On Thu, 16 Nov 2023 08:39:32 GMT, Matthias Baesken <[email protected]> wrote:
> There are a few places in the JDK C coding where the setsocktopt return value
> is not handled but better should be handled.
src/java.base/windows/native/libnio/ch/Net.c line 160:
> 158:
> 159: /* IPV6_V6ONLY is true by default */
> 160: /* attempt to disable IPV6_V6ONLY to ensure dual-socket support;
> ignore errors */
You can the previous comment at L159.
src/java.base/windows/native/libnio/ch/WindowsAsynchronousServerSocketChannelImpl.c
line 134:
> 132: SOCKET s2 = (SOCKET)jlong_to_ptr(acceptSocket);
> 133:
> 134: if (setsockopt(s2, SOL_SOCKET, SO_UPDATE_ACCEPT_CONTEXT, (char
> *)&s1, sizeof(s1)) < 0) {
I think this goes back to when the JDK targeted Windows releases where this
wasn't supported or there were non-MS LSPs installed. It should be okay to
change now.
-------------
PR Review Comment: https://git.openjdk.org/jdk/pull/16684#discussion_r1395369394
PR Review Comment: https://git.openjdk.org/jdk/pull/16684#discussion_r1395368258
