a3005d8c0a
tor-config: MustRead: Make public
...
I think this ought to be exhaustive.
2022-08-25 15:58:29 +01:00
7c0637ad44
config watch: Fix and reduce debounce interval
...
The parameter to FileWatcher::new is not a polling time fallback; it
is a "debounce time". Events are always delayed by at least this
much.
10s is much too long for this. 1s is more appropriate.
2022-08-25 15:58:29 +01:00
7d8b3e2f2f
config sources: Read arti.d as well as arti.toml
...
Fixes #474 aka #271
2022-08-25 15:58:29 +01:00
ba94c4a4fa
config sources tests: Test results of directory scan
2022-08-25 15:58:29 +01:00
b700816eef
config sources tests: Break out sources_nodefaults
2022-08-25 15:58:29 +01:00
e4fea3e1ea
config sources tests: Introduce test of reading directory
2022-08-25 15:58:29 +01:00
08767f59d8
config sources: Supporting reading directories
2022-08-25 15:58:29 +01:00
7d088cf8df
config sources: Introduce scan() and FoundConfigFiles
...
We're going to need to do config file reading in two phases.
Right now this isn't actually necessary, because the set of files
is fixed since we don't support dynamically scanning directories.
But the new API will be needed in a moment.
Code motion and API changes, but no overall functional change.
Review with `git show -b` may be helpful.
The new API also provides for dealing with directories, but right now
that doesn't happen.
2022-08-25 15:58:29 +01:00
2fa75be660
tor-basic-utils: Provide IoErrorExt is_not_a_directory()
...
We're going to want this functionality, which isn't in the stable
stdlib.
2022-08-25 15:58:29 +01:00
587fa5f418
config watch: Provide watch_dir
...
No call site just yet; that will come shortly.
This requires a bit of reorganisation first.
2022-08-25 15:58:29 +01:00
863c66159b
config watch: Re-establish watcher on each iteration
...
This is going to be needed in a moment.
2022-08-25 15:58:29 +01:00
a7bb3a73b4
config watch: Rescan once on startup
...
That way if the config changes after we read it initially, but before
we set up the watcher, we will still pick it up.
Fixes #544
2022-08-25 15:58:29 +01:00
0f9bf12a7f
config watch: Break out prepare_watcher
...
This is going to become more complicated, and gain another call site.
2022-08-25 15:58:29 +01:00
8e86599df4
config watch: Make the mpsc channel part of FileWatcher
...
The previous approach (inherited from the API of notify) was kind of
odd.
Soon we are going to want to be able to drop the watcher and replace
it. That really wants the same object to contain all the things that
ought to be dropped together. (notify's watchers stop generating
events and give EOF on the channel, when dropped.)
2022-08-25 15:58:29 +01:00
25b5a53953
config: Do process hardening on reconfigure even if not watching
...
These blocks were in the wrong order.
Previously, if you tried to turn on process hardening in the config
and then reloaded rather than restarting, it wouldn't take effect.
2022-08-25 15:58:29 +01:00
cd2432474c
tor-config sources: Remove some unneeded .to_string() from tests
2022-08-25 15:58:29 +01:00
920f7a3f53
tor-chanmgr: don't log addresses so much.
...
We now log connection attempts at debug!, and mark relay target
addresses as sensitive.
2022-08-25 10:54:50 -04:00
8fa1568b02
arti: Raise the default console log severity to "info"
...
Previously we logged at "debug", but that's not meant to
user-facing.
2022-08-25 09:10:42 -04:00
504bba2191
Fix setuid testing for MacOS
2022-08-25 12:21:33 +01:00
679c1c9b6a
Merge branch 'no_setuid' into 'main'
...
arti_client: Refuse to build a client if we are setuid.
See merge request tpo/core/arti!689
2022-08-24 19:47:15 +00:00
a8b3e147fe
arti_client: Refuse to build a client if we are setuid.
...
Arti is not designed to be a setuid-safe program.
Part of #523 .
2022-08-24 15:23:24 -04:00
075ec7557f
fix nightly lints
2022-08-24 19:10:12 +02:00
c7d2b5001c
fix test failing due to missing allow_running_as_root
2022-08-24 18:34:02 +02:00
ca99e9f635
add feature annotation to fields made visible through visible::StructFields
2022-08-24 18:22:41 +02:00
426a59b2ba
add feature annotation not added by doc_auto_cfg
2022-08-24 18:22:41 +02:00
7f939fa480
enable doc_auto_cfg feature on every crate when documenting for docs.rs
2022-08-24 18:22:41 +02:00
187c6b48df
Merge branch 'no_root' into 'main'
...
arti: Do not allow running as root.
