Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Use all valid routes during blinded path construction #9334

Open
wants to merge 7 commits into
base: master
Choose a base branch
from

Conversation

MPins
Copy link
Contributor

@MPins MPins commented Dec 5, 2024

Fixes #9076

MaxNumPaths restriction moved from FindBlindedPaths to BuildBlindedPaymentPaths this way we can interact through all possibly routes.

If the reviewers agreed to this approach we will have to modify the TestFindBlindedPathsWithMC because it is considering that MaxNumPaths retriction is applied on FindBlindedPaths function.

Copy link
Contributor

coderabbitai bot commented Dec 5, 2024

Important

Review skipped

Auto reviews are limited to specific labels.

🏷️ Labels to auto review (1)
  • llm-review

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.


Thank you for using CodeRabbit. We offer it for free to the OSS community and would appreciate your support in helping us grow. If you find it useful, would you consider giving us a shout-out on your favorite social media?

❤️ Share
🪧 Tips

Chat

There are 3 ways to chat with CodeRabbit:

  • Review comments: Directly reply to a review comment made by CodeRabbit. Example:
    • I pushed a fix in commit <commit_id>, please review it.
    • Generate unit testing code for this file.
    • Open a follow-up GitHub issue for this discussion.
  • Files and specific lines of code (under the "Files changed" tab): Tag @coderabbitai in a new review comment at the desired location with your query. Examples:
    • @coderabbitai generate unit testing code for this file.
    • @coderabbitai modularize this function.
  • PR comments: Tag @coderabbitai in a new PR comment to ask questions about the PR branch. For the best results, please provide a very specific query, as very limited context is provided in this mode. Examples:
    • @coderabbitai gather interesting stats about this repository and render them as a table. Additionally, render a pie chart showing the language distribution in the codebase.
    • @coderabbitai read src/utils.ts and generate unit testing code.
    • @coderabbitai read the files in the src/scheduler package and generate a class diagram using mermaid and a README in the markdown format.
    • @coderabbitai help me debug CodeRabbit configuration file.

Note: Be mindful of the bot's finite context window. It's strongly recommended to break down tasks such as reading entire modules into smaller chunks. For a focused discussion, use review comments to chat about specific files and their changes, instead of using the PR comments.

CodeRabbit Commands (Invoked using PR comments)

  • @coderabbitai pause to pause the reviews on a PR.
  • @coderabbitai resume to resume the paused reviews.
  • @coderabbitai review to trigger an incremental review. This is useful when automatic reviews are disabled for the repository.
  • @coderabbitai full review to do a full review from scratch and review all the files again.
  • @coderabbitai summary to regenerate the summary of the PR.
  • @coderabbitai generate docstrings to generate docstrings for this PR. (Beta)
  • @coderabbitai resolve resolve all the CodeRabbit review comments.
  • @coderabbitai configuration to show the current CodeRabbit configuration for the repository.
  • @coderabbitai help to get help.

Other keywords and placeholders

  • Add @coderabbitai ignore anywhere in the PR description to prevent this PR from being reviewed.
  • Add @coderabbitai summary to generate the high-level summary at a specific location in the PR description.
  • Add @coderabbitai anywhere in the PR title to generate the title automatically.

CodeRabbit Configuration File (.coderabbit.yaml)

  • You can programmatically configure CodeRabbit by adding a .coderabbit.yaml file to the root of your repository.
  • Please see the configuration documentation for more information.
  • If your editor has YAML language server enabled, you can add the path at the top of this file to enable auto-completion and validation: # yaml-language-server: $schema=https://coderabbit.ai/integrations/schema.v2.json

Documentation and Community

  • Visit our Documentation for detailed information on how to use CodeRabbit.
  • Join our Discord Community to get help, request features, and share feedback.
  • Follow us on X/Twitter for updates and announcements.

@MPins MPins changed the title route blinding: use all valid routes during path construction iteration instead of the currently capped set by MaxNumPaths Use all valid routes during path construction iteration instead of the currently capped set by MaxNumPaths Dec 5, 2024
@MPins MPins changed the title Use all valid routes during path construction iteration instead of the currently capped set by MaxNumPaths use all valid routes during blinded path construction Dec 5, 2024
@MPins MPins changed the title use all valid routes during blinded path construction Use all valid routes during blinded path construction Dec 5, 2024
@ziggie1984 ziggie1984 self-requested a review December 5, 2024 18:06
@ziggie1984
Copy link
Collaborator

ziggie1984 commented Dec 7, 2024

This seems like the simpler solution to implement. Since we’re already querying all routes in FindBlindedPaths, it makes sense to just return them. However, it might be worthwhile to introduce a hard constant cap during the pathfinding process to prevent the array size from growing exponentially. Currently, the algorithm explores all possible paths within the given constraints. Instead, we could stop after finding, say, 100 routes in total, and then apply filtering to select the MaxNumPaths in BuildBlindedPaymentPaths, as proposed in this PR.

