Alexander Færøy [Fri, 23 Nov 2018 04:56:41 +0000 (05:56 +0100)]
Use process_t for managed proxies.
This patch makes the managed proxy subsystem use the process_t data
structure such that we can get events from the PT process while Tor is
running and not just when the PT process is being configured.
Alexander Færøy [Thu, 22 Nov 2018 17:14:03 +0000 (18:14 +0100)]
Make sure we call process_notify_event_exit() as the last thing in different callbacks.
This patch makes sure that we call process_notify_event_exit() after we
have done any modifications we need to do to the state of a process_t.
This allows application developers to call process_free() in the
exit_callback of the process.
Alexander Færøy [Thu, 22 Nov 2018 04:22:24 +0000 (05:22 +0100)]
Add process_get_pid() to the Process subsystem.
This patch adds support for getting the unique process identifier from a
given process_t. This patch implements both support for both the Unix
and Microsoft Windows backend.
Alexander Færøy [Thu, 22 Nov 2018 03:49:23 +0000 (04:49 +0100)]
Add Windows backend for the Process subsystem.
This patch adds support for Microsoft Windows in the Process subsystem.
Libevent does not support mixing different types of handles (sockets,
named pipes, etc.) on Windows in its core event loop code. This have
historically meant that Tor have avoided attaching any non-networking
handles to the event loop. This patch uses a slightly different approach
to roughly support the same features for the Process subsystem as we do
with the Unix backend.
In this patch we use Windows Extended I/O functions (ReadFileEx() and
WriteFileEx()) which executes asynchronously in the background and
executes a completion routine when the scheduled read or write operation
have completed. This is much different from the Unix backend where the
operating system signals to us whenever a file descriptor is "ready" to
either being read from or written to.
To make the Windows operating system execute the completion routines of
ReadFileEx() and WriteFileEx() we must get the Tor process into what
Microsoft calls an "alertable" state. To do this we execute SleepEx()
with a zero millisecond sleep time from a main loop timer that ticks
once a second. This moves the process into the "alertable" state and
when we return from the zero millisecond timeout all the outstanding I/O
completion routines will be called and we can schedule the next reads
and writes.
The timer loop is also responsible for detecting whether our child
processes have terminated since the last timer tick.
Alexander Færøy [Thu, 22 Nov 2018 03:43:27 +0000 (04:43 +0100)]
Add Unix backend for the Process subsystem.
This patch adds the Unix backend for the Process subsystem. The Unix
backend attaches file descriptors from the child process's standard in,
out and error to Tor's libevent based main loop using traditional Unix
pipes. We use the already available `waitpid` module to get events
whenever the child process terminates.
Alexander Færøy [Thu, 22 Nov 2018 03:25:11 +0000 (04:25 +0100)]
Add new Process subsystem.
This patch adds a new Process subsystem for running external programs in
the background of Tor. The design is focused around a new type named
`process_t` which have an API that allows the developer to easily write
code that interacts with the given child process. These interactions
includes:
- Easy API for writing output to the child process's standard input
handle.
- Receive callbacks whenever the child has output on either its standard
output or standard error handles.
- Receive callback when the child process terminates.
We also support two different "protocols" for handling output from the
child process. The default protocol is the "line" protocol where the
process output callbacks will be invoked only when there is complete
lines (either "\r\n" or "\n" terminated). We also support the "raw"
protocol where the read callbacks will get whatever the operating system
delivered to us in a single read operation.
This patch does not include any operating system backends, but the Unix
and Windows backends will be included in separate commits.
Alexander Færøy [Mon, 10 Sep 2018 12:42:29 +0000 (14:42 +0200)]
Add buf_flush_to_pipe() and buf_read_from_pipe().
This patch adds two new functions: buf_flush_to_pipe() and
buf_read_from_pipe(), which makes use of our new buf_flush_to_fd() and
buf_read_from_fd() functions.
Alexander Færøy [Mon, 10 Sep 2018 12:35:38 +0000 (14:35 +0200)]
Refactor buf_read_from_socket() into buf_read_from_fd().
This patch refactors buf_read_from_socket() into buf_read_from_fd(), and
creates a specialized function for buf_read_from_socket(), which uses
buf_read_from_fd().
