[PATCH] Fix AAPCS-VFP non-compliance when returning HFA from variadic function

[Oliver Stannard]

Updated patch attached.

Oliver

-----Original Message-----
From: Tim Northover [mailto:t.p.northover at gmail.com]
Sent: 16 January 2014 16:35
To: Oliver Stannard
Cc: Renato Golin; cfe-commits at cs.uiuc.edu
Subject: Re: [PATCH] Fix AAPCS-VFP non-compliance when returning HFA from variadic function

On 16 January 2014 16:02, Oliver Stannard <oliver.stannard at arm.com> wrote:
> I had another look, and it turns out that there is a very similar bug
> affecting HFA function arguments to variadic functions. I have attached an
> updated version of my patch which fixes this.

I still think this is expressed at the wrong level. It's not that HFAs
are particularly special in variadic functions, but that AAPCS_VFP
rules simply don't apply to variadics. HFAs don't exist as a concept,
just like they don't for APCS or AAPCS.

I think you get away with it because passing them as an HFA would be
(e.g. "double, double, double") works with the backend anyway, but
they should go as any other struct ([3 x i64] by the looks of it in
that case).

Cheers.

Tim.


-- IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium.  Thank you.

ARM Limited, Registered office 110 Fulbourn Road, Cambridge CB1 9NJ, Registered in England & Wales, Company No:  2557590
ARM Holdings plc, Registered office 110 Fulbourn Road, Cambridge CB1 9NJ, Registered in England & Wales, Company No:  2548782
-------------- next part --------------
A non-text attachment was scrubbed...
Name: arm-hfa-variadic-vfp-calling-convention-2.patch
Type: application/octet-stream
Size: 5519 bytes
Desc: arm-hfa-variadic-vfp-calling-convention-2.patch
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20140117/49bc11de/attachment.obj>