[15] RFR (L): 8241436: C2: Factor out C2-specific code from MacroAssembler

Vladimir Kozlov vladimir.kozlov at oracle.com
Tue Mar 24 16:14:17 UTC 2020


Good. Thank you for testing.

Vladimir K

On 3/24/20 8:37 AM, Vladimir Ivanov wrote:
> Thanks for the review, Vladimir.
> 
>> Did you build and test without C2: configure --with-jvm-features=-compiler2 ?
> 
> I didn't perform extensive testing, but I verified that the build works and ran some smoke tests that the JVM isn't 
> completely broken.
> 
> Best regards,
> Vladimir Ivanov
> 
>> On 3/23/20 6:49 AM, Vladimir Ivanov wrote:
>>> http://cr.openjdk.java.net/~vlivanov/8241436/webrev.00/
>>> https://bugs.openjdk.java.net/browse/JDK-8241436
>>>
>>> (The patch is built on top of 8240248 [1].)
>>>
>>> MacroAssembler has C2-only declarations and the number is growing over time.
>>>
>>> C1 and template interpreter already subclass MacroAssembler for their own needs (C1_MacroAssembler and 
>>> InterpreterMacroAssembler respectively).
>>>
>>> I propose to introduce C2_MacroAssembler, move C2-specific code there, and use it exclusively across C2 code base.
>>>
>>> I took the liberty and refactored aarch64, arm32, ppc64, s390x code along the way. I did a cross-build, but I ask 
>>> maintaners to verify that it doesn't break anything.
>>>
>>> Testing: tier1-4, build (product+debug): linux-aarch64,linux-arm32,linux-ppc64le,linux-s390x,linux-x64,solaris-sparcv9
>>>
>>> Thanks!
>>>
>>> Best regards,
>>> Vladimir Ivanov
>>>
>>> [1] 8240248: Extend superword reduction optimizations for x86
>>>      http://cr.openjdk.java.net/~sviswanathan/8240248/webrev.02/


More information about the hotspot-compiler-dev mailing list