Re: [PATCH net-next v3 1/7] overflow: add DEFINE_FLEX() for on-stack allocs

[Przemek Kitszel <przemyslaw.kitszel@intel.com>]

On 8/18/23 12:49, David Laight wrote:
> From: Przemek Kitszel
>> Sent: Friday, August 18, 2023 11:28 AM
> ...
>>>>> I'm not sure you should be forcing the memset() either.
>>>>
>>>> This already got discussed: better to fail safe.
>>>
>>> Perhaps call it DEFINE_FLEX_Z() to make this clear and
>>> give the option for a non-zeroing version later.
>>> Not everyone wants the expense of zeroing everything.
>>
>> per Kees, zeroing should be removed by compiler when not needed:
>> https://lore.kernel.org/intel-wired-lan/202308101128.C4F0FA235@keescook/
> 
> Expect in the most trivial cases the compiler is pretty much never
> going to remove the zeroing of the data[] part.
> 
> I'm also not at all sure what happens if there is a function
> call between the initialisation and any assignments.
> 
> With a bit of effort you should be able to pass the '= {}'
> through into an inner #define.
> Possibly with the alternative of a caller-provider
>   '= { .obj = call_supplied_initialiser }'
> The 'not _Z' form would pass an empty argument.
> 
> 	David

Thanks, makes sense, there could be also DEFINE_FLEX_COUNTED
(or DEFINE_FLEX_BOUNDED) to cover Kees's __counted_by() cases.

Would you like me to cover/convert any existing code/use cases (as with 
other patches in the series, to have some examples/actual usage of newly 
introduced macros)?

> 
> -
> Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
> Registration No: 1397386 (Wales)