Add an option to disable circuit padding

We need an option like ConnectionPadding to disable circuit padding. Do we also need ReducedCircuitPadding for mobile?

changed milestone to %Tor: 0.4.1.x-final

added 041-must actualpoints::0.33 component::core tor/tor milestone::Tor: 0.4.1.x-final padding parent::28634 points::0.5 priority::high resolution::fixed reviewer::asn severity::normal sponsor::2 status::closed tor-cell tor-relay type::defect wtf-pad labels

Trac:
Milestone: Tor: 0.4.0.x-final to Tor: unspecified

If we do a ReducedCircuitPadding, we should have a way to specify which padding machines are acceptable for use when it is set.

Trac:
Keywords: N/A deleted, 041-proposed added

Trac:
Points: N/A to 0.5

Trac:
Priority: Medium to High

Trac:
Sponsor: N/A to Sponsor2

The answer here is "no -- not once #28693 (moved) is implemented".

Trac:
Status: new to closed
Resolution: N/A to fixed

Replying to nickm:

The answer here is "no -- not once #28693 (moved) is implemented".

I don't understand. This ticket is #28693 (moved), and it's not implemented.

I still think we need a CircuitPadding option.

Trac:
Resolution: fixed to N/A
Status: closed to reopened

For example: sbws doesn't want padding on its circuits, because padding interferes with bandwidth measurements.

Trac:
Keywords: 041-proposed deleted, 041-must added

See also #30173 (moved). Both the consensus param and the torrc option should play nice with eachother.

Trac:
Parent: #28632 (moved) to #28634 (moved)

We have two similar options, ConnectionPadding and CircuitPaddingDisabled.

In the current branch, 0 means "connection padding off" but "circuit padding on". Which makes it easier for a developer or user to make a mistake.

But it's hard to make the options entirely consistent, because ConnectionPadding has 3 options (force on, negotiate, force off), but CircuitPadding has 2 (negotiate, force off).

Here's one possible design (let's call it A):

	0	auto	1
ConnectionPadding	off	negotiate	on
CircuitPadding	off	negotiate	can't force on: reject config as invalid

Here is another (let's call it B):

	0	auto	1
ConnectionPadding	off	negotiate	on
CircuitPadding	off	no third state: reject config as invalid	negotiate

In A, ConnectionPadding and CircuitPadding are consistent with each other: 0 means off and auto means negotiate.

In B, CircuitPadding is consistent with other 2-state torrc options, because it uses 0 and 1. But negotiate is auto in ConnectionPadding and 1 in CircuitPadding.

Do we have any existing options that are (0, auto) or (1, auto)?

• If we don't, we shouldn't add one.
• If we do, I guess it's ok, as long as we document it.

Trac:
Status: reopened to needs_revision

Ok, I took approach B, since we don't have any autobools that don't support 1. https://github.com/torproject/tor/pull/965

That PR contains fixes for #28693 (moved), #30173 (moved), and #29203 (moved). I recommend that the same reviewer review all three tickets -- they are extremely similar and related code (hence all in the same branch with the same test; doing these separetely would involve needless extra mental context switching and extra communication).

Trac:
Actualpoints: N/A to 0.33
Status: needs_revision to needs_review

Both Travis and Appveyor CIs failed on https://github.com/torproject/tor/pull/965

Trac:
Status: needs_review to needs_revision

These tickets all have code in the same pull request, so I'm assuming they're all 0.4.1.

Trac:
Milestone: Tor: unspecified to Tor: 0.4.1.x-final

CI issues fixed in the PR.

Trac:
Status: needs_revision to needs_review

Trac:
Reviewer: N/A to asn

Comment left in #30173 (moved).

Trac:
Status: needs_review to needs_revision

Trac:
Status: needs_revision to needs_review

Trac:
Status: needs_review to needs_revision

please add bool options

CircuitPadding

and

ReducedCircuitPadding

Replying to cypherpunks:

please add bool options {{{ CircuitPadding }}} and {{{ ReducedCircuitPadding }}}

Is there some reason you are saying this? This is already implemented in the branch for this ticket.

Trac:
Status: needs_revision to needs_review

Trac:
Status: needs_review to merge_ready

Can I please have a torspec branch for the consensus parameters?

Ah, there is one. NM.

merged1

Trac:
Resolution: N/A to fixed
Status: merge_ready to closed

closed

changed time estimate to 4h

added 2h 38m of time spent

mentioned in issue #28694 (moved)

mentioned in issue #29203 (moved)

mentioned in issue #30173 (moved)

moved to tpo/core/tor#28693 (closed)

mentioned in issue tpo/network-health/sbws#28694 (closed)

mentioned in issue tpo/core/tor#40277