openssl: build fails in test_encoder_decoder with OPENSSL_TEST_RAND_ORDER=0

The builds currently show:


    # ERROR: (int) 'OSSL_ENCODER_CTX_get_num_encoders(ectx) > 0' failed @ test/endecode_test.c:187
    # [0] compared to [0]
    # ERROR: (bool) 'encode_cb(&encoded, &encoded_len, pkey, selection, output_type, output_structure, pass, pcipher) == true' failed @ test/endecode_test.c:127
    # false
    not ok 7 - test_protected_DSA_via_DER
# ------------------------------------------------------------------------------
    # ERROR: (int) 'OSSL_ENCODER_CTX_get_num_encoders(ectx) > 0' failed @ test/endecode_test.c:187
    # [0] compared to [0]
    # ERROR: (bool) 'encode_cb(&encoded, &encoded_len, pkey, selection, output_type, output_structure, pass, pcipher) == true' failed @ test/endecode_test.c:127
    # false
    not ok 12 - test_unprotected_DSA_via_PEM
# ------------------------------------------------------------------------------
    # ERROR: (int) 'OSSL_ENCODER_CTX_get_num_encoders(ectx) > 0' failed @ test/endecode_test.c:187
    # [0] compared to [0]
    # ERROR: (bool) 'encode_cb(&encoded, &encoded_len, pkey, selection, output_type, output_structure, pass, pcipher) == true' failed @ test/endecode_test.c:127
    # false
    not ok 16 - test_params_DSA_via_PEM
[...]
# ------------------------------------------------------------------------------
../../util/wrap.pl ../../test/endecode_test => 1
not ok 1
# ------------------------------------------------------------------------------
#   Failed test at test/recipes/04-test_encoder_decoder.t line 23.
# Looks like you failed 1 test of 1.04-test_encoder_decoder.t .......... 
Dubious, test returned 1 (wstat 256, 0x100)
Failed 1/1 subtests 

About this issue

  • Original URL
  • State: closed
  • Created 4 years ago
  • Comments: 25 (25 by maintainers)

Commits related to this issue

Most upvoted comments

… and would need some help figuring that out.

I actually figured it out.

Found one cause! do_i2b (in crypto/pem/pvkfmt.c) forcedly downgrades the input pkey. That needs being corrected…

I’m running tests now. I can’t see any reason why no-asm should make any difference, so I’m trying without that to avoid the abysmal slowness…