Java 25 Advanced Network (PPSA99871)

[karimibrahim07]

Existing thread

• I searched this category for the title ID and no thread exists for this title on any platform.

Title

Java 25 Advanced Network (PPSA99871)

Content version tested

01.000.002, release Advanced_Network_Probe-v01.000.002 of mateo-cogeanu/ps5-java25-homebrew (PPSA99871-advanced-network-pass-01.000.002.zip, SHA-256 8c95655c87b1433c291170d606c22e740b2099b4e8a28ed2906905ba534c32cc, matching the release's Advanced_Network_Probe-SHA256SUMS.txt). It's one of the project's OpenJDK 25 probe apps, like PPSA99869 (#5501) and PPSA99868 (#5972). The release reports all six rows passing on a PS5 (firmware 13.60). It needs Internet access for the HTTPS row and no helper machine.

AnyPS5 commit

main 44df21632, with #5918, which only changes thread stacks

System tested

Windows 11 Pro (26300), RTX 3060, WinLibs GCC 15.2.0 r7, Release

Furthest point

Menu

What is shown

Its menu with six rows. The first two initialize the runtime and create the VM (JNI_CreateJavaVM returned 0). The HTTPS row (https://example.com/ with the bundled CA store) loads the trust store and initializes TLS, then stops:

STAGE: HTTPS: connect HTTPS with certificate and hostname verification
NETWORK FAILURE: Https25: java.lang.Exception: HTTPS stage connect HTTPS with certificate and hostname verification: java.net.SocketException: Protocol not available

The title doesn't exit; the remaining rows can still be run. Only the menu was used; audio wasn't checked.

Current blocker

setsockopt(IPPROTO_IPV6, IPV6_V6ONLY, 0) on Java's AF_INET6 sockets returns ENOPROTOOPT. #6663 handles IPV6_V6ONLY. With it, the IPv6 TCP/UDP loopback row, the network load row and DestroyJavaVM pass. The HTTPS row then connects to the IPv4 address of example.com through ::ffff:, but its first read fails with Software caused connection abort. Here the interpreter-only VM takes longer between connect and the TLS ClientHello than the server keeps an idle connection open; a plain connection that waits 15 s before sending gets a reset. That row may pass on a faster machine or once the VM's per-call tracing costs less.

How it was run

relinker --windows --to-intel --trap-sdk-syscall-stubs --registry eboot.elf eboot.exe on the unwrapped eboot.bin and sce_module/libc.prx. Rows were run with Enter and the arrow keys posted to the window. The title writes its stages to /download0/java25-startup.log.

AI-assisted.