Re: [wwwdocs] Document in changes.html -fcode-hoisting, -fipa-bit-cp, -fipa-vrp, -fsplit-loops, GCJ removal, x86 ISA additions, -fshrink-wrap-separate etc.

[Martin Jambor <mjambor at suse dot cz>]

Hi,

On Sun, Jan 29, 2017 at 07:55:18PM +0100, Jakub Jelinek wrote:
> On Sun, Jan 29, 2017 at 06:08:32PM +0530, Prathamesh Kulkarni wrote:
> > > +  <li>A new interprocedural bitwise constant propagation optimization
> > > +  has been added, which propagates knowledge about which bits of variables
> > > +  are known to be zero (including pointer alignment information) across
> > > +  the call graph.  It can be enabled by using the <code>-fipa-bit-cp</code>
> > > +  option if <code>-fipa-cp</code> is enabled as well, and is enabled by
> > > +  default at the <code>-O2</code> optimization level and higher.</li>
> > Hi,
> > Just a small comment:  -fipa-bit-cp makes -fipa-cp-alignment deprecated.
> 
> I've committed it as is, feel free to propose changes.  But before that, we
> should figure out what to do with -fipa-cp-alignment.  Should it have
> Ignore flag and
> Does nothing. Preserved for backward compatibility.
> description in common.opt, and
> 
> @item -fipa-cp-alignment
> @opindex -fipa-cp-alignment
> When enabled, this optimization propagates alignment of function
> parameters to support better vectorization and string operations.
> 
> This flag is enabled by default at @option{-O2} and @option{-Os}.  It
> requires that @option{-fipa-cp} is enabled.
> @option{-fipa-cp-alignment} is obsolete, use @option{-fipa-bit-cp} instead.
> 
> removed altogether, or should it say be alias for -fipa-bit-cp?

Well, -fipa-bit-cp is a superset of the old -fipa-cp-alignment in the
sense that the latter worked only for pointers whereas the former now
also operates on integers.

We could still refuse to store results of the analysis to pointers if
user provided -fno-ipa-cp-alignment, but I do not think it is
reasonable.  My preference would be to mark it obsolete and document
that -fipa-bit-cp should be used instead.

> The documentation certainly doesn't document what it does, because nothing
> ever looks at that flag, so it is effectively ignored, not just obsolete.

I'm not sure I understand, it was used by gcc 6 and the documentation
describes what it did.  The documentation of course should be updated.

Martin