(), and put it under a combined guard of self._tls and self.conn.shutdown. The self._tls bit is there to prevent getting stuck on this condition, and self.conn.shutdown is there to prevent the code being called by instances where the patched luasec isn't deployed. The destroy() method can be called from various places and is read by me as the "we give up" error path. To accommodate for these unexpected entrypoints I've added a single call to self.conn:shutdown() to prevent the socket being put into blocking mode. I have no expectations that there is any other use here. Same as previous, the self.conn.shutdown check is there to make sure it's not called on unpatched luasec deployments and self._tls is there to make sure we don't call shutdown() on tcp sockets. I wouldn't recommend logging of the conn:shutdown() error inside close(), since a lot of clients simply close the connection before SSL_shutdown() is done.

-- Prosody IM
-- Copyright (C) 2008-2010 Matthew Wild
-- Copyright (C) 2008-2010 Waqas Hussain
--
-- This project is MIT/X11 licensed. Please see the
-- COPYING file in the source package for more information.
--

local indexedbheap = require "prosody.util.indexedbheap";
local log = require "prosody.util.logger".init("timer");
local server = require "prosody.net.server";
local get_time = require "prosody.util.time".now
local type = type;
local debug_traceback = debug.traceback;
local tostring = tostring;
local xpcall = require "prosody.util.xpcall".xpcall;
local math_max = math.max;
local pairs = pairs;

if server.timer then
	-- The selected net.server implements this API, so defer to that
	return server.timer;
end

local _ENV = nil;
-- luacheck: std none

local _add_task = server.add_task;

local _server_timer;
local _active_timers = 0;
local h = indexedbheap.create();
local params = {};
local next_time = nil;
local function _traceback_handler(err) log("error", "Traceback[timer]: %s", debug_traceback(tostring(err), 2)); end
local function _on_timer(now)
	local peek;
	local readd;
	while true do
		peek = h:peek();
		if peek == nil or peek > now then break; end
		local _, callback, id = h:pop();
		local param = params[id];
		params[id] = nil;
		--item(now, id, _param);
		local success, err = xpcall(callback, _traceback_handler, now, id, param);
		if success and type(err) == "number" then
			if readd then
				readd[id] = { callback, err + now };
			else
				readd = { [id] = { callback, err + now } };
			end
			params[id] = param;
		end
	end

	if readd then
		for id,timer in pairs(readd) do
			h:insert(timer[1], timer[2], id);
		end
		peek = h:peek();
	end

	if peek ~= nil and _active_timers > 1 and peek == next_time then
		-- Another instance of _on_timer already set next_time to the same value,
		-- so it should be safe to not renew this timer event
		peek = nil;
	else
		next_time = peek;
	end

	if peek then
		-- peek is the time of the next event
		return peek - now;
	end
	_active_timers = _active_timers - 1;
end
local function add_task(delay, callback, param)
	local current_time = get_time();
	local event_time = current_time + delay;

	local id = h:insert(callback, event_time);
	params[id] = param;
	if next_time == nil or event_time < next_time then
		next_time = event_time;
		if _server_timer then
			_server_timer:close();
			_server_timer = nil;
		else
			_active_timers = _active_timers + 1;
		end
		_server_timer = _add_task(next_time - current_time, _on_timer);
	end
	return id;
end
local function stop(id)
	params[id] = nil;
	local result, item, result_sync = h:remove(id);
	local peek = h:peek();
	if peek ~= next_time and _server_timer then
		next_time = peek;
		_server_timer:close();
		if next_time ~= nil then
			_server_timer = _add_task(math_max(next_time - get_time(), 0), _on_timer);
		end
	end
	return result, item, result_sync;
end
local function reschedule(id, delay)
	local current_time = get_time();
	local event_time = current_time + delay;
	h:reprioritize(id, delay);
	if next_time == nil or event_time < next_time then
		next_time = event_time;
		_add_task(next_time - current_time, _on_timer);
	end
	return id;
end

return {
	add_task = add_task;
	stop = stop;
	reschedule = reschedule;
};