Probably the constraints of the Min/Max Hop number prevents these high numbers for routes but I think having a cap is kind of a sanity check.

@MPins
Copy link
Contributor Author

MPins commented Dec 9, 2024

This seems like the simpler solution to implement. Since we’re already querying all routes in FindBlindedPaths, it makes sense to just return them. However, it might be worthwhile to introduce a hard constant cap during the pathfinding process to prevent the array size from growing exponentially. Currently, the algorithm explores all possible paths within the given constraints. Instead, we could stop after finding, say, 100 routes in total, and then apply filtering to select the MaxNumPaths in BuildBlindedPaymentPaths, as proposed in this PR.

Probably the constraints of the Min/Max Hop number prevents these high numbers for routes but I think having a cap is kind of a sanity check.

Before the proposed changes we are querying all routes, sorting the routes based on probability and them capping them based on MaxNumPaths before returning from FindBlindedPaths. After the changes we are capping them just one step later after FindBlindedPaths return them to BuildBlindedPaymentPaths. The hard constant cap should be introduced after the sorting! I might be wrong, but it seems that we are getting too little introducing it!

@ziggie1984
Copy link
Collaborator

ziggie1984 commented Dec 9, 2024

The hard constant cap should be introduced after the sorting! I might be wrong, but it seems that we are getting too little introducing it!

I think we need to do way more when creating the path rather than creating the path and later checking the probability this does not seem efficient. I think we should build the path in the first place based on probability not just running a depth first search on all available paths. But I am not sure how difficult that might be because we went with the recusrive approach.

@MPins
Copy link
Contributor Author

MPins commented Dec 10, 2024

The hard constant cap should be introduced after the sorting! I might be wrong, but it seems that we are getting too little introducing it!

I think we need to do way more when creating the path rather than creating the path and later checking the probability this does not seem efficient. I think we should build the path in the first place based on probability not just running a depth first search on all available paths. But I am not sure how difficult that might be because we went with the recursive approach.

I think we might compute the total probability and sort the path list during the creation of it (into processNodeForBlindedPath), but I'm not sure it's worth it. What do you think?

@ziggie1984
Copy link
Collaborator

yes let's keep it as is and focus only on the underlying issue.

routing/router.go Outdated Show resolved Hide resolved
@MPins
Copy link
Contributor Author

MPins commented Dec 19, 2024

Hello @ziggie1984 gently reminder of review 😉

@ziggie1984
Copy link
Collaborator

unit-test fail !

Copy link
Collaborator

@ziggie1984 ziggie1984 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looking good I am missing 2 things here:

  1. Can you add some logging stats how many routes are find in total, then how many are filtered out because of the probability as for example debug information. I am very curious whether routes will fail often at the level where the final route and relay information is accumulated.
  2. I tend to introduce a max route variable for the blinded path finding in general, maybe a default of 50 routes, otherwise we look for everything which with a big number of blinded hops might be wasteful.

routing/blindedpath/blinded_path.go Outdated Show resolved Hide resolved
routing/router_test.go Show resolved Hide resolved
rpcserver.go Outdated Show resolved Hide resolved
@MPins MPins force-pushed the issue-9076 branch 2 times, most recently from 41573d0 to c1251d7 Compare January 20, 2025 23:28
@MPins MPins requested a review from ziggie1984 January 20, 2025 23:44
@MPins
Copy link
Contributor Author

MPins commented Jan 24, 2025

Looking good I am missing 2 things here:

  1. Can you add some logging stats how many routes are find in total, then how many are filtered out because of the probability as for example debug information. I am very curious whether routes will fail often at the level where the final route and relay information is accumulated.
  2. I tend to introduce a max route variable for the blinded path finding in general, maybe a default of 50 routes, otherwise we look for everything which with a big number of blinded hops might be wasteful.

Hello Ziggie, I had addressed your comments, thank you! Ready for your next review!

