PR - Ticket 50308 - Fix memory leaks for repeat binds and replication

[389-ds-bot]

Cloned from Pagure Pull-Request: https://pagure.io/389-ds-base/pull-request/50309

• Created at 2019-03-28 23:18:57 by mreynolds (@mreynolds389)
• Merged at 2019-03-29 14:19:55

---

Description: Fixed two memory leaks:

- If a worker thread had multiple binds the "bind dn"
  thread data was leaked.
- Memory leak when processing changes in the changelog

Resolves: #3367

Reviewed by: ?

[389-ds-bot]

Comment from firstyear (@Firstyear) at 2019-03-29 02:40:31

Looks pretty reasonable to me, ack

[389-ds-bot]

Comment from mreynolds (@mreynolds389) at 2019-03-29 14:19:09

rebased onto 6c2bb66

[389-ds-bot]

Comment from mreynolds (@mreynolds389) at 2019-03-29 14:19:55

Pull-Request has been merged by mreynolds389

[389-ds-bot]

Comment from tbordaz (@tbordaz) at 2019-03-29 14:49:06

Sorry to be late on that review
buf->buf_cscbs[i] was allocated (calloc) few lines above (clcache_new_cscb).
I think we had no leak at this point.

[389-ds-bot]

Comment from tbordaz (@tbordaz) at 2019-03-29 14:54:12

csn_dup_or_init_by_csn does not overwrite the consumer_maxcsn pointer but overwrite the content of it.
I do not see how it can create a leak.

[389-ds-bot]

Comment from tbordaz (@tbordaz) at 2019-03-29 14:58:06

The same here. The content of consumer_maxcsn was overitten but not the pointer.

[389-ds-bot]

Comment from mreynolds (@mreynolds389) at 2019-03-29 15:31:56

@tbordaz, ASAN complains about these though :-/ I will double check if they are all needed, but some of them definitely were leaking. I have to fix a compiler warning anyway so I'll see what I can find...

[389-ds-bot]

Patch
50309.patch