Add CandidateDescriptorV3 support to experimental validator by AlexandruCihodaru · Pull Request #11306 · paritytech/polkadot-sdk

AlexandruCihodaru · 2026-03-09T09:09:01Z

Adds CandidateDescriptorV3 support to the experimental validator-side collator protocol.

… validator side * Enable V3 candidate descriptor handling in the experimental collator * Add tests for V3 acceptance, scheduling parent leaf enforcement, backwards compatibility, and unknown version rejection Signed-off-by: Alexandru Cihodaru <alexandru.cihodaru@parity.io>

AlexandruCihodaru · 2026-03-09T09:21:33Z

/cmd prdoc --audience node_dev --bump minor

…e_dev --bump minor'

sandreim · 2026-03-09T11:55:20Z

+}
+
+#[tokio::test]
+// Test that a V3 descriptor is treated as V1 via backward compatibility


What about v2 ?

it can't be treated as v1, but indeed would be good to test a v2 descriptor is fine with v3 enabled

…on initialization Signed-off-by: Alexandru Cihodaru <alexandru.cihodaru@parity.io>

…idator_candidate_v3

…d slot Signed-off-by: Alexandru Cihodaru <alexandru.cihodaru@parity.io>

…idator_candidate_v3

AlexandruCihodaru · 2026-03-13T09:17:30Z

/cmd fmt

Extract LeafSchedulingInfo, slot extraction, and scheduling parent validation into shared functions in lib.rs. Fix V3 tests that incorrectly asserted scheduling_parent instead of relay_parent for PVD requests. Signed-off-by: Alexandru Cihodaru <alexandru.cihodaru@parity.io>

alindima

LGTM, some suggestions for the tests remain

alindima · 2026-03-16T11:03:04Z

+}
+
+#[tokio::test]
+// Test that a V3 descriptor is treated as V1 via backward compatibility


it can't be treated as v1, but indeed would be good to test a v2 descriptor is fine with v3 enabled

eskimor

Good work!

Some nits, but overall looks good!

Signed-off-by: Alexandru Cihodaru <alexandru.cihodaru@parity.io>

…idator_candidate_v3

Signed-off-by: Alexandru Cihodaru <alexandru.cihodaru@parity.io>

See discussion here: #11306 (comment) also fixes a bug, where for held off advertisements we would assume v3 descriptors are never used.

See discussion here: #11306 (comment) also fixes a bug, where for held off advertisements we would assume v3 descriptors are never used. (cherry picked from commit f02ad9e)

…idator_candidate_v3

AlexandruCihodaru · 2026-03-20T21:50:13Z

/cmd fmt

Signed-off-by: Alexandru Cihodaru <alexandru.cihodaru@parity.io>

eskimor

Looks good on a quick pass, a few nits from last time don't seem addressed yet though.

…idator_candidate_v3

Signed-off-by: Alexandru Cihodaru <alexandru.cihodaru@parity.io>

eskimor · 2026-03-23T16:31:29Z

 						},
 					};

 				let core = match self.get_our_core(sender, leaf, session_index).await {


Pre-existing but this looks wrong. We should determine our core based on the scheduling parent - not the active leaf. Otherwise calling this in the loop is also completely pointless: We enforce the session index to be the same for all ancestors, so this call will return the same core always.

Related: Requesting the session index for each ancestor may be sensible, if we use it to enforce the same session constraint (but that should already be enforced elsewhere), otherwise it is completely pointless to call here either.

Given that this is pre-existing, not opposing a follow-up to keep diffs small.

Pre-existing but this looks wrong. We should determine our core based on the scheduling parent - not the active leaf. Otherwise calling this in the loop is also completely pointless:

Indeed, this should be the ancestor.

Given that this is pre-existing, not opposing a follow-up to keep diffs small.

I agree

polkadot/node/network/collator-protocol/src/validator_side_experimental/collation_manager/mod.rs whould this do the trick?

polkadot/node/network/collator-protocol/src/validator_side_experimental/collation_manager/mod.rs whould this do the trick?

I think so, although the test is setting the same core for all relay parents, so it would pass even without the fix. Moreover, maybe we could add a regression test that exercises the core index check instead of exposing this implementation detail as a test-only getter function.

