Add support for HA Proxy Protocol TLV's by damianorenfer · Pull Request #5972 · eclipse-vertx/vert.x

damianorenfer · 2026-02-13T16:10:48Z

Motivation:

TLV's are supported in netty, but currently not accessible to vert.x users.
Here the TLV's are exposed in NetSocket#tlvs

This also answers to #5452

Conformance:
ECA signed

vietj · 2026-02-15T21:29:36Z

I'd rather expose that has a Map<Buffer, Buffer> instead (that map can be ordered if needed)

damianorenfer · 2026-02-16T08:51:11Z

Sure, I just updated the code accordingly

vietj · 2026-02-16T08:56:52Z

vertx-core/src/main/java/io/vertx/core/net/NetSocket.java

+   * the value contains the TLV value, mainly used for HA Proxy Protocol v2
+   */
+  @GenIgnore()
+  Map<Buffer, Buffer> tlvs();


why do we need that ? Quic does not yet support proxying at the moment and we are not sure it makes sense

that seems at the moment to be a TCP concern only

Do you see a better place to put this one ? The goal is to access TLV's in NetServer#connectHandler

sorry, it is fine here, I meant actually on QuicConnection it should apply to.

vietj · 2026-02-16T08:57:56Z

vertx-core/src/main/java/io/vertx/core/http/HttpConnection.java

+   * the value contains the TLV value, mainly used for HA Proxy Protocol v2
+   */
+  @GenIgnore()
+  Map<Buffer, Buffer> tlvs();


this should return null when it does not apply, e.g. HTTP/3

tmonney · 2026-02-16T09:01:31Z

The protocol technically allows for multiple occurrences of TLVs with the same type. It's probably not that common, but I am wondering if it makes sense to add unnecessary constraints. The underlying Netty decoder also uses a List, presumably for that reason.

damianorenfer · 2026-02-16T10:55:42Z

The protocol technically allows for multiple occurrences of TLVs with the same type. It's probably not that common, but I am wondering if it makes sense to add unnecessary constraints. The underlying Netty decoder also uses a List, presumably for that reason.

@vietj I was also not able to find anything about TLV uniqueness based on type here.
What do you think ?

vertx-core/src/main/java/io/vertx/core/http/HttpConnection.java

vertx-core/src/main/java/io/vertx/core/net/QuicConnection.java

vietj · 2026-02-19T12:48:21Z

The protocol technically allows for multiple occurrences of TLVs with the same type. It's probably not that common, but I am wondering if it makes sense to add unnecessary constraints. The underlying Netty decoder also uses a List, presumably for that reason.

@vietj I was also not able to find anything about TLV uniqueness based on type here. What do you think ?

I think we can use an Iterable<Map.Entry<Buffer, Buffer>> instead

damianorenfer · 2026-02-20T09:10:16Z

The protocol technically allows for multiple occurrences of TLVs with the same type. It's probably not that common, but I am wondering if it makes sense to add unnecessary constraints. The underlying Netty decoder also uses a List, presumably for that reason.

@vietj I was also not able to find anything about TLV uniqueness based on type here. What do you think ?

I think we can use an Iterable<Map.Entry<Buffer, Buffer>> instead

Sure, done!

vietj · 2026-03-04T12:34:03Z

vertx-core/src/main/java/io/vertx/core/http/HttpConnection.java

+   * This is mainly used for HA Proxy Protocol v2
+   */
+  @GenIgnore()
+  default Iterable<Map.Entry<Buffer, Buffer>> tlvs() {


I think the name should be more clear about what is provided, those are coming from HA proxy protocol, e.g. proxyProtocolV2HeaderTLVs seems more descriptive.

Sure, should we rename it everywhere ? also in NetSocket etc ?

yes it should be consistently renamed I think

vietj · 2026-03-16T07:57:38Z

vertx-core/src/main/java/io/vertx/core/http/HttpConnection.java

+   * This is mainly used for HA Proxy Protocol v2
+   */
+  @GenIgnore()
+  default Iterable<Map.Entry<Buffer, Buffer>> proxyProtocolV2HeaderTLVs() {


actually we could use List here as a better type since everything is buffered, no need to use a weaker collection type ?

There you go

TLV's are supported in netty, but currently not accessible to vert.x users. Here the TLV's are exposed in `NetSocket#proxyProtocolV2HeaderTLVs`

vietj · 2026-03-18T10:05:54Z

thanks @damianorenfer

drenferwl · 2026-03-18T10:17:01Z

@vietj Do you think we can backport this to 4.x ?
Would be really appreciated to have it in Quarkus 3

damianorenfer force-pushed the master branch from 87477f4 to c40907d Compare February 16, 2026 08:50

vietj reviewed Feb 16, 2026

View reviewed changes

vietj reviewed Feb 17, 2026

View reviewed changes

vertx-core/src/main/java/io/vertx/core/http/HttpConnection.java Outdated Show resolved Hide resolved

vietj reviewed Feb 17, 2026

View reviewed changes

vertx-core/src/main/java/io/vertx/core/net/QuicConnection.java Outdated Show resolved Hide resolved

damianorenfer force-pushed the master branch 4 times, most recently from c9094eb to c36e061 Compare February 20, 2026 09:09

damianorenfer force-pushed the master branch from c36e061 to 06af1da Compare March 2, 2026 08:58

vietj added this to the 5.1.0 milestone Mar 2, 2026

damianorenfer force-pushed the master branch 2 times, most recently from 3ff869f to a77685a Compare March 2, 2026 15:24

vietj reviewed Mar 4, 2026

View reviewed changes

vietj force-pushed the master branch from a77685a to 12b8562 Compare March 5, 2026 12:12

damianorenfer force-pushed the master branch 2 times, most recently from c292362 to c2c136e Compare March 15, 2026 15:59

damianorenfer requested a review from vietj March 15, 2026 15:59

vietj reviewed Mar 16, 2026

View reviewed changes

Add support for HA Proxy Protocol TLV's

75af8d5

TLV's are supported in netty, but currently not accessible to vert.x users. Here the TLV's are exposed in `NetSocket#proxyProtocolV2HeaderTLVs`

damianorenfer force-pushed the master branch from c2c136e to 75af8d5 Compare March 16, 2026 09:42

damianorenfer requested a review from vietj March 16, 2026 09:44

vietj merged commit 291fadc into eclipse-vertx:master Mar 18, 2026
1 check passed

Conversation

damianorenfer commented Feb 13, 2026

Uh oh!

vietj commented Feb 15, 2026

Uh oh!

damianorenfer commented Feb 16, 2026

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!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

tmonney commented Feb 16, 2026

Uh oh!

damianorenfer commented Feb 16, 2026

Uh oh!

Uh oh!

Uh oh!

vietj commented Feb 19, 2026

Uh oh!

damianorenfer commented Feb 20, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

damianorenfer Mar 6, 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!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

vietj commented Mar 18, 2026

Uh oh!

drenferwl commented Mar 18, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

damianorenfer Mar 6, 2026 •

edited

Loading

drenferwl commented Mar 18, 2026 •

edited

Loading