RFR(M): 8172049: [s390] Implement "JEP 270: Reserved Stack Areas for Critical Sections".
martin.doerr at sap.com
Fri Dec 30 09:19:14 UTC 2016
thank you very much for implementing this change.
I was wondering why you use the compare relation "<=" to compare the stack pointer with the reserved stack activation while all other platforms use "<".
Is it a mistake or is there a reason for it?
Besides that, I only have minor suggestions:
- reserved_stack_check always get Z_R14 as return pc which I think is the only useful argument. So I think it would be better to either remove the argument or to replace the lgr_if_needed by assert(return_pc==Z_R14...).
- In the ad file, the 2 parts of the safepoint poll are directly in juxtaposition with each other. They should better get combined.
Besides this, the change looks very good. I think you have also tested this change in our nightly tests which look good.
Thanks and best regards,
From: hotspot-runtime-dev [mailto:hotspot-runtime-dev-bounces at openjdk.java.net] On Behalf Of Lindenmaier, Goetz
Sent: Mittwoch, 28. Dezember 2016 08:41
To: hotspot-runtime-dev at openjdk.java.net
Subject: RFR(M): 8172049: [s390] Implement "JEP 270: Reserved Stack Areas for Critical Sections".
This implements JEP 270 on s390.
It's s390-only except for enabling the test, thus I need a sponsor.
More information about the hotspot-runtime-dev