See merge request tpo/core/arti!688
2022-08-24 15:21:57 +00:00
25c287636b
arti: Do not allow running as root.
...
This can be overridden with `application.allow_running_as_root`.
Part of #523 .
2022-08-24 10:57:03 -04:00
51eb0e6ca2
Merge branch 'test-old-config' into 'main'
...
Test oldest example configuration file
See merge request tpo/core/arti!684
2022-08-23 14:43:30 +00:00
ee2580face
tor-config: resolve_option_general: Add semver note
2022-08-23 14:21:13 +01:00
0c747b8069
tor-config: resolve_option_general: Add TODO about example
2022-08-23 14:20:49 +01:00
32c14d37bf
Fix syntax in doc comment.
2022-08-23 12:59:41 +00:00
298f6f508c
arti cfg: Write down future plans
...
Mostly cribbed from
https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/684#note_2829937
2022-08-23 13:55:48 +01:00
457c17adb0
tor-config: Provide resolve_option_general, for T: !Default etc.
...
At one point in this MR I thought I was going to want this for
arti::cfg::ListenConfig (which we don't want to be Default).
In fact ListenConfig is being handled specially, but having written
this function it seemed sensible to keep it. Since resolve_option
becomes a wrapper for it, the existing tests exercise it.
2022-08-22 16:26:38 +01:00
086cf3701a
arti cfg: Test parsing of the oldest config file we still support
2022-08-22 16:26:22 +01:00
a1cc2578b3
arti cfg: Fix docs about ARTI_EXAMPLE_CONFIG
...
The defaults are now
2022-08-22 16:26:22 +01:00
46d9dadd02
Merge branch 'channel' into 'main'
...
Channel padding
See merge request tpo/core/arti!657
2022-08-17 10:26:25 +00:00
d4ce701fa4
channel padding: Only do anything with link protocol 5
...
Or rather, if we *didn't* negotiate 4, which is too old.
As per
https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/657#note_2828354
2022-08-17 10:54:41 +01:00
08a7257fe9
channel padding: Properly advertise link protocol 5
...
We already actually send and negotiate the padding, since !657 ,
but we ought to negotiate a protocol version where that's not a
violation!
As per
https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/657#note_2828354
2022-08-17 10:54:41 +01:00
a0b0eca93e
tor-netdoc: if_ito: Add expansion of acronym
2022-08-17 10:54:41 +01:00
8339558ba8
channel padding: Move initial config of ChannelPaddingInstructions
...
Making ChannelPaddingInstructions::default() accurately reflect the
initial state of the reactor's padding timer simplifies the code
somewhat.
(When padding is wanted, parameters are computed and inserted
explicitly, so the only change is that if we start out dormant, we
defer setting the timer parameters until necessary.)
As per
https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/657#note_2827249
2022-08-17 10:54:41 +01:00
8a484a26e6
channel padding: Parameters: default_padding replaces Default impl
...
As per
https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/657#note_2827249
2022-08-17 10:54:41 +01:00
3aed28acd4
channel usage: Rename CU::UserTraffic from CU::Exit
...
As suggested in
https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/657#note_2827246
2022-08-17 10:54:41 +01:00
24e2c717f0
chanmgr padding: Correct docs for padding_parameters*
...
Prompted by
https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/657#note_2827242
These texts were out of date; one of them still mentioned
`all_zeroes`.
2022-08-17 10:54:41 +01:00
cbd759163a
Improve comment for ChannelUsage::Exit
2022-08-17 10:54:41 +01:00
d44c860355
Fix typo in comment
2022-08-17 10:54:41 +01:00
c97ea4fce0
tor-chanmgr tests: Fix a minimal-versions failure (actually)
2022-08-17 10:54:41 +01:00
a892a3ee62
tor-chanmgr tests: Fix a minimal-versions failure
2022-08-17 10:54:41 +01:00
074d08bee8
arti: Fix a minimal-versions failure
2022-08-17 10:54:41 +01:00
d9338293c8
Add semver notes
2022-08-17 10:54:41 +01:00
4111c63bd4
channel reactor: Add some commented-out debug for padding
2022-08-17 10:54:41 +01:00
46885ac14b
channel padding: Correctly always send the timer parameters
...
Fixes "chanmgr configuration: Avoid sending needless initial update(s)"
2022-08-17 10:54:41 +01:00
496563e7cf
channel padding: Abolish ARTI_EXPERIMENTAL_CHANNEL_PADDING env var
...
This was for testing and is no longer needed.
2022-08-17 10:54:41 +01:00
299f302f35
channel engage_padding_activities: swap docs to tor0proto
...
This allow us to make a working cross-reference.
