RFR: 8229971: ARM32: implementation for Thread-local handshakes

Robbin Ehn robbin.ehn at oracle.com
Fri Jan 24 11:10:41 UTC 2020


Thanks, seems good now.

/Robbin

On 1/23/20 1:37 PM, Boris Ulasevich wrote:
> Hi Robin,
> 
> Thanks for review!
> You are right about missing suspend_flags check! Thanks for that.
> Here is an updated review (findings fix, copyrights updated):
> http://cr.openjdk.java.net/~bulasevich/8229971/webrev.05
> 
> Boris
> 
> On 10.01.2020 12:58, Robbin Ehn wrote:
>> Hi Boris, thanks for addressing this!
>>
>> On 12/11/19 12:59 PM, Boris Ulasevich wrote:
>>> http://cr.openjdk.java.net/~bulasevich/8229971/webrev.04
>>
>> #
>> sharedRuntime_arm.cpp line 1230:
>> As we have not replaced suspend_flags, yet, with handshakes.
>> So you still need to check them, or am I misunderstanding something here?
>> As in e.g. open/src/hotspot/cpu/x86/sharedRuntime_x86_64.cpp line 2615.
>>
>> #
>> Small nit, same file line 1802:
>> 1802   if (!cause_return) {
>> 1803     if (SafepointMechanism::uses_thread_local_poll() && !cause_return) {
>> cause_return checked twice.
>>
>> #
>> templateInterpreterGenerator_arm.cpp line 974
>> Same here you still need to check suspend_flags.
>> As in e.g. 
>> open/src/hotspot/cpu/aarch64/templateInterpreterGenerator_aarch64.cpp line 1387.
>>
>> Thanks, Robbin
>>
>>>
>>> Boris
>>>
>>>> Best regards,
>>>> Martin
>>>>
>>>>
>>>>> -----Original Message-----
>>>>> From: hotspot-runtime-dev <hotspot-runtime-dev-
>>>>> bounces at openjdk.java.net> On Behalf Of Boris Ulasevich
>>>>> Sent: Dienstag, 10. Dezember 2019 13:03
>>>>> To: hotspot-runtime-dev at openjdk.java.net
>>>>> Subject: RFR: 8229971: ARM32: implementation for Thread-local handshakes
>>>>>
>>>>> Hi,
>>>>>
>>>>> Please review the implementation of Thread-local handshakes [1] support
>>>>> for ARM32. The implementation basically follows the similar update in
>>>>> aarch64 port [2].
>>>>>
>>>>> http://cr.openjdk.java.net/~bulasevich/8229971/webrev.03
>>>>> http://bugs.openjdk.java.net/browse/JDK-8229971
>>>>>
>>>>> Tested with JTReg with no regressions in release mode (fastdebug mode
>>>>> testing is still in progress, but I am pretty sure it is Ok).
>>>>>
>>>>> thanks,
>>>>> Boris
>>>>>
>>>>> [1] http://bugs.openjdk.java.net/browse/JDK-8185640
>>>>> [2]
>>>>> http://mail.openjdk.java.net/pipermail/aarch64-port-dev/2017-
>>>>> November/005104.html


More information about the hotspot-runtime-dev mailing list