RFR 8043954: Fix behavior difference of connect() for AIX (original) (raw)

Volker Simonis volker.simonis at gmail.com
Tue Jun 3 12:48:32 UTC 2014

Hi Jonathan,

thanks for fixing this! I've looked at the change and it looks good to me (but I'm not a reviewer). The only minor flaw I found is that you declare the helper variable 'int rc = -1' but never assign it. Instead you could just return '-1' directly where you currently return 'rc' and remove 'rc' altogether.

I'm currently still doing test build and I'll run some tests. I'll let you know if I should see any problems.

By the way - does this change fix a real problem or is it just an improvement of the current implementation (just curious)?

Thank you and best regards, Volker

On Tue, Jun 3, 2014 at 11:51 AM, Jonathan Lu <luchsh at linux.vnet.ibm.com> wrote:

Hello,

Could I have following patch reviewed for bug 8034954 ? http://cr.openjdk.java.net/~luchsh/JDK-8043954/ The patch is to fix a behavior difference of connect() API for AIX platform, according to the documentation, http://www-01.ibm.com/support/knowledgecenter/sswaix71/com.ibm.aix.commtrf2/connect.htm?lang=en On AIX, when connect() got interrupted by signal, the underlying connection will be made asynchronously, "EINTR The attempt to establish a connection was interrupted by delivery of a signal that was caught; the connection will be established asynchronously." This fix tries to poll() for successfully established connection or error in the similar way as NETTimeout(). Thanks Jonathan

More information about the core-libs-dev mailing list