Formats failing on the new Intel OpenCL CPU intel-oneapi (2024.0)

[claudioandre-br]

FTR:

• I'm documenting here formats that fail on the new Intel OpenCL CPU driver/runtime.
• BTW: they were working on previous driver version 18.1.0.013 (for Ubuntu 16).

"$JTR_BIN" -test-full=0 --format=opencl

2023-11-26 16:39:13 (13.4 MB/s) - 'run_tests.sh' saved [14372/14372]

---------------------------- TESTING -----------------------------
Binary: ../run/john
=> ../run/john --list=build-info
Version: 1.9.0-jumbo-1+bleeding-09f03ea13e 2023-11-25 10:47:56 -0300

[...]
Testing: krb5tgs-opencl, Kerberos 5 TGS-REP etype 23 [MD4 HMAC-MD5 RC4 OpenCL]... Build log: Compilation started
Compilation done
Linking started
Linking done
Device build started
Options used by backend compiler: -I opencl -cl-mad-enable -DDEVICE_INFO=36 -D__SIZEOF_HOST_SIZE_T__=8 -DDEV_VER_MAJOR=2023 -DDEV_VER_MINOR=16 -D_OPENCL_COMPILER -DPLAINTEXT_LENGTH=27 -DDATA_LEN=1145 -DUCS_2 -DCONST_CACHE_SIZE=131072 -DENC_RAW -DENC_RAW -DLOC_0=-1 -DLOC_1=-1 -DLOC_2=-1 -DLOC_3=-1 -DNUM_INT_KEYS=1 -DIS_STATIC_GPU_MASK=0
Device build done
Kernel "krb5tgs_init" was successfully vectorized (16)
Kernel "krb5tgs_crypt" was successfully vectorized (16)
Done.
FAILED (crypt_all(49) zero return)
[...]
Testing: pfx-opencl, (.pfx, .p12) [PKCS#12 PBE (SHA1/SHA-256/512) OpenCL]... Build log: Compilation started
Compilation done
Linking started
Linking done
Device build started
Options used by backend compiler: -I opencl -cl-mad-enable -DDEVICE_INFO=36 -D__SIZEOF_HOST_SIZE_T__=8 -DDEV_VER_MAJOR=2023 -DDEV_VER_MINOR=16 -D_OPENCL_COMPILER -DPLAINTEXT_LENGTH=48 -DMAX_DATA_LENGTH=8192 -DENC_RAW -DENC_RAW
Device build done
Kernel "pfx" was successfully vectorized (16)
Done.
FAILED (cmp_all(49))
[...]
2 out of 73 tests have FAILED

[solardiz]

Is the OpenCL "driver" upgrade the only change? I made relevant changes in July, so if your prior testing was before July, then those changes I made could also be suspect.

[claudioandre-br]

It is CI.

[Edited]
It was tested and passed yesterday. See the download date of run_tests.sh and --list=build-info output

Rebased yesterday (it includes NEWS: Add Keplr wallet support):

2023-11-25 16:20:07 (40.5 MB/s) - 'run_tests.sh' saved [14251/14251]

---------------------------- TESTING -----------------------------
Binary: ../run/john
=> ../run/john --list=build-info
Version: 1.9.0-jumbo-1+bleeding-09f03ea13e 2023-11-25 10:47:56 -0300

[...]
Testing: krb5tgs-opencl, Kerberos 5 TGS-REP etype 23 [MD4 HMAC-MD5 RC4 OpenCL]... Build log: Compilation started
Compilation done
Linking started
Linking done
Device build started
Device build done
Kernel <krb5tgs_init> was successfully vectorized (16)
Kernel <krb5tgs_crypt> was not vectorized
Done.
PASS
[...]

Testing: pfx-opencl, (.pfx, .p12) [PKCS#12 PBE (SHA1/SHA-256/512) OpenCL]... Build log: Compilation started
Compilation done
Linking started
Linking done
Device build started
Device build done
Kernel <pfx> was successfully vectorized (16)
Done.
PASS

[solardiz]

In other words, it worked yesterday and stopped working today with only the OpenCL "driver" having changed, correct?

[claudioandre-br]

In other words, it worked yesterday and stopped working today with only the OpenCL "driver" having changed, correct?

The exact same (last) commit was tested yesterday and today.

[solardiz]

Somehow we saw even more/different formats fail in: #5382 (comment)

[claudioandre-br]

mscash (OpenCL) formats are disabled during my tests. They were problematic somewhere and I keep them disabled (only inside CI).