Open pshipton opened 1 year ago
https://openj9-jenkins.osuosl.org/job/Test_openjdk17_j9_extended.functional_s390x_linux_Nightly_testList_1/415 - ub20-390-3 CryptoTests_0 cryptotest/tests/SignatureTests.java
72) SunRsaSign: SHA384withRSA~OID.1.2.840.113549.1.1.12 (Signature)
"MainThread" prio=5 Id=22 RUNNABLE
at java.base@17.0.6-internal/java.math.BigInteger.implSquareToLen(BigInteger.java:2202)
at java.base@17.0.6-internal/java.math.BigInteger.squareToLen(BigInteger.java:2124)
at java.base@17.0.6-internal/java.math.BigInteger.square(BigInteger.java:2093)
at java.base@17.0.6-internal/java.math.BigInteger.square(BigInteger.java:2076)
at java.base@17.0.6-internal/java.math.BigInteger.lucasLehmerSequence(BigInteger.java:1070)
at java.base@17.0.6-internal/java.math.BigInteger.passesLucasLehmer(BigInteger.java:998)
at java.base@17.0.6-internal/java.math.BigInteger.primeToCertainty(BigInteger.java:978)
at java.base@17.0.6-internal/java.math.BitSieve.retrieve(BitSieve.java:203)
at java.base@17.0.6-internal/java.math.BigInteger.largePrime(BigInteger.java:847)
at java.base@17.0.6-internal/java.math.BigInteger.probablePrime(BigInteger.java:785)
at java.base@17.0.6-internal/sun.security.rsa.RSAKeyPairGenerator.generateKeyPair(RSAKeyPairGenerator.java:187)
at java.base@17.0.6-internal/java.security.KeyPairGenerator$Delegate.generateKeyPair(KeyPairGenerator.java:721)
at java.base@17.0.6-internal/java.security.KeyPairGenerator.genKeyPair(KeyPairGenerator.java:495)
at app//cryptotest.utils.KeysNaiveGenerator.getRsaKeyPair(KeysNaiveGenerator.java:165)
at app//cryptotest.utils.KeysNaiveGenerator.getRsaPrivateKey(KeysNaiveGenerator.java:169)
at app//cryptotest.tests.SignatureTests.checkAlgorithm(SignatureTests.java:85)
at app//cryptotest.utils.AlgorithmTest.mainLoop(AlgorithmTest.java:79)
at app//cryptotest.tests.SignatureTests.main(SignatureTests.java:68)
@0xdaryl it seems OpenJ9 is too slow to complete this newly added test. I notice jdk17 implSquareToLen() has an @IntrinsicCandidate annotation, while in jdk11 it's @HotSpotIntrinsicCandidate.
@llxia can we increase the timeout setting for this test?
timeout set to 120000ms
https://openj9-jenkins.osuosl.org/job/Test_openjdk19_j9_extended.functional_aarch64_linux_Nightly_testList_0/99 - cent8-aarch64-1 CryptoTests_0 cryptotest/tests/SignatureTests.java
34) SUN: SHA3-256withDSA~OID.2.16.840.1.101.3.4.3.6 (Signature)
"MainThread" prio=5 Id=27 RUNNABLE
at java.base@19.0.2-internal/java.math.BigInteger.implMulAdd(BigInteger.java:3320)
90) SunRsaSign: SHA512withRSA~OID.1.2.840.113549.1.1.13 (Signature)
40) SUN: SHA256withDSA~OID.2.16.840.1.101.3.4.3.2 (Signature)
"MainThread" prio=5 Id=35 RUNNABLE
at java.base@19.0.2-internal/java.math.BigInteger.implSquareToLen(BigInteger.java:2316)
46) SUN: NONEwithDSA~RawDSA (Signature)
"MainThread" prio=5 Id=35 RUNNABLE
at java.base@19.0.2-internal/java.math.BigInteger.implMulAddCheck(BigInteger.java:3296)
29) SUN: SHA512withDSA~2.16.840.1.101.3.4.3.4 (Signature)
"MainThread" prio=5 Id=23 RUNNABLE
at java.base@19.0.2-internal/java.math.BigInteger.montReduce(BigInteger.java:3241)
25) SUN: SHA3-512withDSA~OID.2.16.840.1.101.3.4.3.8 (Signature)
"MainThread" prio=5 Id=23 RUNNABLE
at java.base@19.0.2-internal/java.math.BigInteger.implSquareToLen(BigInteger.java:2316)
91) SunRsaSign: SHA512withRSA~1.2.840.113549.1.1.13 (Signature)
cryptotest/CryptoTest.java is also timing out https://openj9-jenkins.osuosl.org/job/Test_openjdk19_j9_extended.functional_ppc64_aix_Nightly_testList_1/100 https://openj9-jenkins.osuosl.org/job/Test_openjdk19_j9_extended.functional_ppc64le_linux_Nightly_testList_1/101 https://openj9-jenkins.osuosl.org/job/Test_openjdk19_j9_extended.functional_s390x_linux_Nightly_testList_1/102 https://openj9-jenkins.osuosl.org/job/Test_openjdk19_j9_extended.functional_x86-64_linux_Nightly_testList_1/100
and cryptotest/tests/SecureRandomTests.java https://openj9-jenkins.osuosl.org/job/Test_openjdk19_j9_extended.functional_x86-64_linux_Nightly_testList_1/100
The jdk19 timeouts are also happening on Adoptium builds.
I can set -timeoutFactor:2
(instead of 1). But I am not sure if it is the best way to do it.
https://openj9-jenkins.osuosl.org/job/Grinder/2022/
Increase timeout for CryptoTests: https://github.com/adoptium/aqa-tests/pull/4339
I notice jdk17 implSquareToLen() has an @IntrinsicCandidate annotation, while in jdk11 it's @HotSpotIntrinsicCandidate.
Interesting. I believe our intrinsics matching code is only looking for the former, not the latter. But it is the former that might be the issue in this case. We'll investigate the slowness, but not for 0.38.
@BradleyWood : please investigate the slowdown here. I moved this to 0.40.
@0xdaryl it seems OpenJ9 is too slow to complete this newly added test. I notice jdk17 implSquareToLen() has an @IntrinsicCandidate annotation, while in jdk11 it's @HotSpotIntrinsicCandidate.
I don't think implSquareToLen
intrinsic was ever implemented in OpenJ9 on any platform. I think we should open an issue to implement it, and any other unimplemented intrinsics.
Reassigning to @jmesyou, as he's working on issue #17778.
I'm not sure whether we need both issues, or if #17778 will act as an epic issue with this as a subtask. I'll leave it for him to decide.
I'm assuming this won't be ready for the 0.41 code split, so I'll move it out to 0.43.
https://openj9-jenkins.osuosl.org/job/Test_openjdk17_j9_extended.functional_x86-64_linux_Nightly_testList_0/394/ - ub16x64j96 CryptoTests_0 cryptotest/tests/SignatureTests.java
timeout, no diagnostic files.