Hotspot segfaulting on Linux SPARC (original) (raw)

David Holmes david.holmes at oracle.com
Fri Apr 13 01:45:38 UTC 2018


On 13/04/2018 7:04 AM, Kim Barrett wrote:

On Apr 9, 2018, at 11:03 AM, Thomas Stüfe <thomas.stuefe at gmail.com> wrote:

VMVersion::determinefeatures() needs a code blob, and hence ResourceMark, on sparc and ppc and s390 I think.

VM_Version::determine_features calls the os-cpu specific VM_Version::platform_features. I don't see where a code blob is being used on Solaris sparc, or Linux sparc. ??

David

But I think the reason that only sparc fails is that VMVersion::determinefeatures() is called too early on Sparc (as part of os::initbeforeergo()). The other platforms call that function later in initialization (VMVersioninit() -> initglobals()). Could that be the error, and if yes, could VMVersion::determinefeatures() be delayed? This sounded familiar to me, so I looked into the history of that placement. initbeforeergo was created specifically to call determinefeatures that early: https://bugs.openjdk.java.net/browse/JDK-8133023



More information about the hotspot-dev mailing list