[RFC][PATCH] malloc: Fix test for malloc_usable_size

[Matheus Castanho]

On 12/16/19 4:55 PM, DJ Delorie wrote:
> Matheus Castanho <msc@linux.ibm.com> writes:
>> We have seen some tests on recent builds on ppc64le failing with
>> the following:
>>
>>   ---=== ./malloc/tst-malloc-usable-tunables.out ===---
>>   malloc_usable_size: expected 7 but got 24
>>   FAIL: malloc/tst-malloc-usable-tunables.c
> 
> The root cause here is much more insidious.
> 
> See https://bugzilla.redhat.com/show_bug.cgi?id=1659276
> 
> Basically, tunables has a bug-er-design-issue, and without the right
> tunables, the malloc test can't work right anyway.
> 
> See also https://sourceware.org/bugzilla/show_bug.cgi?id=25035
> 
> Fix imminent :-)
> 

Great! Thanks for working on that.

But I'm still puzzled by why the test on malloc/tst-malloc-usable.c does
not consider the possibility of the return value being greater than the
requested alloc size, as the comment on the source indicates that could
happen.

--
Matheus Castanho