Hello,
Could you review the updated fix:
http://cr.openjdk.java.net/~alexsch/8252248/webrev.01/
- moving shared code to net_util_md.h is avoided
- INTERRUPT_SIGNAL is defined as SIGRTMAX - 2 instead of __SIGRTMAX -
2 for Linux in NativeThread.c
- "#include <pthread.h>" is moved out from "#ifdef" in NativeThread.c
- sigWakeup is changed to "#define WAKEUP_SIGNAL (SIGRTMAX - 2)" in
linux_close.c
At least test/jdk/java/net/Socket/asyncClose/AsyncClose.java touches all
four methods where the signal changes are used:
NativeThread.c
- Java_sun_nio_ch_NativeThread_init(JNIEnv *env, jclass cl)
- Java_sun_nio_ch_NativeThread_signal(JNIEnv *env, jclass cl, jlong thread)
linux_close.c
- __attribute((constructor)) init()
- closefd(int fd1, int fd2)
I run test/jdk/java/net tests on Ubuntu 18.04.4 and Alpine Linux 3.11.6
with musl libc 1.1.24.
There are 10 failed tests but they fail without the fix as well. I
believe that it is because of some network settings of my machine.
java/net/MulticastSocket/IPMulticastIF.java
java/net/MulticastSocket/NoSetNetworkInterface.java
java/net/MulticastSocket/Promiscuous.java
java/net/MulticastSocket/PromiscuousIPv6.java
java/net/MulticastSocket/SetGetNetworkInterfaceTest.java
java/net/MulticastSocket/SetLoopbackMode.java
java/net/MulticastSocket/SetOutgoingIf.java
java/net/MulticastSocket/Test.java
java/net/NetworkInterface/NetworkInterfaceRetrievalTests.java
java/net/ipv6tests/UdpTest.java
Thanks,
Alexander.
On 26.08.2020 13:42, Alan Bateman wrote:
On 25/08/2020 18:00, Alexander Scherbatiy wrote:
Hello,
Could your review the fix for the issue:
Bug: https://bugs.openjdk.java.net/browse/JDK-8252248
Fix: http://cr.openjdk.java.net/~alexsch/8252248/webrev.00/
:
The fix has been discussed on the portola-dev alias [2] where it was
pointed out that the fix can be reviewed in the mainline and it was
suggested to rename the INTERRUPT_SIGNAL and move its definition to
net_util_md.h.
The xxx_close.c sources are for the old SocketImpl and
DatagramSocketImpl implementations; they are aren't used by default
and will eventually go away. Maybe an option for the musl port is to
just not compile in the old socket implementations? It might be best
to start a discussion on nio-dev and net-dev about this as we
shouldn't be changing NativeThread.c (which is used for signalling in
the FileChannel implementation) to include net_util.h.
-Alan
