RFR(s): 8149036: Add UL tracing for thread related events at os level
coleen.phillimore at oracle.com
Mon Feb 22 23:09:45 UTC 2016
Oh, also, I am willing to sponsor this.
On 2/22/16 5:45 PM, Coleen Phillimore wrote:
> On 2/22/16 3:53 PM, Coleen Phillimore wrote:
>> Hi Thomas,
>> I think the logging conversion looks really good. The errno stuff is
>> very ugly though. Can you call strerrno_r - the thread safe
>> version? I didn't see discussion around this.
> Hi Thomas, I was talking to Kim. You should use strerror() and not
> add this code to os.cpp. The rationale is in this bug:
> It seems like the errors would have standard errno at this point and
> wouldn't be susceptible to the race in strerror().
> + log.warning("Failed to start thread - pthread_create failed
> (%s) for attributes: %s.",
> + os::errno_name(errno),
> os::Posix::describe_pthread_attr(buf, sizeof(buf), &attr));
> It seems the chance of a race creating different errno at this early
> point in the jvm is very low.
>> On 2/22/16 11:54 AM, Thomas Stüfe wrote:
>>> Dear all,
>>> please take a look at this proposed addition to UL. This adds a
>>> number of
>>> trace points to thread creation. In detail:
>>> - it traces thread creation and thread creation errors, including
>>> attributes (for Posix platforms)
>>> - it traces stack location and creation/removal of stack guard pages.
>>> This all was first AIX-only tracing, but I converted this to UL and
>>> made it
>>> available on all platforms.
>>> Bug: https://bugs.openjdk.java.net/browse/JDK-8149036
>>> Note also that I added a helper function, os::errno_name(), which is
>>> a very
>>> simple replacement for strerror() without its problems (thread safety,
>>> unwanted localizations...).
>>> What do you think?
>>> Kind Regards, Thomas
More information about the hotspot-runtime-dev