This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [patch] new portlet: cirrus-maverick


On Wed, Aug 14, 2002 at 01:26:54PM -0400, Phil Edwards wrote:
> On Wed, Aug 14, 2002 at 10:01:54AM -0700, Aldy Hernandez wrote:
> > > I thought the convention was to only document the option that was /not/
> > > the default?  Or at least to describe which of the flags was on by default.
> > 
> > News to me.  I can do either one.  Which is preferred?
> > 
> > Btw, the default is no-maverick-fix-invalid-insns.
> 
> Check out the opening description of the manual's warning options.
> (Or maybe it's optimization options.)  Somewhere, when describing the
> -f/-fno and -W/-Wno duality, it adds something to the effect of, "only
> the non-default options are documented here, so you can figure out the
> default behavior by inverting the listed option."
> 
> If It Were Me[tm], I'd only document maverick-fix-invalid-insns, with a
> sentence or two about what the problem is and what the default behavior is.
> I would give an example but I don't have a good understanding of the
> default behavior here.  :-)

Fair enough.  Fixed thusly:

@item -mmaverick-fix-invalid-insns
@opindex mmaverick-fix-invalid-insns
Some combinations of Maverick instructions are invalid if they occur
in sequence.  If you use this option, GCC will insert nops in the
relevant places to avoid generating these sequences.  This option is
turned off by default and is only needed for revisions of Maverick
chips for which certain combinations are invalid.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]