[Bug tree-optimization/104060] [11/12 Regression] -Wmaybe-uninitialized false alarm on address of local array

["rguenth at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104060

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P2
                 CC|                            |hubicka at gcc dot gnu.org,
                   |                            |msebor at gcc dot gnu.org,
                   |                            |rguenth at gcc dot gnu.org

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
Note we can eventually use modref info in the "wrong way" to query whether the
argument is dereferenced.  Note it can only alswer the question whether it
may be dereferenced or whether it kills (via stores) a specific range based on
it, both are not an exact match here since modref will conservatively say
the argument is dereferenced.

But I agree with Andrew - with the above we could make the warning only trigger
for local functions GCC saw the body for (or with LTO bodies analyzed during
IPA but now in another LTRANS unit).  That should eventually reduce the number
of false positives.