SIGFPE on MacOSX
Bob Vandette
bob.vandette at oracle.com
Tue Oct 27 14:37:12 UTC 2015
> On Oct 26, 2015, at 10:21 PM, David Holmes <David.Holmes at oracle.com> wrote:
>
> Hi Bob,
>
> On 21/10/2015 3:02 AM, Bob Vandette wrote:
>> Is there a fix for this __commpage_gettimeofday issue? According to this post, It may be related to time running
>> backwards on MacOSX.
>
> I'm unclear what the issue actually is. If external native code raises SIGFPE then the process will abort. In this case the VM detects and reports it, but of course still aborts.
>
I don’t know what the real issue is either. The only reason I posted this question was to find out if anyone
could provide more insight into the comment from the mail archives.
"This happens from time to time. It looks like wrong usage of syscalls in the MacOSX libc,
because it seems to happen if the clock on Mac goes backwards. This happens more often in
virtual machines, because time synchronization of virtual machine and host clock happens more
often because of virtual processor scheduling issues, so clocks differ more often and need
to be corrected by NTP. As far as I remember, there was already a patch on hotspot mailing
list to use more modern APIs."
Bob.
> David
>
>> http://mail-archives.apache.org/mod_mbox/lucene-dev/201405.mbox/%3C00bc01cf6524$143920b0$3cab6210$@thetaphi.de%3E <http://mail-archives.apache.org/mod_mbox/lucene-dev/201405.mbox/%3C00bc01cf6524$143920b0$3cab6210$@thetaphi.de%3E>
>>
>> According to the reporter (CC’d), he doesn’t believe his system is virtualized and has seen backward time reporting with
>> and without NTP turned on.
>>
>>
>> Bob.
>>
>>
>>
>>> On Oct 20, 2015, at 12:24 AM, Michael Elges <michael.elges at oracle.com <mailto:michael.elges at oracle.com>> wrote:
>>>
>>> Hello Bob,
>>>
>>> I moved to Java 8 and it still crashes.
>>>
>>> Thread 42 Crashed:: Java: ContinuationBayeux-0
>>> 0 libsystem_kernel.dylib 0x00007fff93e1d286 __pthread_kill + 10
>>> 1 libsystem_c.dylib 0x00007fff957a59b3 abort + 129
>>> 2 libjvm.dylib 0x000000010606e14f os::abort(bool) + 25
>>> 3 libjvm.dylib 0x000000010618ec2c VMError::report_and_die() + 2250
>>> 4 libjvm.dylib 0x0000000105de0237 report_vm_error(char const*, int, char const*, char const*) + 84
>>> 5 libjvm.dylib 0x00000001060c91dc SharedRuntime::continuation_for_implicit_exception(JavaThread*, unsigned char*, SharedRuntime::ImplicitExceptionKind) + 530
>>> 6 libjvm.dylib 0x000000010606fc77 JVM_handle_bsd_signal + 872
>>> 7 libjvm.dylib 0x000000010606c057 signalHandler(int, __siginfo*, void*) + 47
>>> 8 libsystem_platform.dylib 0x00007fff943b0f1a _sigtramp + 26
>>> 9 libsystem_kernel.dylib 0x00007fff93e186b3 __commpage_gettimeofday + 67
>>
>>
More information about the hotspot-dev
mailing list