[aarch64-port-dev ] Fwd: 8u-aarch64 : Backport 8163363: AArch64: Stack size in tools/launcher/Settings.java needs to be adjusted

[Andrew Dinn]

On 06/08/2019 18:29, Aleksey Shipilev wrote:
> On 8/6/19 6:48 PM, Andrew Dinn wrote:
>> On 06/08/2019 17:35, Aleksey Shipilev wrote:
>>> On 8/6/19 10:49 AM, Andrew Dinn wrote:
>>>> The backport requires a minor tweak to the original patch as per the
>>>> following webrev:
>>>>
>>>>   http://cr.openjdk.java.net/~adinn/8163363-jdk8u/webrev.00/
>>>
>>> *) I don't understand, why "ppc64le" case was added?
>>>
>>>   78         if (getArch().equals("ppc64") || getArch().equals("ppc64le")) {
>>
>> That is because I back-ported the original patch and it was included in
>> that patch and went in automatically. Is ppc64le not a thing in jdk8u?
>> If so I can remove it.
> 
> It is not in the 8u code, backport should not add it.

Ah, ok, I see what is wrong here. This is my cock-up. I updated the root
tree (hg pull+update) rather than pulling all subtrees (run
get_source.sh). So, I ended up applying and tweaking the patch against a
stale checkout. The pp64cle check /is/ supposed to be present before and
after patching (it is present in upstream jdk8u).

Even with the latest tree the original patch still does not apply
cleanly even after path demangling because of:

  1) stack size arguments supplied as strings not numbers
  2) different local Strings used in comparison checks.

So, I corrected for both of these differences. Here is a second webrev
as a sanity check (it includes the omitted comment):

  http://cr.openjdk.java.net/~adinn/8163363-jdk8u/webrev.01/

Testing:

'make test core_tools' passes ok with this new version.

regards,


Andrew Dinn
-----------
Senior Principal Software Engineer
Red Hat UK Ltd
Registered in England and Wales under Company Registration No. 03798903
Directors: Michael Cunningham, Michael ("Mike") O'Neill, Eric Shander