On 07/09/2018 03:23 AM, Alexey Brodkin wrote:
> Hi David,
>
> On Mon, 2018-07-09 at 10:18 +0000, David Laight wrote:
>> From: Alexey Brodkin
>>> Sent: 09 July 2018 11:00
>>
>> ...
>>> That's a good idea indeed but it doesn't solve the problem with
>>> struct devres_node. Consider the following snippet:
>>> -------------------------------->8-------------------------------
>>> struct mystruct {
>>> atomic64_t myvar;
>>> }
>>>
>>> struct mystruct *p;
>>> p = devm_kzalloc(dev, sizeof(*p), GFP_KERNEL);
>>> -------------------------------->8-------------------------------
>>>
>>> Here myvar address will match address of "data" member of struct devres_node.
>>> So if "data" is has offset of 12 bytes from the beginning of a page then
>>> myvar won't be 64-bit aligned regardless of myvar's attribute, right?
>>
>> ...
>>>>> - unsigned long long data[]; /* guarantee ull alignment */
>>
>> Ahh, that line should be:
>> u8 data[] __aligned(8); /* Guarantee 64bit alignment */
>
> And that pretty much what I suggested in my initial patch :)
>
> For the record x86 has exactly the same atomic64_t as you suggested,
> see https://elixir.bootlin.com/linux/latest/source/arch/x86/include/asm/atomic64_32.h#L13:
> ---------------------->8------------------
> typedef struct {
> u64 __aligned(8) counter;
> } atomic64_t;
> ---------------------->8------------------
And so does the ARC version since when the atomic64_t support was added by commit
ce6365270ecd
Also, you should consider using the pre-canned type aligned_u64.
typedef struct {
aligned_u64 counter;
^^^^^^^^^^^
} atomic64_t;
