talloc: talloc_set_memlimit causes all reallocs to fail when used on pools. talloc_set_memlimit not enforced correctly on pools.

[Arran Cudbard-Bell]

> On Oct 20, 2020, at 4:34 PM, Jeremy Allison <jra at samba.org> wrote:
> 
> On Tue, Oct 20, 2020 at 02:50:05PM -0500, Arran Cudbard-Bell wrote:
>> 
>> 
>>> On Oct 20, 2020, at 2:16 PM, Jeremy Allison <jra at samba.org> wrote:
>>> 
>>> On Tue, Oct 20, 2020 at 01:42:00PM -0500, Arran Cudbard-Bell wrote:
>>>> 
>>>> 
>>>>> On Oct 20, 2020, at 1:33 PM, Jeremy Allison <jra at samba.org> wrote:
>>>>> 
>>>>> On Tue, Oct 20, 2020 at 01:03:14PM -0500, Arran Cudbard-Bell wrote:
>>>>>> 
>>>>>> Then there's another issue with object_count ending
>>>>>> up off by one, which means talloc_free_children doesn't
>>>>>> actually return memory to the pool, and that messes up
>>>>>> some of the other tests I'm adding.  Just tracking down
>>>>>> when and why this happens now.... It might have been a
>>>>>> pre-existing issue and not related to this patch, I'm just
>>>>>> seeing it because of using talloc_free_children to reset
>>>>>> the pool between some tests.
>>>> 
>>>> Apologies, there was a steal I didn't spot in the tests.
>>>> One of the chunks was moved out of the pool into the root ctx.
>>>> 
>>>> Explicitly freeing the chunk or stealing it back into the pool
>>>> means talloc_free_children works as expected.
>>> 
>>> FYI, once you've gotten everything working can you
>>> send your test cases to me so I can add them into
>>> the regression test suite for talloc ?
>> 
>> Attached.
>> 
>>> Then we'll add a new bug in bugzilla, update the minor library
>>> version number and create a gitlab MR.
>> 
>> Sounds good, thanks!
> 
> BUG: https://bugzilla.samba.org/show_bug.cgi?id=14540
> 
> is what we'll use to track this !

Posted regression tests there too.

Thanks!

-Arran