This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Long Double Register Spillage in PA Port
- To: Mark Klein <mklein at dis dot com>
- Subject: Re: Long Double Register Spillage in PA Port
- From: Jeffrey A Law <law at cygnus dot com>
- Date: Mon, 31 Jul 2000 05:57:27 -0600
- cc: gcc at gcc dot gnu dot org
- Reply-To: law at cygnus dot com
In message <4.2.0.58.20000730182644.00bc18c0@garfield.dis.com>you write:
> At 08:58 PM 7/23/00 -0600, Jeffrey A Law wrote:
> You had long double successfully working on a pa32 port? Hmm ...
Yes.
> The culprit in this case is HARD_REGNO_MODE_OK. I don't understand
> why yours would work for HP-UX on PA32 and not for MPE unless the
> HP-UX calling conventions are different?
The conventions are the same.
> The main difference is the following that I added in my port
> for GET_MODE_SIZE > 8 (e.g. TF):
>
> || ((REGNO) & 3) == 0
>
> (Remember, I'm !TARGET_PA_11)
Could be. I never test !TARGET_PA_11 since there are so few machines
of that era still in service. It's entirely possible the problem you are
seeing is specific to !TARGET_PA11.
jeff