RFR(M): 8172049: [s390] Implement "JEP 270: Reserved Stack Areas for Critical Sections".

[Lindenmaier, Goetz]

Hi Martin 

thanks for the review in the holiday season :)
New webrev:
http://cr.openjdk.java.net/~goetz/wr16/8172049-s390_reservedStack/webrev.02/

> 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?
I changed it to <. Tests are running.

> 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...).
Changed to assert. I think the argument makes it obvious that the
existing return pc must be set when branching to the exception.

> - In the ad file, the 2 parts of the safepoint poll are directly in juxtaposition
> with each other. They should better get combined.
Fixed. Yes, looks better.

Best regards,
  Goetz


> 
> 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,
> Martin
> 
> 
> -----Original Message-----
> 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".
> 
> Hi,
> 
> This implements JEP 270 on s390.
> It's s390-only except for enabling the test, thus I need a sponsor.
> Please review.
> http://cr.openjdk.java.net/~goetz/wr16/8172049-
> s390_reservedStack/webrev.01/
> 
> Best regards,
>   Goetz.
>