RFR(XS) : 8155192 : [TESTBUG] SHA tests assumes that x86 does not have SHA intrinsics

Gustavo Romero gromero at linux.vnet.ibm.com
Fri Jun 22 19:03:13 UTC 2018 

Hi Igor,

On 06/22/2018 03:41 PM, Igor Ignatyev wrote:
> thanks for checking. I'll push webrev.01 + your diff later today.

Sounds good, thanks!


Regards,
Gustavo

> -- Igor
> 
>> On Jun 22, 2018, at 5:42 AM, Gustavo Romero <gromero at linux.vnet.ibm.com> wrote:
>>
>> Hi Igor,
>>
>> On 06/21/2018 08:41 PM, Igor Ignatyev wrote:
>>> http://cr.openjdk.java.net/~iignatyev//8155192/webrev.01/index.html
>>>> 33 lines changed: 1 ins; 10 del; 22 mod;
>>> I've removed Platforms::* from UseSHAIntrinsicsSpecificTestCaseForUnsupportedCPU and UseSHASpecificTestCaseForSupportedCPU. all compiler/intrinsics/sha tests passed in our labs on sparc and x86 hosts.
>>> new version of webrev -- http://cr.openjdk.java.net/~iignatyev//8155192/webrev.01/index.html
>>> Gustavo, could you please test this version of path on your platforms?
>>
>> Sure.
>>
>> So after applying it I got the following state:
>>
>> FAILED: compiler/intrinsics/sha/cli/TestUseSHA512IntrinsicsOptionOnSupportedCPU.java
>> Passed: compiler/intrinsics/sha/cli/TestUseSHAOptionOnUnsupportedCPU.java
>> FAILED: compiler/intrinsics/sha/cli/TestUseSHA256IntrinsicsOptionOnSupportedCPU.java
>> Passed: compiler/intrinsics/sha/cli/TestUseSHA256IntrinsicsOptionOnUnsupportedCPU.java
>> FAILED: compiler/intrinsics/sha/cli/TestUseSHA1IntrinsicsOptionOnUnsupportedCPU.java
>> Passed: compiler/intrinsics/sha/cli/TestUseSHA512IntrinsicsOptionOnUnsupportedCPU.java
>> FAILED: compiler/intrinsics/sha/cli/TestUseSHAOptionOnSupportedCPU.java
>> Passed: compiler/intrinsics/sha/cli/TestUseSHA1IntrinsicsOptionOnSupportedCPU.java
>> Passed: compiler/intrinsics/sha/sanity/TestSHA1Intrinsics.java
>> Passed: compiler/intrinsics/sha/sanity/TestSHA1MultiBlockIntrinsics.java
>> Passed: compiler/intrinsics/sha/sanity/TestSHA512Intrinsics.java
>> Passed: compiler/intrinsics/sha/sanity/TestSHA256Intrinsics.java
>> Passed: compiler/intrinsics/sha/sanity/TestSHA512MultiBlockIntrinsics.java
>> Passed: compiler/intrinsics/sha/sanity/TestSHA256MultiBlockIntrinsics.java
>> Passed: compiler/intrinsics/sha/TestSHA.java
>> Test results: passed: 11; failed: 4
>>
>> due to (Please see full log at http://cr.openjdk.java.net/~gromero/logs/sha_jtreg_failure.log):
>>
>> java.lang.Error: Support for CPUs different fromn AARCH64, S390x, SPARC, and X86 is not implemented
>>         at compiler.intrinsics.sha.cli.SHAOptionsBase.getWarningForUnsupportedCPU(SHAOptionsBase.java:92)
>>         at compiler.intrinsics.sha.cli.testcases.GenericTestCaseForSupportedCPU.verifyWarnings(GenericTestCaseForSupportedCPU.java:50)
>>         at compiler.intrinsics.sha.cli.SHAOptionsBase$TestCase.test(SHAOptionsBase.java:152)
>>         at compiler.intrinsics.sha.cli.SHAOptionsBase.runTestCases(SHAOptionsBase.java:129)
>>         at jdk.test.lib.cli.CommandLineOptionTest.test(CommandLineOptionTest.java:537)
>>         at compiler.intrinsics.sha.cli.TestUseSHA512IntrinsicsOptionOnSupportedCPU.main(TestUseSHA512IntrinsicsOptionOnSupportedCPU.java:47)
>>         at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
>>         at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
>>         at java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
>>         at java.base/java.lang.reflect.Method.invoke(Method.java:566)
