JDK-8133416: [TESTBUG] Remove @ignore for closed/runtime/4345157/Prog.java (original) (raw)

Daniel D. Daugherty daniel.daugherty at oracle.com
Fri Nov 6 19:07:40 UTC 2015

On 11/4/15 3:33 AM, Stas Smirnov wrote:

Hello,

please review this fix for JDK-8133416. The test has been fixed and rewritten in Java to start using native support mechanism in Jtreg. Also, its native part has been modified to avoid compilation on inappropriate platforms and also includes improvements. Unfortunately the JBS issue is not visible to the community as well as the original version of the test sources. JBS: https://bugs.openjdk.java.net/browse/JDK-8133416 webrev: http://cr.openjdk.java.net/~stsmirno/8133416/webrev.00

make/test/JtregNative.gmk No comments.

test/runtime/ThreadSignalMask/Prog.java No comments.

test/runtime/ThreadSignalMask/ThreadSignalMask.java No comments.

test/runtime/ThreadSignalMask/exeThreadSignalMask.c General: I think the indent standard for 'C' was also two spaces, but I could be wrong...

 L39 char path[PATH_MAX];
     Recommend '+1' for NULL termination space.

 L45: extern void exit(int);
     Hard to believe this isn't in a header file...

 L50: fputs("Error occurred while closing handle", stderr);
     Will this output include a newline?

 L62:     char lib[PATH_MAX];
     Recommend '+1' for NULL termination space.

 L63: snprintf(lib, sizeof (lib), "%s/lib/sparcv9/server/libjvm.so",

path); I don't think snprintf() will NULL terminate if the buffer is full. Add: lib[PATH_MAX] = '\0';

 L180:     strncpy(path, argv[1], PATH_MAX);
     strncpy() won't NULL terminate if buffer is full.
     Add: path[PATH_MAX] = '\0';

I think that's it...

Dan

Tested: the fix has been tested on all platforms with the hotspot/test/runtime testset

More information about the hotspot-runtime-dev mailing list