Software / code / prosody
File
util/error.lua @ 13652:a08065207ef0
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.
| author | Martijn van Duren <martijn@openbsd.org> |
|---|---|
| date | Thu, 06 Feb 2025 15:04:38 +0000 |
| parent | 13079:e7a5e5a0dc02 |
line wrap: on
line source
local id = require "prosody.util.id"; local util_debug; -- only imported on-demand -- Library configuration (see configure()) local auto_inject_traceback = false; local error_mt = { __name = "error" }; function error_mt:__tostring() return ("error<%s:%s:%s>"):format(self.type, self.condition, self.text or ""); end local function is_error(e) return getmetatable(e) == error_mt; end local function configure(opt) if opt.auto_inject_traceback ~= nil then auto_inject_traceback = opt.auto_inject_traceback; if auto_inject_traceback then util_debug = require "prosody.util.debug"; end end end -- Do we want any more well-known fields? -- Or could we just copy all fields from `e`? -- Sometimes you want variable details in the `text`, how to handle that? -- Translations? -- Should the `type` be restricted to the stanza error types or free-form? -- What to set `type` to for stream errors or SASL errors? Those don't have a 'type' attr. local function new(e, context, registry, source) if is_error(e) then return e; end local template = registry and registry[e]; if not template then if type(e) == "table" then template = { code = e.code; type = e.type; condition = e.condition; text = e.text; extra = e.extra; }; else template = {}; end end context = context or {}; if auto_inject_traceback then context.traceback = util_debug.get_traceback_table(nil, 2); end local error_instance = setmetatable({ instance_id = id.short(); type = template.type or "cancel"; condition = template.condition or "undefined-condition"; text = template.text; code = template.code; extra = template.extra; context = context; source = source; }, error_mt); return error_instance; end -- compact --> normal form local function expand_registry(namespace, registry) local mapped = {} for err,template in pairs(registry) do local e = { type = template[1]; condition = template[2]; text = template[3]; }; if namespace and template[4] then e.extra = { namespace = namespace, condition = template[4] }; end mapped[err] = e; end return mapped; end local function init(source, namespace, registry) if type(namespace) == "table" then -- registry can be given as second argument if namespace is not used registry, namespace = namespace, nil; end local _, protoerr = next(registry, nil); if protoerr and type(next(protoerr)) == "number" then registry = expand_registry(namespace, registry); end local function wrap(e, context) if is_error(e) then return e; end local err = new(registry[e] or { type = "cancel", condition = "undefined-condition" }, context, registry, source); err.context.wrapped_error = e; return err; end return { source = source; registry = registry; new = function (e, context) return new(e, context, registry, source); end; coerce = function (ok, err, ...) if ok then return ok, err, ...; end return nil, wrap(err); end; wrap = wrap; is_error = is_error; }; end local function coerce(ok, err, ...) if ok or is_error(err) then return ok, err, ...; end local new_err = new({ type = "cancel", condition = "undefined-condition" }, { wrapped_error = err }); return ok, new_err, ...; end local function from_stanza(stanza, context, source) local error_type, condition, text, extra_tag = stanza:get_error(); local error_tag = stanza:get_child("error"); context = context or {}; context.stanza = stanza; context.by = error_tag and error_tag.attr.by or stanza.attr.from; local uri; if condition == "gone" or condition == "redirect" then uri = error_tag:get_child_text(condition, "urn:ietf:params:xml:ns:xmpp-stanzas"); end return new({ type = error_type or "cancel"; condition = condition or "undefined-condition"; text = text; extra = (extra_tag or uri) and { uri = uri; tag = extra_tag; } or nil; }, context, nil, source); end return { new = new; init = init; coerce = coerce; is_error = is_error; is_err = is_error; -- COMPAT w/ older 0.12 trunk from_stanza = from_stanza; configure = configure; }