>>         at com.sun.javatest.regtest.agent.MainWrapper$MainThread.run(MainWrapper.java:115)
>>         at java.base/java.lang.Thread.run(Thread.java:832)
>>
>> which seems resolved by the following tiny change:
>>
>> diff -r 027c909decaf test/hotspot/jtreg/compiler/intrinsics/sha/cli/SHAOptionsBase.java
>> --- a/test/hotspot/jtreg/compiler/intrinsics/sha/cli/SHAOptionsBase.java        Fri Jun 22 08:10:43 2018 -0400
>> +++ b/test/hotspot/jtreg/compiler/intrinsics/sha/cli/SHAOptionsBase.java        Fri Jun 22 08:33:42 2018 -0400
>> @@ -75,7 +75,7 @@
>>       */
>>      public static String getWarningForUnsupportedCPU(String optionName) {
>>          if (Platform.isAArch64() || Platform.isS390x() || Platform.isSparc() ||
>> -            Platform.isX64() || Platform.isX86()) {
>> +            Platform.isX64() || Platform.isX86() || Platform.isPPC()) {
>>              switch (optionName) {
>>              case SHAOptionsBase.USE_SHA_OPTION:
>>                  return SHAOptionsBase.SHA_INSTRUCTIONS_ARE_NOT_AVAILABLE;
>> @@ -89,7 +89,7 @@
>>                  throw new Error("Unexpected option " + optionName);
>>              }
>>          } else {
>> -            throw new Error("Support for CPUs different fromn AARCH64, S390x, SPARC, and X86 "
>> +            throw new Error("Support for CPUs different from AARCH64, S390x, SPARC, X86, and PPC "
>>                              + "is not implemented");
>>          }
>>      }
>>
>> Thus with your fix plus that fix above ^ the current state looks ok:
>>
>> Passed: compiler/intrinsics/sha/cli/TestUseSHA512IntrinsicsOptionOnSupportedCPU.java
>> Passed: compiler/intrinsics/sha/cli/TestUseSHAOptionOnUnsupportedCPU.java
>> Passed: compiler/intrinsics/sha/cli/TestUseSHA256IntrinsicsOptionOnUnsupportedCPU.java
>> Passed: compiler/intrinsics/sha/cli/TestUseSHA512IntrinsicsOptionOnUnsupportedCPU.java
>> Passed: compiler/intrinsics/sha/cli/TestUseSHA1IntrinsicsOptionOnUnsupportedCPU.java
>> Passed: compiler/intrinsics/sha/cli/TestUseSHA1IntrinsicsOptionOnSupportedCPU.java
>> Passed: compiler/intrinsics/sha/cli/TestUseSHA256IntrinsicsOptionOnSupportedCPU.java
>> Passed: compiler/intrinsics/sha/sanity/TestSHA1Intrinsics.java
>> Passed: compiler/intrinsics/sha/cli/TestUseSHAOptionOnSupportedCPU.java
>> Passed: compiler/intrinsics/sha/sanity/TestSHA1MultiBlockIntrinsics.java
>> Passed: compiler/intrinsics/sha/sanity/TestSHA512Intrinsics.java
>> Passed: compiler/intrinsics/sha/sanity/TestSHA256Intrinsics.java
>> Passed: compiler/intrinsics/sha/sanity/TestSHA512MultiBlockIntrinsics.java
>> Passed: compiler/intrinsics/sha/sanity/TestSHA256MultiBlockIntrinsics.java
>> Passed: compiler/intrinsics/sha/TestSHA.java
>> Test results: passed: 15
>>
>>
>> Thanks,
>> Gustavo
>>
>>> Thanks,
>>> -- Igor
>>>> On Jun 21, 2018, at 2:43 PM, Igor Ignatyev <igor.ignatyev at oracle.com> wrote:
>>>>
>>>> Hi Gustavo,
>>>>
>>>> please see my answers inline.
>>>>
>>>> Thanks,
>>>> -- Igor
>>>>
>>>>> On Jun 21, 2018, at 1:34 PM, Gustavo Romero <gromero at linux.vnet.ibm.com <mailto:gromero at linux.vnet.ibm.com>> wrote:
>>>>>
>>>>> Hi Vladimir, Igor
>>>>>
>>>>> On 06/21/2018 02:47 PM, Vladimir Kozlov wrote:
>>>>>> But I find it strange that PPC64 (Platform::isPPC) is only listed in one files. According to vm_version_ppc.cpp it supports SHA256 and SHA512.
>>>>>
>>>>> Yes, that's correct.
>>>>>
>>>>> Igor, do you mind to confirm which tests exactly must be enabled on PPC64 (or
>>>>> disabled if the list is smaller...)?
