This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [RHBZ#150150] emit aliased static functions in -fno-unit-at-a-time
On Mar 4, 2005, Andrew Pinski <pinskia@physics.uc.edu> wrote:
> On Mar 4, 2005, at 8:52 AM, Jakub Jelinek wrote:
>> On Fri, Mar 04, 2005 at 12:01:51AM -0300, Alexandre Oliva wrote:
>>> We refrained from telling the cgraph infrastructure that a symbol is
>>> referenced through an emitted alias just because we were not in
>>> -funit-at-a-time, but the assumption that this is only needed in
>>> unit-at-a-time mode is incorrect: static functions are not emitted at
>>> the point they're encountered even in -fno-unit-at-a-time, and if we
>>> don't add the edges to the graph and no other references to the static
>>> symbol exist, we may end up not emitting it at all.
> This was declared not a bug.
> See PR 20167 and PR 15700.
No. What was declared not a bug was to refrain from following the
symbol names. Problem is we *do* follow them now, but only in
-funit-at-a-time mode, and that gets us inconsistent behavior between
-f{no-,}unit-at-a-time. See the chunk of code that was conditional in
the patch I proposed.
--
Alexandre Oliva http://www.ic.unicamp.br/~oliva/
Red Hat Compiler Engineer aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist oliva@{lsd.ic.unicamp.br, gnu.org}