What is up with glibc in rawhide?

[John Reiser]

Jakub Jelinek wrote:

> If you have issues with the new malloc code, first verify you aren't
> allocating anything from within async signal handlers.

Would it be possible (and similarly inexpensive) for glibc to detect
some such cases, somewhat like glibc detects some cases of double free()
or some overwriting of its internal data structures which support malloc?
Particularly as long as the new malloc code is considered to be experimental,
then the cost of extra checking may be worthwhile.

--