Copy link
Contributor

@Abdulkbk Abdulkbk left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for the PR. I looked through the changes and left some comments/questions.

routing/router.go Outdated Show resolved Hide resolved
routing/router.go Outdated Show resolved Hide resolved
routing/router.go Outdated Show resolved Hide resolved
routing/router.go Outdated Show resolved Hide resolved
routing/router.go Outdated Show resolved Hide resolved
routing/blindedpath/blinded_path_test.go Outdated Show resolved Hide resolved
Copy link
Contributor Author

@MPins MPins left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thank you @Abdulkbk for your review!

routing/router.go Outdated Show resolved Hide resolved
routing/router.go Outdated Show resolved Hide resolved
@MPins
Copy link
Contributor Author

MPins commented Jan 31, 2025

@ziggie1984 and @Abdulkbk , it is ready for your review. 🤝

Copy link
Collaborator

@ziggie1984 ziggie1984 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looking good, I want to get @ellemouton opinion on the recursive abort criteria in case we have a maximum number of paths.

Missing release notes for LND 20.

routing/blindedpath/blinded_path.go Outdated Show resolved Hide resolved
for _, route := range routes {
if len(bestRoutes) >= int(restrictions.MaxNumPaths) {
if len(allRoutes) >= maxRoutes {
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

that is not what I meant, my idea was to exit the recursion early when enough paths are found,

I was more thinking of restricting the recursion in processNodeForBlindedPath but I would like to get @ellemouton opinion on this and also this can probably done in another PR.

Currently our DFS exhausts all Paths which exist in the graph, so I was thinking we could abort the DFS after we already have a hopSet of let's say 100 ?

Imagine we have a minRouteLen of 2 and a maxRouteLen of 6, each node has more than 100 channels, when reaching the depth of 6, we could easily accumulate > 1000 routes which seems unreasonable.

But the above restrictions we do not need, because really makes not difference we already have all the paths in memory.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You right, thanks for the explanation.

The point is that, in the current implementation, we first retrieve all the routes and then sort them to prioritize those with the highest success probability.

Instead, we should modify the process to evaluate the probability of each route during the iteration, keeping only the best ones.

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

seems like this is removed but not in this commit, can you make sure this is removed also in this commit

routing/router_test.go Show resolved Hide resolved
routing/blindedpath/blinded_path_test.go Outdated Show resolved Hide resolved
@yyforyongyu
Copy link
Member

There is no release notes for LND 20 yet. Should I create it?

Yeah feel free to open a PR!

@ellemouton ellemouton requested review from ellemouton and removed request for Abdulkbk February 10, 2025 06:27
@ziggie1984
Copy link
Collaborator

please make sure the check commits passes, you can run this script locally as well:

scripts/check-each-commit.sh upstream/master

Copy link
Collaborator

@ziggie1984 ziggie1984 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It's close let's bring it over the finish line 💪

@@ -100,6 +100,9 @@ type BuildBlindedPathCfg struct {
// route.
MinNumHops uint8

// MaxNumPaths is the maximum number of blinded paths to select.
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The MaxNumPath can now be removed from BlindedPathRestrictions we are not using it anymore there. Super confusing to have all this MaxPaths variables.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You right, thanks!

for _, route := range routes {
if len(bestRoutes) >= int(restrictions.MaxNumPaths) {
if len(allRoutes) >= maxRoutes {
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

seems like this is removed but not in this commit, can you make sure this is removed also in this commit

@@ -640,6 +640,8 @@ func (r *ChannelRouter) FindBlindedPaths(destination route.Vertex,
probability float64
}

totalRoutesFound := 0
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We don't need this counters, we just get the information by comparing the different lengths of the arrays.

routes and paths

@@ -640,6 +640,8 @@ func (r *ChannelRouter) FindBlindedPaths(destination route.Vertex,
probability float64
}
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

can you be more verbose in the commit msg: Something like:

We now return all possible paths found by the pathfinder and only remove those which do have a low probability.

@@ -708,9 +708,6 @@ func (r *ChannelRouter) FindBlindedPaths(destination route.Vertex,
// Return all routes, capped by the maxNumberOfRoutes.
allRoutes := make([]*route.Route, 0, len(routes))
for _, route := range routes {
if len(allRoutes) >= maxRoutes {
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

this change should be in the previous commit

@@ -3249,27 +3245,12 @@ func TestFindBlindedPathsWithMC(t *testing.T) {
"alice,charlie,dave",
})

// Change the MaxNumPaths to 1 to assert that only the best route is
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

in this file there is a typo: fluke => flake

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It is not a typo ... "fluke" and "flake" have different meanings, though they sound similar. Fluke is a lucky or random event that happens by chance.

rpcserver.go Outdated
@@ -6168,6 +6168,7 @@ func (r *rpcServer) AddInvoice(ctx context.Context,
MaxHTLCMsat: 0,
},
MinNumPathHops: blindingRestrictions.NumHops,
MaxNumPaths: blindingRestrictions.MaxNumPaths,
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Given my understanding we can use the config values here directly without adding a middle step assigning it to the restrictions struct.