>>>> all the tests are already enabled on PPC64, but they don't do much testing, as they will be skipped by the tests themselves.
>>>>>
>>>>> For instance, it's not clear to me yet why the suggested fix in [1] is adding PPC to
>>>>> UseSHAIntrinsicsSpecificTestCaseForUnsupportedCPU.java.
>>>> UseSHAIntrinsicsSpecificTestCaseForUnsupportedCPU is supposed to be run to hosts which *do* support some sha instructions, but don't support a specific one. I don't know if it's a case for PPC or not. in any case, it might be better to get rid of Platform::* predicates in this class and just use IntrinsicPredicates and SHAOptionsBase.getPredicateForOption, so its c-tor will be:
>>>>>          super(optionName, new AndPredicate(
>>>>>                          IntrinsicPredicates.ANY_SHA_INSTRUCTION_AVAILABLE,
>>>>>                          new NotPredicate(SHAOptionsBase.getPredicateForOption(optionName))));
>>>>>     < ...>
>>>>
>>>>
>>>> and UseSHASpecificTestCaseForSupportedCPU's c-tor can be changed to:
>>>>>          super(SHAOptionsBase.USE_SHA_OPTION, IntrinsicPredicates.ANY_SHA_INSTRUCTION_AVAILABLE);
>>>>
>>>>>     < ...>
>>>>
>>>> I'll retest the patch for 8155192 w/ these changes.
>>>>
>>>>>
>>>>> Current status on PPC64 is:
>>>>>
>>>>> Passed: compiler/intrinsics/sha/cli/TestUseSHA256IntrinsicsOptionOnUnsupportedCPU.java
>>>>> Passed: compiler/intrinsics/sha/cli/TestUseSHA1IntrinsicsOptionOnSupportedCPU.java
>>>>> Passed: compiler/intrinsics/sha/cli/TestUseSHAOptionOnUnsupportedCPU.java
>>>>> Passed: compiler/intrinsics/sha/cli/TestUseSHA512IntrinsicsOptionOnSupportedCPU.java
>>>>> Passed: compiler/intrinsics/sha/cli/TestUseSHA512IntrinsicsOptionOnUnsupportedCPU.java
>>>>> Passed: compiler/intrinsics/sha/cli/TestUseSHA256IntrinsicsOptionOnSupportedCPU.java
>>>>> Passed: compiler/intrinsics/sha/cli/TestUseSHAOptionOnSupportedCPU.java
>>>>> Passed: compiler/intrinsics/sha/cli/TestUseSHA1IntrinsicsOptionOnUnsupportedCPU.java
>>>>> Passed: compiler/intrinsics/sha/sanity/TestSHA1Intrinsics.java
>>>>> Passed: compiler/intrinsics/sha/sanity/TestSHA1MultiBlockIntrinsics.java
>>>>> Passed: compiler/intrinsics/sha/sanity/TestSHA512Intrinsics.java
>>>>> Passed: compiler/intrinsics/sha/sanity/TestSHA256Intrinsics.java
>>>>> Passed: compiler/intrinsics/sha/sanity/TestSHA512MultiBlockIntrinsics.java
>>>>> Passed: compiler/intrinsics/sha/sanity/TestSHA256MultiBlockIntrinsics.java
>>>>> Passed: compiler/intrinsics/sha/TestSHA.java
>>>>> Test results: passed: 15
>>>> as I explained above, all the tests are run on all platforms, but some of their parts are just skipped on PPC64.
>>>>
>>>>>
>>>>> Thanks.
>>>>>
>>>>> Best regards,
>>>>> Gustavo
>>>>>
>>>>> [1] https://bugs.openjdk.java.net/browse/JDK-8205492
>>>>>
>>>>>> Thanks,
>>>>>> Vladimir
>>>>>> On 6/20/18 8:28 PM, Igor Ignatyev wrote:
>>>>>>> http://cr.openjdk.java.net/~iignatyev//8155192/webrev.00/index.html
>>>>>>>> 29 lines changed: 10 ins; 3 del; 16 mod;
>>>>>>>
>>>>>>> Hi all,
>>>>>>>
>>>>>>> could you please review this tiny fix for SHA tests? the fix adjusts definitions of supported/unsupported CPUs in sha tests.
>>>>>>>
>>>>>>> JBS: https://bugs.openjdk.java.net/browse/JDK-8155192
>>>>>>> webrev: http://cr.openjdk.java.net/~iignatyev//8155192/webrev.00/index.htm
>>>>>>> testing: test/hotspot/jtreg/compiler/intrinsics tests
>>>>>>>
>>>>>>> Thanks,
>>>>>>> -- Igor
>>
>

More information about the hotspot-compiler-dev mailing list