This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: egcs-1.1 release schedule
- To: Dave Love <d dot love at dl dot ac dot uk>
- Subject: Re: egcs-1.1 release schedule
- From: Jeffrey A Law <law at cygnus dot com>
- Date: Mon, 22 Jun 1998 23:29:50 -0600
- cc: egcs at cygnus dot com
- Reply-To: law at cygnus dot com
In message <rzqiultp90m.fsf@djlvig.dl.ac.uk>you write:
> >>>>> "Jeff" == Jeffrey A Law <law@cygnus.com> writes:
>
> Jeff> You can't do this with STACK_BOUNDARY since that says we
> Jeff> will 100% always have a properly aligned stack,
>
> I guess we've just been lucky with the previous g77. (BTW, perhaps
> the documentation of the alignment macros could be reviewed?) I
> thought there was an agreed patch available to do things right, too,
> which presumably I misunderstod about.
I went back and reviewed Marc's patch. It had ABI issues to contend
with. It was my mistake to say it didn't have ABI isssues a few months
ago.
Well, the optimizations which eliminate useless masks were in older
versions of gcc, but they were much less aggressive. Kenner spent
quite some time making them more aggressive (since I think it really
helps the alpha).
> Jeff> * The ABI is still going to mandate that some doubles in
> Jeff> argument lists are going to be mis-aligned.
>
> Fortunately that isn't much of an issue for Fortran :-).
True, but it is an issue for C :-)
> Jeff> Note that some non-ABI breaking changes to align doubles and
> Jeff> other values have gone into the x86 compiler. In particular we
> Jeff> should be properly aligning all data in the static store.
>
> Does that make -malign-double redundant?
I don't think so. I think -malign-double also effects stuff like alignment
of doubles in structures and the like.
jeff