Re: [PATCH] RFC: use C11 alignas instead of GCC attribute aligned

["Mattias Rönnblom" <hofors@lysator.liu.se>]

On 2024-01-26 11:18, Morten Brørup wrote:
>> From: Mattias Rönnblom [mailto:hofors@lysator.liu.se]
>> Sent: Friday, 26 January 2024 11.05
>>
>> On 2024-01-25 23:53, Morten Brørup wrote:
>>>> From: Tyler Retzlaff [mailto:roretzla@linux.microsoft.com]
>>>> Sent: Thursday, 25 January 2024 19.37
>>>>
>>>> ping.
>>>>
>>>> Please review this thread if you have time, the main point of
>>>> discussion
>>>> I would like to receive consensus on the following questions.
>>>>
>>>> 1. Should we continue to expand common alignments behind an
>> __rte_macro
>>>>
>>>>     i.e. what do we prefer to appear in code
>>>>
>>>>     alignas(RTE_CACHE_LINE_MIN_SIZE)
>>>>
>>>>     -- or --
>>>>
>>>>     __rte_cache_aligned
>>>>
>>>> One of the benefits of dropping the macro is it provides a clear
>> visual
>>>> indicator that it is not placed in the same location or get applied
>>>> to types as is done with __attribute__((__aligned__(n))).
>>>
>>> We don't want our own proprietary variant of something that already
>> exists in the C standard. Now that we have moved to C11, the __rte
>> alignment macros should be considered obsolete.
>>
>> Making so something cache-line aligned is not in C11.
> 
> We are talking about the __rte_aligned() macro, not the cache alignment macro.
> 

OK, in that case, what is the relevance of question 1 above?

>>
>> __rte_cache_aligned is shorter, provides a tiny bit of abstraction, and
>> is already an established DPDK standard. So just keep the macro. If it
>> would change, I would argue for it to be changed to rte_cache_aligned
>> (i.e., just moving it out of __ namespace, and maybe making it
>> all-uppercase).
>>
>> Non-trivial C programs wrap things all the time, standard or not. It's
>> not something to be overly concerned about, imo.
> 
> Using the cache alignment macro was obviously a bad example for discussing the __rte_aligned() macro.
> 
> FYI, Tyler later agreed to introducing the RTE_CACHE_ALIGNAS you had proposed in an earlier correspondence.
> 
>>
>>>
>>> Note: I don't mind convenience macros for common use cases, so we
>> could also introduce the macro suggested by Mattias [1]:
>>>
>>> #define RTE_CACHE_ALIGNAS alignas(RTE_CACHE_LINE_SIZE)
>>>
>>> [1]: https://inbox.dpdk.org/dev/dc3f3131-38e6-4219-861e-
>> b31ec10c08bb@lysator.liu.se/
>>>
>>>>
>>>> 2. where should we place alignas(n) or __rte_macro (if we use a
>> macro)
>>>>
>>>> Should it be on the same line as the variable or field or on the
>>>> preceeding line?
>>>>
>>>>     /* same line example struct */
>>>>     struct T {
>>>>         /* alignas(64) applies to field0 *not* struct T type
>> declaration
>>>> */
>>>>         alignas(64) void *field0;
>>>>         void *field1;
>>>>
>>>>         ... other fields ...
>>>>
>>>>         alignas(64) uint64_t field5;
>>>>         uint32_t field6;
>>>>
>>>>         ... more fields ...
>>>>
>>>>     };
>>>>
>>>>     /* same line example array */
>>>>     alignas(64) static const uint32_t array[4] = { ... };
>>>>
>>>>     -- or --
>>>>
>>>>     /* preceeding line example struct */
>>>>     struct T {
>>>>         /* alignas(64) applies to field0 *not* struct T type
>> declaration
>>>> */
>>>>         alignas(64)
>>>>         void *field0;
>>>>         void *field1;
>>>>
>>>>         ... other fields ...
>>>>
>>>>         alignas(64)
>>>>         uint64_t field5;
>>>>         uint32_t field6;
>>>>
>>>>         ... more fields ...
>>>>
>>>>     };
>>>>
>>>>     /* preceeding line example array */
>>>>     alignas(64)
>>>>     static const uint32_t array[4] = { ... };
>>>>
>>>
>>> Searching the net for what other projects do, I came across this
>> required placement [2]:
>>>
>>> uint64_t alignas(64) field5;
>>>
>>> [2]:
>> https://lore.kernel.org/buildroot/20230730000851.6faa3391@windsurf/T/
>>>
>>> So let's follow the standard's intention and put them on the same
>> line.
>>> On an case-by-case basis, we can wrap lines if it improves
>> readability, like we do with function headers that have a lot of
>> attributes.
>>>
>>>>
>>>> I'll submit patches for lib/* once the discussion is concluded.
>>>>
>>>> thanks folks
>>>