Adaptive padding 20181218 rebased again #622
Closed
Conversation
We need this for padding negotiation so that we can have later machine revisions supercede earlier ones.
This is a good code review start point, to get an overview of the interfaces and types used in circuit padding.
This helps us to determine if a middle node can pad to us or not.
These event callbacks allow circuit padding to decide when to attempt to launch and negotiate new padding machines, and when to tear old ones down.
These callbacks allow the padding state machines to react to various types of sent and received relay cells.
This implements all of the event handling, state machines, and padding decisions for circuit padding. I recommend reviewing this after you look at the call-in points into it from the rest of Tor.
* Tighten and justify cutoff for approximation in GPD sampler.
* Use expm1(-xi log U) to evaluate U^{-xi} - 1 without precision loss.
* Use log1p(x) to evaluate log(1 + x) precisely even for x near 0.
* Avoid 0 by rejection sampling, not by choosing to return 0 instead.
(This avoids the problematic 0 return from crypto_rand_double()
into the log() without introducing distribution bias).
Hope is this will make it easier to test on the live tor network. Does not need to be merged if we don't want to, but will come in handy for researchers.
Note to self/others: don't merge this.
This project introduces the prob_distr.c subsystem which implements all the probability distributions that WTF-PAD needs. It also adds unittests for all of them. Code and tests courtesy of Riastradh. Co-authored-by: Taylor R Campbell <campbell+tor@mumble.net>
Also add some basic tests for sampling delays from the uniform distribution.
Tighten up integer-like type usage.
Tighten up integer type usage.
Tighten up type usage.
Use new type for bin.
Fix a return type issue.
Fix a return type issue.
Fix a return type issue.
In binary80 arithmetic, the bins for the log-logistic distribution, which has a very fat tail, were computed with a value that overshot the bounds, leading to a NaN intermediate. One defensive way to avoid this would be to change the cdf_* and sf_* functions for distributions of bounded support -- like log-logistic, which is supported only on (0, +\infty) -- to return -/+inf for points outside the bounds. But for testing purposes it might be better not to work defensively like that because it might mask upstream problems.
Test a couple shutdown conditions and wrapping coverage.
Squeeze some more coverage out of existing tests.
Update test section.
Reorganize TODO file for 0.3.6/0.3.7
Fix crash and double-free bugs found during testing. And some other bugs too.
Add wronghop tests and some additional negotiation coverage.
Update TODO.
Be a little more ambitious with compat-breaking features. We should try to do as many as possible for 0.3.6..
Reorder and group machine compat-breaking changes so we can look at all of them as a unit. We shouldn't break compatibility for research machines. So we must solidify all of the machine behaviors ASAP.
This way we can remove a redundant function declaration.
No need to do rejection sampling in circuitpadding.c anymore, which was not covered by tests.
Add test for closest hist bin.
Rest of tokens tests.
Tokens tests are done.
Pull Request Test Coverage Report for Build 3397
|
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
No description provided.
The text was updated successfully, but these errors were encountered: