JDK 9 fails to build on MIPS

David Holmes david.holmes at oracle.com
Fri Dec 2 01:05:08 UTC 2016

FYI I have filed an enhancement request for this:


On 2/12/2016 5:08 AM, Thomas Stüfe wrote:
> On Thu, Dec 1, 2016 at 2:36 PM, Matthias Klose <doko at ubuntu.com
> <mailto:doko at ubuntu.com>> wrote:
>     On 01.12.2016 14:20, David Holmes wrote:
>     > That said, given that this patch does nothing to fix the supported OpenJDK 9
>     where can I find the list of supported platforms? Is zero a
>     supported platform?
> Well, there is that:
> https://wiki.openjdk.java.net/display/Build/Supported+Build+Platforms
> But this seems to be a bit outdated, I do not see arm nor the new s390
> port. Should this be updated? If yes, should this be updated by the port
> maintainers?

The Build Group owns that wiki so they would have to update it. And of 
course the set of build platforms is not the same as the set of execute 
platforms. But yes Aarch64 and S390 should be added.

There doesn't seem to be a list of official OpenJDK runtime platforms, 
but it can be inferred by looking at the various Port projects, their 
JEPS and status.

>>     > platforms it would have to be considered an enhancement and we are closed to new
>>     > enhancements for JDK 9. The JDK 10 repos will be opening very soon though,
>>     same here:
>>     http://mail.openjdk.java.net/pipermail/jdk9-dev/2016-September/004819.html
>>     <http://mail.openjdk.java.net/pipermail/jdk9-dev/2016-September/004819.html>
>> How can these porting patches find it's way into jdk9? I see that
>> JEPs for three new hotspot ports were accepted after September.

The JDK 9 schedule was modified after that post [1] and we did not in 
fact enter RDP1 at that time. Additional features were allowed to be 
added under an extension mechanism, the deadline for completion of such 
is Dec 22. However that is for final code to reach the master forest and 
so we are effectively "closed" in the hotspot forest in terms of 
starting anything new. It may not be impossible but it would require 
significant justification and very quick work.

>> Is
>> this the right way to write a JEP for each zero port, or would these
>> fixes considered for backporting into jdk9?>
> I always thought the point of the zero port was to just work on any
> platform with a C++ compiler, and therefore does not constitute a real
> porting platform. So, any error on zero for any architecture would be
> handled as just a bug. I may be wrong here, so please correct me if I am
> wrong.

Zero "ports" are not considered OpenJDK ports so do not normally need 
JEPs (unless Zero itself is undergoing some major enhancement to support 
a new platform).

For the most part we rely on the Zero maintainers to report any issues 
that other changes cause with Zero, or where Zero needs to be updated to 
support platform X and we do work together to get things fixed. That 
said, if someone tries to build Zero for platform X, and that in turn 
requires changes to platform specific VM code unrelated to Zero itself, 
then that is a different matter.

A change, like has been requested, which simply changes to using more 
standard mechanisms is obviously an acceptable request (unfortunately 
just not now for JDK 9) - however if it required an "ifdef MIPS" then 
that would not be acceptable.

The question of "backporting to 9" can't be answered until there is a "9 
update" project and it defines what its back porting policy is.


[1] http://mail.openjdk.java.net/pipermail/jdk9-dev/2016-October/005092.html

> ..Thomas
>     Matthias

More information about the hotspot-dev mailing list