Considering that this PR is already open for quite a while and ready for merge, let's merge it and open a subsequent PR later (this bug has already been here for a while)

@eskimor can you approve?

eskimor

I found two issues, which should be addressed in a followup:

Pre-existing, mixing up leaf with ancestor fetches.
We validate the pvd on advertisements - but we use the scheduling parent. We should either just skip that check or use the relay parent if available (v3 protocol as introduced by Alin).

eskimor · 2026-03-24T10:13:12Z

 		futures::join!(
 			state.handle_fetched_collation(&mut sender, (adv, res)),
-			test_state.assert_pvd_request(adv, None)
+			test_state.assert_pvd_request(adv, None, adv.scheduling_parent)


This will fail with v3. We should either skip this check entirely or use the relay parent Alin introduced in the advertisement.

alindima · 2026-03-24T10:42:07Z

We validate the pvd on advertisements - but we use the scheduling parent. We should either just skip that check or use the relay parent if available (v3 protocol as introduced by Alin).

Where?

alindima · 2026-03-24T10:43:38Z

We validate the pvd on advertisements - but we use the scheduling parent. We should either just skip that check or use the relay parent if available (v3 protocol as introduced by Alin).

Where?

You mean as part of the CanSecond request? I'll add the relay parent there on the prospective parachains PR

eskimor · 2026-03-24T10:46:24Z

Sorry looked deeper, we only check that for v2. So it is fine. @AlexandruCihodaru

Adds CandidateDescriptorV3 support to the experimental validator-side collator protocol. Fixes: #11084 --------- Signed-off-by: Alexandru Cihodaru <alexandru.cihodaru@parity.io> Co-authored-by: cmd[bot] <41898282+github-actions[bot]@users.noreply.github.com> (cherry picked from commit a1eb95e)

paritytech-release-backport-bot · 2026-03-24T11:13:08Z

Successfully created backport PR for stable2603:

[stable2603] Backport #11306 #11478

Backport #11306 into `stable2603` from AlexandruCihodaru. See the [documentation](https://github.com/paritytech/polkadot-sdk/blob/master/docs/BACKPORT.md) on how to use this bot.  Signed-off-by: Alexandru Cihodaru <alexandru.cihodaru@parity.io> Co-authored-by: Alexandru Cihodaru <40807189+AlexandruCihodaru@users.noreply.github.com> Co-authored-by: cmd[bot] <41898282+github-actions[bot]@users.noreply.github.com>

Adds CandidateDescriptorV3 support to the experimental validator-side collator protocol. Fixes: #11084 --------- Signed-off-by: Alexandru Cihodaru <alexandru.cihodaru@parity.io> Co-authored-by: cmd[bot] <41898282+github-actions[bot]@users.noreply.github.com>

AlexandruCihodaru requested review from alindima, eskimor, sandreim and tdimitrov March 9, 2026 09:09

AlexandruCihodaru self-assigned this Mar 9, 2026

AlexandruCihodaru added T8-polkadot This PR/Issue is related to/affects the Polkadot network. T20-low-latency for issues and PRs related to Low Latency work labels Mar 9, 2026

Update from github-actions[bot] running command 'prdoc --audience nod…

02055b4

…e_dev --bump minor'

sandreim reviewed Mar 9, 2026

View reviewed changes

eskimor added the A4-backport-stable2603 Pull request must be backported to the stable2603 release branch label Mar 9, 2026

AlexandruCihodaru added 3 commits March 9, 2026 22:47

[Collator Protocol] Move CandidateDescriptorV3 feature check to sessi…

bd07bd7

…on initialization Signed-off-by: Alexandru Cihodaru <alexandru.cihodaru@parity.io>

Merge remote-tracking branch 'origin' into acihodaru/experimental_val…

0331aa2

…idator_candidate_v3

Merge remote-tracking branch 'origin' into acihodaru/experimental_val…

ea2f122

…idator_candidate_v3

eskimor reviewed Mar 12, 2026

View reviewed changes

Comment thread polkadot/node/network/collator-protocol/src/validator_side_experimental/common.rs Outdated

