Re: [PATCH] drivers: staging: lustre: lustre: include: add "__attribute__((packed))" for the related union

[Chen Gang <gang.chen.5i5j@gmail.com>]

On 02/03/2014 06:22 PM, James Hogan wrote:
> On 03/02/14 10:05, David Laight wrote:
>> From: Dan Carpenter
>>> On Sat, Feb 01, 2014 at 09:57:39PM +0800, Chen Gang wrote:
>>>> It seems, our kernel still stick to treate 'pack' region have effect
>>>> with both 'align' and 'sizeof'.
>>>
>>> It's not about packed regions.  It's about unions.  It's saying the
>>> sizeof() a union is a multiple of 4 unless it's packed.
>>>
>>> union foo {
>>> 	short x;
>>> 	short y;
>>> };
>>>
>>> The author intended the sizeof(union foo) to be 2 but on metag arch then
>>> it is 4.
>>
>> The same is probably be true of: struct foo { _u16 bar; };
> 
> Yes indeed.
> 
>> Architectures that define such alignment rules are a right PITA.
>> You either need to get the size to 2 without using 'packed', or
>> just not define such structures.
>> It is worth seeing if adding aligned(2) will change the size - I'm
>> not sure.
> 
> __aligned(2) alone doesn't seem to have any effect on sizeof() or
> __alignof__() unless it is accompanied by __packed. x86_64 is similar in
> that respect (it just packs sanely in the first place).
> 
> Combining __packed with __aligned(2) does the trick though (__packed
> alone sets __aligned(1) which is obviously going to be suboptimal).
> 

Oh, thank you for your explanation.

And hope this feature issue can be fixed, and satisfy both kernel and
ABI. :-)


Thanks.
-- 
Chen Gang

Open, share and attitude like air, water and life which God blessed