RFR(S): 8193443: [s390]: EncodeISOArray generates wrong vector code

[Schmidt, Lutz]

Thank you, Goetz and Martin!
That’s what I call “no unnecessary delay”. 

Regards, Lutz


 

On 14.12.2017, 12:07, "Doerr, Martin" <martin.doerr at sap.com> wrote:

    Hi Lutz,
    
    reviewed and pushed to jdk/jdk.
    
    Best regards,
    Martin
    
    
    -----Original Message-----
    From: hotspot-compiler-dev [mailto:hotspot-compiler-dev-bounces at openjdk.java.net] On Behalf Of Lindenmaier, Goetz
    Sent: Donnerstag, 14. Dezember 2017 10:31
    To: Schmidt, Lutz <lutz.schmidt at sap.com>; hotspot-compiler-dev at openjdk.java.net
    Subject: RE: RFR(S): 8193443: [s390]: EncodeISOArray generates wrong vector code
    
    Hi Lutz, 
    
    change looks good. (Had to search a bit to find the essential change :))
    
    If you want to get it into jdk10, you should push this to jdk/jdk today.
    
    Best regards,
      Goetz.
    
    > -----Original Message-----
    > From: hotspot-compiler-dev [mailto:hotspot-compiler-dev-
    > bounces at openjdk.java.net] On Behalf Of Schmidt, Lutz
    > Sent: Mittwoch, 13. Dezember 2017 17:17
    > To: hotspot-compiler-dev at openjdk.java.net
    > Subject: RFR(S): 8193443: [s390]: EncodeISOArray generates wrong vector
    > code
    > 
    > Dear all,
    > 
    > 
    > 
    > I would like to request reviews for this s390-only bux fix:
    > 
    > 
    > 
    > Bug:    https://bugs.openjdk.java.net/browse/JDK-8193443
    > 
    > Webrev: http://cr.openjdk.java.net/~lucy/webrevs/8193443.00/index.html
    > 
    > 
    > 
    > The condition code setting from vector compare instructions differs slightly
    > from that of scalar instructions. Distinct condition code enum values were
    > introduced, related conditional branch aliases were defined and, of course,
    > the buggy code was fixed.
    > 
    > 
    > 
    > The previously failing jtreg tests now pass.
    > 
    > 
    > 
    > I would very much appreciate if this fix would make it into Java 10.
    > 
    > 
    > 
    > Thank you!
    > 
    > Lutz
    > 
    >