[11u] RFR(M): 8213754: PPC64: Add Intrinsics for isDigit/isLowerCase/isUpperCase/isWhitespace and 8215317: [GRAAL] unit test CheckGraalIntrinsics failed after 8213754

[Gustavo Romero]

Hi Goetz,

On 01/18/2019 06:34 AM, Lindenmaier, Goetz wrote:
> I'm fine with downporting this feature. I also will sponsor this.
> But I would prefer that you request downport for the two
> changes as-is, without merging them. I'll make sure they are
> pushed together so that nothing breaks.

Thanks a lot for sponsoring the changes!

Sure, I sent them separated to hotspot-compiler-dev ML, where the
were first reviewed:

https://mail.openjdk.java.net/pipermail/hotspot-compiler-dev/2019-January/032266.html
https://mail.openjdk.java.net/pipermail/hotspot-compiler-dev/2019-January/032267.html

I confess I was a bit in a paradox, since I understood from Vladimir [0]
that a combined version should be sent. But now we have both versions,
so I think it's fine either way :) Personally, now that I did both
ways, I think that requesting them separated is better, even for any
future bisecting, keeping the original authors and reviewers, etc.


> This also answers your questions wrt. the authors and reviewers:
> just leave them as-is.

Yep, it's automatically solved by the separated requests.


> You don't need to update the copyright.

Got it.


> So please send two RFRs. Send them to the lists where
> the changes were originally reviewed. Point out
> the tiny changes you made.
> 
> Once you sent the review, you can do the normal
> downport request as you did for 8214205, just add
> a link to your RFR mail in the pipermail archive as in
> https://bugs.openjdk.java.net/browse/JDK-8211856
> (where I also used the wrong mailing list :) )

One day I'll use the right list for backporting at the
first time I send the RFR! :)


Best regards,
Gustavo  

[0] https://mail.openjdk.java.net/pipermail/hotspot-compiler-dev/2019-January/032243.html

> Best regards,
>    Goetz.
> 
> 
>> -----Original Message-----
>> From: jdk-updates-dev <jdk-updates-dev-bounces at openjdk.java.net> On
>> Behalf Of Gustavo Romero
>> Sent: Freitag, 18. Januar 2019 07:11
>> To: jdk-updates-dev at openjdk.java.net; vladimir.kozlov at oracle.com; Doerr,
>> Martin <martin.doerr at sap.com>
>> Cc: Michihiro Horie <HORIE at jp.ibm.com>
>> Subject: [11u] RFR(M): 8213754: PPC64: Add Intrinsics for
>> isDigit/isLowerCase/isUpperCase/isWhitespace and 8215317: [GRAAL] unit
>> test CheckGraalIntrinsics failed after 8213754
>>
>> Hi,
>>
>> Could the following backport be reviewed, please?
>>
>> Webrev: http://cr.openjdk.java.net/~gromero/8213754_8215317_jdk11u/v1/
>>
>> This backport combines the following two changes in one:
>>
>> 8213754: PPC64: Add Intrinsics for
>> isDigit/isLowerCase/isUpperCase/isWhitespace
>> https://bugs.openjdk.java.net/browse/JDK-8213754
>> http://hg.openjdk.java.net/jdk/jdk/rev/7384e00d5860
>>
>> 8215317: [GRAAL] unit test CheckGraalIntrinsics failed after 8213754
>> https://bugs.openjdk.java.net/browse/JDK-8215317
>> http://hg.openjdk.java.net/jdk/jdk/rev/108a161aed93
>>
>> 8213754 adds intrinsics that use new instructions introduced by POWER9
>> CPUs
>> to speed up methods isDigit, isLowerCase, isUpperCase, and isWhitespace.
>> Change 8213754 is mostly PPC64-only but it does touch shared code, for
>> instance, in order to introduce/adapt in the libraries the methods in
>> question so they can be properly intrinsified. In doing so it can break the
>> Graal test HotspotTest.java/CheckGraalIntrinsics.java that checks for
>> intrinsics,
>> hence change 8215317, which adjusts that Graal test by including the new
>> intrinsics, is also backported and included in this change with 8213754.
>>
>> 8213754 applies almost cleanly: only a small tweak is necessary because the
>> hunk for ppc.ad file relies on some absent text around the change to be
>> applied. The absent text is from the Superword feature (a non-related
>> feature) that is not backported yet to 11u.
>>
>> 821531 was only adapted to match the inclusion of the intrinsics when JDK
>> is 11 or higher, instead of 12 or higher (original patch).
>>
>> The copyright dates were updated in all touched files to include 2019.
>>
>> This backportt was tested on POWER8 and POWER9 and no regressions were
>> observed.
>>
>> This backport was also tested on x86_64 with
>> ./test/hotspot/jtreg/compiler/{c1,c2,intrinsics} and
>> ./test/hotspot/jtreg/compiler/graalunit (with Graal compiler enabled) and
>> no regressions were observed too.
>>
>> I'm not sure which author I should keep since it's a combined changeset. Is
>> it ok to use me as the author?
>>
>> I also merged the reviewers. Is that ok as well?
>>
>> Thank you.
>>
>> Best regards,
>> Gustavo
>