Issue 4563 - Failure on s390x: 'Fails to split RDN "o=pki-tomcat-CA" …

[tbordaz]

…into components'

Bug description:
SLAPI_OPERATION_TYPE is a stored/read as an int (slapi_pblock_get/set).
This although the storage field is an unsigned long.
Calling slapi_pblock_get with an long (8 btyes) destination creates
a problem on big-endian (s390x).

Fix description:
Define destination op_type as an int (4 bytes)

relates: #4563

Reviewed by:

Platforms tested: F31 (little endian)

[slyon]

Applying this patch to our 389-ds-base package and re-running the (previously) regressed dogtag-pki test suite makes it PASS:
https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-hirsute-slyon-testing/hirsute/s390x/d/dogtag-pki/20210127_140849_b37ed@/log.gz

[mreynolds389]

I know this fixes it, but should we be using strong int types here? int32_t?

[tbordaz]

@slyon thanks for precious help.
@mreynolds389 there are many 'int' set/get from pblock. I do not think it worth changing it (and all of them)

[mreynolds389]

@slyon thanks for precious help.
@mreynolds389 there are many 'int' set/get from pblock. I do not think it worth changing it (and all of them)

I was not suggesting changing all of them, just where it might be critical, but the problem here was simply the wrong type being used. Anyway, ACK from me

[Firstyear]

Ack from me too.