RFR: 8372159: HttpClient SelectorManager thread could be a VirtualThread [v2]

[Jaikiran Pai]

On Wed, 19 Nov 2025 19:03:28 GMT, Daniel Fuchs <dfuchs at openjdk.org> wrote:

>> Each HttpClient instance creates an additional platform thread for its SelectorManager. With recent updates to NIO/VirtualThreads that thread could now become a VirtualThread. This would avoid having each HttpClient instance use up one platform thread.
>> This is similar to what was done for the HttpClient QuicSelectorThread in [JDK-8369920](https://bugs.openjdk.org/browse/JDK-8369920).
>> This should be transparent for users of the API. 
>> An undocumented internal system property is introduced that can revert the change in case of unforeseen trouble.
>
> Daniel Fuchs has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Revert changes to SelectorManager::shutdown

src/java.net.http/share/classes/jdk/internal/net/http/HttpClientImpl.java line 131:

> 129: 
> 130:     static final UseVTForSelector USE_VT_FOR_SELECTOR =
> 131:         Utils.useVTForSelector("jdk.internal.httpclient.tcp.selector.useVirtualThreads", "default");

Hello Daniel, I vaguely remember we discussed whether or not to use a value called "default" for this property. But I don't remember what we decided. Looking at this now, I am wondering whether we should just do something like:

System.getProperty("jdk.internal.httpclient.tcp.selector.useVirtualThreads", "true")

to keep it simple?

test/jdk/java/net/httpclient/ReferenceTracker.java line 393:

> 391:     }
> 392: 
> 393:     private boolean isSelectorManager(Thread t) {

The copyright year on this file will need an update.

test/jdk/java/net/httpclient/http2/H2SelectorVTTest.java line 145:

> 143:             throw new AssertionError("Unexpected null sslContext");
> 144:         }
> 145:         // create an H3 only server

The use of "H3" here looks like a typo.

test/jdk/java/net/httpclient/http2/H2SelectorVTTest.java line 175:

> 173:             final HttpRequest req1 = reqBuilder.copy().GET().build();
> 174:             System.out.println("\nIssuing request: " + req1);
> 175:             final HttpResponse<?> resp1 = client.send(req1, BodyHandlers.ofString());

Nit - this line and the other 3 send() lines could perhaps use `BodyHandlers.discarding()`? If you prefer to use `ofString()`, that's fine too - it wasn't clear to me if this was an oversight.

test/jdk/java/net/httpclient/http2/H2SelectorVTTest.java line 195:

> 193:     }
> 194: 
> 195:     private static void assertSelectorThread(HttpClient client) {

I think a comment on what this method checks would be useful for future.

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/28395#discussion_r2545102282
PR Review Comment: https://git.openjdk.org/jdk/pull/28395#discussion_r2545125815
PR Review Comment: https://git.openjdk.org/jdk/pull/28395#discussion_r2545108417
PR Review Comment: https://git.openjdk.org/jdk/pull/28395#discussion_r2545116878
PR Review Comment: https://git.openjdk.org/jdk/pull/28395#discussion_r2545124094