Alexander Færøy [Mon, 10 Sep 2018 12:29:27 +0000 (14:29 +0200)]
Refactor buf_flush_to_socket() into buf_flush_to_fd().
This patch refactors buf_flush_to_socket() into buf_flush_to_fd() and
creates a specialization function for buf_flush_to_socket() that makes
use of buf_flush_to_fd().
Alexander Færøy [Tue, 27 Nov 2018 01:35:44 +0000 (02:35 +0100)]
Explicitly specify path to OpenSSL for Appveyor.
This patch explicitly specifies the path to our OpenSSL dependency and
disables the installation of an external OpenSSL version and instead
uses the OpenSSL version available from the MinGW environments.
Nick Mathewson [Thu, 15 Nov 2018 18:16:58 +0000 (13:16 -0500)]
Add options to control dormant-client feature.
The DormantClientTimeout option controls how long Tor will wait before
going dormant. It also provides a way to disable the feature by setting
DormantClientTimeout to e.g. "50 years".
The DormantTimeoutDisabledByIdleStreams option controls whether open but
inactive streams count as "client activity". To implement it, I had to
make it so that reading or writing on a client stream *always* counts as
activity.
Nick Mathewson [Wed, 21 Nov 2018 12:38:38 +0000 (07:38 -0500)]
Fix a fun heisenbug in memoize_protover_flags()
After we clear the protover map for getting full, we need to
re-create it, since we are about to use it.
This is a bugfix for bug 28558. It is a bugfix for the code from
ticket 27225, which is not in any released Tor. Found by Google
OSS-Fuzz, as issue 11475.
Nick Mathewson [Tue, 28 Aug 2018 15:34:31 +0000 (11:34 -0400)]
Backend for compact node-family representation.
This representation is meant to save memory in microdescriptors --
we can't use it in routerinfo_t yet, since those families need to be
encoded losslessly for directory voting to work.
This representation saves memory in three ways:
1. It uses only one allocation per family. (The old way used a
smartlist (2 allocs) plus one strdup per entry.)
2. It stores identity digests in binary, not hex.
3. It keeps families in a canonical format, memoizes, and
reference-counts them.
David Goulet [Thu, 15 Nov 2018 19:36:39 +0000 (14:36 -0500)]
man: ClientOnionAuthDir can't be reloaded with Sandbox 1
By adding a file to the ClientOnionAuthDir and sending a HUP signal, tor would
load the new file and use it. However, that doesn't work with the Sandbox
since post initilization, nothing can be changed.
Document in the manpage that limitation within the Sandbox description.
Closes #28128
Signed-off-by: David Goulet <dgoulet@torproject.org>
Nick Mathewson [Tue, 13 Nov 2018 20:51:53 +0000 (15:51 -0500)]
Add a periodic event to become dormant.
This event makes us become dormant if we have seen no activity in a
long time.
Note that being any kind of a server, or running an onion service,
always counts as being active.
Note that right now, just having an open stream that Tor
did not open on its own (for a directory request) counts as "being
active", so if you have an idle ssh connection, that will keep Tor
from becoming dormant.
Many of the features here should become configurable; I'd like
feedback on which.
Nick Mathewson [Tue, 13 Nov 2018 14:04:11 +0000 (09:04 -0500)]
Move the responsibility for delayed shutdown into the mainloop
This is part of 28422, so we don't have to call
consider_hibernation() once per second when we're dormant.
This commit does not remove delayed shutdown from hibernate.c: it
uses it as a backup shutdown mechanism, in case the regular shutdown
timer mechanism fails for some reason.
Nick Mathewson [Tue, 6 Nov 2018 12:27:31 +0000 (07:27 -0500)]
Add new "ALL" and "NET_PARTICIPANT" roles for periodic events
The previous "ALL" role was the OR of a bunch of other roles,
which is a mistake: it's better if "ALL" means "all".
The "NET_PARTICIPANT" role refers to the anything that is actively
building circuits, downloading directory information, and
participating in the Tor network. For now, it is set to
!net_is_disabled(), but we're going to use it to implement a new
"extra dormant mode".