@@ -547,35 +547,58 @@ func genBlindedRouteData(rand *rand.Rand) *record.BlindedRouteData {
// an example mentioned in this spec document:
// https://github.com/lightning/bolts/blob/master/proposals/route-blinding.md
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Try keeping the commit msg below 50 chars, so its not overflowing like you see here.

@@ -547,35 +547,58 @@ func genBlindedRouteData(rand *rand.Rand) *record.BlindedRouteData {
// an example mentioned in this spec document:
// https://github.com/lightning/bolts/blob/master/proposals/route-blinding.md
// This example does not use any dummy hops.
// Added Dave to test MaxNumPaths restriction.
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

can we add a isolated test which does nothing more than testing the maxPaths. Now the test feels a bit confusing, we can just add a new test, make sure the paths are returned but not check all the values which we expect.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think we can make it clear with better comments.

Comment on lines 23 to 27
* [Fixed a bug](https://github.com/lightningnetwork/lnd/pull/9334) when
constructing blinded payment paths for an invoice, we select up to MaxNumPaths
candidate routes and pass them forward. This process may result in running out
of paths if some of them are discarded for any reason later. This PR return all
possible routes and then do the MaxPath cap later on.
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
* [Fixed a bug](https://github.com/lightningnetwork/lnd/pull/9334) when
constructing blinded payment paths for an invoice, we select up to MaxNumPaths
candidate routes and pass them forward. This process may result in running out
of paths if some of them are discarded for any reason later. This PR return all
possible routes and then do the MaxPath cap later on.
[Fixed a bug](https://github.com/lightningnetwork/lnd/pull/9334) in constructing blinded payment paths for invoices. The restriction on the maximum number of paths is now applied later during pathfinding, ensuring that all potential paths are considered before filtering out ineligible ones. This allows for better utilization of the available paths for payments.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Much better thanks 👍

@MPins
Copy link
Contributor Author

MPins commented Feb 14, 2025

please make sure the check commits passes, you can run this script locally as well:

scripts/check-each-commit.sh upstream/master

Thanks ... now it is fixed.

Copy link
Contributor Author

@MPins MPins left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks, everything addressed!

Comment on lines 23 to 27
* [Fixed a bug](https://github.com/lightningnetwork/lnd/pull/9334) when
constructing blinded payment paths for an invoice, we select up to MaxNumPaths
candidate routes and pass them forward. This process may result in running out
of paths if some of them are discarded for any reason later. This PR return all
possible routes and then do the MaxPath cap later on.
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Much better thanks 👍

@@ -100,6 +100,9 @@ type BuildBlindedPathCfg struct {
// route.
MinNumHops uint8

// MaxNumPaths is the maximum number of blinded paths to select.
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You right, thanks!

@@ -547,35 +547,58 @@ func genBlindedRouteData(rand *rand.Rand) *record.BlindedRouteData {
// an example mentioned in this spec document:
// https://github.com/lightning/bolts/blob/master/proposals/route-blinding.md
// This example does not use any dummy hops.
// Added Dave to test MaxNumPaths restriction.
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think we can make it clear with better comments.

routes, err := ctx.router.FindBlindedPaths(
dave, 1000, probabilitySrc, &BlindedPathRestrictions{
MinDistanceFromIntroNode: 2,
NumHops: 2,
MaxNumPaths: 3,
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I didn't modify the test logic, but as I understand it, there's no probability difference affecting the expected order. So there is no reason to assert it.

@@ -3249,27 +3245,12 @@ func TestFindBlindedPathsWithMC(t *testing.T) {
"alice,charlie,dave",
})

// Change the MaxNumPaths to 1 to assert that only the best route is
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It is not a typo ... "fluke" and "flake" have different meanings, though they sound similar. Fluke is a lucky or random event that happens by chance.

@MPins
Copy link
Contributor Author

MPins commented Feb 14, 2025

@ziggie1984 notice that the last changes broke some tests ... I'll be fixing today! And let you know.

@MPins MPins force-pushed the issue-9076 branch 10 times, most recently from a92cc61 to 1d6fc66 Compare February 15, 2025 02:36
We now return all possible paths found by the pathfinder and only
remove those which do have a low probability.
MaxNumPaths restriction moved from FindBlindedPaths to BuildBlindedPaymentPaths
this way we can interact through all possibly routes when creating a new
blinded path.
MaxNumPaths restriction moved from FindBlindedPaths to BuildBlindedPaymentPaths
this way we have to fill the MaxNumPaths parameter when calling this function.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

route blinding: use all valid routes during path construction iteration instead of the currently capped set
5 participants