[openssl-dev] [openssl.org #4581] [1.0.2] Running tests in parallel results in failure

[Richard Levitte via RT]

Like Rich says, our build system in 1.0.2 doesn't support parallell building or
testing. For upcoming 1.1.0, the build system has been remade from the ground
up, with parallell building in mind. Parallell testing hasn't been tested there
either, though... it might work, it might not. However, the '-j' option for
make till not affect the parallellism of testing in any way.

Closing this ticket.

Cheers,
Richard

On Tue Jun 21 13:14:12 2016, pmenzel at molgen.mpg.de wrote:
> Dear OpenSSL folks,
>
>
> downloading the latest 1.0.1t release [1], building it, and running the
> tests in parallel I get the failure below.
>
> I am able to reproduce this, with the branch.*OpenSSL_1_0_2-stable* [2],
> but not with the branch *master*.
>
> With `-j1` and `-j2` the failure is unreproducible. With `-j3` it
> sometimes works and sometimes fails. With `-j4` and greater it reliably
> fails.
>
> ```
> ...long/negative scalar tests allowing precomputation ... without
> precomputation ... ok
> combined multiplication ....p -> p
> ........d -> d
> .camellia-128-cbc
> ......n -> d
> ..........p -> d
> .....camellia-128-cbc base64
> .d -> n
> ...n -> n
> ok
>
> ..testing internal curves: ..............camellia-128-ecb
> ......................p -> n
> ............d -> p
> ..............................n -> p
> .............p -> p
> ... ok
>
> camellia-128-ecb base64
> .....rsa
> .testing rsa conversions
> p -> d
> .Parsing test certificates
> ...camellia-192-cbc
> p -> p
> ....OK
> .camellia-192-cbc base64
> d -> d
> .testing crl conversions
> p -> d
> ..p -> d
> ..p -> p
> .camellia-192-ecb
> .d -> p
> ..p -> p
> d -> d
> ...p -> d
> camellia-192-ecb base64
> ..d -> p
> .../util/shlib_wrap.sh ./rsa_test
> ..p -> p
> PKCS #1 v1.5 encryption/decryption ok
> OAEP encryption/decryption ok
> ..camellia-256-cbc
> PKCS #1 v1.5 encryption/decryption ok
> OAEP encryption/decryption ok
> .PKCS #1 v1.5 encryption/decryption ok
> OAEP encryption/decryption ok
> ...testing session-id conversions
> p -> d
> ......p -> p
> ......d -> d
> ..camellia-256-cbc base64
> .....p -> d
> .......PKCS #1 v1.5 encryption/decryption ok
> d -> p
> OAEP encryption/decryption ok
> .camellia-256-ecb
> ...PKCS #1 v1.5 encryption/decryption ok
> OAEP encryption/decryption ok
> p -> p
> ...PKCS #1 v1.5 encryption/decryption ok
> OAEP encryption/decryption ok
> .......camellia-256-ecb base64
> Generate and verify a certificate request
> .generating certificate request
> ...rsa
> .There should be a 2 sequences of .'s and some +'s.
> There should not be more that at most 80 per line
> This could take some time.
> ....cast
> ..Generating a 1024 bit RSA private key
> .....................cast base64
> ....................++++++
> ......................cast-cbc
> ..................testing req conversions
> ..p -> d
> .......................cast-cbc base64
> .Makefile:237: recipe for target 'test_req' failed
> make[1]: *** [test_req] Error 1
> make[1]: *** Waiting for unfinished jobs....
> ```
>
>
> Thanks,
>
> Paul
>
>
> PS: I am sorry, if Mozilla Thunderbird wraps the pasted lines.
>
>
> [1] https://www.openssl.org/source/openssl-1.0.1t.tar.gz
> [2] OpenSSL_1_0_2h-88-g4824496 (Fix missing opening braces)
> [3] 6feb3c5 Avoid using latest clang since repo not available
>


--
Richard Levitte
levitte at openssl.org

-- 
Ticket here: http://rt.openssl.org/Ticket/Display.html?id=4581
Please log in as guest with password guest if prompted