2022-08-17 10:54:41 +01:00
88dc816770
channel fake_channel_details: Use precise cfg
...
As per
https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/657#note_2826169
2022-08-17 10:54:41 +01:00
97363ba250
tor-netdir: testnet: construct_custom_netdir_with_params: Fix doc
...
As per
https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/657#note_2826168
2022-08-17 10:54:41 +01:00
52848a41c6
ChannelPaddingInstructions: Update and expand docs
2022-08-17 10:54:41 +01:00
2a4fd97229
ChannelPaddingInstructions: Rename `params` field to `insns`
2022-08-17 10:54:41 +01:00
949e5395e0
Rename ChannelsParams types to ChannelPaddingInstructions (fmt)
...
Run rustfmt; no other changes.
2022-08-17 10:54:41 +01:00
94dd3361e2
Rename ChannelsParams types to ChannelPaddingInstructions
...
As per
https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/657#note_2826167
This makes some lines too long; I will run rustfmt in a separate
commit for clarity.
2022-08-17 10:54:41 +01:00
d1a709f46d
chanmgr: Fix typo
...
Reported here
https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/657#note_2826166
2022-08-17 10:54:41 +01:00
590e174d3a
chanmgr parameterize: Correct doc comment
...
As per
https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/657#note_2826164
2022-08-17 10:54:41 +01:00
2404357282
Rename PaddingParameters::all_zeroes constructor to disabled
2022-08-17 10:54:41 +01:00
bb2ae00217
chamgr NetParamsExtract: Replace direct nf_ito access
...
It turns out that the TryFrom was infallible, which wasn't obvious
with the previous code structure.
2022-08-17 10:54:41 +01:00
03d8a09c0e
Channel: Make mutable() and engage_padding_activities infallible
...
As per
https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/657#note_2826151
This gets rid of quite some Bug error paths.
2022-08-17 10:54:41 +01:00
5db974ef72
Move ChannelUsage from tor_proto to tor_chanmgr
...
Replace Channel::note_usage with Channel::engage_padding_activities,
which unconditionally causes the channel to (start to) do netflow
padding things.
The condition now lives in chanmgr.
Addresses
https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/657#note_2826094
2022-08-17 10:54:41 +01:00
49dad4bd13
tor-chanmgr: Expand docs for map::Inner dormancy field
...
As per
https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/657#note_2826093
2022-08-17 10:54:41 +01:00
4cf418f421
tor-cell: PaddingNegotiate::default_reduced
...
Get rid of unneeded constructor.
We never need to use hardcoded reduced padding parameters during
negotiation cell construction. If we are using reduced padding
parameters, the layers which decide this have netparams to use.
Prompted by
https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/657#note_2826092
2022-08-17 10:54:41 +01:00
14c6e17797
tor-cell: PaddingNegotiate::start: take IntegerMilliseconds
2022-08-17 10:54:41 +01:00
18827b7fd0
chanmgr padding_parameters: Use Option
...
The network consensus parameters use (0,0) to mean "no padding"
(which is not the same as (0,0) means in a PADDING_NEGOTIATE cell).
Representing "no padding" this way is actually quite convoluted and
un-Rustic. Ensure that we convert (0,0) to None, and do the primary
logic in Option.
2022-08-17 10:54:41 +01:00
bdaee82187
tor-cell: PaddingNegotiate: give better spec xrefs
2022-08-17 10:54:41 +01:00
6fdaca8eb9
channel: Clarify (and in some places replace) "frontend" terminology
2022-08-17 10:54:41 +01:00
481bf0bbe9
tor-proto channel padding: Document PADDING_NEGOTIATE overwriting
...
As per
https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/657#note_2825979
2022-08-17 10:54:41 +01:00
1104235ee5
chanmgr: padding: Split IEFI into named function
...
Roughly as suggested in
https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/657#note_2826160
IMO this is less clear, because of the way it splits the logic up.
2022-08-17 10:54:41 +01:00
58fb118480
chanmgr: NetParamsExtract: extend docs for nf_ito
2022-08-17 10:54:41 +01:00
84d8c32543
chanmgr: Get rid of final Result relating to netdir/netparams
...
No functional change. (The error tyep was Void.)
2022-08-17 10:54:41 +01:00
48fc5d3c39
chanmgr: Use NetDirProvider::params and Arc<dyn..Netparameters>
...
This gets rid of many Result(). Many parameters are renamed.
Test cases of the now-impossible branch are removed.
Deleting the match from padding_parameters will come in a moment.
I've split off that commit since it has much whitespace noise.
for now, change the error type to Void.
