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 12:42:41 UTC 2018 

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