[llvm-commits] [PATCH] ARM ABI: handle varargs with vector types.

[Manman Ren]

On Oct 15, 2012, at 5:34 PM, Eli Friedman wrote:

> On Mon, Oct 15, 2012 at 5:26 PM, Manman Ren <mren at apple.com> wrote:
>> 
>> On Oct 15, 2012, at 5:11 PM, Eli Friedman wrote:
>> 
>>> On Mon, Oct 15, 2012 at 4:46 PM, manman ren <mren at apple.com> wrote:
>>>> 
>>>> The initial patch was updated and separated to two patches (attached).
>>>> 
>>>> The first patch will fix passing legal vector types as varargs:
>>>> We make sure the vector is correctly aligned before casting it to the vector type.
>>> 
>>> +  uint64_t Size = CGF.getContext().getTypeSize(Ty) / 8;
>>> +  uint64_t TyAlign = CGF.getContext().getTypeAlign(Ty) / 8;
>>> +
>>> +  // The ABI alignment for vectors is 8 for AAPCS and 4 for APCS.
>>> +  if (Ty->getAs<VectorType>() && Size >= 8) {
>>> +    if (getABIKind() == ARMABIInfo::AAPCS_VFP ||
>>> +        getABIKind() == ARMABIInfo::AAPCS)
>>> +      TyAlign = 8;
>>> +    else
>>> +      TyAlign = 4;
>>> +  }
>>> 
>>> We should be using the same rules here we do for argument passing.  In
>>> particular, for APCS, the argument-passing type alignment is
>>> unconditionally 4.  (This can have effects for structs marked with
>>> "__attribute__((aligned(16)))", etc.)
>> Is there an interface to query the argument-passing alignment?
> 
> Nothing cares about it other than the calling convention code, so
> there isn't a general API.  It's basically just "match whatever
> ARMABIInfo::classifyArgumentType does".

I don't see code in ARMABIInfo::classifyArgumentType that checks the argument-passing alignment.
The code in the patch is trying to match what the calling convention does in the backend:
>>> +  // The ABI alignment for vectors is 8 for AAPCS and 4 for APCS.
>>> +  if (Ty->getAs<VectorType>() && Size >= 8) {
>>> +    if (getABIKind() == ARMABIInfo::AAPCS_VFP ||
>>> +        getABIKind() == ARMABIInfo::AAPCS)
>>> +      TyAlign = 8;
>>> +    else
>>> +      TyAlign = 4;
>>> +  }

Thanks,
Manman

> 
> -Eli