2022-08-17 10:54:41 +01:00
3497d8723b
chanmgr: Add a TODO for renaming NetDirExtract
2022-08-17 10:54:41 +01:00
36e3529e34
chanmgr ChannelUsage: Fix and clarify docs
2022-08-17 10:54:41 +01:00
e4cd01732f
chanmgr set_dormancy: Remove obsolete comment
2022-08-17 10:54:41 +01:00
b5176a700b
chanmgr: Extend doc comments for Dormancy
...
Prompted by
https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/657#note_2825962
2022-08-17 10:54:41 +01:00
e5396a35c8
Fix typos in comments.
2022-08-17 10:54:41 +01:00
7ed983800d
channel padding: Rename low_ms and high_ms
...
These have the unit in the type. Putting that in the field name too
is otiose.
2022-08-17 10:54:41 +01:00
18a6234101
channel padding: Test through most of the layers
2022-08-17 10:54:41 +01:00
bd4204a42e
Merge branch 'transient_circmgr_errors' into 'main'
...
Mark some circuit-building errors as "transient".
Closes #517
See merge request tpo/core/arti!676
2022-08-16 18:33:17 +00:00
8862eb1e34
Additionally, mark the LostUsablityRace error as a "reset".
2022-08-16 14:03:43 -04:00
4c9bf4803c
Mark some circuit-building errors as "transient".
...
A "transient" error is one that does not indicate a true failure,
but rather an _expected_ need to retry. When we hit one of these,
we do not count it against the total number of permitted failures.
(We do impose a higher limit on "real failures plus transient
failures", though, to prevent infinite loops in the event of a
programming error.
Closes #517 .
2022-08-16 14:03:43 -04:00
e4e06f66ec
chanmgr testing: FakeChannel: retain the ChannelsParamsUpdates
2022-08-16 18:43:23 +01:00
7ae60c229b
tor-proto, testing: Provide accessors for ChannelsParamsUpdates
2022-08-16 18:43:23 +01:00
9310ee2930
tor-proto, testing: Provide new_fake_channel
...
To test the padding control we will want this.
2022-08-16 18:43:23 +01:00
182b9775bb
tor-proto, testing: Make fake_channel_details available
...
Now it's not just cfg(test), but feature testing.
2022-08-16 18:43:23 +01:00
db3fafb0d4
tor-proto: Make "testing" feature that exports some things
...
We are going to want this for through-the-layers padding control
testing.
2022-08-16 18:43:23 +01:00
76403563e5
tor-cell, testing: Provide PaddingNegotiate::from_raw
...
This allows test cases to describe precisely the contents of the
negotiation cell ought to be generated.
2022-08-16 18:43:23 +01:00
2385d43279
tor-netdir, testnet: Provide construct_custom_netdir_with_params
...
We're going to want this for testing netdirs with padding parameters
that diff ere from the spec's default.
2022-08-16 18:43:23 +01:00
cd8d41fd2e
tor-netdoc: NetParams: impl FromIterator
...
Otherwise a caller who has a set of parameters must open-code the
loop.
2022-08-16 18:43:23 +01:00
50ca64218e
channel padding: Send negotiation cells
2022-08-16 18:43:23 +01:00
0a1bffb047
tor-proto channel: Make arrangements to send PADDING_NEGOTIATE
...
This is actually a general facility for inserting locally-generated
cells into the outgoing stream.
It doesn't seem to be possible to do this without adding an additional
condition check to the reactor, since we need to insert it into the
right place in the stream, giving it priority over data, and only
using it up if there was room in the output.
We don't engage this machinery yet, because nothing sets
special_outgoing.
2022-08-16 18:43:23 +01:00
67072de75c
channel padding: Send padding iff required
...
Now all the information is plumbed to the right place, and we can
actually decide if we're sending padding.
Additionally, we conditionalise sending timing parameters on whether
padding is actually enabled, so in dormant mode we do not generate
updates (broadcast to all channels) just to reconfigure unused timing
parameters.
2022-08-16 18:43:23 +01:00
d8972e2cde
chanmgr configuration: Avoid sending needless initial update(s)
...
Change ChannelsParams::initial_update to compare fields with their
default values, and, if they're the same as the default, not to
include them in the returned update.
And if that update is then empty, return None.
The overall effect is to avoid the call to chan.reparameterize if
we're using the builtin default parameters, which is usual.
2022-08-16 18:43:23 +01:00
ede2510695
chanmgr configuration: Initialise correctly
...
This arranges that the ChannelsParams we have retain, and which we
send to every newly created channel, actually has the right
parameters, even if they're not the default.
2022-08-16 18:43:23 +01:00
Ian Jackson
Nick Mathewson
trinity-1686a
Ian Jackson
eta