Re: RFC/RFA: patch for PR 22156: improve SRA for bit-fields

[Alexandre Oliva <aoliva at redhat dot com>]

On Apr 20, 2007, "Andrew Pinski" <pinskia@gmail.com> wrote:

> On 4/20/07, Alexandre Oliva <aoliva@redhat.com> wrote:
>> I can think of a few possible ways to do it, but I don't understand
>> why this should be necessary.  Could anyone give me an idea of why
>> these (D)s should cause any problem?

> Yes because they are using uninitialized variables, anything can
> happen so later passes are causing that weird stuff to happen.

But how is this different from setting a bit-field in a
formerly-uninitialized word?  What you're seeing is the same, just
expressed in a slightly different way.

> Bernd's reload patch is just one example where GCC assumes
> uninitialized variables have an undefined value so it can use any
> register

And that would be correct behavior.

> and cause undefined behavior after that.

Now this would be a problem.  But if it was, why wasn't it a problem
before?  Just because the variable used to be in memory?

-- 
Alexandre Oliva         http://www.lsd.ic.unicamp.br/~oliva/
FSF Latin America Board Member         http://www.fsfla.org/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}