AlexandruCihodaru added 2 commits March 13, 2026 09:46

collator-protocol: Validate V3 scheduling parent against last finishe…

2eda45f

…d slot Signed-off-by: Alexandru Cihodaru <alexandru.cihodaru@parity.io>

Merge remote-tracking branch 'origin' into acihodaru/experimental_val…

4374b00

…idator_candidate_v3

Update from github-actions[bot] running command 'fmt'

6ee0b01

alindima reviewed Mar 13, 2026

View reviewed changes

alindima approved these changes Mar 16, 2026

View reviewed changes

eskimor approved these changes Mar 16, 2026

View reviewed changes

alindima mentioned this pull request Mar 17, 2026

add relay parent to advertisement v3 #11393

Merged

AlexandruCihodaru added 3 commits March 17, 2026 13:41

fix some typos

4d8b0bb

Signed-off-by: Alexandru Cihodaru <alexandru.cihodaru@parity.io>

Merge remote-tracking branch 'origin' into acihodaru/experimental_val…

1ff933c

…idator_candidate_v3

test that the parent of the sp from the current slot is accepted

da1f1a6

Signed-off-by: Alexandru Cihodaru <alexandru.cihodaru@parity.io>

pgherveou added this to [preview] release tracker Mar 18, 2026

github-project-automation Bot moved this to Todo in [preview] release tracker Mar 18, 2026

github-merge-queue Bot pushed a commit that referenced this pull request Mar 18, 2026

add relay parent to advertisement v3 (#11393)

f02ad9e

See discussion here: #11306 (comment) also fixes a bug, where for held off advertisements we would assume v3 descriptors are never used.

pgherveou removed this from [preview] release tracker Mar 19, 2026

Merge remote-tracking branch 'origin' into acihodaru/experimental_val…

b1b02ae

…idator_candidate_v3

github-actions Bot and others added 2 commits March 20, 2026 21:53

Update from github-actions[bot] running command 'fmt'

c6e225e

address comments

5dc858e

Signed-off-by: Alexandru Cihodaru <alexandru.cihodaru@parity.io>

eskimor reviewed Mar 23, 2026

View reviewed changes

AlexandruCihodaru added 2 commits March 23, 2026 13:57

Merge remote-tracking branch 'origin' into acihodaru/experimental_val…

3ed7d90

…idator_candidate_v3

address feedback

354ae36

Signed-off-by: Alexandru Cihodaru <alexandru.cihodaru@parity.io>

eskimor reviewed Mar 23, 2026

View reviewed changes

Comment thread polkadot/node/network/collator-protocol/src/validator_side_experimental/common.rs Outdated

AlexandruCihodaru and others added 2 commits March 23, 2026 17:40

remove manual hash implementation for Advertisement

3c184d8

Signed-off-by: Alexandru Cihodaru <alexandru.cihodaru@parity.io>

Merge branch 'master' into acihodaru/experimental_validator_candidate_v3

e25b50f

eskimor requested changes Mar 23, 2026

View reviewed changes

eskimor approved these changes Mar 24, 2026

View reviewed changes

AlexandruCihodaru added this pull request to the merge queue Mar 24, 2026

Merged via the queue into master with commit a1eb95e Mar 24, 2026
257 of 258 checks passed

AlexandruCihodaru deleted the acihodaru/experimental_validator_candidate_v3 branch March 24, 2026 11:12

paritytech-release-backport-bot Bot mentioned this pull request Mar 24, 2026

[stable2603] Backport #11306 #11478

Merged

pgherveou added this to [preview] release tracker Mar 31, 2026

github-project-automation Bot moved this to Done in [preview] release tracker Mar 31, 2026

Conversation

AlexandruCihodaru commented Mar 9, 2026

Uh oh!

AlexandruCihodaru commented Mar 9, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

AlexandruCihodaru commented Mar 13, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

alindima left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

eskimor left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

AlexandruCihodaru commented Mar 20, 2026

Uh oh!

eskimor left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AlexandruCihodaru Mar 23, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

eskimor left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

alindima commented Mar 24, 2026

Uh oh!

alindima commented Mar 24, 2026

Uh oh!

AlexandruCihodaru Mar 23, 2026 •

edited

Loading