Re: Warn when returning the address of a temporary (middle-end) v2

[Jeff Law <law@redhat.com>]

On 07/02/14 06:19, Alan Modra wrote:
> On Mon, Jun 30, 2014 at 11:37:50PM +0200, Marc Glisse wrote:
>> On Mon, 30 Jun 2014, Jeff Law wrote:
>>
>>> On 06/29/14 03:22, Marc Glisse wrote:
>>>>
>>>> After looking at PR 61597, I updated the 2 conditions to:
>>>>
>>>> +          if ((TREE_CODE (valbase) == VAR_DECL
>>>> +               && !is_global_var (valbase))
>>>> +              || TREE_CODE (valbase) == PARM_DECL)
>>>>
>>>> a PARM_DECL is a local variable and returning its address is wrong,
>>>> isn't it?
>>> Right.  It can live in either a caller or callee allocated slot.
>>
>> The "caller" case scares me a bit. Is it really wrong to return the
>> address in that case? The slot still exists after returning if the
>> caller is responsible for it.
>
> At least on powerpc64, which uses a caller allocated parameter save
> area, returning the address of something in the parameter save area
> merits a warning.  The ABIs explicitly state that the parameter save
> area is not preserved over function calls.
>
> Also note that anything left in a caller allocated parameter save area
> will potentially be trashed by arguments written for the next call.
Similarly for the PA ABIs.  If something returned the address of one of 
those slots, we'd definitely want a warning.

I'd be amazed if that wasn't the case for every ABI with a caller 
allocated parameter save-back area.

jeff