[PATCH] Revert "linux: Implement mremap in C" (bug 31968)

[Florian Weimer]

* Andreas K. Huettel:

>> >
>> > I'd like to see testcases first.  They may have
>> > impacts on how the fix will look like.  Testcases should cover all known
>> > mremap constants.
>> > The function prototype should be consistent
>> > with the implementation.  Can we keep the
>> > C implementation and add a compile time
>> > check for unknown constants?
>> >
>> 
>> I uploaded a different patch at
>> 
>> https://sourceware.org/bugzilla/show_bug.cgi?id=31968#c3
>> 
>
> Practical question, since 2.40 is coming closer...  *If* this fix is still 
> supposed to go in (and especially if not all tests are ready in time?)- 

It's not a regression, so I don't think this should block the 2.40
release.  I thought the revert would be uncontroversial, but that does
not seem to be the case.

Thanks,
Florian